GAS TURBINES IN SIMPLE CYCLE & COMBINED CYCLE APPLICATIONS* Gas Turbines in Simple Cycle Mode Introduction The gas turbine is the most versatile item of turbomachinery today. It can be used in several different modes in critical industries such as power generation, oil and gas, process plants, aviation, as well domestic and smaller related industries. A gas turbine essentially brings together air that it compresses in its compressor module, and fuel, that are then ignited. Resulting gases are expanded through a turbine. That turbine’s shaft continues to rotate and drive the compressor which is on the same shaft, and operation continues. A separate starter unit is used to provide the first rotor motion, until the turbine’s rotation is up to design speed and can keep the entire unit running. The compressor module, combustor module and turbine module connected by one or more shafts are collectively called the gas generator. The figures below (Figures 1 and 2) illustrate a typical gas generator in cutaway and schematic format.
Fig. 1. Rolls Royce RB211 Dry Low Emissions Gas Generator (Source: Process Plant Machinery, 2nd edition, Bloch & Soares, C. pub: Butterworth Heinemann, 1998) * Condensed extracts from selected chapters of “Gas Turbines: A Handbook of Land, Sea and Air Applications” by Claire Soares, publisher Butterworth Heinemann, BH, (for release information see www.bh.com) Other references include Claire Soares’ other books for BH and McGraw Hill (see www.books.mcgraw-hill.com) and course notes from her courses on gas turbine systems. For any use of this material that involves profit or commercial use (including work by nonprofit organizations), prior written release will be required from the writer and publisher in question. Please note that several topics in the gas turbine handbook, for instance Turbine Controls, Instrumentation and Diagnostics; as well as Performance Optimization and Environmental issues are not covered in this author’s material on this CD. The “Gas Turbines” book in question is several hundred pages long and besides the basics, covers some of the more complex and lengthy work in recent gas turbine development. Condensing it all here was not practical. What is here however, does give the reader the basic theory and practice of gas turbines in simple cycle and combined cycle mode, in power generation service.
1
Fig. 2. Schematic of modules: f: fan section, ag: low pressure compressor, bg: high pressure compressor, nd c: turbine, e: shaft, h: combustor (Source: Process Plant Machinery, 2 edition, Bloch & Soares, C. pub: Butterworth Heinemann, 1998)
Figure 3 below shows a gas turbine cutaway with its basic operating specification. Note this particular turbine model can be used for both 50 and 60Hz power generation.
Fig. 3. Alstom’s GT-8C2, 50/60Hz gas turbine with basic specification (Table 1: base load at ISO conditions) (Source: Alstom Power) Table 1. GT8C2 (60Hz) (ISO 2314: 1989) Fuel Natural Gas Frequency 60 Hz Gross electrical output 56.2 MW Gross electrical efficiency 33.8% Gross Heat Rate 10,098 Btu / kWh Turbine speed 6204 rpm Compressor pressure ratio 17.6:1 Exhaust gas flow 197 kg/s Exhaust gas temperature 508 °C NOx emissions, gas dry (corr. to 15% O2, dry) < 25 vppm
2
Claire Soares
Figure 4 shows another cutaway of another gas turbine. This gas turbine is used in 60Hz power generation service.
Fig. 4. Siemens V84.3A, 60Hz gas turbine. Note partial hybrid burner (24 burners) ring
Fig. 5. The basic gas turbine cycle
(Source: The Aircraft Engine Book, Rolls Royce UK)
The basic gas turbine cycle is illustrated (PV and T-s diagrams) in Figure 5. A comparison can be drawn between the gas turbine’s operating principle and a car engine’s. See Figures 5 and 6. A car operates with a piston engine (reciprocating motion) and typically handles much smaller volumes than a conventional gas turbine.
3
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 6. Comparison of the gas turbine and the reciprocating engine cycles (Source: The Aircraft Engine Book, Rolls Royce UK)
GT Applications (Simple Cycle) Direct drive and mechanical drive With land-based industries, gas turbines can be used in either direct drive or mechanical drive application. With power generation, the gas turbine shaft is coupled to the generator shaft, either directly or via a gearbox “direct drive” application. A gearbox is necessary in applications where the manufacturer offers the package for both 60 and 50 cycle (Hertz, Hz) applications. The gear box will use roughly 2 percent of the power developed by the turbine in these cases.
Fig. 7. A simple cycle gas turbine plant, 100 MW simple cycle power plant, Charleston, South Carolina USA, powered by Siemens gas turbines. (Source: Siemens Westinghouse)
Power generation applications extend to offshore platform use. Minimizing weight is a major consideration for this service and the gas turbines used are generally “aeroderivatives” (derived from lighter gas turbines developed for aircraft use). For mechanical drive applications, the turbine module arrangement is different. In these cases, the combination of compressor module, combustor module and turbine module is termed the gas generator. Beyond the turbine end of the gas generator is a freely rotating turbine. It may be one or more stages. It is not mechanically connected to the gas generator, but instead is mechanically coupled,
4
Claire Soares
sometimes via a gearbox, to the equipment it is driving. Compressors and pumps are among the potential “driven” turbomachinery items. See Figure 8 below.
Fig. 8. A typical free power turbine. (Source: Rolls-Royce, UK)
In power generation applications, a gas turbine’s power/ size is measured by the power it develops in a generator (units watts, kilowatts, Megawatts). In mechanical drive applications, the gas turbine’s power is measured in horsepower (HP), which is essentially the torque developed multiplied by the turbine’s rotational speed. In aircraft engine applications, if the turbine is driving a rotor (helicopter) or propeller (turboprop aircraft) then its power is measured in horsepower. This means that the torque transmission from the gas turbine shaft is, in principle, a variation of mechanical drive application. If an aircraft gas turbine engines operates in turbothrust or ramjet mode, (i.e. the gas turbine expels its exhaust gases and the thrust of that expulsion, propels the aircraft forward), its power is measured in pounds of thrust. See Figure 9 below.
Fig. 9. Propulsive efficiency is high for a propeller and low for a jet. (Source: Rolls-Royce, UK)
5
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 10. Gas turbines in offshore service: Offshore platforms produce their own power. Power plant selection is generally an aeroderivative (for weight considerations) gas turbine in simple cycle operation. (Source: GE Power Systems)
In marine applications, the gas turbine is generally driving the ship’s or ferry’s propellers, via a gear box.
Fig. 11. Gas turbines in marine service: SGT-500 Industrial Gas Turbine – 17 MW, Application: Two SGT-500 power packages for FPSO vessel in the Leadon oilfields (Note the SGT-500 was Alstom’s, formerly ABB’s GT-35, designation changed after Siemens acquisition). The Global Producer III from the Swan Hunter shipyards at Tyneside, UK, heads for the Leadon oil field in the UK Sector of the North Sea. This vessel is an FPSO (Floating Production, Storage and Offloading) vessel, and power on board is provided by two SGT-500 gas turbines. One WHRG (Waste Heat Recovery Generator) for each gas turbine heats process water. The SGT-500 is a light-weight, high-efficiency, heavy-duty industrial gas turbine. Its special design features are high reliability and fuel flexibility. It is also designed for single lift, which makes the unit suitable for all offshore applications. The modular, compact design of the GT35C facilitates onsite modular exchange. (Source: Siemens Westinghouse)
6
Claire Soares
GT24/26
GT11N2
Fig. 12a. Pictorial Examples of gas turbines, some with main operational parameters (Source: Alstom) Table 2. Alstom’s GT 24/ GT 26 (188MW 60Hz, 281MW 50Hz). Both used in simple cycle, combined cycle and other co-generation applications.
GT24 (ISO 2314:1989) Fuel Frequency Gross Electrical output Gross Electrical efficiency Gross Heat rate Turbine speed Compressor pressure ratio Exhaust gas flow Exhaust gas temperature NOx emissions (corr. to 15% O2,dry)
Natural gas 60 Hz 187.7 MW* 36.9 % 9251 Btu/kWh 3600 rpm 32:1 445 kg/s 612 °C < 25 vppm
GT26 (ISO 2314:1989) Fuel Frequency Gross Electrical output Gross Electrical efficiency Gross Heat rate Turbine speed Compressor pressure ratio Exhaust gas flow Exhaust gas temperature NOx emissions (corr. to 15% O2, dry)
Natural gas 50 Hz 281 MW* 38.3 % 8910 Btu/kWh 3000 rpm 32:1 632 kg/s 615 °C < 25 vppm
In combined cycle, approximately 12 MW (GT26) or 10 MW (GT24) is indirectly produced by the steam turbine through the heat released in the gas turbine cooling air coolers into the water steam cycle.
7
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Table 3. Alstom’s GT 11N2, either 60Hz or 50 Hz (with a gear box). Used in simple cycle, combined cycle and other cogeneration applications.
GT11N2 (50Hz) Fuel Frequency Gross Electrical output Gross Electrical efficiency Gross Heat rate Turbine speed Compressor pressure ratio Exhaust gas flow Exhaust gas temperature NOx emissions (corr. to 15% O2,dry)
Fuel Frequency Gross Electrical output Gross Electrical efficiency Gross Heat rate Turbine speed Compressor pressure ratio Exhaust gas flow Exhaust gas temperature NOx emissions (corr. to 15% O2,dry)
Natural Gas 50 Hz 113.6 MW 33.1% 10,305 Btu/kWh 3600 rpm 15.5:1 399 kg/s 531 °C < 25 vppm
GT11N2 (60Hz) Natural gas 60 Hz 115.4 MW 33.6% 10,150 Btu/kWh 3600 rpm 15.5 : 1 399 kg/s 531 °C < 25 vppm
Fig. 12b. SGT-600 Industrial Gas Turbine - 25 MW (former designation, Alstom’s GT10) (Source: Siemens Westinghouse) Technical Specifications Dual Fuel
natural gas and liquid
Frequency
50/60 Hz
Electrical output
24.8 MW
Electrical efficiency
34.2%
Heat rate
10,535 kJ/kWh
Turbine speed
7,700 rpm
Compressor pressure ratio
14.0:1
Exhaust gas flow
80.4 kg/s
Exhaust gas temperature NOx emissions (corr. to 15% O2, dry)
543 deg C <25 vppm
Figures 13 and 14 depict a cutaway and an external view respectively, of two aeroderivative engine models.
8
Claire Soares
Fig. 13. The GE LM6000 (aeroderivative of the CF6-80C2). (Source: GE Power Systems)
Fig. 14. The GE LM2500 (aeroderivative of the CF6-80C2). (Source: GE Power Systems)
9
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Figure 15 shows an industrial gas turbine during assembly at the OEM’s facility.
Fig. 15. GE-9H gas turbine is prepared for testing (Source: GE Power Systems)
Fig.16. A GE Frame 9H during test/manufacture. (Source: GE Power Systems)
Figure 17 shows an industrial gas turbine on a trestle in preparation for shipping.
10
Claire Soares
Fig. 17. A GE Frame 9F ready for shipping. (Source: GE Power Systems)
Figure 18 shows a large GE Frame 7F industrial gas turbine on a test bed in the OEM’s facility.
Fig.18. GE Frame 7F during manufacture/test showing rotor in half the casing (Source: GE Power Systems)
Applications versatility of the gas turbine The gas turbine’s operational mode gives it unique size adaptation potential. The largest gas turbines today are over 200 MW (megawatts) which then places gas turbines in an applications category that until recently, only steam turbines had owned. The smallest gas turbines are microturbines. The smallest commercially available microturbines are frequently used in small power generation (distributed power) applications and can be as small as 50 kW (kilowatts). Work continues on developing microturbines that will be thumbnail size. The world of “personal turbines” where one might plug this turbine into a “drive slot” in their car, come home from work and plug it into a “household slot” for all one’s household power, is a discernable, if as yet unpredictable, target. The content on this CD deals mainly with power generation, however with the gas turbine, understanding its origins and other applications, gives the gas turbine community a better handle on optimized design, operation and maintenance. Gas turbines came into
11
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
their own in the Second World War In peacetime; NASA took over the research that led to better alloys, components, and design techniques. This technology was then handed down to military aviation, and eventually commercial aviation. However, since the same manufacturers also make gas turbines for land and marine use, aeroderivative gas turbines were a natural offshoot of their flying forerunners. However, the same manufacturers also make gas turbines for land and marine use. So aeroderivative gas turbines were a natural offshoot of their flying forerunners. Aeroderivative gas turbines are essentially aviation gas turbines that are installed on a light frame and installed on a flat surface (ground based, marine craft or offshore platform). Aeroderivatives are commonly used in power generation service, particularly where a relatively light package is required, such as in offshore service. The Rolls Royce Spey and Olympus engines for instance, are both aero engines but are also popular when packaged as aeroderivatives in land based and offshore platform service. Pratt and Whitney’s (PW) JT- 8D was once the largest (in terms of fleet size) aircraft engine family in existence. The engine first made its appearance in the 1950s and delivered about 10,000 pounds of thrust, then. Several variations on the basic core produced a version that delivered roughly 20,000 pounds of thrust about twenty years later. This incremental power development around the same basic design is common and saves on development costs, spares stocking costs and maintenance. PW’s FT- 8D is their aeroderivative equivalent used in both power generation and mechanical drive application. Similarly General Electric’s (GE’s) LM2500 and LM6000 family (aero derivative) are essentially CF6-80C2 (aero) engines that have been adapted for land based use. What was ABB’s GT35 (land based), then Alstom’s GT35 (change of corporate ownership), then Siemens Westinghouse’s SGT500 (yet another corporate purchase) is another example of an aeroderivative. Most aeroderivatives can also be used in marine (ferry, ship) applications. Some of them are also used on mobile land applications, such as in military tanks. Aero and aeroderivative gas turbine engines are likely to be built in modular construction. This means that one module of the gas turbine engine may be removed from service and the other modules left in place. A substitute module may be inserted in place of the removed module so the gas turbine can resume service. An industrial engine is more likely to be constructed in a non-modular format. If part of an industrial engine has serious problems, it is likely that the entire engine will be “down for maintenance”. The term “industrial” gas turbine implies a heavier frame and a gas turbine model that was not intended for service where the mass (weight) to power ratio (in other words weight minimization for the power plant) was of paramount concern. That said, the metallurgical selections for contemporary industrials reflect the best developments in metallurgical selections. The gas turbine field is a highly competitive one, and the highest turbine inlet temperatures (TITs) that can be tolerated by the metallurgical and fuel selections, are sought as this optimizes the gas turbine’s peak power rating. In other words, GE’s industrial Frame 7s and 9s (be they “- F”, “- G” or “- H” technology) may incorporate similar metallurgy to that used on their aircraft engines. The letters F, G and H refer to temperature ceilings and therefore imply higher power (with “later” alphabet letters). Some turbine model designations can appear confusing due to several changes in corporate ownership. This is partly due to the fact that the OEM (original equipment manufacturer) gas turbine scene changes constantly with corporate mergers, partial mergers, buyouts of specific divisions and joint ventures. This section and the one on combined cycles therefore have several notes about specific engines’ model designation history and previous ownership. This has considerable relevance when it comes to noting the finer points of any gas turbine’s design. This is critical to operators as they can then make better decisions regarding the overhaul, performance optimization, component updates and retrofit systems on their turbine systems. Any application of a gas turbine could have a great deal to offer end-users in other industrial sectors. Power generation is often the least demanding application for a given gas turbine, unless it used in variable load/ peaking service. Mechanical drive units are more likely to experience load swings. One example would be turbines driving pumps that injects (into the soil) varying volumes of sea water that accompany “mixed field” (oil, gas and seawater deposits) oil and gas production. Aircraft engine turbines may see varying stresses depending on their service. If for instance, one considers an aerobatic squadron, one needs to be aware that the engines on the planes trying to stay a fixed distance from the wing tip of the formation’s leader may accumulate life cycle losses of twenty times that of the formation leader’s engines. In other words, the variations in all parameters that pertain to a gas turbine’s overall life, component lives or time between overhauls (TBOs) offer insight to gas turbine operators regardless of whether that turbine operates in “their” industry or not. Lessons which are
12
Claire Soares
learned in one sector of industry on gas turbine metallurgy and operating systems, such as controls or condition monitoring, can be applied in some way, to other gas turbine applications.
The History of the Gas Turbine The development of the gas turbine took place in several countries. Several different schools of thought and contributory designs led up to Frank Whittle’s 1941 gas turbine flight. Despite the fact that NASA’s development budget now trickles down to feed the improvement of flight, land based and marine engines, the world’s first jet engine owed much to early private aircraft engine pioneers and some lower profile land-based developments. The development of the gas turbine is a source of great pride to many engineers world wide and, in some cases takes on either industry sector fervor (for instance the aviation versus land based groups) or claims that are tinged with pride with one’s national roots. People from these various sectors and subsectors can therefore get selective in their reporting. So for understanding the history of the gas turbine, one would have to read several different papers and select material written by personnel from the aviation, and land-based sectors. At that point, one can “fill in the gaps”. What follows therefore are two different accounts of the gas turbine’s development. Neither of them is wrong. The first of these presents an aircraft engine development perspective. * * Reference: “The History of Aircraft Gas Turbine Development in the United States”, St. Peter, J., Published IGTI, ASME, 1999. Attempts to develop gas turbines were first undertaken in the early 1900’s, with pioneering work done in Germany. The most successful early gas turbines were built by Holzwarth, who developed a series of models between 1908 and 1933. The first industrial application of a gas turbine was installed in a steel works in Hamborn, Germany, in 1933. In 1939 a gas turbine was installed in a power plant in Neuchâtel. 1931 U.S. army awards GE a turbine-powered turbosupercharger development contract 1935 U.S. Army, Northrop, TWA, and GE combine to test fly a Northrop Gamma at 37,000 feet from Kansas City to Dayton. This led to a production contract for GE to build 230 units of the “Type B” supercharger and led to establishment of the GE Supercharger Department in Lynn, Massachusetts (later the site of the I-A development based on the Whittle engine). 1938 Wright Aeronautical Corporation designs its own vaned superchargers for its own engines, although the superchargers were manufactured for Wright by GE. 1940 NACA joins with Wright, Allison and P&W to standardize turbo supercharger testing techniques.
1.1 Simple and Combined Cycles 1925 R.E. Lasley of Allis-Chalmers receives the first of several patents on gas turbines. Around 1930 he forms the Lasley Turbine Motor Company in Waukegan, IL. with the goal of producing a gas turbine for aircraft propulsion. 1934 U.S. Army personnel from Wright Field visit Lasley’s shop and inspected his hardware and the engine which he had filmed in operation earlier that year. However, neither the Army nor Navy would fund Lasley. 1939 GE studies gas turbine aircraft propulsion options and concludes the turbojet is preferable to the turboprop. Note, however, that two years later they changed their minds and proposed a turboprop to the Durand Committee. 1941 GE Steam Turbine Division (Schenectady) participates in the Durand Special Committee on Jet Propulsion and proposes a turboprop, designated the TG-100 (later the T31), which ran successfully in May 1943 under Army sponsorship. 1941 GE Turbo Supercharger Division (Lynn, Massachusetts) receives the Whittle W.1.X engine and drawings for the W.2.B improved version. A top secret effort begins to build an improved version, known as the I-A, for flight test in the Bell P-59. 1941 Durand Committee also awards Navy contracts to Allis-Chalmers and Westinghouse. The Westinghouse W19, a small booster turbojet, resulted from this but Allis-Chalmers dropped out of the “gas turbine race” in 1943.
13
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
1942 In April, the GE I-A runs for the first time in a Lynn test cell. In October, it powers the Bell P-59 on its first flight at Muroc Dry Lake, CA. 1929 Haynes Stellite develops Hastelloy alloy for turbine buckets, allowing operation up to gas temperatures of over 1800 F. This superior alloy was later crucial to the successful operation of the I-A and it gave U.S. turbine manufacturers the ability to use uncooled designs rather than include the complexity of blade cooling. By the latter part of 1942, the following “native” aircraft gas turbine efforts were proceeding. These projects included: 1. Northrop Turbodyne turboprop 2 .P&W PT-1 turboprop 3. GE/Schenectady TG-100 turboprop 4. Allis-Chalmers turbine-driven ducted fan 5. NACA piston-driven ducted fan 6. Westinghouse 19A turbojets 7. Turbo Engineering Corporation’s booster-sized turbojet The following timeline contains many of the relevant land based gas turbine design developments. * Note that this also contains some timeline references to aircraft engine development. * Reference: ASME 2001 - GT- 0395 “Advanced gas turbine technology – ABB/ BBC historical firsts” by Eckardt, D., and Rufli, P., ALSTOM Power Ltd. Note: BBC = Brown Boveri Company, ABB = Asea Brown Boveri
Switzerland (& Swiss Abroad) 1921 J. Ackeret, high-speed aerodynamics scientist at ETH Zurich, arrives at L. Prandtl’s AVA Aerodynamische Versuchs-Anstalt Gottingen; stays seven years 1925 CEM (G. Darrieus) - a French subsidiary of BBC (Brown Boveri Company) produces a series of windmills, using airfoil design theory. 1926 BBC’s 4 stage axial test compressor designed, first with untwisted blades, later swirl adapted. 1932 BBC sold a number of 11 stage axial compressors, PR= 3.4, for the Mondeville project and high-speed windtunnels at ETHZurich and Rome. 1934 C. Keller, assistant to J. Ackeret at ETH Zurich, designed one of the windtunnel blowers (2nd blower for high speed tunnel came from BBC). 1939 A. Meyer, BBC’s Technical Director, presents a comprehensive paper on GT design achievements (including GT usage for compact & lightweight ship/destroyer propulsion) at the Institute of Mechanical Engineering, London • First commercial industrial GT from BBC is operational at Neuchatel • BBC delivers 1st Industrial GT to RAE, 1.6 MW 20 stage axial compressor • In 1940 BBC delivers axial aircraft superchargers, 190 hp, PR=2.5 to complete a RR purchase order
Germany (& Germans Abroad) 1922 W. Bauersfeld suggests the use of airfoil theory for fluid machinery 1935 At AVA Gottingen, a 4 stage axial turbocharger, 7 stage compressor design undergoes development (Encke et al. design), PR=3.8 [in production PR=3.1]. 1935 H.P. von Ohain gets a secret turbo-engine patent no.317/38 1937 H.P. von Ohain’s test engine HeS313 runs 1939 The first jet-powered flight He 178 aircraft with HeS313, on Sunday Aug 27, 1939.
14
Claire Soares
• R. Friedrich, Junkers Magdeburg, design the 14 st. axial compressor for the “RTO” engine (Riickstoss-Turbine ohne Leistungsabgabe with a Propeller), for the Helium aircraft S30 engine, based on Gottingen airfoil design 1942 Me 262 fighter aircraft entered service with two Jumo 004 engines, first test flight on August 18.
England (& English Abroad) 1926 A.A. Griffith releases “An Aerodynamic Theory of Turbine Design”, which discusses a GT as an aircraft’s power plant 1930 F. Whittle gets the first patent for a turbo aeroengine • Tizard, Gibson & Glauert committee denies that the gas turbine could be superior to the piston engine 1937 F. Whittle, radial compr. engine, test run on December 4 1938 A delegation at BBC decides that “Exclusivity on the BBC (axial) compressor design would not be granted” 1941 The Whittle engine has its first flight Note that in the early 1930s, BBC designed components used for the Velox project boilers. They developed a turbine that had enough power to drive the compressor, and could also generate excess power through the inverse operation of the electric starter motor. Also, in 1936, BBC’s '34MW all-axial process gas turbine/ blower train with a PR = 4, was supplied to a US refinery. In July 1939, BBC commissioned the world’s first utility gas turbine at Neuchatel, Switzerland. The gas turbine had one 23 stage axial compressor, one single-can combustor, one 7 stage axial turbine, and a synchronously operated generator on the same shaft.
Fig. 19. BBC - First Utility GT Power Plant, 4 MW, Neuchatel, Switzerland, 1939 (courtesy Alstom Power)
15
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Gas Turbine Major Components, Modules, and Basic Systems
Fig. 20. Modules in a gas turbine Source: Courtesy of Butterworth nd Heinemann, from “Process Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce UK
Primary Modules The primary modules in a gas turbine are the: • Compressor module • Combustion module • Turbine module A gas turbine also has an inlet section/ module and an exhaust section/ module. See Figure 20. Most advanced and large gas turbines have compressors that are the axial design type. Some of the earlier, smaller or deliberately compact gas turbines have centrifugal compressors. See Figures 21, 22, 23. Each compressor stage provides an opportunity for stepping up the overall compressor pressure ratio (PR), so although an axial stage may not offer as much of a PR as a centrifugal stage of the same diameter, a multistage axial compressor offers far higher PR (and therefore mass flow rates and resultant power) than a centrifugal design.
16
Claire Soares
Fig. 21. Centrifugal-compressor flow, pressure, and velocity changes (a) Airflow through a typical centrifugal compressor, (b) Pressure and velocity changes through a centrifugal compressor. Courtesy Rolls Royce UK
Fig. 22. A modern high-performance compressor assembly. (General Electric) (Source: “Aircraft Gas Turbine Engine Technology” McGraw Hill)
17
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 23. Stator case for the General Electric 179 engine. (Source: “Aircraft Gas Turbine Engine Technology” McGraw Hill)
Therefore most gas turbine designs incorporate axial compressors. In newer designs, the compressor and turbine modules may be split into further submodules, to lessen the stress on individual components and achieve better efficiencies. So a compressor may have a low pressure (LP) module or LPC, and a high pressure module (HPC). In this case there will be equivalent high and low pressure turbine modules (LPT and HPT). The LPC and LPT will operate on one long shaft at the same speed. The HPC and HPT will operate on a shorter shaft that fits around and concentric to, the low pressure shaft, and at a higher speed than the low pressure module. See Figure 21. Some contemporary gas turbines have three modules, designated low, intermediate and high pressure, each with their own shaft. This modular concept allows for module replacement or exchange, if maintenance to a module is required, without taking the entire gas turbine out of service.
Module component aerodynamic and thermodynamic basics Air inlet section A gas turbine takes in many multiples of what an equivalent size reciprocating engine can. The air inlet is generally a smooth, bell shaped, aluminum alloy duct. It leads air into the compressor with minimized turbulence. Typically, struts brace the outer shell of the front frame to minimize air flow vibration. An anti-icing system directs compressor air (at discharge or some pressure higher than atmospheric) that is bled off an appropriate compressor stage, into these struts. The temperature of this air prevents ice formation. Ice ingestion can and has destroyed many gas turbine engines. Compressor module The compressor is made up of rotating blades on discs and stationary vanes that direct the air to the next row of blades. The first stage compressor rotor blades accelerate the air towards their trailing edges and towards the first stage vanes. The first stage vanes slow the air down and direct it towards the second stage compressor rotor blades, and so on through the compressor rotor stages (each stage is one rotating stage and one stationary stage).
18
Claire Soares
Then air enters the diffuser section. The highest total air velocity and maximum compressor pressure is at the inlet of the diffuser. Air moves through the diffuser, which presents the air with an increasing cross sectional area, so the air’s velocity decreases and the static pressure increases. The highest static pressure is at the diffuser outlet. The compressor rotor can be described as an air swallower. The volume of air swallowed by the compressor rotor is proportional to the lowest pressure (in a multiple shaft gas turbine) rotor rpm. However, the altitude at which the gas turbine is located will alter the horsepower (for a mechanical drive) or the power (in watts, kilowatts or Megawatts) that the gas turbine develops. This is because air density decreases with altitude and with increasing air temperature and humidity. That means when the compressor swallows a certain volume of air, that air will be a smaller weight of air, if the gas turbine is at a high altitude, still less if it is a hot day, and still less if it is also a humid day. This smaller weight of air requires a smaller weight of fuel to combine with, and the mixture then produces less power when burned. Note however, that humidity, in comparison with temperature, and pressure altitude, has a much smaller effect on density. In aircraft engine applications, with increased forward speed, ram air pressure increases and air temperature and pressure increase. Ram air pressure is defined as the free stream air pressure created by the forward motion of the aircraft engine. The effect of rise in air intake temperature on power developed by a gas turbine can be noted in figure 24.
Fig. 24. Variation of shaft power with inlet air temperature for different configurations of the Rolls Royce Avon nd Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2 edition, Bloch, H. and Soares C., 1998
There are many different compressor designs that result from the manufacturer’s balance of several design factors, including target gas turbine power developed, cost of manufacture, anticipated serviceability factors and so forth. As previously discussed, when the gas turbine is started up, the turbine section will keep the compressor section rotating. The compressor’s efficiency is a key factor in
19
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
determining the power necessary to create the pressure rise of a given airflow. This pressure rise will in turn affect the temperature difference between the compressor inlet and outlet. As mentioned previously, the main types of compressor design are centrifugal and axial flow. The axial-centrifugal-flow compressor is a combination of both and operates with a combination of their characteristics. It is a less common design.
Centrifugal-flow compressor As the rotor turns, air is drawn into the blades near the center of the front rotor stage. Centrifugal force accelerates this air as it moves outward from the axis of rotation towards the edge of the rotor. It is then forced through the diffuser section at high velocity (high kinetic energy). A pressure rise results when the air slows in the diffuser (some velocity energy becomes pressure energy). One centrifugal compressor stage is capable of a relatively high compression ratio per stage. It is not practical to use on larger engines because of its size and weight, relative to axial stages. Because of the high tip speeds it develops, the centrifugal compressor is most used on smaller engines where simplicity, flexibility of operation, and ruggedness outweigh its characteristics of less overall pressure ratio than that developed by an axial compressor. Axial-flow compressor The air is compressed, in a direction parallel to the longitudinal axis of the engine. Axial flow compressors consist of several stages that collectively create high compression ratios with high efficiencies. The streamlined shape of this type of compressor make is suitable for use on high speed (ram jet) aircraft. Its design is less rugged than that of the centrifugal compressor though, making it more susceptible to foreign object damage (FOD). The required efficiency and power rating then mean that the design parameters that govern its design, such as rotor dynamics characteristics, clearances and fits, also make it more expensive to manufacture. With the rising cost of fuel, most gas turbine designers use axial compressors, as features such as power delivered per unit weight of the gas turbine outweigh initial manufacture costs. Axial-centrifugal-flow compressor The axial-centrifugal-flow compressor, also called the dual compressor, is a combination of the two types. Its operating advantages and characteristics are also a combination of both rotor types. It is useful is specialized application designs, such as those for US Army helicopters. Typically the compressor is five- to seven-stage axial-flow compressor and one centrifugal-flow compressor. The compressors are mounted on the same shaft and therefore turn in the same direction and at the same speed. The centrifugal compressor is situated aft of the axial compressor stages. Most high performance gas turbines today also have inlet guide vanes (IGVs) and/ or variable inlet guide vanes (VIGVs) at the compressor inlet. This is to ensure that the air flow hitting the rotor blades does so at an acceptable angle of attack that does not cause the blade to stall. If we consider a cross section through the wing of an aircraft, we note that the section is similar in shape (if not size) to that of an airfoil in a gas turbine. All airfoils provide lift by producing a lower pressure on the convex (suction) side of the airfoil than on the concave (pressure) side. With any airfoil, lift increases with an increasing angle of attack, but only up to a critical angle. Beyond this critical angle of attack, lift falls off rapidly. This is due mostly to the separation of the airflow from the suction surface of the airfoil. In simpler terms, we know that when the cushion of air under the aircraft wing is reduced to a certain level, the wing has inadequate lift. It (and the aircraft) tend to drop from their existing level. The airfoils in a gas turbine can stall in exactly the same way, one blade at a time. If a whole row of blades stalls, we have a condition called rotating stall, at which point surge occurs. Surge causes a rotor to go back on itself, in an attempt to regain the lift under the airfoil. In flight, the pilot then pushes the nose down to recover from stall, as this then restores the air cushion under the wing.
20
Claire Soares
Combustor module There are three main combustion chamber types in use today (See figures 25-29): • annular combustor chambers • can (multican) combustor chambers • can-annular combustor chambers
Fig. 25. A combustion chamber. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd Edition, Bloch, H. and Soares, C., 1998
Fig. 26. Flame stabilizing and general airflow pattern. Source: Courtesy of Butterworth Heinemann, from nd “Process Plant Machinery” 2 Edition, Bloch, H. and Soares, C., 1998,
21
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 27. Flame tube cooling methods Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd Edition, Bloch, H. and Soares, C., 1998
Fig. 28. Multiple Combustion Chambers Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd Edition, Bloch, H. and Soares, C., 1998
22
Claire Soares
Fig. 29. Annual Combustion Chambers Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd Edition, Bloch, H. and Soares, C., 1998
Some variations on these basic designs occur in specialized applications. Again, one example is US Army helicopters that use the annular reverse-row type. The combustor module contains the combustion chambers, igniter plugs, and fuel nozzles. The combustor burns a fuel-air mixture and delivers the products of combustion to the turbine at temperatures within design range. Fuel is injected at the upstream end of the burner in a highly atomized spray. Fuel nozzles may be simplex type (delivering gaseous fuel or liquid fuel) or they may be designed to be dual fuel (delivering gas or liquid at different times in the operation). Some gas turbines are “bi-fuel”. They may burn a mixture of gas and liquid fuel. 1.1 Simple and Combined Cycles Combustion air, with the help of swirler vanes, flows in around the fuel nozzle and mixes with the fuel. This air is called primary air and represents approximately 25 percent of total air ingested by the engine. The fuel-air mixture by weight is roughly 15 parts of air to 1 part of fuel. The remaining 75 percent of the air is used to form an air blanket around the burning gases and to lower the temperature. Flame temperatures in excess of 3600° F (roughly 200 degrees C) are not uncommon in high performance aircraft engines. Cooling air drops this temperature to a value that the turbine inlet guide vanes can withstand. The air used for burning fuel is “primary” air. Cooling air is “secondary” air and is controlled and directed by holes and louvers in the combustion chamber liner. Certain aircraft engines are termed high bypass ratio fan engines. With this design, the gas turbine has an inlet fan upstream of the low pressure compressor (LPC). That fan’s diameter is far larger than that of the LPC. Much of the air ingested by the fan is directed through an annular sleeve type casing that fits around the compressor. This bypass air provides still more cooling but also helps with other gas turbine performance characteristics like power developed, total mass flow and FOD ingestion capabilities.
23
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
As we read previously, igniter plugs function during start up and are cut out of the circuit as soon as combustion is self supporting (the turbine has developed design speed and is driving the compressor on its own). On engine shutdown, or, if start failure occurs, the combustion chamber drain pressure-actuated valve, automatically drains any raw fuel from the combustion chamber. The material suitable for fabricating the combustion chamber liner is typically welded high-nickel steel. The hottest zone is about the first upstream third of its length (flame zone). The most severe operating periods in combustion chambers are during engine idle (reduced air) and maximum rpm (power) operation. Sustained operation is generally unnecessary. “Base load” with a ground based gas turbine is generally a power setting lower than this value. With aircraft engine operation, maximum rpm generally corresponds to maximum (take off) thrust. The annular combustion chamber can enhance a geometrically compact design. Instead of individual combustion chamber cans, compressed air is introduced into an annular space formed by a chamber liner that may be situated in some designs, around the turbine assembly. Annular space left between the outer liner wall and the combustion chamber housing conducts the flow of compressor secondary cooling air. Primary air is mixed with the fuel for combustion. Secondary (cooling) air reduces the temperature of the hot gases seen by the turbine first stage inlet nozzle guide vanes (IGVs). An annular combustion chamber provides a larger combustion volume per unit of exposed metal area and therefore of metal weight. The can combustion chamber design has individual combustion chambers. Air from the compressor enters each individual chamber through a transition section. Each individual can has two cylindrical tubes, concentric in most locations, the combustion chamber liner and the outer combustion chamber. Combustion occurs within the inner liner. Louvers and holes control airflow into the combustion area. Continuous airflow helps prevent carbon from forming on the inside of the liner. Carbon deposits can cause hot spots or block cooling air passages, which then shortens burner life. Ignition occurs during the start cycle. The igniter plug(s) is (are) located in the combustion liner adjacent to the start fuel nozzle. Two is a typical number. The flame lights off in the can closest the igniter and cross tubes rapidly conduct the flame to the other combustion cans. Some engines use a single can combustor. In the case of the illustration below, because of the size of the single can, the design is referred to as a “silo” burner. This design can be vulnerable to one or more of the heat shield tiles that line the inside of the silo, breaking loose and potentially proceeding downstream into the turbine’s gas path.
Fig. 30. GE 9H partial combustion module during manufacture (Source: GE Power Systems)
24
Claire Soares
Can-annular combustion chamber This combustion chamber is a combination of both annular and can-type designs. The can-annular combustion chamber consists of an outer shell (annular), with a number of cans (can type) mounted about the engine axis. The combustion chambers are cooled by air that enters the liners through various holes and louvers. This air is mixed with fuel from the fuel nozzles. The fuel-air mixture is ignited by igniter plugs, and the flame is then carried through the crossover tubes to the remaining liners. The inner combustion chamber casing serves as structural support and a heat shield. Bearing oil supply lines run through it. Low NOx combustors As previously noted, raising the temperature at which the combustion gases enter the turbine (turbine inlet temperature or TIT), will also raise the efficiency of the gas turbine cycle. Care has to be taken however, that an increase in TIT does not cause other operational problems, such as overheating of turbine components and turbine lubrication oil. If the TIT increase is not accompanied with sufficient additional cooling, this could happen. Also, one needs to consider that the amount of oxides of nitrogen (NOx) produced by a combustor increases with the value of the flame temperature in the combustor and the corresponding value of TIT. NOx emissions contribute to acid rain and legislation against NOx production has become increasingly stringent. Hence lower TITs, to the extent permitted by optimized efficiency, are desirable. Fig. 31. DLN (dry low NOx) combustor
This fact needs to be kept in focus when selecting and/or specifying gas (Source: GE Power Systems) turbines for particular applications and specific demographics (i.e. country or state concerned and their particular legislation). Two examples of low NOx combustors are shown in Figures 33 and 34.
Fig. 32. The SGT-600 dry, low-emission (DLE) combustion system Source: Siemens Westinghouse
25
The SGT-600 fleet has clocked up one and a half million operating hours with its dry, low-emission (DLE) combustion system, which significantly reduces environmental impact. The DLE combustion system was developed for the SGT-600 in 1990 (original design developed by ABB, later Alstom Power, then acquired by Siemens Westinghouse). The SGT-600 burner lowers NOx by reducing the flame temperature in its combustion chamber. The SGT-600 annular combustor has a total of 18 burners. Each burner consists of a cone split in two halves, which are slightly offset to form two slots for the combustion air to enter (original Alstom designation was the “EV” burner). The main gas supply also enters through these slots, via tubes fitted along them. Primary fuel is injected at the tip of the cone. This results in a richer fuel mixture, enabling a control feature to stabilize the flame over a range of load conditions. Further combustion control can be provided by means of an optional bypass system that allows the amount of dilution air to be varied. The current design achieves NOx emission levels of less than 25 ppmv (at 15% O2), operating on natural gas in 50-100% load range. Singledigit NOx levels have been measured in some plants. The DLE system for the SGT-600 has been operating successfully in a variety of applications, including mechanical drives for pipeline and gas storage compressors; cogeneration for industrial duty as well as municipal district-heating systems; and power generation, in both combined-cycle and simple-cycle operation. Installations cover a range of environments, including offshore, from arctic to tropical, at altitudes of up to 1500 meters. Although DLE technology is suitable for dual-fuel combustion, water injection is required to reduce NOx emissions when burning liquid fuel. Emission levels for operation on liquid fuel are below 42 ppmv, at full load, with a modest water-to-fuel ratio of 0.8.
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Low NOx combustors are designed, optimized and promoted extensively for both performance and profit-based reasons. The extent of the profit they represent varies with the demographics of the location in question. Specifically: 1. In the U.S., as flameless combustor designers are quick to point out, their ultra low single digit NOx designs succeed in getting their operators legally permitted (to commence power production) in some cases a few months ahead of their rivals, who may have quite respectable NOx levels ranging from 9 to 15 ppm. This represents a considerable amount of revenue. 2. Both the US and Canada deal with emissions trading. Regardless of any opinion on the technical wisdom of such measures with respect to the overall atmospheric load, low NOx abilities represents revenue to an operator who can then sell his “spare” credits. 3. In Scandinavian countries, operators pay taxes per unit weight of NOx and SOx emissions. This source of revenue method may spread through the western world. 4. Low NOx means that other emissions such as CO and CO2 are also lowered. CO2 taxes may soon be reality in global, particularly western world terms. In this aspect, once again Scandinavian countries point the way for other operators. 5. End users may also note that reduced NOx generally means lower TITs, hence reduced wear on hot section components and therefore reduced costs per fired hour.
Fig. 33. A GE LM6000 partially assembled. (Source: GE Power Systems)
Turbine module The kinetic energy of the gases entering the turbine is transformed into shaft horsepower (see Figures 34 through 38) which is then used to drive the compressor and other support systems (via accessory system gears. Note that this turbine, combustor and compressor modules form an assembly that is termed the “gas generator”. In power generation applications, the entire gas turbine is a gas generator that is then mechanically coupled either directly or via a gear box, to the generator that in turn is coupled to the grid or power supply system. However, in land based mechanical drive applications, we read earlier that a free power turbine rotates downstream of the gas generator at the turbine end and that it is on a different shaft system (with or without a gear box) together with the machinery (typically compressors or pumps) it turns. Aviation turboprop or helicopter applications have a transmission system (gearbox) that may be located at the compressor end of the gas turbine, to conduct torque to the propellers or helicopter rotors. The main turbine airfoil design type used in gas turbines today is axial flow design. Some manufacturers however, use a radial inflow design. The radial inflow turbine is rugged, less complex, less expensive and easier to manufacture than the axial-flow turbine. This radial flow turbine design is a “backwards” version of the centrifugal flow compressor. Similarly, radial turbine rotors used in small engines have a high efficiency relative to their weight and the space they occupy.
26
Claire Soares
The axial flow turbine consists of stages, each made up primarily of a set of stationary vanes followed by a row of rotating blades, also on a disc. Turbine blades are either impulse or reaction type. Typically modern aircraft gas turbine blades have both impulse and reaction sections.
Fig. 34. Comparison Between a pure impulse turbine and an impulse/reaction turbine. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Fig. 36. Gas flow pattern through nozzle and blade. Source: Courtesy of Butterworth Heinemann, from nd “Process Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
27
Fig. 35. A typical turbine blade showing twisted contour. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Fig. 37. Typical nozzle guide vanes showing their shape and location. Source: Courtesy of Butterworth Heinemann, from “Process Plant nd Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 38. Various methods of attaching blades to turbine disks. Source: Courtesy of Butterworth Heinemann, nd from “Process Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
The stationary part of the turbine assembly consists of a row of contoured vanes set at a predetermined angle to form a series of small nozzles which direct the gases onto the blades of the turbine rotor. For this reason, the stationary vane assembly is usually called the turbine nozzle, and the vanes are called nozzle guide vanes.
Exhaust module The gas turbine’s hot gases exit via the exhaust section or module. Structurally, this section supports the power turbine and rear end of the rotor shaft. The exhaust case typically has an inner and outer housing. Hollow struts locate its position. The inner housing typically has a cone shape or cover that encloses a chamber for cooling the thrust bearing at the end of the shaft. When we consider aircraft engine applications, we note that turboshaft engines (such as those used in helicopters) do not develop thrust with the use of the exhaust duct, as they must be capable of stationary hover. So helicopters use divergent ducts that dissipate energy in exhaust gases. On fixed wing aircraft, the exhaust duct could be convergent in design. That would accelerate exhaust gases and produce thrust which adds additional power to the engine. Combined thrust and shaft horsepower give equivalent shaft horsepower (ESHP).
Other Gas Turbine Systems Cooling system Air for cooling the hot sections of the turbine are drawn (bleed air) from various stages in the compressor. Most OEMs prefer to use air only for cooling, even if they have a combined cycle operation and therefore a source of steam derived from water that is boiler feed water quality. Steam cooling can be very effective, both in closed loop and open loop configurations, as some OEMs, such as MHI have proven. If the steam quality stays uniform, deposits will not form on the insides of fine laser drilled cooling holes in the turbine airfoils. However, if there is a divergence in required water/ steam quality, this could prove a problem. Other OEMs, such as Rolls Royce prefer not to worry about this possibility, however remote. They configure their designs so that they only need air cooling. See Figs 39 and 40. For a description on a successful steam cooling design, see the section on Design Development. Air “spent” on cooling will cost the OEM in terms of nominal efficiency, so some designers are less keen to spend more than they absolutely have to. In power generation machinery that generally operates at base load, this may not be of major concern. It is, in applications with severe swings in load and/ or speed.
28
Claire Soares
Fig. 39. Air system flow in the turbine. Source: Courtesy of Butterworth Heinemann, from “Process nd Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Bearing and lubrication system Basically, sleeve bearings locate the turbine modules concentrically around the shaft(s) during operation and when the turbine is not running. They provide the rotor with support. The thrust developed by the overall rotor is absorbed by thrust bearings at the end of the rotor. One arrangement of key bearing positions is shown in figure 40.
Fig. 40. Main internal air system flows. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Oil flow to the bearings is regulated. The bearings in the hot section require far more oil flow than those in the cooler compressor section. Thermocouples or RTDs measure oil flow temperature. Sudden temperature rises in the oil trigger an alarm or shutdown. The section on design development refers to varying design philosophies between OEMs. The lubrication system is one area where it shows as much as anywhere else. Certain OEMs have a preference for greater lubrication flows than others, at a given temperature range.
29
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fuel system As we will see in the section on design development, the gas turbine can run on a very wide variety of fuels that are gaseous, liquid, atomized solid (coal) suspended in gas, or semi-solid (biomass waste liquor). Each of these different fuel types requires its own customized fuel delivery systems with varying combustor residence times. However, most OEMs have standard fuel systems for natural gas, liquid fuel (such as LNG or diesel), dual fuel (gas or liquid) and in the case of manufacturers such as Rolls Royce bi-fuel (both gas and liquid at the same time). See figures 41 through 45. The basic principle of most gas turbines power development revolves around “temperature topping”. A fuel control unit, which in earlier gas turbines is a mechanical device with several cams and contours, controls the fuel flow. In newer engines the control is more like an electronic brain where electronic functions take the place of the mechanical cams. In its simplest form, temperature topping works as follows. Exhaust gas temperature readings tell the turbine’s control system whether the gas turbine needs to be hotter or cooler, for a given operational requirement. That reading then is compared to the fuel flow set point and that set point raised or lowered, as required. Other systems essential to the gas turbine’s operation but not covered in this summary, include: • Compressor wash system (on-line and / or offline) • Engine Condition Monitoring System which incorporates subsystems such as Vibration Analysis, Pulsation Monitoring, and Life Cycle Assessment. • Fuel treatment (see case study in Design Development section)
Fig. 41. Liquid fuelled phase 1 gas generator schematic. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
30
Claire Soares
Fig. 42. Dual fuelled gas generator schematic. Source: Courtesy of Butterworth Heinemann, from “Process Plant Machinery” 2nd edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Fig. 43. Gas fuel burner. Source: Courtesy of Butterworth nd Heinemann, from “Process Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
31
Fig. 44. Liquid fuel burner. Source: Courtesy of Butterworth nd Heinemann, from “Process Plant Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 45. Dual fuel burner. Source: Courtesy of Butterworth Heinemann, from “Process Plant nd Machinery” 2 edition, Bloch, H. and Soares, C., 1998, original source Rolls Royce
Design Development with Gas Turbines 1,2 This section deals with some of the predominant trends in gas turbine and gas turbine system projects and design development.
Maximizing Rotor Component Commonality The field of gas turbine technology increases in sophistication daily. Every manufacturer has a unique design philosophy. Primarily, design development work concentrates on improving the core of already established designs. The market entry of a totally new gas turbine model with a substantially different core, represents a major capital investment and is usually only done if there is a substantial gap in that original equipment manufacturer’s (OEM’s) product line that the specific OEM intends to cover (See Figure 46 below. >>. Even then, an OEM takes this step only if potential revenues from the new turbine justify the development funds.
1
“Gas Turbines: An applications handbook for land, air and sea”, Soares, C. publisher Butterworth Heinemann. Notes from the annual panel session “Engine Condition Monitoring used to extend the life of gas turbine engine components”, 1995 through 2003, Chair: Soares, C.
2
32
Claire Soares
Fig. 46. Siemens gas turbine power plants range from 65 MW to 814 MW (simple cycle and combined cycle. Power output according to applied turbine and plant type. (Source: Siemens Westinghouse)
Several gas turbines have dual frequency capability. A dual frequency power generation package is illustrated in Figure 47.
Siemens’ SGT-200 Industrial Gas Turbine for Power Generation (ISO) 6.75MW (e) Power Generation Package is of light modular construction, 50Hz or 60Hz, and suitable for small power generation, especially in locations where power to weight ratio is important (offshore applications) and small footprint is required. The SGT-200 is available as a factory assembled packaged power plant for utility and industrial power generation applications. It incorporates the gas turbine, gearbox, generator and all systems mounted on a base. The package is available for either multi-point or three-point mounting for onshore or offshore use as required. An option for acoustic treatment reduces noise levels to 80dB (A) and is available in carbon steel and stainless steel. Doors and panels are incorporated to provide access for servicing. Fig. 47. SGT-200 Modular Package - Generator Set. (Source: Siemens Westinghouse)
The cases below include an illustration of a similar rotor [Mitsubishi Heavy Industries, MHI]5, shape using different metallurgy, aerodynamic (including bleed air modifications) or cooling techniques to increase the power developed by that rotor. To increase compressor discharge pressure (and therefore mass of compressor air delivered and in turn power developed by the turbine), additional compressor stages can be added to give higher compression ratios. This can be done while leaving the core diameter the same. Note 5: See footnote on page 38
The After-Sales Market Generally, much of an OEM’s revenue is made from the sale of new or reconditioned spare parts. End users exert pressure on OEMs to optimize component designs and thus reduce their operational cost per fired hour. Therefore, design development that perfects component design within economically practical limits and develops repairability strategies, is continual. Design development also aims at offering a larger power range with models that have essentially the same core
33
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
geometry. This is done by optimizing metallurgical selections and improving cooling. The design development process is best illustrated by case studies (see below) drawn from OEM authored papers. OEM strategy with respect to repair development varies, sometimes even within their own divisions. Factors such as end-user group pressure (to develop specific repairs), international economics (end users do not always pay the same rate in dollars per fired hour for “power by the hour” contracts) and other reasons unrelated to the gas turbine system itself. So when OEMs merge or acquire divisions of another OEM, this may prove very beneficial to the end user, if certain technology areas improve. It may also prove a logistical problem with spare parts stocking and changing codes, if the “new” OEM also changes model numbers.
Acquisitions and Different Design Philosophies OEM strategy with respect to repair development varies, sometimes even within that OEM’s own divisions. Factors such as end-user group pressure (to develop specific repairs), international economics (end users do not always pay the same rate in dollars per fired hour for “power by the hour” contracts) and other reasons unrelated to the gas turbine system itself. So when OEMs merge or acquire divisions of another OEM, this may prove very beneficial to the end user, if certain technology areas improve. It may also prove a logistical issue with spare parts stocking and changing codes, if the “new” OEM also changes model numbers. Other facts that affect design development are the continuous acquisitions that occur among gas turbine manufacturers. Totally different design philosophies merge when this happens. Consider for instance Siemens’ acquisition of Westinghouse. The latter’s newest models at the time had strong evidence of design methods that originated with Mitsubishi (MHI) design methods, because of the technology cooperation they had previously had with Westinghouse. Later Siemens acquired a subsidiary of what was Alstom Power (formerly ABB) in Sweden. ABB’s Swedish developed turbines had designs that had been independently developed in Sweden and were not always a scaled version of ABB’s Switzerland designs, although they drew on specialized knowledge that had been developed in Switzerland. At one point ABB Alstom (before the “ABB” was dropped from the name) had acquired what was European Gas Turbines (EGT) which formerly was Ruston, an English manufacturer. Joint ventures from component suppliers’ previous programs tend to add to the technology pool at an OEM’s disposal. Following an acquisition in 2003, the original EGT models, and the former ABB Stal (Sweden) models, are now part of Siemens. Siemens has renamed all of their turbines, including turbines that she had originated, such as the V series (V94.3, V64.3, V84.3 and so forth). End users can benefit if they watch corporate evolution of this nature as it may extend, or reduce, their own constant drive to reduce their costs per fired hour. Fleet size can also impose design development requirements. The larger OEMs, such as General Electric, tend to have several licensees that assemble their gas turbines. Designs that specify assembly methods which promote uniformity in terms of how a gas turbine is assembled save money, but may have to evolve with experience. At times, as with the introduction of the GE Frame 9F in the mid 1990s, a new design can prove vulnerable to inconsistencies in quality control systems between licensees. The 9F fleet went through a period of severe vibration suffered, on an inconsistent basis, by certain members of the fleet; some units were relatively free of this problem. Changes in rotor assembly methodology removed the potential for the compressor stack to be inaccurately assembled. As personnel migrate between countries and different OEMs, design variations tend to follow. The wide chord fan blade was pioneered by Rolls Royce and featured on engines such as their Tay and the IAE joint venture V2500. Several years later, the GE 90 featured a wide chord fan blade, which is constructed and manufactured differently from the Rolls design, but shares its performance characteristics. OEM methodology to solve the same issue may differ. For specific gas turbine plants sold in SE Asia in the early 1990s, both Siemens and Alstom, then ABB, used a silo combustor design. The Siemens design had several fuel nozzles, that had a fuel distribution pattern that reduced NOx levels to the level they had targeted. ABB chose to use a single fuel nozzle for essentially the same NOx level target as Siemens, but used water injection to reduce NOx levels. Later, to give their client base an option that would eliminate the need for boiler feed water quality for water injection; ABB developed a retrofit with multiple fuel nozzles.
34
Claire Soares
As we will see in one of the case studies below, MHI (Mitsubishi) use a combination of air cooling, steam cooling open cycle and steam cooling closed cycle for their hottest airfoils. This then provides their customer base with a wide range of power developed values, all with essentially the same core geometry.
Metallurgy Limits TIT The limiting factor to the maximum power to weight ratio a gas turbine can reach is the metallurgical tolerance of the alloys used in the hot section of the gas turbine. Ceramic coatings on the surfaces of the turbine airfoils can increase the peak temperatures these airfoils can tolerate, however ceramics have brittleness characteristics that have not been totally overcome yet. So the “ruling parameter” is turbine inlet temperature (TIT). TIT in turn is a function of the turbine flame/ firing temperature, compression ratio, mass flow, and centrifugal stress. So these factors limit size and ultimately, efficiency. A rough rule of thumb is that 55°C (100°F) increase in firing temperature gives a 10 to 13 percent power output increase and a 2 to 4 percent efficiency increase. The combustion chambers and the turbine first stage stationary nozzles and blades are therefore the most critical areas of the turbine that determine its power output and efficiency.
Fuel Options Fuel selection also plays a major part in determining cost per fired hour, depending on its physical state and purity level. Natural gas is the most desirable fuel, as it takes least toll of the gas turbine’s component surfaces. Diesel oil (distillate) is a liquid fuel and also takes minimal (if not quite as good as natural gas’) toll of the gas turbine components. However, residual, also called “bunker” or crude oil is a viable fuel. Because of its high salt levels (sodium and potassium based), water washing is required. Also because of its Vanadium content, fuel treatment additives are required. The Vanadium salts that result take the Vanadium “out of solution” and the salts deposit on the surfaces of the turbine blades. The turbine can be washed, typically every 100 to 120 hours, and the salts are then removed. Were it not for the fuel treatment additives, the vanadium compounds that would form would form a hard coating on the turbine blades that could not be removed. For this entire system to work, TITs are kept down below 900 degrees Celsius. That TIT may be valid as base load and therefore part of a design or it may be run “derated” at the appropriate temperature until a “cleaner” fuel can be used.
Aeroderivative versus Industrial Gas Turbines Industrial gas turbine compression ratios are in the order of 16:1 and aeroderivative (like their aeroengine parents) have compression ratios of about 30:1 and higher. About 50 percent of the total turbine power in any gas turbine is used to drive the compressor. Aero (and therefore aeroderivative) gas turbine designs have weight and size limitations depending on their mission profile. The minimized weight feature makes aeroderivatives highly suitable for offshore platform use, both in power generation and mechanical drive applications. Efficiency translates into fuel burn and this is a major and increasingly pertinent selling point. Rival OEMs in a specific engine size category vie for even 0.5% efficiency margin over their rivals. Design features that ultimately affect operator safety such as cooling air mass, are trimmed to the extent possible. Design engineers have been known to fight their management to get the pilots who fly their engines more cooling air, at the cost of efficiency. In severe service, such as aerobatic combat, that small margin of cooling air, can make the difference between the pilot getting home or not, especially if his engine is already severely stressed. War time conditions can and has included factors such as much heavier fuel than the aeroengines were designed for, being used. Such was the case with part of the Pegasus fleet (that power the VSTOL Harrier) during the Falklands war. In that particular case, the fleet survived the heavier fuel well, despite the fact that its TIT was higher than the industrial and marine engines that typically use the heavier fuel. Industrial gas turbines have none of the weight limitations imposed on their aero counterparts. Like the LM2500, General Electric’s (GE’s) 40 MW LM6000 is an aeroderivative based on GE’s CF6-80C2. The LM6000 has 40 percent simple-cycle efficiency and weighs 6 tons. If we consider GE’s Frame 9F, we note an output of about 200 MW with a weight of 400 tons. The contrast in power delivered to mass weight ratios between the aeroderivative and the industrial model is evident. Further the 9F is only about 34 percent efficient. High thermal efficiency (over 40 % on simple cycle and over 60 % on combined cycle are now common values for most new gas turbine systems) contributes to minimizing fuel burn and therefore minimizing environmental emissions. Even if an engine is “officially” an industrial engine, aero technology is likely to, at some point, contributed to its design. For instance, a contributor to V84.3 (Siemens Westinghouse) efficiency is the 15 stage compressor and 3 stage turbine which use aeroengine technology to optimize
35
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
circumferential blade velocities. The turbine aeroengine technology is partially courtesy of Ansaldo, Italy (who contract manufacture turbine sections for Siemens Westinghouse) and therefore from Pratt and Whitney, on whose design technology some elements of Ansaldo’s turbine manufacture are modeled. With a land based turbine, the designer does not have to aim for maximum pressure rise across a stage together with weight minimization (or maximum power per pound of turbine). The priority in this case is maximum efficiency.
Repowering Repowering is a growing trend in Europe and the US. The incentives include adherence to Kyoto objectives. Although emissions taxes are not yet reality worldwide, they are in some European countries. The higher efficiencies available with gas turbine options are another incentive, as they make IPPs far more competitive. A case in point is the Peterhead station in Scotland. The 2 boiler, 2 GE 115 MW Frame 9E station had been designed to operate on heavy fuel oil, LNG, sour gas, and natural gas. In 1998, the decision was made to increase plant capacity with three Siemens V94.3 combined cycle units. The V94.3 is a scaled up version of the V84.3 which can run at both 60 and 50 cycles. The economics of the situation are heavily influenced by the UK’s gas supplies. Thus that station’s efficiency went from 38% to between 50 and 55%. NOx emissions will be reduced by 85 %. Another major reason for repowering is that what was thought to be 60 years worth of natural gas left in global supply terms was “updated” to 70 years plus recently. Evidence from ongoing exploration indicates that this figure will climb. Despite China’s anxiety to use its coal, and the Middle East’s desire to use its residual oil, the trend towards gas turbines burning cleaner fuels will continue as lending agencies increasingly tie their loans up with environmental standards as conditions. However, as gas turbines get better at burning pulverized coal dust, residual fuel and other erosive or corrosive fuels, use of the gas turbine will increase.
Environmental Factors Legislation pressure on environmental emissions has created extensions on OEMs design staff to lower gas turbine emissions, particularly oxides of nitrogen or NOx (See Table 4.). NOx production will tend to increase with higher flame temperatures. NOx control techniques include a variety of techniques, such as adding cooling air, or extending the combustion process with a two stage combustor (see Alstom’s sequential burner design in the case studies), which results in lowered overall maximum combustor temperatures. Unburned hydrocarbons, particularly carbon monoxide or CO, are undesirable. Greenhouse gases, such as carbon dioxide and methane, are also the subject of increasing attention, as they contribute to global warming. NOx emissions (like oxides of sulphur or SOx emissions) are now taxed in a growing number of global locations and carbon dioxide (CO2) tax will soon be widespread. So the design development cases that follow include work on low NOx combustor development and CO2 sequestering projects.
36
Claire Soares
Table 4. Emissions Factors for Utility and Industrial Combustion Systems 3 (based on fuel energy input rather than output, i.e. not taking account of combustion efficiency)
Emissions Source
Emissions Factors (g/GJ energy input) CO2 CO CH4 NO2 N2O
Utility Application Natural gas boilers
56,100
19
0.1
267
n/a
Gas turbine, combined cycle
56,100
32
6.1
187
n/a
Gas turbine, simple cycle
56,100
32
5.9
188
n/a
Residual oil boilers
77,350
15
0/7
201
n/a
Distillate oil boilers
74,050
15
0.03
68
n/a
Municipal solid waste (mass feed)
n/a
98
n/a
140
n/a
Coal, spreader stoker
94,600
121
0.7
326
0.8
Coal, fluidized bed
94,600
n/a
0.6
255
n/a
Coal, pulverized
94,600
14
0.6
857
0.8
Coal, tangentially fired
94,600
14
0.6
330
0.8
Coal, pulverized, wall fired
94,600
14
0.6
461
0.8
Wood-fired boilers
26,260
1,473
18
112
n/a
Coal-fired boilers
94,600
93
2.4
329
n/a
Residual-fired boilers
77,350
15
2.9
161
n/a
Natural gas-fired boilers
56,100
17
1.4
67
n/a
Wood-fired boilers
26,260
1,504
15
115
n/a
Bagasse/agricultural waste boilers
n/a
1,706
n/a
88
n/a
Municipal solid waste, mass burn
n/a
96
n/a
140
n/a
Municipal solid waste, small modular
n/a
19
n/a
139
n/a
Industrial Applications
Figure 48 is another example of a gas turbine and its primary operational data in simple cycle mode. Note NOx ppm value.
Fig 48. Alstom’s GT13E2 (operating data below is for simple cycle) Gas Turbine (Source: Alstom) 3
Figures quoted in “Greenhouse Gas Abatement Investment Project Monitoring and Evaluation Guidelines” World Bank, Global Environment Coordination Division, Early Release Version, June 1994.
37
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
GT13E2 Fuel Frequency Gross Electrical output Gross Electrical efficiency Gross Heat rate Turbine speed Compressor pressure ratio Exhaust gas flow Exhaust gas temperature NOx emissions (corr.to 15% O2, dry)
Natural gas 50 Hz 172.2 MW 36.4% 9376 Btu/kWh 3000 rpm 15.4 :1 537 kg/s 522 °C < 25 vppm
The following cases are typical of, but cannot fully represent, the end results of contemporary OEM design development projects. The author’s handbook covers several dozen, and they cannot all be condensed or repeated here. These cases feature some OEM methods of maximizing operational convenience and efficiency, while staying within legislative environmental guidelines. They also demonstrate how end-user requirements may shape the course of design development and can moderate an OEM’s focus. Case 1. Gas turbine system features that allow the use of residual oil as a fuel. 4 Case 2. MHI steam cooling design for their highest temperature zones. The steam circuit can be either a closed or an open system type. In the latter case, the steam is released into the gas path of the products of combustion after it has completed its cooling task. 5 Case 3. The use of low BTU “waste liquid” fuel. This case involved what was originally designated Alstom’s GT-10 gas turbine model at the Petrochemical Corporation of Singapore (PCS) plant, Singapore. Highly pertinent to the operation was the use of a “stepper” valve in the fuel system supply. 6 Case 4. Cycle modifications, involving water injection for power augmentation, to boost gas turbine performance.
7
Case 1: Gas turbine system features that allow the use of residual oil as a fuel.4 Mixed fields (that produce both gas and oil) often want to use their oil as fuel. These mixed fields are common in many areas of the world including the offshore fields in Malaysia and the North Sea in Europe. The answer for some owners, who have a grade of oil that is better than residual oil, is to use that as fuel for reciprocating engines that burn crude oil for pipeline mechanical drives. The penalty for using this fuel in gas turbine power generation however, must be carefully weighed for the individual model in question. With or without special design features, gas turbines designed for a (high grade) liquid fuel burn capability, can burn any liquid fuel with a consequential penalty in parts life. It can be done for emergencies as NATO studies for contingency measures in wartime conditions proved. However, gas turbines with oil fuel as an option (to gas), are increasingly popular in many areas of the world. If they can burn residual fuel, they are still more popular. The world, China included, can cheaply import the Middle East’s glut of residual oil. Light oil: There are some gas turbines that can run on light oil with very little penalty in performance versus natural gas. Consider the following data on what was the Alstom GT10, which burns both gas and oil.
4
The author’s Power Generation course notes, extracts from the author’s articles for Asian Electricity and Modern Power Systems and extracts from her book “Environmental Technology and Economics”, (publisher Butterworth Heinemann), on the installation of Alstom (formerly ABB) 13-Ds at the Shunde power plant in Guangdong province, China. Note that the 13-D application range is now fulfilled by their 11N-2 model, primarily a 60Hz model, which also serves the 50Hz market with inclusion of a gear box 5 “Cooling steam application in industrial gas turbines and field experience”, Kallianpur V., et al, Mitsubishi Power Systems 6 Author’s notes, Power Generation Systems course and author’s articles in European Power News, Middle East Electricity and Independent Power Generation magazines 7 The power of water in gas turbines: “Alstom’s experience with inlet air cooling”, Lecheler S., et al (Alstom power)
38
Claire Soares
In British units:
Light oil
Natural gas
Power (hp)
31,641
33,022
Thermal efficiency
33.1
34.2
Heat rate (BTU/hp-hr)
7,685
7,440
Exhaust gas temperature (degrees Fahrenheit)
998
993
In Metric units: Light oil
Natural gas
Power (MW)
23,100
24,630
Thermal efficiency
As above
Heat rate (kJ/KW-hr)
10,880
10,518
Exhaust gas temperature (degrees Centigrade)
537
534
This option of running on oil versus natural gas is also available for newer, more sophisticated gas turbine models, such as Alstom’s 13E2 which powers several SE Asian plant locations. The choice of using oil as a gas turbine fuel is normally decided on the answers to three questions: i) What will the efficiency penalty be? ii) What will the TBO (time between overhauls) and parts longevity penalties be? iii) Which fuel is inexpensively and abundantly available? The answer to ii) is probably the more critical one to operators in terms of their cost per fired hour figures. Some OEMs (original engine manufacturers) therefore have a separate design to minimize the impact on ii) if the answer to iii) is “residual oil” (no. 4 or no. 6 oil). China has an in-country steam turbine manufacturer, with coal reserves that outweigh its oil or gas resources, so gas turbine (or combined cycle, CC) territory within China is hard won. A CC operation powered by residual fuel is a design and operations achievement, due to hot section and fuel additive technology required. The ideal turbine for this application is a relatively low temperature, sturdy, preferably cast, simple design that then results in minimal maintenance. What was the 50Hz Alstom GT13D (and their 60Hz 11N2, which, with a gearbox, can replace the 13D) has a proven track record in these applications where far greater turbine sophistication with respect to alloys and turbine inlet temperatures would be self-defeating. These machines’ track record thus far indicates that operations have been satisfactory to the owners and could indicate further such inroads into a difficult market. China needs to run on as cheap a fuel as possible with maximum efficiency and time between overhauls. Production economics dictated that the -11N2 replace the -13D. They were very similar: the -11N2 package was adapted, so it could be substituted for the earlier model. The -11N2 can run on 50Hz or 60 Hz, produces about 109 MW at base load, and can handle the same dismal fuel quality as the -13D. The observations made in this case involve the Shunde power plant in Guangdong province China which uses Alstom residual fuel technology in its 13D2s. The GT13D gas turbine operates under the critical firing temperature of 1015 degrees C (Celsius) without much derating. Each turbine develops about 90MW. 18 compressor stages and 5 turbine stages are lightly loaded, at a 43.6% gross combined cycle (LHV) efficiency, for longer time between overhauls. (The -11N2 was previously equipped only for 60 Hz generation. With an optional gearbox, it can also run at 50 Hz, and the -13D was phased out of production)
39
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Residual oil as a fuel is not possible without specialized gas turbine design features. Corrosion, plugging and fouling will occur. Higher firing temperatures in most contemporary high performance gas turbines require complex blade cooling, expensive super alloys and substantial derating. The -13D has integrally cast blade and vane cooling passages, with relatively simple geometry (versus a high performance aerofoil which normally has laser produced cooling passages) and a large flow cross section. This provides better resistance against plugging. Cooling air is extracted after the last compression stage, at the blade root. The air is routed to the first stage turbine blades below the rotor surface. The single piece welded rotor supported by two bearings is a simple, less vibration prone design. No through bolts are used: another useful maintenance feature. This design has only one silo combustor, a solid cast design. It has one large bore fuel nozzle, which helps avoid clogging and erosion. No air atomization is required, which means no compression air stream is required. The nature of the burner design means that water injection is required. At Shunde, water injection is 1.3 times the fuel flow rate (maximum 10.5litres/s). Water injection adds 9 to 10MW of power. No flow divider is required in this design, so no consequential temperature unbalance is observed. This also helps cut down on maintenance costs. The generator is driven from the cold end, which means turbine exhaust end inspections are easier. All bearings are accessible without disassembly and no elbow conduits are required. As the generator is air cooled, no hydrogen system or hazards have to be allowed for. The cooling loop is closed and maintenance free. The boiler, a vertical assisted circulation, single pressure design type, has a preheating loop. It delivers 44kg/s of 37.5 bar steam at 475 degrees C. Sodium phosphate (Na2SO4) is used for anticorrosion measures in steam treatment. Although the primary focus for this case is gas turbine system design modifications, these gas turbines are part of a combined cycle operation. (The steam turbine is a single cylinder design with a single flow low pressure section. Its gross output is 92MW. The steam turbine at Shunde runs with 472 degree C steam (480 degrees maximum) at 36 bars. The exhaust is condensed. Total gross power output then is 280 MW nominal. At Shunde 273 MW is guaranteed. The gross efficiency (LHV) at Shunde is 43% (43.8% nominal), based on a guaranteed heat rate of 8376 kJ/kWh (8221 nominal). Slow roll to running speed with the gas turbine takes 5 minutes. Getting the steam turbine running takes approximately two hours). Combustion and fuel economics are as follows. Sodium (Na), Sulfur (S), and Vanadium (V) content in the fuel are the major problems. Na is removed by mixing preheated fuel with water and demulsifier and then centrifuging. Potassium (K) impurities are removed in the same manner and at the same time as the sodium down to 0.5 ppm total (for both the Na and K). The sulphur left in the fuel becomes SOx upon combustion. The 120 meter stack at Shunde provides dispersal for the SOx. In areas where legislated SOx limits are tighter, flue gas desulphurisation or other methods can be used. Magnesium additives combine with the vanadium to form salts that deposit onto the blade surfaces. When the turbine is shut down, the salt levels fall off with the drop in temperature. Remaining salts are washed off with plain water. In Shunde, the wash is done every 100 operating hours for heavy oil. If gas or diesel fuel (back up fuel) is used, no wash is required. For inspection of the hot gas path, the inspector visually inspects the tiles on the inside of the combustor, the transition piece, and the first stage vanes. He uses a mirror to check the first stage blades. The other turbine and compressor stages can be observed by borescope. For major inspections every 16 to 24,000 hours, the burner is lifted off in one piece. The limit for magnesium addition is 1105 degrees C, as at 1120 degrees C, MgO (magnesium oxide) solidifies to the extent it can only be chiseled off, and V2O5 (vanadium oxide) with its low melting point corrodes. (Both MgO and V2O5 are formed from the safe additive compound after 1120 degrees C). The turbine inlet temperature of the Shunde units is maintained at 990 degrees C. When starting the gas turbines, diesel fuel is used until synchronous speed and then heavy fuel is used. This helps prevent clogging. The turbines are run for 5 minutes on diesel when shutting down. Again this prevents clogged nozzles and ignition problems. The 11N2 can also handle the same rough fuel as the -13D. Peak metal temperatures, internal metallurgy and fuel treatment requirements are all quite similar. The single burner design for this model can get NOx down to 42 ppm with water injection. An EV silo combustor (several fuel nozzles) option is available if the end user has gas or diesel fuel. NOx can then be reduced to 15 ppm when at base load on natural gas. A gas turbine inlet filtration system is also necessary in this location. This particular inlet filtration system has three stages. In the first stage the air flow direction is changed. The second stage consists of mats. The third stage is for fine filtration. The gas turbine compressors are still washed off-line every 300 to 400 operating hours. Cheap fuel more than offsets the capital expenditure required for fuel treatment and additives, washing the fuel and other costs. This cost savings increases with the power capacity of a plant. Using a difference in residual oil and diesel prices of $50 per ton, a 300 MW
40
Claire Soares
facility similar in design to Shunde’s could save $22 million at 0.5 capacity factor and $36 million at 0.75 capacity factor. Savings of $264 million and $432 million respectively are indicted over the life of the plant, (US dollar figure expressed at 1995 values). Case 2: MHI steam cooling design for their highest temperature zones.5 The art of steam cooling has proved a valuable asset in the drive to maximize power per unit weight in gas turbine technology. The current limiting factor to maximum horsepower for a given rotor size is turbine inlet temperature (TIT). Internal cooling to the gas turbine vanes and blades, as well as the combustion liner, keeps those airfoils cooler for a given fuel flow rate. The steam cooling circuit can be either a “closed” or an “open” design. In the latter, the steam coolant is allowed to enter the gas path, which provides a further horsepower boost to the gas turbine. The major manufacturers compete with design modifications like steam cooling to produce effective turbines in the various horsepower size categories, “effective” in this context meaning that the turbine in question delivers its rated horsepower (and other deliverables) without leaks or other operational problems. Table 7a shows -D, -F, -G and –H category gas turbine parameters for the Mitsubishi Heavy Industries (MHI) range of gas turbines. These parameters vary for different manufacturers, but the table nevertheless provides an illustration of the effectiveness of steam cooling in raising TITs. For illustrative purposes, this article references parameters with MHI gas turbines. Readers may use this as a template for queries on or comparisons with other manufacturers’ designs. Note also that Table 5 mentions subcategories of the major horsepower size categories. These occur due to individual customer requirements or conditions that “create” a subcategory that can then be offered to other clients. For instance, the G1 is an upgraded G, with cooling steam applied to the blade ring in addition to the combustion liners. Table 5. Categories of gas turbines for the Mitsubishi Gas Turbine product line 5
GT type
TIT
Cooling Type
deg C
Turbine Combustor
Performance (ISO: LHV) Gas turbine
NOx
Combined Cycle
ppm
M501DA
1250
Air
Air
114MW
34.9%
167MW
51.4%
9
M501F
1350
Air
Air
153MW
35.3%
229MW
52.8%
25
M501F3
1400
Air
Air
185MW
37.0%
285MW
57.1%
9
M501G
1500
Air
Steam
254MW
38.7%
371MW
58.0%
25
M501G1
1500
Air
Steam
267MW
39.1%
399MW
58.4%
15
M501H
1500
Steam
Steam
-
-
403MW
60.0%
15
Steam cooling, like any other cooling technology helps alleviate the potential life cycle cost incurred with partial load cycling operation and frequent starts and stops. As of March 2004, MHI had 150,000 operating hours of steam cooling experience with their G units, logged. This figure includes both 50Hz and 60Hz applications. Both their G and H models have steam cooled combustion liners. The H model also has blades and vanes in the first two rows of its turbine rotor and the blade rings, steam cooled. Material selection With steam cooling, as with any design feature, wear limits and future repairability are major concerns. The steam cooling feature merits concern about corrosion rate and electrochemical reaction strength levels, which would depend on the mating materials in question and the steam purity. Although many steam cooling designers would like to claim that the steam supply conditions are no more stringent than the steam required for their steam turbines, higher steam quality standards make good economic sense at the design conditions in G and H gas turbines. Stress corrosion cracking is accelerated by long term steam exposure, particularly at high stress concentration locations like disc dovetails, bolt holes and spigots. MHI were able to use the same low alloy steel as for their F design with their G and H models which gave them a wealth of data. Further, they used scaled up but similar geometry for the hotter models. With respect to scale size after steam exposure, the actual engine tests confirmed earlier laboratory prognoses closely. See Figure 49. In MHI’s design, expensive aircraft engine type alloys such as Inconel (for the rotors) and single crystal castings (for blades and vanes) are avoided. This enhances reliability, initial capital costs and life cycle costs.
41
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 49. Scale size after steam exposure 5
Operation at load With the H model, steam is delivered at about 5 Mpa (megapascals). Maximum steam temperature can reach around 600 degrees C. Load testing in 1999 revealed a leakage point at 60 percent load. A redesigned connector got the model up to full load conditions with no leaks. Active Clearance Controls (ACC) The term ACC was originally coined around aircraft engine design where the cooling medium was air. In this land based application, MHI supply the steam cooling stream to the blade rings for better blade tip clearance at different load conditions. Originally developed for the H model, this feature has also been added to the G model as an upgrade.
Fig. 50. Blade tip Active Clearance Control
42
5
Claire Soares
Closed loop reliability Steam flow is monitored continuously. Three main monitored parameters are linked to the control system via a redundant interlock. See Figure 51.
Fig. 51. Steam cooling continuous monitoring and interlock
5
The interlock allows for both alarm and shut down functions depending on the parameter readings. The three main parameters are (with reference to Figure 51): 1. Cooling steam temperature at the combustion liner outlet, which gives an indication of steam overheating (interlock: alarm and runback). 2. The control system keeps the steam cooling pressure at higher than the combustor shell pressure, so low differential between these two parameters indicates steam leaks (interlock: alarm and trip). 3. Differential pressure across the liner can indicate inadequate steam flow (interlock: alarm and trip). Blade path temperature (BPT) spread monitoring provides a back-up indicator to this system and helps pinpoint where a combustion liner, for instance, may have an integrity problem, such as a crack. There are redundant steam supply strainers with continuous monitoring of the differential pressure across them, to check of obstruction of the steam cooling passages with solid carry over from the heat recovery steam generator (HRSG) or auxiliary boiler. On shutdown, an air purge sequence eliminates the potential for condensate accumulation in the steam cooling circuit. Combustion liner design To allow for steam passage and for better heat transfer properties, the combustion liner design is a double walled structure. Flame temperatures for the F, G and H turbines is the same, however with the G and H designs, the combustor exit temperature is higher. See Figure 52. There is no cooling air mixing with the cooling steam design.
43
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 52. Schematic of an F and G combustor. 5
To date, there has been no delamination experienced with the G model liners. All 18 (as of March 2004) G models operate with varying external temperature conditions, fuel type and other variables. Figure 53 shows the condition of a combustor liner at the combustor interval inspection. The TBC (thermal barrier coating) is intact. Protective monitoring systems have proved effective in ensuring the steam reliability and flow characteristics for the closed-loop cooling-steam.
Fig. 53. Condition of the Operated Combustor Liner. 5
Application case for a steam cooled G model MHI’s 501G model was installed in combined cycle (CC) application at Korean Electric Power Corporation’s (KEPCO’s) Ilijan’s power plant in the Philippines. There are 2, 600 MW blocks, each with two gas turbines and a steam turbine. Performance test results indicated 57.8 percent efficiency (natural gas), at a net rated capacity of 1285.7MW. At Ilijan, an auxiliary boiler is used to supply the combustor cooling of the first gas turbine unit. The gas turbine is started, run up to synchronization speed and loaded to 50MW. At this speed, the cooling steam supply is switched to the intermediate pressure (IP) superheater (normal combustor steam cooling supply). All water requirements for this plant are met with sea water using reverse osmosis desalination. Water quality needs to be with in required parameters for steam to be admitted to the steam turbine, which happens between 50 and 100MW load. When the first gas
44
Claire Soares
turbine is in combined cycle operation, the second gas turbine can be started, again using IP steam for combustor cooling. The second gas turbine is synchronized at 100MW. Loading on the train continues at 11MW/minute up to full rating.
Case 3: The use of low BTU “waste liquid” fuel.6 Deregulation is now a major feature in the power production industry’s development. The incentive for “small” power users, such as process and petrochemical plants, to produce their own power (become small power producers or SPPs), increases. Thailand provides an excellent illustration of this. Thailand has difficulty producing all the power the country needs with just the efforts of their national power company. For several years now, she has allowed bids from large independent power producers (IPPs) to better match her power demand curve growth. What she has also done is provided incentives for process plants to produce their own power, and sell the excess back to the national grid. The amount that can be sold back is often limited by distribution line size, which is as small as 15 kV, in the case of the grid adjacent to Esso’s Sriracha refinery for instance, but nevertheless the scheme is in place. Most countries in SE Asia are “a work in progress” in terms of their power supply and tariff infrastructure. The Petrochemical Corporation of Singapore (PCS) decided to take advantage of “pool rules for small generators” which covered generators of less than 10 MW and industrial in-house generators (“auto-generators”), which were instituted in Singapore as of April 1, 1998. An SPP such as PCS does not have the luxury of a known steady load for its power needs. Also, the quality, type and heating value of their fuels will vary. This is because they use process gases and fluids for fuel whenever they can, especially if that is the most cost effective use for what would otherwise be a waste process fluid. Due to the variations in the different characteristics of these fuels which are in essence different process streams, two things are required: - A gas turbine design that will accommodate fuels with a wide range of heating values. Such a turbine generally also has a more conservative design with turbine inlet temperatures (TITs) that will not be the highest for that turbine’s power range. - A very fast response valve (for cut-off of the fuel supply) is required. Without such a valve the exhaust gas thermocouples on the gas turbine would note large swings in turbine exhaust temperature. The key to PCS’s successful use of process fluids - which it didn’t have much other use for - as fuel, is valve response time and actuation characteristics. An ideal valve for this type of application is a “stepper” valve or its equivalent. The “stepper” valve and functional equivalents: The stepper valve is a fast response electrically operated valve which was pioneered by Vosper Thornycroft, UK (aka HSDE, UK) in the mid 1960’s. The term “stepper” actually refers to the motor type that drives the valve as opposed to the valve itself. The motor is a stepper motor, as opposed to a torque or AC or DC motor. Its self-integrating function ensures that the valve will proceed to a desired position and then the motor will stop. With other motors, the motor has to continue to run in order to keep the valve in that position - such valves need signals to cue them: run, stop running, then start running again, and so forth. If something were to happen causing the valve to fail, the stepper-type valve position would still lock and the system would continue running. The valve then makes the system fault tolerant, which is critical in applications such as emergency power supply generators. It also provides the fast response required by aeroderivative and some industrial gas turbines. This is useful for both power generation and mechanical drive service. Before the stepper valve was introduced in the mid 1960s, hydraulic and pneumatic actuation valves were used to provide the required response time. This increased the overall complexity of the fuel system. As always with instances where system complexity is heightened, system cost rose, but mean time between failures (MTBF) and availability decreased. The valve takes up very little space on the installation and service people unused to this new design spend frustrated time looking for the extensive “old” equivalent control system. Development of valves that could compete with HSDE’s original stepper arose from competition with that early design. As a result, there are now many manufacturers who produce functional equivalents on the market, for use in gas turbine fuel systems, high resolution controls for robots, automatic machining controls and so forth. In PCS’s application, they use a Moog (German manufacturer) valve which uses a DC motor. To get the same “stay in position” feature as a stepper type valve would have, manufacturers typically use a spring to hold a position. Design aims of fast response valves: The original design aims of the stepper type valve generally include the following safety considerations: • A fail freeze or fail closed option, depending on whether the operator is a power generation facility (“freezing” at the last power setting is then required) or a pipeline (in which case turbine shut down on valve failure is required). • The liquid fuel version of the valve incorporates a pressure relief valve protecting the system against over pressure and the fuel pump running on empty or “deadheading”, caused by closure of valves downstream of the fuel valve during system operation.
45
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
• High speed response of less than 60 ms required by aeroderivative gas turbines to prevent overspeed in block off-load conditions. • Explosion proof actuation to appropriate specification standards, allows operation in hazardous methane service. • Resistance to fuel contaminants including tar, shale, water, sand and so forth. • 24 volts DC is the maximum drive voltage which ensures personnel safety • Corrosion resistance in components exposed to wet fuel and corrosion resistance to all parts if the service is sour gas. Other operational objectives that dictate design features are operator’s requirements for: • Low mean time to repair (LMTR). The target of 1 hour, achieved with modular design, together with the target MTBF provided an availability of 99.998% for HSDE’s original stepper. • Higher Mean time between failures (MTBF). In HSDE’s case, a target of 50,000 hours was set and achieved. • Low maintenance costs, since the modular design can be repaired by an individual with relatively low expertise. Service intervals are 12 months. • Large control ratio which allows control over the ignition to full load as well as full speed ranges to be possible with one fuel valve. Fuel pressure variation compensation is provided. The additional speed ratio type control valve found in many other industrial gas fuelled installations is not required here. • Low power consumption since an electric motor of less than 100 watts is used. This also eliminates the need for additional hydraulic or pneumatic systems. Also black starting is more reliable if the fuel system is powered by the same batteries as the controller. PCS applications experience with fast response valves: Power production in Phase II of the Petroleum Corporation of Singapore or PCS, was commissioned in June 1997. PCS is part of a massive petrochemical plastics conglomerate in Singapore. Power production was an afterthought, as when they were built, their design did not include provision for them becoming an SPP. PCS chose a nominally 25 MW (23 MW in their normal ambient conditions) ABB GT10, although their power needs are roughly 26MW. This was because while SP were pleased to sell them their residual requirement; they would not buy any power from SPPs at the time of original power plant design. The turbine is fuelled by three different types of fuel, depending on the state of the plant. The BTU for each type varies, so again the fast response time for the stepper valve is critical. As PCS operations found, their fast response valve proved as useful as the stepper valve has been for power generation on the North Sea oil and gas platforms. The fast response time of the Moog (and other stepper valve manufacturers’) design helps the valve avoid the sudden burst of excess temperatures that accompany higher heating value fuel. (North Sea platform users frequently operate gas, liquid or gas & liquid fuel mixtures). Not all gas turbines are tolerant of a wide range of fuel types in a single application. Some of them require a whole different fuel system - nozzles, lines and all components - to be able to handle a totally different heating value fuel. In this application in Singapore, the ABB machine shows no sign of distress, which is interesting since the heating value of the fuel types varies as much as 50 percent. The exact fuel composition data is proprietary to PCS. PCS’s GT10 heat recovery steam generator (HRSG) provides a reliable source of steam. The plant exports steam to the nearby Seraya Chemicals plant in addition to fulfilling their needs. Emissions and steam supply: The original ABB EV burner design - a low NOx burner which can be fitted and retrofitted on the GT10, fuel types permitting - was not fitted in this case. The EV burner will handle clean natural gas and clean diesel fuel. It was not suitable for the high hydrogen content and variations in fuel composition that this application involves. Such fuels need a more forgiving fuel system, as well as water or steam injection to keep the NOx down. The PCS Singapore application uses steam for NOx reduction purposes. The steam is piped in through nozzles that are adjacent to the fuel nozzles on the fuel manifold of the GT10’s annular combustor. The source of the steam is the heat recovery steam generator (HRSG) that is packaged as part of the GT-10 system. If and when required, the plant also can draw high pressure steam from their process cracker. In PCS’ case, one boiler has been found to suffice. This is noteworthy as in applications like this, a redundant “packaged boiler” (running hot and on minimum load) is often found essential. This is so that it is possible to pick up the steam load should the turbine trip or be unavailable due to maintenance. A common subject for debate is whether uninterrupted steam supply during the switch from HRSG mode to fresh air firing is possible without flame out on the boiler supplementary burners.
46
Claire Soares
The PCS plant is part Japanese owned, so the specifications the installation had to meet matched those of environmentally particular Singapore, as well as the Japanese, who are the most environmentally strict practitioners in Asia. Steam injection reduces NOx levels from 300 to 400 mg/MJ fuel to just below 100 mg/MJ fuel. In this and similar cases, the GT system footprint may be of prime concern, if space comes at a high premium. The figure below outlines what the layout for the application (and similar applications) above may look like.
Fig. 54. SGT-600 Industrial Gas Turbine - 25 MW, Power Generation Application Layout (Note: Siemens SGT-600 was Alstom’s, formerly ABB’s GT-10) Dimensions in millimeters, mm (Source: Siemens Westinghouse)
In summary: The GT10’s ability to use three different “waste” petrochemical fluids as fuel, despite the 50 % variance in these three fluids’ heating value, is significant to process plants who could similarly become SPPs. Note that NOx emissions stayed below legislated limits for countries such as environmentally strict Singapore.
Case 4. Water and/ or steam injection for power augmentation and NOx reduction 7 Gas turbines swallow air and therefore are sensitive to ambient temperature and pressure. To increase the power output of gas turbines, especially in hot, humid (air density decreases with rising temperature and humidity) climates, water injection is used. (See Figure 55). The location of injection is commonly the filter plane and the compressor inlet.
47
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 55. Air Inlet Cooling Principle 7
The power gain is achieved due to 3 factors: i) The water which evaporates in the air intake increases relative humidity of the air from ambient conditions to nearly saturation. The evaporation of water reduces the air temperature hence density and the GT swallows a higher air mass flow. Higher power generation per unit volume of air swallowed and better efficiency result. ii) The water which evaporates inside compressor reduces the compressor work and increases GT net power output and GT efficiency as well. iii) The turbine power output is increased proportionally to the increased mass flow of air and water. Maximum power gain is achieved, if water is added at 2 locations in the air intake: just after the fine filter and additionally near the compressor intake as shown in fig. 58. After the fine filter an evaporative cooler or a fogging nozzle rack saturates the air and near the compressor intake a high fogging nozzle rack injects additional water, which evaporates inside the compressor.
Fig. 56. Evaporative Cooler System
7
Air chiller: An air intake chiller system consists of a heat exchanger, which is located in the air intake downstream of the filter. The heat exchanger cools the compressor inlet flow by the transfer of heat energy to a closed cooling water circuit. The closed cooling water is re-cooled in plate heat exchangers by one or more chillers. The closed loop cooling water is forwarded by one or more chilled water pumps. Load control regulates the cooling energy of each chiller to the desired plate heat exchanger outlet temperature of the cooling water. Outlet temperatures for each chiller correspond to a set point to the local control. The chillers are usually installed in
48
Claire Soares
the gas turbine air intake downstream of the air filter together with a droplet separator. The latter is needed to take out water droplets from condensation of humid air. Evaporative coolers: Generally, they are installed in the gas turbine air intake downstream of the air filter together with a droplet separator (see Figure 56). The evaporative cooler increases humidity close to saturation. The amount of evaporated water depends on ambient temperature and humidity. The water evaporates mostly before entering the compressor and the air is cooled down before compressor inlet. Thus, the air mass flow through the gas turbine is increased, which increases the power output of the unit. The evaporative cooler is only switched on and off. The cooler media and the droplet separator produce a pressure drop between 1.5 to 3 mbar and need an axial extension of the filter-house (see Figure 57a). The major components of an evaporative cooler are: • the evaporative cooler media (cellulose or fiber-glass, see Fig. 57b) • a water distribution manifold • a water sump tank with a recycle pump • a droplet separator (see Fig. 57c) Water requirements The water must be at least potable or flocculated and filtrated water quality or can be de-mineralized water. The water consumption is higher if tap water is used. Maximum total capacity is 25,000 l/h for a GT26 or GT13 and 17,000 l/h for a GT24 or GT11, where only 1 1,000 l/h and 7,500 l/h are evaporated and the remaining blow-down water is re-circulated.
Fig. 57. a) Evaporative Cooler Location, b) Evaporative Cooler, c) Droplet Separator
7
Inlet fogging Like evaporative coolers, this OEM’s fogging systems ALFog (an Alstom trademark) are typically installed in the gas turbine air intake downstream of the air filter (Fig.58).
Fig. 58. Fogging System Arrangement 7
The fogging system injects small water droplets into the air by nozzles to increase humidity close to saturation (90-95%). The amount of injected water depends on ambient temperature and humidity and is controlled by logic. The water evaporates and the air is cooled down before entering the compressor. In contrast to evaporative coolers, fogging systems have negligible pressure losses and do not need an axial extension of the filter house and are therefore ideal for retrofitting.
49
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 59. a) Fogging Nozzle Rack, b) Fogging Pump Skid 7
The major components of a fogging unit are: • the nozzle rack with nozzles (fig. 59a) • the pump skid including a control unit and a valve skid (fig. 59b) • a water drain system for the air intake and the intake manifold. The nozzles are mounted on tubes which are installed in the air intake downstream of the filter. Swirl nozzles are used in Alstom’s fogging system (trade name ALFog). They provide the required droplet size. Small droplets promote good evaporation in the air intake, high power augmentation and low risk of erosion. A high pressure piston pump feeds de-mineralized water at constant pressure (typically 140 bars) to the valve skid. The valves allow the sequencing of the water flow rate into sub-groups (typically 15 or 31, depending on design conditions). These subgroups are switched on and off by the control logic in order to adjust the water mass flow to ambient conditions. At lower ambient humidity and the higher ambient temperature, higher water quantities are needed to saturate the air, so more sub-groups are switched on. Typically 3 additional drain lines are installed in the air intake before and after the silencer and in the manifold. This is to ensure that water films and large secondary droplets, which might be generated on obstacles inside the air flow, are extracted from the air-stream flow. Water must be de-mineralized and 2 standard fogging systems are used, one for a design ambient humidity of 45% (design capacity 8,000 l/h or 2.2 kg/s for GT26 or GT13) and one for a design ambient humidity of 30% (design capacity 12,000 l/h or 3.3 kg/s for a GT26 or GT 13). High Fogging System: In order to increase power augmentation further, an additional nozzle rack is installed near the compressor intake. These systems are called high fogging, wet compression, over-spray or over-fogging systems. ALSTOM’s high fogging system ALFog is installed horizontally in the gas turbine air intake (fig 60). The system sprays small water droplets (<50µm) through nozzles into the air. These droplets evaporate mainly inside the compressor as the air is heated up during compression.
Fig. 60. High Fogging System in Combination with Fogging or Evap Cooler 7
The power of the gas turbine is increased mainly by 2 effects: • Compressor inter-cooling, which reduces compression work and compressor discharge temperature. • The mass flow through the turbine is increased.
50
Claire Soares
While fogging and evaporative cooler power increase depends on ambient conditions, the high fogging power increase is nearly independent of ambient humidity and temperature.
Fig. 61. a) High Fogging Nozzle Rack, b) High Fogging Pump Skid
7
The major components of a high fogging unit are • the nozzle rack with nozzles (fig. 61a) • the pump skid including a control unit (fig. 61b) and a water filtration system • the valve skid with staging valves • a water drain system for intake manifold Swirl nozzles are used in Alstom’s high fogging system for the same reasons as with the regular fogging system. The high-pressure pump operation is also similar. The valves, located at the valve rack, allow the sequencing of the water flow rate into subgroups (typically 5 or 10), that are switched on according to the power demand. Drains in the air intake manifold ensure that water films and large secondary droplets are extracted from the air-steam flow. The total water mass flow capacity of the high fogging system for a GT24 and GT26 is currently 1.2% of the air intake mass flow of the specific engine at ISO conditions. Accordingly, the demand of de-mineralized water is about 18,000 l/h or 5 kg/s for a GT24 and about 25,000 l/h or 7 kg/s for a GT26. If the control system is not adjusted to take into account the effect of the water content due to high fogging the pulsation levels of the combustion system and CO emissions may increase. Steady state cycle simulations confirmed that high fogging leads to a slight shift in the hot gas temperature if dry TIT (turbine inlet temperature) formulas are applied without any adoption. As countermeasure a modified TIT formula analogue to those used for oil operation with NOx water injection or operation with steam injection for power augmentation was implemented. This takes into account the amount of water injected for High Fogging. When using the adjusted TIT formulas high fogging has a negligible influence on CO emissions under base load operating conditions where the CO emissions are small (typically < 5 ppm). NOx typically appears to decrease with increasing high fogging water mass flow.
Gas Turbine Performance Ambient Condition Effects, Performance Optimization, and Extending Application Range Certain atmospheric conditions have a critical impact on any given gas turbine’s available power: a) Ambient temperature: As this rises, a gas turbine may swallow the same volume of air, but that air will weigh less with increasing atmospheric temperature. Less air mass means less fuel mass is required to be ignited with that air and consequential lower power developed. b) Altitude: Increasing altitude means lower density air, so that is turn decreases power developed by the turbine. c) Humidity: Water vapor is less dense than air, so more water vapor in a given volume means less weight of that air than if it had less water vapor. The effect is the same as with the two above factors.
51
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
The figure below provides graphical representation of how external conditions can affect gas turbine performance. The following conditions apply to figures 62 (a) through (d): • intake losses 10 mbar / 4" H2O • exhaust losses 25 mbar / 10" H2O • relative humidity 60% • altitude sea level
a) Generator output and heat rate versus compressor inlet air temperature
c) Exhaust gas flow and exhaust temperature versus compressor inlet air temperature
b) Heat rate and efficiency versus load
d) Nominal steam production (in combined cycle application) capability:
Fig. 62. Performance Data: SGT-600 Industrial Gas Turbine - 25 MW (Source: Siemens Westinghouse)
The subject of performance optimization is a vast one which would include several subtopics. Inlet cooling and water/ steam injection for power augmentation can be methods which are used to supplement power “lost” by factors such as high ambient temperatures, and high altitude. See the section on Design Development. The table below on performance for the Siemens SGT6-5000F (formerly Siemens Westinghouse W501F Econopac) indicates the difference water injection and steam injection can make to nominal power ratings.
52
Claire Soares
Table 6. Net Ref. Performance for the Siemens SGT6-5000F.
The following figures also demonstrate the effect of atmospheric conditions on the power developed, this time for a much larger turbine model than the SGT-600 depicted previously in this section.
Fig. 63. SGT6-5000F (formerly the W501F) Estimated Performance (Source: Siemens Westinghouse)
53
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 64. Combined Cycle Diagram with Drum-Type Boiler Source: Siemens Westinghouse
Fig. 65. SGT6-5000F (formerly the W501F) 2x1 Combined Cycle Source: Siemens Westinghouse
54
Claire Soares
Depending on how one defined performance optimization, the term could include cycle modifications and support systems that are external to the gas turbine core. Some examples are: • cycle modifications (which may also include, but are not limited to, inlet cooling systems, that are discussed under “Design Development”) • engine condition monitoring systems • life cycle counters/ assessment In the interests of space, these topics are not discussed here but they are exhaustively covered in the author’s book on Gas Turbines. As discussed in the section on design development, performance optimization is frequently attained by maximizing the power available using modifications to the base core. This allows the OEM to use proven technology that has long emerged from prototype growing pains, to fulfill a broader mandate in terms of power requirements and other operational needs. A case in point, Siemens’ SGT-700 (29MW) is an uprated SGT-600 (24MW), which then fills a broader range of applications.
Fig. 66. SGT-700 Gas Turbine - 29MW (Improved power output and efficiency over the SGT-600) (Source: Siemens Westinghouse) The SGT-700 has simple cycle shaft output of 29.1 MW and a thermal efficiency of 36% at base load on gas. This two-shaft machine can be used for both power generation and mechanical drive in both combined cycle and cogeneration applications. As a skid-mounted package with single-lift capacity and standard anti-corrosion materials and coatings, the SGT-700 is also suitable for offshore applications. The updated machine has full dry lowemission (DLE) capability. It can operate on both gas and liquid fuels with on-line switchover between fuels. To optimize performance, the SGT-700 power turbine is equipped with advanced profile blades that improve gas flow. Its overall design ensures easy service access to the combustor and burners. The revised 11-stage compressor produces a higher pressure ratio and an increase in mass flow through the engine. This results in greater power output and higher efficiency. Direct drive of pipeline or process compressor is provided for by the free high-speed power turbine, eliminating the need for a gearbox. The digital control unit is based on the proven design of the SGT-600.
An application case for the SGT-700 illustrates an example of extending the application of a basic gas turbine core (in this case the SGT-600) design. We noted in the section on design development, and the Mitsubishi case study (see Case Study 2) which listed several variations on the same GT core that additional power was added with essentially the same gas turbine core, with the addition of design features (for instance steam cooling instead of air cooling in certain hot section areas). Frequently, these developments result from a customer’s request: “I really could use another “x” MW in that plant, if you can make that happen” or “I’d rather have a slightly larger version of your “y” model rather than two of the “z” model, as I only have “w” amount of space and I can run the larger “y” at base load anyway, most of the time”. This “core growth” design is really an extension of design development work, as any such design modification has to be full load tested. Some air or steam leaks may not show up at 60% load, but may appear at close to 100% load. So the OEM goes through the expense of rigorous testing to minimize the risk of warranty-period costs.
55
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
The application example below, which illustrates application of an SGT-700, is also another “repowering” (see the section on Combined Cycles) case illustration. The very first SGT-800 gas turbine was delivered to Helsingborg Energi AB (now called Öresunds Kraft). This municipal utility in the southern Swedish town of Helsingborg is using this gas turbine to extend its Vasthamn coal-fired power station. The SGT-800 gas turbine has been integrated with an existing steam turbine system to create a combined cycle, CHP plant. The project is supported by the State Energy Authority and DESS (the Delegation for Energy Supplies for Southern Sweden). The turbine was ordered in August 1998 and connected to the grid at 100% load in November 1999. It burns natural gas from the pipeline which passes through Helsingborg. Fitted with AEV burners, it provides emissions of NOx and CO below 15 ppmv (at 15% O2).The electrical generating capacity at Vasthamn went from 64 MW to 126 MW, and the heat production capacity from 132 MW to 186 MW. Fig. 67. Municipal utility in the southern Swedish town of Helsingborg (Source: Siemens Westinghouse)
Combined Cycles Combined Cycle Basic Components, Terminology and Heat Cycle(s) The term combined cycle (CC) refers to a system that incorporates a gas turbine (GT), a steam turbine (ST), a heat recovery steam generator (HRSG), where the heat of the exhaust gases is used to produce steam and a generator. The shaft power from the gas turbine and that developed by the steam turbine both run the generator that produces electric power. The term “cogeneration” means generation of both work (shaft power) and heat (steam, in the case of a CC). So a combined cycle is a form of cogeneration. Fig. 68. Single and Multi shaft arrangements for CC plants (Reference: The World Bank)
The following figure shows a single shaft CC cycle block diagram in more detail.
56
Claire Soares
Fig. 69. A schematic diagram for a single shaft combined cycle. Source: Courtesy McGraw Hill, from “Power Generation Handbook”, Kiameh, P.
57
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
The following figure shows a schematic for a dual pressure combined cycle.
Fig. 70. A schematic diagram for a dual-pressure combined cycle. Source: Courtesy McGraw Hill, from “Power Generation Handbook”, Kiameh, P.
58
Claire Soares
Combined cycle plants are generally open cycle systems, however CC closed systems are possible if not that common. The plant system may also incorporate other accessories, such as a gear box (often used to “convert” 60 Hz models to 50Hz models), and/ or subsystems (that may themselves be closed or open systems) such as: • condensing units, intercooling heat exchangers (for the GT compressor air), • a regeneration (heat addition) heat exchanger to preheat the GT compressor discharge air, • reheat heat exchangers (for adding heat to the GT turbine module products of combustion), • inlet cooling and/or water or steam injection on the GT for power augmentation and/or NOx reduction, • a closed or open steam (and/ or air) cooling system (for the hottest areas of the GT turbine module), and • a supplementary firing system positioned downstream of the GT exhaust to maximize combustion of the exhaust gases (which will include unburned fuel hydrocarbons). See the block diagram figures below for a representation of GT closed systems, one with regeneration and intercooling, and one with reheat and regeneration. They are followed by a figure that represents a GT open system with water injection and regeneration.
Fig. 71. A Schematic of a GT closed system with regeneration and intercooling. Source: Courtesy McGraw Hill, from “Power Generation Handbook”, Kiameh, P.
59
Fig. 72. A Schematic of a GT closed system with regeneration and reheat. Source: Courtesy McGraw Hill, from “Power Generation Handbook”, Kiameh, P.
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 73. A Schematic of a GT open system with water injection and regeneration combined cycle plant efficiencies are now typically up to between 58% and 60%. Source: Courtesy McGraw Hill, from “Power Generation Handbook”, Kiameh, P.
Gas flapper valves allow the gas turbine exhaust to bypass the heat recovery boiler (HRSG) allowing the gas turbine to operate if the steam unit is down for maintenance. In earlier designs supplementary oil or gas firing was also included to permit steam unit operation with the gas turbine down. This is not generally included in contemporary combined-cycle designs, as it adds to capital cost, complicates the control system, and reduces efficiency. Sometimes as many as four (but most frequently two) gas turbines, each with individual boilers may be associated with a single steam turbine. As stated previously, the gas turbine, steam turbine, and generator may be arranged as a single-shaft design. A multi-shaft arrangement can also be used: Each gas turbine drives a generator and has its own HRSG, and steam turbine, which in turn, may also add power to the generator. In areas such as Scandinavia, additional criteria such as cogeneration in combined heat and power plants (CHP) or district heating, as well as demanding conditions (e.g. available space, emissions, noise level, architecture, environmental permits) associated with existing sites and available infrastructure must also be considered. A customer’s preferences regarding fuel election, personnel training level required and service requirements must also be accommodated.
Combined Cycle Module Flexibility With combined cycles, capacity can be installed in modules or module stages. Gas turbines can be commissioned initially (1 to 2 years project construction) and then the HRSG(s) and steam turbine(s) (an additional 6 months to 1 year). For instance, an Alstom 13E2 CC module can consist of 2, 13E2 gas turbines with heat recovery steam generation and one steam turbine, as in their Kuala Langat, Malaysia plant. In this way, combined cycle capacity can be installed in segments. This further assists generation dispatching, as each gas turbine can be operated with or without the steam turbine. This then provides better efficiency at partial load than operating one large machine with the total capacity equal to the gas turbine(s) and the steam turbine.
60
Claire Soares
Another case88 illustrating application of the 2, GT and 1, ST module is Alstom’s contract for Sohar Aluminum Company for the turn key construction of a 1000 MW gas-fired combined cycle power plant in Oman. The power plant, which will supply electricity to power a new aluminum smelter, will include four 13E2s, four heat recovery steam generators, two steam turbines, six generators. The size of the modules then provides the option for Sohar to add an additional 500 MW of capacity in the future (two GT13E2 gas turbines, two heat recovery steam generators, steam turbine and three generators). Gas turbine (GT) or combined-cycle (CC) construction cost per kilowatt cost does not increase much for smaller turbines. With steam turbines, it would to a far greater extent, because of the high additional construction work that comes with a steam turbine plant. A CC unit can typically be installed in two to three years, and a steam plant often takes four to five years, with no incremental power available until the complete plant is commissioned. An application case that illustrates the availability of power in increments is Alstom’s recent project award99 from Australian energy company Alinta Ltd, to supply 2, 172 MW GT13E2 gas turbines for the first stage of a major cogeneration facility at Alcoa’s Wagerup alumina refinery in Australia. That power plant will also provide reserve capacity to the new wholesale electricity market in the state of Western Australia. The Alstom turbines will operate initially in open cycle (Wagerup Stage 1). At a later stage, (Wagerup Stage 2), the turbines will be part of a cogeneration plant, operating as a base load power station providing both steam and electricity. A project1010database (developed by Siemens KWU) was used to analyze all combined, open cycle and steam power plants globally with respect to capacity (MW), fuel requirements, power system frequency and regional location. The database lists projected orders through 2005. Specific areas of the analysis are summarized as follows: In terms of overall plant size, 300-600 MW combined cycle plants are the most favored plant size in both 50 and 60 Hz markets (Figure 75). A combination of more than one block improves economics, and 300-600 MW fits well with the demand curve of most power grids in well developed countries. Financiers are also familiar with these economies of scale.
Fig. 74. The 395 MW Combined-Cycle Power Plant Otahuhu, New Zealand uses the modular concept Source: Siemens Westinghouse
Countries with large grids and high power demand growth prefer combined cycle plants in the range 600 to 2,500 MW. For this combination 2 to 6 parallel units (single shaft or multi-shaft) will suffice. Power systems in countries with relatively small generating capacity, which require smaller capacity additions, need combined cycle power plants in the range 100 to 300 MW. A large gas 8
Alstom Power Press Release 14 Dec 2005 Alstom Power Press Release 9 Dec 2005 10 Tailor-made Off the Shelf: Reducing the Cost and Construction Time of Thermal Power Plants. Paul I., (Siemens Power), Karg J. (KWU), O’Leary, Sr. D, (World Bank) 9
61
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
turbine and a steam turbine located on a single shaft can deliver this range. Countries with smaller or specialized grids buy multi-shaft combined cycle plants with several smaller gas turbines with one or more steam turbines. Dirty fuels, for instance residual promote requests for stolid, highly reliable trains that may run derated, over higher efficiency turbines. For peaking power or power systems with very low cost fuels, gas turbines in an open cycle system serve the power range between 50 and 300 MW. New order forecasts show the market evenly divided between 50 Hz or 60 Hz customers. Rising gas and oil prices everywhere, including the USA, will mean renewed strength in technologies that use alternative fuels, such as pulverized coal, paper liquor waste and steel mill flue gas. Steam-only (coal fired) Power Plant: The forecast projects 10% of the new orders will be steam power plants in 60 Hz market from 1999 to 2003 (Figure 76). In the 50 Hz market, the key ranges are 300 to 500 MW and 500 to 700 MW. Above 700 MW, supercritical technology represents a small but growing market share.
OEM Modular Strategy As previously discussed, to save on costs to both OEMs and end users, OEMs have developed modular plants. Siemens has twelve basic power plant combinations (Figure 79); four for open cycle gas turbine plants, six for combined cycle plants and two for coalfired steam power plants (with sub- and supercritical technology). Each combination covers a specific power range, efficiency, and fuel specification, with allowance for cogeneration system additions. For design flexibility, options to the reference version for each major functional unit (Figure 80) are provided. For example, “via-ship” is the reference for the functional unit “coal supply” with delivery “via rail” as an option. Flexible design requires breaking down the power plant into functional units, each of which will only directly affect one or two other modules. For a combined cycle plant, the functional units are arranged around the gas turbine and steam turbine. With the gas turbine, as we saw earlier, OEMS strive to maintain core feature commonalities.
Fig. 75. Markets for Gas Turbines (1995-2005) 10
62
Claire Soares
Fig. 76. Markets for Steam Turbines (1999-2005) 10
Fig. 77. Reference Power Plant Data 10
63
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 78. 2x700 MW Steam Reference Power Plant 10
Fuels for Combined Cycles11 Gas turbine operators prefer to burn natural gas and light oil (diesel, No.2). As we saw previously, crude oil, residual and “bunker fuel contain corrosive components. They require fuel treatment equipment. Also, ash deposits from these fuels can result in gas turbine derating of up to 15 percent. As we also saw previously (in the case of the Shunde plant in south China), they may still be economically attractive fuels, particularly in combined-cycle plants. Sodium and potassium are removed from residual, crude and heavy distillates by a water washing procedure. A simpler and less expensive purification system will do the same job for light crude and light distillates. A magnesium additive system reduces vanadium. Note that reduced availability will result due to water cleaning shutdowns to remove blade deposits, as on-line washing, even at reduced speeds, is not effective. A shutdown with a crank soak every 100 to 120 hours is required. Reduced component life due to hot gas path corrosion caused by vanadium deposits and other corrosion is another factor to consider. Table 7 provides a sample of naphtha- and heavy oil-fired power plants in operation and in the planning stage. As this table shows, some plants (e.g., Kot Addu and Valladolid) have accumulated 30-60,000 hrs of successful operation over their first five years plus.
64
Claire Soares
Table 7. Naphtha- and heavy oil-fired power plants in operation and planning stage
1111
Design and operation of these plants requires more attention than natural gas fired plants particularly in relation to fuel variables such as calorific content, density, composition, concentration of contaminants and emissions, as well as different burning behaviors (e.g. ignitability, flame velocity and stability). To overcome these difficult fuel properties, technological adaptation, additional equipment and operational requirements are necessary. These include GT layout (compressor, turbine) for the changed mass flows, different burner technology (burner design, burner nozzles), additional startup/shutdown fuel system, and safety measures. Performance, availability and operation & maintenance (O&M) expenses can be affected. To illustrate this, Table 11 shows some key non-standard fuels and their effect on a standard fuel system.
11
Gas Turbine Power Plants: A Technology of Growing Importance for Developing Countries. Taud R., (Siemens Power), Karg J. (KWU), O’Leary, Sr. D, (World Bank)
65
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Table. 8. Gas Turbines for None Standard Fuels Critical Fuel Properties11
An example of gas turbine combined cycle plant burning a non-conventional is the 220 MW Valladolid plant in Mexico. This plant, commissioned in 1994, burns heavily contaminated fuel oil, containing 4.2% sodium and up to 300 ppm vanadium. Fuel impurities (sodium, potassium and vanadium), tend to form ash particles in the combustion process, form deposits and corrode the gas turbine blades. In the case of the Valladolid plant, “Epsom salts”, consisting mainly of magnesium sulfate (MgSO4 7 H20), is dissolved in water injected into the gas turbine combustor through special orifices. This converts the vanadium into a stable water-soluble product (magnesium vanadate). This is deposited downstream of the combustor on the gas turbine blades, and causes only minor blade corrosion. To prevent major performance loss with salt build up (as with the Shunde, China plant that we read about previously); washing every 150 hrs was necessary to restore aerodynamic performance and plant efficiency. Good manhole access was a critical success factor for this project as servicing and maintenance during turbine washing shutdowns are simplified. (The plan is to eventually convert the Valladolid plant to natural gas operation).
Factors that affect Costs per Fired Hour11 Fuel type and mode of operation (steady load/ partial load) will determine maintenance intervals and the maintenance work items required. Some estimate that burning residual or crude oil will increase maintenance costs by a factor of 3, (assuming a base of 1 for natural gas, and by a factor of 1.5 for distillate fuel) and that those costs will be three times higher for the same number of fired hours if the unit is started every fired hour, instead of starting once very 1000 .fired hours. “Peaking” at 110 percent rating will increase maintenance costs by a factor of 3 relative to base-load operation at rated capacity, for any given period. The control system on combined cycle units is automatic. When an operator starts the unit, it accelerates, synchronizes and loads “by itself”. Fewer operators are required than in a steam plant.
Trends in Global Combined Cycle Installations11 A few hundred power generation plants are ordered from about a dozen OEMs every year. This means the market is exceptionally competitive. Given that most of the new plants are going into newly developing countries, the biggest factor in determining the winner of each project (bid on by several OEMs or not) is the financial deal the OEM can put together for the end user.
66
Claire Soares
As one might expect, maintenance costs are higher for any type of plant in countries that have not had as much exposure to the OEMs technology. As a significant extension of their revenue, OEMs offer overall “power by the hour” maintenance contracts. These costs vary, even for the same basic modular configuration and mechanical design, depending on the location’s demographics. So then will the actual and contractually set “cost per fired hour” figures. There would be a significant difference between what actual operational costs are for the same OEM’s CC block in a well developed area of the USA and a remote area in Azerbaijan, for instance. Demographics also alter construction costs. (As an illustration, in 1990s figures, costs varied from $592/kW for a new 1,080 MW combined-cycle plant in Egypt to $875/kW for a steam addition to convert four gas turbines in Pakistan to a combined-cycle plant, according to World Bank data). OEMs are aware that end users compare cost data at various meetings and forums, and that price variations are a sore and much negotiated point. Therefore OEMs continually strive to optimize designs and assembly methods to minimize the steepness of new operators’ learning curve.
Fig. 79. Schematic Diagram of a Parallel Combined Cycle Block with Full Flue Gas Cleaning 11
“Modularization” (for instance the Siemens Westinghouse GUD block which is 2, V94.3 gas turbines, their HRSG boiler capacity and a steam turbine) reduces construction costs. Compared with the customized design and construction, modularization can reduce project costs of detailed engineering, material price contingencies and financial loan interest during construction. Downsizing power delivery (to the grid) requirements will change overall operational cost figures. “Repowering” will change operational statistics significantly. Repowering is a term used to define the reconfiguration of a power station. It may mean replacing a steam turbine with a gas turbine or combined cycle. One example of a repowering option offered by an OEM is Alstom’s combining their 181 MW GT24 gas turbine with a dual pressure reheat cycle consisting of a 70 MW LP/IP steam turbine and a 20MW HP steam turbine, to generate a total of 270 MW. The most common configuration is called (Figure 79) parallel powering, where the gas turbine exhausts are used in the existing steam cycle. This is achieved by feeding the exhausts into a heat-recovery steam generator (HRSG) which provides additional steam to the existing steam turbine. Typically, parallel powering requires the addition of a gas turbine, associated electrical and instrumentation and control equipment, civil engineering, HRSG, additional piping and pumps as well upgrading the steam turbine. Generally, parallel powering can be undertaken fairly separately from the existing part of the plant, with a final integration phase and a plant down time of 1.5 to 2 months. The typical cost range is $US$ 300-500/kW. In some cases, national or international markets alter a power plant’s budget by changing available fuels. An example would be the United Kingdom’s temporary moratorium on their indigenous natural gas (which promoted coal for that period). When the decision was made to allow North Sea petrochemical liquid deposits to vaporize and be delivered as gas instead, that move created operational ripples in all industries that used petrochemical fuel, including power generation. Since the late eighties10, market growth in plant additions/ optimization technology retrofits has shifted in part, from Europe, North America and Japan to newly industrializing countries in Asia and Latin America. Financial means keep many of the end users in these
67
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
regions from using newer technologies that would extend their power generation capacity and reduce their costs per fired hour. Nevertheless, they are becoming increasingly aware of these design developments and do seek to incorporate them where and when possible. For the OEM, the main challenges are minimizing project cost, construction time and risk guarantees (Figure 80). Between the 1980s and 2000, project cost and construction time of coal and gas fired units have dropped by 50%. However, to compete, OEMs must offer better warranty packages. So the standardization of core design to minimize spares costs, make factory assembly methods and repair and overhaul methods “foolproof” increases in importance.
Fig. 80. Driving Forces in Power Plant Construction 10
Figure 81 shows11 the cost breakdown for combined cycle plants (350 MW-700 MW capacity) based on Siemens experience into the following categories: integrated services (project management/subcontracting; plant and project engineering/project management software, plant erection/commissioning /training; transport/insurance) and lots (civil works; gas- and steam-turbine and generator sets; balance of plant; electrical systems; instrumentation and control systems; and the boiler island).
68
Claire Soares
Fig. 81. Cost Breakdown for CC Power Plants 11
The Changing Power Generation Market11 From 1994 to 1999, power plant contract awards for fossil fuel-fired power plants (above 50 MW) averaged 63 GW per year. In 1999, sales were forecasted to average about 67 GW per year over the period 1999-2004. The market in the Asia Pacific Basin was declining, while showing a moderate growth in Europe and (starting from a low level), strong growth in North America. In comparison with coal fuelled power plants, open and closed cycle power plants are characterized by lower investment costs. However, USA fuel related costs (i.e. fuel price and plant efficiency) have changed with the rise in oil and gas prices in the USA that was precipitated by the Iraq war and hurricane Katrina. At the turn of the century gas prices ranged from about US$ 2.0/GJ to US$ 4.5/GJ, with the North American prices being at the lower end of the range. The only fact that anyone will sign their name to, in terms of oil and gas prices in 2006, is that they will go up. One Canadian forecast agency suggests that gas prices in 2006 in Canada will stay at about C$8/GJ. This then means that the fierce inter-OEM rivalry with respect to fuel efficiency will escalate. Three major infrastructure changes continue to drastically alter the face of the power generation industry and directly or indirectly promote technological innovation. They are: Deregulation: This then means that independent power producers (IPPs), some of them small power producers (SPPs), help make large plant new construction or expansion unnecessary. Consider the earlier examples of the PCS Company’s use of what were Alstom Power GT10s (this model was part of the Siemens Westinghouse acquisition of Alstom’s smaller engine divisions) in combined cycle operation. The waste hydrocarbon fluids they used as fuel, helped further develop low BTU fuel technology experience. Many SPPs can sell their excess power back to the utility grid. OEMs as IPPs: Most of the major OEMs have joint ventures all over the world that involve power generation. They provide training to their local partners and thus promote employment and technology to newly industrialized countries. Two examples are the SiemensYTL partnership for power stations in Malaysia and the Alstom Power-Genting joint venture for the Kuala Langat station. The Kuala Langat station also provided a good example of cogeneration as it sells its excess steam to a nearby mill.
69
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Oil companies as IPPs: Shell in the United Kingdom is a good example of a growing trend. As IPPs, oil companies can be their own customer for their oil and gas. This then short circuits much of the Fuel Purchase Agreement contractual formalities that other IPPs have to negotiate.
Integrated Gasification Combined Cycle (IGCC) Plants11 IGCC plants consist of three main sections: • the "gas island" for conversion of coal and/or refinery residues (such as heavy fuel oil, vacuum residues or petroleum coke). This includes gasification and downstream gas purification (removal of sulfur and heavy metal compounds in accord with required emissions levels), • the air separation unit and • the combined cycle plant. The modular design (gas generation, gas turbine system, HRSG and the steam turbine system) allows phased construction as well as retrofitting of the CC plant with a gasification plant. This replaces the "standard" gas turbine fuels (natural gas or fuel oil) by syngas produced from coal or refinery residues. IGCC is a combination of two proven technologies, however proper integration depends on using the lessons learned form several demonstration projects in Europe and the USA. Currently, there are more that 350 gasifiers operating commercially worldwide and at least seven technology suppliers. There are about 100 CC units plants ordered per year, but there is limited experience of IGCC commercial operation. Currently, we refer to operating experience at five IGCC plants: the 261 MW Wabash River plant; the 248.5 MW Tampa plant; the 253 MW Buggenum plant; the 99.7 MW Pinon Pine plant and the 318 to 300 MW Puertollano plant. IGCC will see commercial application in developed countries, such as Italy, for residual refinery fuels and gasified coal. However, great care needs to be taken in implementing a commercialization strategy for developing countries. Through 2015, the potential for refinery-based integrated coal gasification combined cycle (IGCC) plants is estimated to be 135 GW. Currently over 6GW of coal and refinery residue based IGCC projects are either, under construction or are planned. Figure 85 shows some of the IGCC plants that are planned or under construction.
Technology/Performance/Environment /Demand Trends11 Figure 82 compares the supply flows, emissions and by products of different 600 MW-class plants. With environmental emissions (including greenhouse gases, GHG), IGCC plants compare well with pulverized coal-fired steam power plants. Depending on the degree of integration between the gas turbine and the air separation unit (ASU), either standard gas turbine/compressor configurations can be applied. If not, the mismatch between turbine and compressor mass flows which results from the application of gases with low heating values, limited modifications are required to compensate. Three options are available. The selection of the appropriate air and nitrogen integration concept depends on a number of factors to be considered on a case by case basis. A summary of the important criteria is provided in Figure 83 Figure 84 sets out the principal criteria for selection of the different IGCC integration concepts. The “fully integrated approach” (selected for the European coal-based demonstration plants) results in the highest efficiency potential, but it can prove more difficult to operate. Nevertheless, after some initial operational problems, the Buggenum IGCC facility has demonstrated that design can provide good availability.
70
Claire Soares
Fig. 82. Comparison of Supply Flows, Emissions and Byproducts of Different 600 MW - Class Power Plants 11
Fig. 83. Main Criteria for Selection of the IGCC Integration Concept
71
11
1.1 Gas Turbines in Simple Cycle and Combined Cycle Applications
Fig. 84. Integrated Options for IGCC Power Plants
11
The non-integrated concept with a completely independent ASU is simpler in terms of plant operation and possibly in achievable availability. However, the loss in overall IGCC net plant efficiency compared with the fully integrated concept is 1.5 to 2.5 percent. So this concept is of interest for applications where efficiency is not the key factor (e.g. for the gasification of refinery residues). The concept with partial air-side integration is a compromise, with an only moderate loss in efficiency but improved plant flexibility, when compared with the fully integrated concept.
Fig. 85. Siemens SGT6-5000F (198MW, 60Hz) Simple cycle, Combined cycle, and other cogeneration applications (Source: Siemens Westinghouse)
72
The SGT5-2000E is used for simple or combined-cycle processes with or without combined heat and power, and for all load ranges, particularly peakload operation. For Integrated Coal Gasification Combined Cycle (IGCC) applications, Siemens Westinghouse provide the SGT5-200E (LCG) machine - the 2-type machine with modified compressor. The SGT5-2000E has more than 120 units in operation accounting for approximately 70,000 starts and more than 4,000,000 operating hours.
(Source: Siemens Westinghouse)
73
A professional engineer registered in Texas, and a Fellow of the American Society of Mechanical Engineers, Claire Soares has worked on rotating machinery for over twenty years. Soares’ extensive experience includes the specification of new turbomachinery systems, retrofit design, installation, commissioning, troubleshooting, operational optimization, and failure analysis of all types of turbomachinery used in power generation, oil & gas, petrochemical & process plants and aviation. The land-based turbines (gas, steam or combined cycle) in question were typically made by General Electric, Alstom power, Siemens Westinghouse, Rolls Allison, Solar and the companies they formerly were, before some of them merged. Her career experience also includes intensive training programs for engineers and technologists in industry. Her specialty areas include turbomachinery diagnostic systems as well as failure analysis and troubleshooting. In her years spent with large aircraft engine overhaul and aircraft engine fleet programs in the USA and Canada, Soares worked on turbine metallurgy and repair procedures, fleet asset management and aeroengine crash investigation. She also was engineering manager for the first overhaul program in the USA for the V2500 engine (commissioned 1991). Gas turbines (land, air and sea) are Ms. Soares’ primary area within the turbomachinery field. Her perspective with respect to gas turbines is that of an operations troubleshooter with extensive design experience in gas turbine component retrofits/ repair specification and retrofit system design development. Claire has authored/ co-authored six books for Butterworth Heinemann and McGraw Hill on rotating machinery (**See the links below for book details). She also writes as a freelancer, for various technical journals, such as Independent Power Generation and European Power News (U.K. based publications). Ms. Soares has an MBA in International Business (University of Dallas, TX), and a B. Sc. Eng. (University of London, external). She is a commercial pilot. Her scuba diving certification and training were in high altitude conditions. She has lived and worked on four continents. Her “non-engineering” time is partly spent on cinematography and still photography. **http://books.elsevier.com/bookscat/search/results.asp?country=United+States&ref=&community=listing &mscssid=0589M7CKL658H5QPFMW2650RBQ26XGD **http://books.mcgrawhill.com/search.php?keyword=claire+soares&template=&subjectarea=113&search= Go
Claire M. Soares P.E.; Fellow ASME; MBA Email:
[email protected]
74
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines
1.2.1-1 Introduction What is gasification? Gasification is a process in which combustible materials are partially oxidized or partially combusted. The product of gasification is a combustible synthesis gas, or syngas. Because gasification involves the partial, rather than complete, oxidization of the feed, gasification processes operate in an oxygen-lean environment. As figure 1 indicates, the stoichiometric oxygen-to-coal ratio for combustion is almost four times the stoichiometric oxygen-to-coal ratio for gasification of Illinois #6 coal.
Fig. 1. Diagram showing the products of reaction as a function of oxygen-tocoal ratio ( Reprinted from M. Ramezan, “Coal-based Gasification Technologies: An Overview” NETL Gasification Technologies Training Course, Sept. 2004.)
Jeffrey Phillips EPRI / Advanced Coal Generation P.O. Box 217097 Charlotte, NC 28221 phone: (704) 595-2250 email:
[email protected]
67
Just as most combustion-based processes such as power plants operate with excess oxygen to ensure complete conversion of the fuel, gasification processes also typically operate above their stoichiometric oxygen-to-fuel ratio to ensure near complete conversion to syngas. The amount of oxygen used in gasification, however, is always far less than that used in combustion and typically is less than half. In addition to coal, gasification processes can use petroleum coke, biomass, heavy oil, or even natural gas as the feedstock; however, this document will focus on coal gasification processes.
Comparison to coal combustion (advantages/disadvantages) As indicated in figure 1, the products of reaction change significantly as the oxygen-to-fuel ratio changes from combustion to gasification conditions. These changes are summarized in table 1. Because the mixture under gasifying conditions is fuel-rich, there are not enough oxygen atoms available to fully react with the feed. Consequently, instead of producing CO2, the carbon in the feed is converted primarily to CO, and the hydrogen in the fuel is converted mostly to H2 rather than H2O. Both CO and H2 are excellent fuels for use in a combustion turbine; however, their combustion characteristics are significantly different from natural gas.
the hydrogen in the fuel is converted mostly to H2 rather than H2O. Both CO and H2 are excellent fuels for use in a combustion turbine; however, their combustion characteristics are significantly different from natural gas. The implications of this will be fully covered in Section 3.1. Table 1 Comparison of the primary products created by the
����� � Comparison of the primary products created by the main fuel constituents in combustion and main fuel constituents in combustion and gasification gasification
���������� ������������ CO2 Carbon CO H2O H2 Hydrogen NO, NO2 HCN, NH3 or N2 Nitrogen SO2 or SO3 H2S or COS Sulfur H2O H2 Water The fate of the fuel’s nitrogen and sulfur in a gasification process has important and beneficial consequences on the environThe fate of the fuel’s nitrogen and sulfur in a gasification process has important and beneficial mental performance of an IGCC. Fuel-bound nitrogen, which is predominantly converted to NOx in combustion, is converted to N2, consequences on the of anRequirements” IGCC. Fuel-bound nitrogen, which is NH and HCN can be NH3 or HCN in gasification. As environmental discussed in the performance “Syngas Clean-up section of this chapter, both 3 predominantly converted to NOx in combustion, is converted to N2, NH3 or HCN in gasification. As limits NO removed to very low levels with the resulting cleaned syngas having essentially no fuel-bound nitrogen. This significantly x discussed in the “Syngas Clean-up Requirements” section of this chapter, both NH3 and HCN can be emissions of an IGCC. removed to fuel veryproduces low levels the resulting cleaned having noinfuel-bound The sulfur in SOxwith in combustion processes butsyngas is converted to essentially H2S and COS gasificationnitrogen. conditions. As will Thisfurther significantly NOx of an IGCC. be described on, both limits H2S and COSemissions can be removed from the syngas using technology developed for the natural gas industry to levels of less than 20 ppm, which means that more than 99% of the sulfur can be removed from the fuel and will not be emitted as The sulfur in fuel produces SOx in combustion processes but is converted to H2S and COS in gasification SOx. conditions. will be between described further on, H2S and COS can be the by syngas using reactions. Another majorAs difference combustion andboth gasification is the amount of removed heat that isfrom released the chemical In combustion, all of the fuel’s chemical is released as heatto(assuming is fully but in gasification technology developed for theenergy natural gas industry levels of itless thanconverted), 20 ppm, which means that most moreof the fuel’s chemical energy is not released as heat. In fact, an important measure of the efficiency of a gasification process is the fraction of the than 99% of the sulfur can be removed from the fuel and will not be emitted as SOx. feedstock’s chemical energy, or heating value, which remains in the product syngas. This fraction is termed the “cold gas efficiency,” and most commercial-scale gasificationbetween processes have a cold and gas efficiency of atisleast 65% andof some Another major difference combustion gasification the amount heatexceed that is80%. released by the Because significantly less heat is released by the gasification process, it is important to limit the amount of heat is transchemical reactions. In combustion, all of the fuel’s chemical energy is released as heat (assuming it isthat fully ferred outconverted), of the zone where the gasification reactions are occurring. If not, the temperatures within the gasification zone could be too but in gasification most of the fuel’so chemicalo energy is not released as heat. In fact, an low to allow the reactions to go forward (a minimum of 1000 C or 1800 F is typically needed to gasify coal). Consequently, unlike a important measure of the efficiency of a gasification process is the fraction of the feedstock’s chemical boiler where the entire firebox is lined with water-filled tubes that capture the heat released by the process and produce steam, many energy, or heating value, which remains in the product syngas. This fraction is termed the “cold gas gasifiers are refractory-lined with no water cooling to ensure as little heat loss as possible. Gasifiers also typically operate at elevated efficiency,” and most commercial-scale gasification processes have a cold gas efficiency of at least 65% pressure, sometimes as high as 6.2 MPa (900 psia), which allows them to have very compact construction with minimum surface area and some exceed 80%. and minimal heat loss.
Because significantly less heat is released by the gasification process, it is important to limit the amount of If not, the temperatures within the gasification zone could be too low to allow the reactions to go forward (a minimum MovingofBed 1000ºC or 1800ºF is typically needed to gasify coal). Consequently, unlike a boiler where the entire Afirebox diagramisoflined a generic bed gasifier shown in figure Moving bed gasifiers are countercurrent flowsteam, reactors in which withmoving water-filled tubes isthat capture the 2. heat released by the process and produce the coal enters at the top of the reactor and air or oxygen enters at the bottom. As the coal slowly moves down through the reactor, it many gasifiers are refractory-lined with no water cooling to ensure as little heat loss as possible. Gasifiers is gasified and the remaining ash drops out of the bottom of the reactor. Because of the countercurrent flow arrangement, the heat of also typically operate at elevated pressure, sometimes as high as 6.2 MPa (900 psia), which allows them to reaction from the gasification reactions serves to pre-heat the coal before it enters the gasification reaction zone. Consequently, the have very compact construction with minimum surface area and minimal heat loss.
1.2.1-2heatGeneric Typesoutof that is transferred of Gasifiers the zone where the gasification reactions are occurring.
temperature of the syngas exiting the gasifier is significantly lower than the temperature needed for complete conversion of the coal.
������� ������� ����� �� ��������� ������ ��� A diagram of a generic moving bed gasifier is shown in Fig. 2. Moving bed gasifiers are countercurrent flow reactors in which the coal enters at the top of the reactor and air or oxygen enters at the bottom. As the coal slowly moves down through the reactor, it is gasified and the remaining ash drops out of the bottom of the reactor. Because of the countercurrent flow arrangement, the heat of reaction from the gasification reactions serves to pre-heat the coal before it enters the gasification reaction zone. Consequently, the temperature of the syngas exiting the gasifier is significantly lower than the temperature needed for complete conversion of the coal. Fig. 2. Diagram of a generic moving bed gasifier
68
Jeffrey Phillips The residence time of the coal within a moving bed gasifier may be on the order of hours. Moving bed gasifiers have the following characteristics:1 • • • • • •
Low oxidant requirements; Relatively high methane content in the produced gas; Production of hydrocarbon liquids, such as tars an oils; High “cold gas” thermal efficiency when the heating value of the hydrocarbon liquids are included; Limited ability to handle fines; and Special requirements for handling caking coal.
Fluidized Bed
A diagram of a generic fluidized bed gasifier is shown in figure 3. A fluidized bed gasifier is a back-mixed or well-stirred reactor in which there is a consistent mixture of new coal particles mixed in with older, partially gasified and fully gasified particles. The mixing also fosters uniform temperatures throughout the bed. The flow of gas into the reactor (oxidant, steam, recycled syngas) must be sufficient to float the coal particles within the bed but not so high as to entrained them out of the bed. However, as the particles are gasified, they will become smaller and lighter and will be entrained out of the reactor. It is also important that the temperatures within the bed are less than the initial ash fusion temperature of the coal to avoid particle agglomeration. Typically a cyclone downstream of the gasifier will capture the larger particles that are entrained out and these particles are recycled back to the bed. Overall, the residence time of coal particles in a fluidized bed gasifier is shorter than that of a moving bed gasifier.
Fig. 3. Diagram of a generic fluidized bed gasifier
Generic characteristics of fluidized bed gasifiers include:2 • • •
Extensive solids recycling; Uniform and moderate temperature; and Moderate oxygen and steam requirements.
Entrained Flow
A diagram of a generic entrained flow gasifier is shown in figure 4. Finely-ground coal is injected in co-current flow with the oxidant. The coal rapidly heats up and reacts with the oxidant. The residence time of an entrained flow gasifier is on the order of seconds or tens of seconds. Because of the short residence time, entrained flow gasifiers must operate at high temperatures to achieve high carbon conversion. Consequently, most entrained flow gasifiers use oxygen rather than air and operate above the slagging temperature of the coal. Generic characteristics of entrained flow gasifiers include:3 • • • • •
69
High-temperature slagging operation; Entrainment of some molten slag in the raw syngas; Relatively large oxidant requirements; Large amount of sensible heat in the raw syngas; and Ability to gasify all coal regardless of rank, caking characteristics or amount of fines.
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines
Fig. 4. Diagram of a generic entrained flow gasifier
Hybrid & Novel Gasifiers
In addition to the three main classifications of gasifier types (moving bed, fluidized bed, and entrained flow) there are also gasifiers that are based on either hybrid combinations of those three classifications or novel processes such as a molten metal bath. The transport reactor-based gasifier developed by Kellogg Brown & Root (KBR) is an example of a hybrid gasifier as it has characteristics of both a fluidized bed and an entrained flow gasifier. The KBR gasifier will be described in more detail in the sub-section covering “pre-commercial” gasifiers.
1.2.1-3 Other Design Options In addition to the generic reactor designs of a gasification process, there are several other design options that a gasification process can have. Each of these options can have important impacts on the downstream processes in an IGCC including the combined cycle.
Atmospheric vs Pressurized
Gasifiers can operate at either atmospheric pressure or at pressures as high as 62 bar (900 psia). Pressurized gasifiers are better suited for IGCC operation since the pressure of product syngas will be sufficient to be fed directly into the GT fuel control system. Low pressure or atmospheric pressure gasifiers will require a fuel gas compressor after the syngas clean-up processes. High pressure gasifiers also have a positive impact on the cost and performance of the syngas clean-up section. Because the volumetric flow of the syngas is much smaller than it would be for an atmospheric process, the size of the clean-up equipment is smaller. For example, Hg capture can be accomplished by passing the syngas through a sulfur-impregnated, activated carbon bed. The size of the bed is dictated by the residence time of the syngas in the bed. Therefore, a smaller volumetric flow of syngas will result in a smaller carbon bed. If CO2 capture is required in future IGCCs, high pressure gasifier operation will improve the performance of physical absorption processes that can remove CO2 from the syngas.
Dry Feed vs Slurry Feed
Coal is typically fed into a pressurized gasifier either pneumatically as a dry solid or pumped as coal-water slurry. Slurry-fed feed systems have a lower capital cost, but result in less efficient conversion of coal to syngas (referred to as the “cold gas efficiency” of the gasifier). This is because some of the syngas must be “burned” in order to generate the heat needed to vaporize the water in the slurry. Consequently, the syngas produced by a slurry-fed gasifier typically has more CO2 in it than a dry-fed gasifier. This is not detrimental to GT operations since the CO2 can act as an effective diluent for NOx control; however, it does impact the design of the “acid gas removal” section of the IGCC as that process must use a solvent which allows the CO2 to pass through with the syngas rather than being stripped out with the sulfur species.
Air-blown versus Oxygen-blown
Oxygen for the gasification reactions can be provided by either air or high purity oxygen produced by a cryogenic air separation unit (ASU). Air-blown gasifiers avoid the large capital cost of an ASU but produce a much lower calorific value syngas than oxygenblown gasifiers. The nitrogen in the air typically dilutes the syngas by a factor of 3 compared to oxygen-blown gasification. Therefore, while a syngas calorific value of 300 Btu/scf might be typical from an oxygen-blown gasifier, an air-blown gasifier will typically produce syngas with a calorific value of 100 Btu/scf. This has a significant impact on the design of the combustion system of the GT. Because the nitrogen in air must be heated to the gasifier exit temperature by burning some of the syngas, air-blown gasification is more favorable for gasifiers which operate at lower temperatures (i.e. non-slagging).
70
Jeffrey Phillips Air-blown gasifiers also have a negative impact on CO2 capture. Because of the dilution effect of the nitrogen, the partial pressure of CO2 in air-blown gasifier syngas will be one-third of that from an oxygen-blown gasifier. This increases the cost and decreases the effectiveness of the CO2 removal equipment.
Quench versus Heat Recovery
A final design option involves the method for cooling the syngas produced by the gasifier. Regardless of the type of gasifier, the exiting syngas must be cooled down to approximately 100oC in order to utilize conventional acid gas removal technology. This can be accomplished either by passing the syngas through a series of heat exchangers which recover the sensible heat for use in the steam cycle of the IGCC, or by directly contacting the syngas with relatively cool water. This latter process is called a “quench” and it results in some of the quench water being vaporized and mixed with the syngas. The quenched syngas is saturated with water and must pass through a series of condensing heat exchanges which remove the moisture from the syngas (so it can be recycled to the quench zone). Quench designs have a negative impact on the heat rate of an IGCC as the sensible heat of the high temperature syngas is converted to low level process heat rather than high pressure steam. However, quench designs have much lower capital costs and can be justified when low cost feedstock is available. Quench designs also have an advantage if CO2 capture is desired. The saturated syngas exiting a quench section has near the optimum H2O/CO ratio for feed into a water-gas shift catalyst which will convert the CO in the syngas to CO2. Non-quench designs that require CO2 capture will have to add steam to the syngas before it is sent to a water-gas shift reactor.
1.2.1-4 Integrating with a Combined Cycle Syngas Clean-Up Requirements
Before syngas can be fed to a GT, it must pass through a series of clean-up steps in order to remove species that would either harm the environment or the GT itself. Particulate Matter qualifies as being both bad for the environment and the gas turbine. Particulates can be removed by “dry” processes such as a cyclone and rigid barrier filter or by “wet” processes such as a venturi scrubber. Gas turbines are fairly tolerant of sulfur species in the fuel. A typical OEM specification will call for no more than 250 ppmv of sulfur in the fuel. Concentrations of sulfur greater than that will cause unacceptably rapid corrosion of hot section parts. Even 250 ppmv of sulfur in the fuel would result in a relatively high sulfuric acid dewpoint in the exhaust gas. This would limit the amount of heat that could be recovered from the exhaust in the HRSG. To allow stack gas temperatures that are typical of natural gas fired combined cycles, the sulfur content of the syngas should be no more than 30 ppmv. A coal with 1wt. % sulfur will typically produce a syngas with 3000 ppmv sulfur. Obviously a sulfur removal step is required in order to meet the OEM’s fuel sulfur specification. As described earlier, well-proven technologies from the natural gas industry exist that can remove more than 99% of the H2S and COS in the syngas. This also allows IGCCs to beat US New Source Performance Standards for sulfur emissions from coal-fueled power plants by a comfortable margin. Nitrogen-containing species other than N2 must be removed to a low level to prevent formation of “fuel-bound NOx” in the GT combustor. HCN and NH3 concentrations typically are on the order of hundreds of ppmv in raw syngas. Because of its high solubility in water, NH3 can be easily washed out of the syngas by contacting it with water in a counter-flow packed column (the water then must be treated with neutralizing chemicals). Catalysts have been developed which promote the reaction of HCN with water vapor to produce NH3 and CO. Therefore, a HCN conversion catalyst is typically installed upstream of the NH3 water wash column. Chlorides must be removed from the syngas to prevent corrosion in the GT hot section. HCl is quite soluble in water, and therefore the same water wash which removes NH3 for NOx control, will also remove chlorides to an acceptable level. While it is not necessary for protection of the GT, mercury vapor in the syngas can be removed by passing the syngas through an activated carbon bed. This bed will also remove heavy metal species such as arsenic, which can cause problems in the GT.
Air-Side Integration
For oxygen-blown gasifiers in an IGCC, the air for the ASU can be supplied by a stand-alone, electric motor-driven compressor or it can be extracted from the discharge of the GT compressor. It could also be supplied by a combination of those two options. Extracting the air from the GT has efficiency advantages (the electrical losses in the GT generator and compressor motor are avoided), but experience at two European IGCCs which rely solely on GT compressor discharge air for their ASUs has shown that it can be operationally difficult, particularly during start-up. Consensus is growing that using a combination of a stand-alone, electric motor-driven compressor and GT air extraction is the best option for supplying the ASU. Typical designs call for at least 50% of the air to be supplied by the stand-alone compressor.
Steam-Side Integration
71
For IGCC which have a syngas cooler to recover heat from the hot syngas, there are two options for the steam that is produced by the syngas cooler. The first option is to send the steam to the HRSG of the combined cycle for superheating and reheating. The steam is combined with the steam produced by the HRSG and drives a single steam turbine. The second option is to provide some level of superheat within the syngas cooler and send the steam to a separate steam turbine which only accepts the steam from the gasification block.
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines To date, all IGCCs have utilized the first option. However, with growing interest in retrofitting existing natural gas fired combined cycles, the second option may be of interest because an HRSG and steam turbine of an existing combined cycle will have to be modified to accommodate the additional steam produced by the syngas cooler. The separate steam turbine option may also be attractive for applications in which the gasification block must be located considerable distance away from the combined cycle.
1.2.1-5 Commercially Available Large-Scale Gasifiers Four gasification technologies have been proven at the large scale (>1250 tpd) needed for IGCC applications: Lurgi Lurgi gasifiers have gasified more coal than any other commercially available gasification process. Lurgi gasifiers use a moving bed design and operate below the ash melting point of the feed. The coal does not have to be finely milled, only crushed. In fact, one of the disadvantages of the Lurgi process is its inability to handle coal fines. The coal is fed into the top of the gasifier via lockhoppers. Oxygen is injected at the bottom of the gasifier and reacts with the coal which has been pre-heated by the hot syngas rising through the coal bed. Ash drops off the bottom of the bed and is depressurized via a lockhopper. The process was originally developed by Lurgi GmbH in the 1930s in Germany. In total more than 150 Lurgi gasifiers have been built with the largest being able to process 1000 tpd of coal on a moisture & ash-free basis. The two most prominent applications of the Lurgi process are not IGCCs: the coal-to-gasoline refineries of Sasol in the Republic of South Africa and the Dakota Gasification Synthetic Natural Gas plant in North Dakota. Both applications feature a series of oxygen-blown gasifiers and use low rank coal from nearby mines as the feedstock. A key disadvantage of the Lurgi process in IGCC applications is the production of hydrocarbon liquids in addition to syngas. The liquids represent about 10% of the heating value of the feed and therefore must be utilized in order to achieve competitive efficiencies. At Sasol and Dakota Gasification the size of the gasification operations (14,000 tpd at Dakota) makes it economical to recover the liquids and convert them into high value products such as naphtha, phenols and methanol. It is not clear that the economies of scale for a 500 to 600 MW IGCC (4000 to 5000 tpd of coal) would allow those by-products to be recovered competitively. The other drawback of the Lurgi process is the need to stay below the ash melting point of the coal. For lower reactivity coals such as bituminous coal from the eastern US, this will result in lower carbon conversion due to the lower gasification temperatures. Conversely, low rank, high ash coals provide a competitive advantage for the Lurgi process versus its higher operating temperature competitors. GE Energy The GE Energy gasification process has the most extensive track record in IGCC applications. Originally developed by Texaco in the 1950s, the technology was purchased by GE from Chevron-Texaco in 2004. The process uses an entrained flow, refractory-lined gasifier which can operate at pressures in excess of 62 bara (900 psia). The coal is fed to the gasifier as coal-water slurry and injected into the top of the gasifier vessel. Syngas and slag flow out the bottom of the gasifier. Three options are available for heat recovery from the GE Energy process: Quench, Radiant, and Radiant & Convective. In the Quench option, both the syngas and slag are forced into a water bath where the slag solidifies and the syngas is cooled and saturated with water vapor. The slag is removed from the bottom of the quench section via a lockhopper while the saturated syngas is directed to gas clean-up equipment. In the radiant option, the syngas and slag enter a long, wide vessel which is lined with boiler tubes. The vessel is designed to cool the syngas below the melting point of the slag. At the end of the radiant vessel both the syngas and slag are quenched with water. In the radiant plus convective option instead of having a water quench for both the syngas and slag, only the slag drops into a water bath at the bottom of the radiant vessel. The syngas exits at the side of the vessel and enters a convective syngas cooler. Both fire tube boiler and water tube boiler designs have been used for the convective cooler. More than 100 commercial applications of the GE Energy gasification process have been licensed since the 1950s. Some of the most notable examples are described below. It was used in the first US IGCC, the Cool Water project, which was built in the early 1980s by a consortium of power industry organizations including Southern California Edison and EPRI. It also received production subsidies from the federal synfuels program operated by the Treasury Department. The Cool Water IGCC operated for five years starting in 1984 and gasified a total of 1.1 million tons of bituminous coal while proving the IGCC concept. After the demonstration period, the gasification block was sold, dismantled and moved to Kansas where it became the heart of a petroleum coke-to-ammonia plant. The GE Energy process was also used at Tampa Electric Company’s Polk County IGCC. That plant received DOE funding in Round III of the Clean Coal Technology program. The Polk plant is designed to produce approximately 250 MW of power and began commercial operations in July 1996. It has continued to operate after the end of the DOE-subsidized demonstration period and currently has the lowest dispatch cost of electricity in the Tampa Electric fleet.
72
Jeffrey Phillips The GE Energy process has also been used in several coal and petroleum coke-to-chemicals applications. Besides the Kansas coke-to-ammonia plant, GE Energy gasifiers featuring the water quench design have been installed at the Ube Industries coke-to-ammonia plant in Japan and the Eastman Chemical coal-to-chemicals facility in Tennessee. These plants have operated with the very high availability factors expected in the chemicals industry for more than 20 years. The GE Energy gasifier can also be used to gasify petroleum refinery liquid by-products such as asphalt residues. Several IGCC projects based on these feedstocks have been built at refineries around the world. Shell Coal Gasification Process (SCGP) The Royal Dutch Shell group of companies (Shell) has developed two different gasification processes. The first, called the Shell Gasification Process or SGP, was developed to gasify liquid and gaseous feedstocks. It features a refractory-lined gasifier with a single feed injection point at the top of the gasifier. The gasification products pass through a syngas cooler before entering a wet scrubber. The second process, called the Shell Coal Gasification Process (SCGP), was developed specifically to gasify solid feeds. The SCGP gasifier features a water-cooled “membrane wall” similar to the membrane walls used in conventional coal boilers. There are four feed injectors oriented horizontally in the mid-section of gasifier vessel. Slag flows out of a slag tap at the bottom of the vessel where it falls into a water bath and syngas flows out the top of the vessel. As the syngas exits the gasifier it is quenched with cool, recycled syngas to a temperature well below the ash melting point of the coal. The quenched syngas is still quite warm (typically 900°C) and passes through a syngas cooler and a dry solids filter before a portion of the gas is split off for recycle to the quench zone. The coal is fed to the SCGP gasifier pneumatically using high pressure nitrogen as the transport medium. The coal must first be dried and finely ground in a roller mill where warm, inert gas flows through the mill to remove the coal’s moisture. The dried coal is then pressurized via a system of lockhoppers. SCGP gasifiers operate at pressures up to approximately 40 bar. Shell began development of the SGP process in the 1950s, and work on the SCGP process started as a joint project with Krupp Koppers in the mid-1970s. Both companies agreed to go their separate ways in the development of coal gasification in 1981, and Krupp Koppers developed a competing dry-feed, membrane wall gasifier with the trade name PRENFLO. The only commercial application of the PRENFLO process has been the 280 MW Elcogas IGCC in Puertollano Spain. In 1999, Shell and Krupp Uhde agreed to join forces again in coal gasification. However, now only SCGP is being offered commercially by the two organizations. The first commercial application of SCGP was the 250 MW Demkolec IGCC built in 1994 in Buggenum, The Netherlands. The plant was originally owned by a consortium of Dutch electric utilities, but was sold to Nuon in the late 1990s. It is now operating as an independent power producer in the deregulated Dutch electricity market. Shell has also sold licenses for 12 SCGP gasifiers which will be used in coal-to-chemicals projects in China. The first of those projects is expected to begin operations in 2006. Perhaps the greatest advantage of Shell’s coal gasification process is its feed flexibility. The 240 tpd SCGP demonstration built at Shell’s refinery in Deer Park, Texas in the 1980s was able to process a full range of feedstocks including lignite, sub-bituminous coal, bituminous coal and pet coke. The reason for SCGP’s flexibility is the coal milling and drying process which eliminates the impact of moisture on the gasifier performance (however, the fuel for the drying process has a negative impact on thermal efficiency). The biggest disadvantage of the SCGP has been its higher capital cost which is inherent in the more expensive nature of the gasifier design (boiler tubes are more expensive than refractory brick) and its dry feed system. ConocoPhillips E-Gas
73
ConocoPhillips owns the E-Gas gasification technology which was originally developed by Dow Chemical. The E-Gas process features a unique two-stage gasifier design. The gasifier is refractory-lined and uses coal-water slurry feed. The first stage of the gasifier has two opposed, horizontally-oriented feed injectors. The syngas exits the top of the first stage and slag flows out of the bottom into a water bath. The syngas produced by the first stage enters the second stage at temperatures comparable to the exit temperatures of the other two entrained flow gasifiers, GE Energy and SCGP. Additional coal-water slurry is injected into this hot syngas in the second gasifier stage, but no additional oxygen is injected. Endothermic gasification reactions occur between the hot syngas and the second stage coal feed. This lowers the temperature of the syngas and increases the cold gas efficiency of the process. Upon exiting the top of the second stage of the gasifer, the syngas passes through a syngas cooler which features a firetube design. The cooled syngas then enters a rigid barrier filter where any unconverted char from the second stage is collected and recycled back to the first stage of the gasifier where the hotter temperatures ensure near complete carbon conversion. Dow began development of the E-Gas process in 1976 with a bench scale reactor. The work progressed to a 36 tpd pilot plant and then a 550 tpd “proto plant” located at Dow’s chemical manufacturing complex in Plaquemine, Louisiana. The main feedstock tested in these early gasifiers was lignite. In 1984 Dow entered an agreement with the federal US government’s Synthetic Fuels Corporation in which Dow received a price guarantee for syngas to be produced from a commercial scale E-Gas gasification plant built in Plaquemine. The plant began operation in 1987 and was operated by the Dow subsidiary Louisiana Gasification Technology Inc. (LTGI). The LGTI facility was designed to process 1600 tpd (dry basis) of sub-bituminous coal from the Powder River Basin. The clean syngas was sent to two Westinghouse 501D gas turbines which were already operating on natural gas at the Plaquemine complex. The total power output from the two turbines was 184 MW.
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines LGTI operated from 1987 through 1995 and received a total of $620 million in price support payments from the Synthetic Fuels Corporation (SFC). It was shutdown after SFC support ended. In total, more than 3.7 million tons of sub-bituminous coal was gasified at the LGTI facility. The E-gas process has also been used at the Wabash River IGCC repowering project in West Terre Haute, Indiana. In 1991 the project was selected to receive partial funding from the US DOE as part of the Clean Coal Technology program. The plant featured a new air separation unit, gasifier, clean-up system, gas turbine and heat recovery steam generator, but it utilized an existing, 30-year old, 100 MW steam turbine in Public Service of Indiana’s coal-fired Wabash River Generation Steam. The coal boiler that was originally built to supply steam to the steam turbine was retired when the IGCC equipment started up. The Wabash River IGCC began operation in 1995 on bituminous coal from the Illinois basin. However, today it operates exclusively on petroleum coke. Unlike the Polk County IGCC owned by TECO, ownership of the Wabash IGCC is split in two. The electric utility (Cinergy PSI) owns and operates the combined cycle plant while SG Solutions LLC owns and operates the gasification plant including the air separation unit. A commercial dispute between the previous owners of the gasification plant and Cinergy PSI led to a prolonged shutdown of the gasifier in 2004. However, with the recent change in ownership to SG Solutions, gasification operations began again in May 2005. ConocoPhillips is actively developing several new IGCC projects; among those are the Mesaba IGCC and the Steelhead Energy project.. The Mesaba project is being developed by Excelsior Energy in northern Minnesota. The project was awarded $36 million by the US DOE in 2004 as part of Round Two of the Clean Coal Power Initiative (CCPI). The money will support the cost of the Front End Engineering Design (FEED) and Permitting activities for the project. The Steelhead project, located in southern Illinois, will produce 600 MW of electricity as well as synthetic natural gas. The project recently received $2.5 million in funding from the State of Illinois to support its FEED effort.
1.2.1-6 Near-Commercial Gasifiers of Interest Several other gasification technologies, which have not yet been tested at sizes of 1250 tpd or larger, have the potential for being used in IGCCs or have been proposed for IGCCs that are under development. They are reviewed in this sub-section. KBR Transport Gasifier The Kellogg Brown and Root (KBR) Transport Gasifier has been under development at the Power Systems Development Facility (PSDF) in Wilsonville, Alabama. PSDF is a clean coal technology test facility built with US DOE sponsorship and has been operated by Southern Company Services since 1996. The KBR Transport Gasifier is a hybrid gasifier that has characteristics of both an entrained flow and a fluidized bed reactor. The technology is derived from KBR’s catalytic cracker technology which has been used for decades in petroleum refineries. The KBR gasifier operates at considerably higher circulation rates, velocities, and riser densities than a conventional circulating fluidized bed, resulting in higher throughput, better mixing, and higher mass and heat transfer rates. What is now the PSDF transport gasifer was mechanically completed in 1996; however, it originally operated as a combustor. In 1999 it was modified to operate as an air-blown gasifier. The PSDF also includes a small gas turbine which has been modified to burn the low Btu syngas produced by the air gasification process. Figure 5 is a diagram of the KBR gasifier. Dried, pressurized, pulverized coal is fed into the mixing zone of the gasifier. The oxidant is added at the bottom of the mixing zone. The gasifier temperature is maintained below the ash melting point of the coal, and this favors the use of air rather than oxygen as the nitrogen in the air serves to moderate the temperatures within the fluidized bed, while also supplying the velocity needed to entrain the solids. Oxygen-blown operation has also been tested at the PSDF and would be the preferred mode for polygeneration applications in which chemical products were produced as well as power.
Fig. 5. Schematic Diagram of the KBR Transport Gasifier
74
Jeffrey Phillips The riser section provides sufficient residence time at hot temperatures to crack any tars which may be produced during devolatilization of the coal. This gives the design a distinct advantage over fixed bed gasifiers which also operate below the ash melting point. Solids which are entrained out of the mixing zone of the gasifier are captured in a hot cyclone and returned to the fluidized bed. To maintain constant reactor inventory, gasifier ash is removed periodically from the lower region of the standpipe. If required, sand can be fed to increase reactor solids inventory. Because of its lower operating temperature and its dry feed arrangement, the KBR reactor is most attractive for lower rank, high moisture coals. The lower temperatures also eliminate the need for refractory lining of the gasifier vessel. The PSDF Transport Gasifier can process 38 tpd of coal in air-blown mode. Coals which have been gasified at the PSDF include Powder River Basin Sub-bituminous, Illinois #6 and North Dakota Lignite. The KBR Transport Gasifier will be used in the 285 MW Stanton IGCC project which was selected by the US Department of Energy as part of Round Two of the CCPI in 2004. Southern Company Services, KBR, and the Orlando Utilities Commission (OUC) are sponsors of the project, which will be located at an existing combined cycle power plant owned by the OUC. Nonetheless, the IGCC will be a Greenfield project rather than a retrofit of the existing combined cycle. Powder River Basin coal will be the feedstock for the gasifier. Future Energy The Future Energy entrained flow gasification process, formerly known as the Noell process, employs a single stage, downward firing gasifier that operates on a variety of liquid and solid feedstocks. The reactants are fed in at the top of the gasifier, which is cylindrical in shape and operates at temperature above ash fusion temperatures. At the bottom of the reaction chamber is a quench section, in which water is injected to cool the slag and the syngas. The solidified and granulated slag accumulates and is discharged via a lockhopper; the cooled syngas then exits the gasifier for further processing. There are two variations of the gasifier, depending on the ash content of the feeds. For high ash content feeds, the gasification chamber is enclosed by a tube screen that carries cooling water to ensure a long life of the gasifier. The screen wall is coated with a thin layer of protective coating; as the slag flows down the wall towards the quench section, a layer of solid slag is formed next to the wall due to the cooling effect. In applications with low ash content feeds, the cooling screen is replaced with a refractory lining cooled with a water jacket screen. Unlike other gasification processes, the Future Energy process allows the option of using either dry feed or slurry feed. The process is currently used in a 440 MW Vresova IGCC plant in the Czech Republic. The plant has 24 Lurgi gasifiers processing brown coal and 360 tpd of tars produced by those gasifiers are pumped to a single Future Energy gasifier for conversion into syngas. BGL Slagging Gasifier The British Gas/Lurgi (BGL) coal gasifier is a dry-feed, pressurized, fixed-bed, slagging gasifier. The reactor vessel is water cooled and refractory lined. Each gasifier is provided with a motor-driven coal distributor/mixer to stir and evenly distribute the incoming coal mixture. Oxygen and steam are introduced into the gasifier vessel through sidewall-mounted tuyeres (lances) at the elevation where combustion and slag formation occur. The coal mixture (coarse coal, fines, briquettes, and flux) which is introduced at the top of the gasifier via a lock hopper system gradually descends through several process zones. Coal at the top of the bed is dried and devolatilized. The descending coal is transformed into char, and then passes into the gasification (reaction) zone. Below this zone, any remaining carbon is oxidized, and the ash content of the coal is liquified, forming slag. Slag is withdrawn from the slag pool by means of an opening in the hearth plate at the bottom of the gasifier vessel. The slag flows downward into a quench chamber and lock hopper in series. The pressure differential between the quench chamber and gasifier regulates the flow of slag between the two vessels. Product gas exits the gasifier at approximately 1050°F (566°C) through an opening near the top of the gasifier vessel and passes into a water quench vessel and a boiler feed water (BFW) preheater designed to lower the temperature to approximately 300°F (150°C). Entrained solids and soluble compounds mixed with the exiting liquid are sent to a gas-liquor separation unit. Soluble hydrocarbons, such as tars, oils, and naphtha are recovered from the aqueous liquor and recycled to the top of the gasifier and/or reinjected at the tuyeres. A 720 tpd BGL gasifier has been built recently in Germany by SVZ as part of a waste-to-methanol plant. The BGL technology was originally developed by British Gas, which built two demonstration gasifiers in the 200 to 500 tpd range in Westfield, Scotland. Those gasifiers are now owned by Global Energy of Cincinnati, OH, which for a time owned the rights to the BGL technology. The rights were recently acquired by the Allied Syngas Corporation (ASC) based in Wayne, PA. ASC is currently pursuing projects based on a 1000 tpd BGL gasifier.
75
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines
Fig. 6. BGL Slagging Gasifier
Fig. 7. Diagram of the MHI air-blown gasifier (with permission from Clean Coal Power R&D Co., Ltd.)
MHI Air-Blown Gasifier MHI has developed a two-stage, air-blown gasifier which will be used in the 250 MW Clean Coal Power R&D Co. Ltd. IGCC being built near Iwaki City in Japan. Construction began in 2004, and the plant is scheduled to begin operation in 2007. The MHI gasifier operates in slagging conditions and, like the E-Gas process, injects coal without oxidant in the second stage. However, the gasifier features a water-cooled membrane wall rather than refractory lining and uses dry feed of coal rather than coalwater slurry. Unconverted char exiting the second stage is captured by a dry solids filter and returned to the first stage where coal and air are injected (see figure 7). The MHI gasifier has also been tested in the 1990s at a 200 tpd demonstration unit located at the same site where the Clean Coal Power R&D IGCC plant is being built.
1.2.1-7 Conclusions Four gasification technologies have been developed and demonstrated at sizes compatible with large scale IGCCs. Three of these technologies are based on entrained flow reactors which rapidly convert coal to a hot syngas. The fourth is based on a moving bed reactor which uses long residence times to convert the coal and produces a more moderate temperature syngas along with liquid hydrocarbons. Several other gasification technologies are nearing demonstrations of large scale gasifier which could be used in IGCCs. The most appropriate gasifier to use in an IGCC is probably more a function of the type of coal to be gasified than anything else. Lower rank, high moisture coals are more capable with dry-fed gasifiers, while high temperature slagging gasifiers are best for high rank coals which are less reactive.
1.2.1-8 Notes _____________________________ 1. D.R. Simbeck, et al., “Coal Gasification Guidebook: Status, Applications, and Technologies,” EPRI Final Report TR 102034 (Electric Power Research Institute). 2. Ibid. 3. Ibid.
76
BIOGRAPHY
1.2.1 Different Types of Gasifiers and Their Integration with Gas Turbines
Jeffrey Phillips EPRI / Advanced Coal Generation P.O. Box 217097 Charlotte, NC 28221 phone: (704) 595-2250 email:
[email protected]
Dr. Phillips began his involvement with IGCCs when he did his PhD research at Stanford on the off-design performance of IGCCs. He then spent 10 years working in the Royal Dutch/Shell group assisting in the development of the Shell Coal Gasification Process and then became an independent consultant, specializing in combustion turbine and combined cycle performance analysis. Among his consulting projects was an analysis of the suitability of current combustion turbine technology for oxy-fuel cycles. He recently accepted a position as project manager at EPRI where he directs projects related to IGCCs.
1.2.2 Implications of CO2 Sequestration for Gas Turbines
1.2.2-1 Introduction A variety of industrial processes such as power plants, oil refineries, cement works, and iron and steel production emit large amounts of CO2. Approximately a third of all the CO2 emissions due to human activity, however, come from fossil fuel-based power plants, with each power plant capable of emitting several million tonnes of CO2 annually. These emissions could be reduced substantially by capturing and storing the CO2 while other sources of emissions, such as transport and domestic buildings, cannot be tackled in the same way because of the large number of small sources of CO2. The capture of CO2 in an IGCC power plant consists of gasifying the feedstock in an O2 blown gasifier system and shifting the CO to H2 by catalytic reaction with steam1: CO + H2O = CO2 + H2.
(1)
The CO2 is then removed for sequestration from the syngas to produce a “decarbonized” fuel gas for combustion in a gas turbine. There are primarily two schemes for these processing steps consisting of shifting, CO2 removal and desulfurization of the syngas for current or near term technology plants, i.e., plants incorporating cold gas cleanup: • Employ sour shifting of the syngas followed by desulfurization and CO2 recovery within the same acid gas removal unit; or, • Desulfurize the syngas first followed by shifting and then removal /recovery of the CO2. The choice of either scheme depends primarily on the gasification heat recovery system employed (i.e., the extent to which cooling of the raw gasifier effluent is accomplished in a syngas cooler before the syngas is quenched / scrubbed with water to remove particulate matter).
1.2.2-2 Implications for Gas Turbines
Ashok Rao, Ph.D., P.E. Chief Scientist, Power Systems Advanced Power and Energy Program University of California Irvine, CA 92697-3550 phone: (949) 824-7302 ext 345 email:
[email protected]
77
Syngas Composition and Thermal Diluents Table 1 shows the composition of a decarbonized syngas (shifted and then 90% of the CO2 removed) as well as an “un-decarbonized” syngas leaving the acid gas removal unit of a high-temperature slurry-fed entrained-bed gasifier fed with a bituminous coal2. The H2 content of both the syngas streams is considered high enough to preclude the use of current design pre-mixed gas turbine combustors to limit the formation of NOx. Diluent addition is required to the syngas in order to reduce the NOx generation when utilizing “diffusion” type combustors, the amount of diluent addition required by decarbonized syngas, however, being higher than that required for the un-decarbonized syngas. Two types of diluents are available in an IGCC plant, water vapor introduced into the syngas stream by direct contact of the syngas with hot water in a counter-current column while recovering low temperature waste heat and / or N2 supplied by an elevated pressure air separation unit. The choice of the diluent depends on a number of factors such as • amount of low temperature waste heat available for the humidification operation, and • amount of excess N2 available from the air separation unit.
Table 1 Typical Clean Syngas Compositions (Dry and Sulfur Free Basis)
Component CO H2 CO2 CH4 Ar + N2 Total
Non-decarbonized
Decarbonized
50.1 37.4 10.2 0.1 2.2 100.0
2.8 94.1 0.6 0.1 2.4 100.0
The amount of low temperature waste heat available in a gasification plant in turn depends primarily on the gasification heat recovery system employed (i.e., the extent to which cooling of the raw gasifier effluent is accomplished in a syngas cooler before the syngas is quenched / scrubbed with water). On the other hand, the amount of N2 available as a diluent for the gas turbine depends on • •
the specific O2 consumption of the gasifier - the amount of N2 produced by the air separation unit is lower when the specific O2 consumption of the gasifier is lower; and, the type of gasifier feed system - dry feed systems utilize significant portions of the N2 as lock hopper pressurization gas as well as in the drying and transport of the coal into the gasifier and only the remaining amount of N2 is available for gas turbine injection.
A combination of the two diluents, i.e. water vapor and N2, may also be utilized, the relative amounts depending on the overall plant integration scheme and the trade-offs between efficiency and capital cost. In such cases, an option available consists of introducing the moisture into the N2 stream instead of the syngas. When N2 or moisturized N2 is utilized as a diluent, it may be either premixed with the decarbonized syngas before supplying it to the combustor of the gas turbine or it may be introduced into the combustor through a separate injector. Premixing the diluents with the syngas versus keeping them separately upstream of the combustor will have implications on the effectiveness of the diluent in lowering the local combustion temperature; a diluent entering the combustor premixed with the syngas would be more effective in lowering the NOx than if it entered the combustor through a separate nozzle. On the other hand, some savings in the N2 compressor horsepower may be realized in the case where the diluent is introduced into the combustor separately if the pressure drop associated with the fuel control valve is much higher than that for the diluent. It should be noted that the specific heat of the triatomic H2O molecule is significantly higher than that of the diatomic N2 molecule on a mole basis and thus the relative amounts of diluents required, i.e. water vapor versus N2 on a volumetric or mole basis by a given amount of syngas are quite different. Thus, in summary, the composition of the syngas / diluent are dependent on the type of gasifier, heat recovery and energy integration options and the type of air separation unit, i.e., whether it is an elevated pressure air separation unit which can supply high pressure N2 for use as a thermal diluent for NOx control. Gas Turbine Pressure Ratio The pressure ratio of the gas turbine increases when firing syngas, which is a much lower heat content gas than natural gas. The increase in pressure ratio is dependent upon the amount and nature of the diluent added and the degree to which the compressor inlet guide vanes are closed. The surge margin available in the compressor could thus constrain the amount of diluent that may be added and the resulting NOx emissions, in addition to the constraints set by the combustor design with respect to achieving stable combustion while limiting the CO emissions. Air extraction from the compressor may be required in order to limit the increase in the engine pressure ratio, in which case the extracted air (after cooldown / heat recovery) may be used efficiently in an elevated pressure air separation unit. Gas Turbine Firing Temperature The H2O vapor content of the working fluid flowing through the turbine, especially in the case when decarbonized syngas is the fuel and while utilizing water vapor as the diluent, is significantly higher than that in the case when natural gas is the fuel (i.e., compared to the case when natural gas is fired in dry low NOx combustors). The following implications exist for the gas turbine in such applications: • •
Derating of the turbine firing temperature due the different aero-heat transfer characteristics3 and Life of the thermal barrier coatings, and any ceramics that may be utilized in advanced gas turbines in the future.
Additionally, a gas turbine designed for a certain firing temperature on natural gas would see derating of the firing temperature not only due to the increased concentration of H2O vapor in the working fluid but also due to the increase in the pressure ratio since the temperature of the cooling air increases as the pressure ratio is increased. In the case of a steam-cooled gas turbine, however, derating of the firing temperature due to the increase in pressure ratio may be less significant (since the cooling steam temperature may be maintained independently of the gas turbine pressure ratio), unless the low pressure air-cooled stages of the gas turbine become the bottleneck.
78
Ashok Rao, Ph.D., P.E. Thus, the choice of the diluent to be utilized, i.e., H2O vapor versus N2 or their relative amounts should be included in the tradeoff / optimization studies, i.e., take into account not only the gasification island heat recovery options but also the accompanying extent of the gas turbine firing temperature reduction. Bottoming Steam Cycle The effect of lowering the firing temperature while increasing the pressure ratio significantly reduces the gas turbine exhaust temperature which has implications on the steam bottoming cycle. With lower steam superheat and reheat temperatures as compared to those corresponding to a natural gas fired combined cycle, the optimum steam cycle pressures would tend to be lower than those for the steam cycle in a natural gas fired combined cycle. Use of Selective Catalytic Reduction (SCR) At the present time, gas turbine manufacturers are willing to guarantee 15 ppmV NOx (dry, 15% O2 basis) for gas turbines in IGCC applications with the requirement of the thermal diluent addition. More stringent NOx emission requirements [e.g., 3 ppmV NOx (dry, 15% O2 basis)] in the future may require the installation of an SCR for post combustion control of the NOx or advanced gas turbine combustors that generate less NOx such as the trapped vortex combustor4. Development of low NOx combustors has a number of technical challenges to overcome due to the presence of a large concentration of H2 in the syngas (auto-ignition and flash-back being the challenges with pre-mixed combustor designs). These technical challenges will be even more severe for the more advanced gas turbines which will have higher pressure ratios (to take full advantage of the higher firing temperatures) and thus higher combustion air temperatures than current gas turbines in syngas applications. Decarbonized syngas will make it more challenging since the H2 content of the decarbonized syngas is significantly increased. Although SCRs have been utilized in coal fired boiler plants, the application of SCRs in IGCC plants poses special challenges. The NH3 slip from the SCR is known to react with the SO3 formed during the combustion process as well as some formed in the SCR itself (depending on the vanadium content of the catalyst) to form salts (ammonium bisulfate, sulfate and bisulfite) as the gases are cooled during heat recovery. Ammonium bisulfate tends to be especially sticky and can deposit on the cooler surfaces of heat transfer equipment causing fouling as well as corrosion. Any unreacted NH3 that may be emitted to the atmosphere is by itself a pollutant. In the case of a boiler, these problems are less severe since the NH3 slip from the SCR is preferentially adsorbed onto the flyash while any ammonium salts formed are captured in the particulate control devices. Furthermore, the air preheater in a boiler plant is cleaned periodically by on-line “soot” blowers. Operating combined cycle plants fired with sulfur bearing fuel oil have shown fouling of the low temperature boiler feed water heater in the HRSG when equipped with an SCR. Note that the salts that do not deposit within the HRSG will be emitted as particulate matter. The use of SCRs in IGCC applications thus requires a syngas that is very low in sulfur content to reduce the SO3 content in the gas turbine exhaust. SCRs have been installed in IGCC plants, the API Falconaro plant in Italy and the Negishi plant in Japan5. The ISAB IGCC plant in Italy which also has an SCR, uses it only on fuel oil operation and bypasses it during syngas operation6. For the Negishi plant, a syngas with a sulfur content of 8 ppmV is produced with a design maximum of 30 ppmV. No problems associated with salt deposition in the HRSG have been experienced in this plant. Both capital cost and thermal penalties are associated however, with deep sulfur removal in an IGCC for the following reasons: • •
A COS hydrolysis unit may be required to convert the COS (which is more difficult to scrub out in the acid gas removal unit) to H2S; and, A large circulation rate is required in the acid gas removal unit for deep sulfur removal.
In the case of an IGCC plant designed for producing a decarbonized syngas using sour shift and an acid gas removal unit to capture the CO2 and also perform desulfurization of the syngas, most of the COS is hydrolyzed to H2S in the shift reactors, while due to the very large solvent circulation rate maintained in the acid gas removal unit to capture the CO2, the sulfur content of the treated syngas is very low. In such cases, the incremental heat rate and cost penalties associated with producing a low sulfur syngas suitable for firing in a gas turbine equipped with an SCR are not significant. Engine Output The gas turbine when fired with syngas with diluent addition can be fully loaded to maximize its power output, the limits being the surge margin of the compressor (pressure ratio being increased) and the shaft torque. A nearly “flat rating” of the engine output with respect to the ambient temperature may be realized by opening up the guide vanes as the ambient temperature increases, the compressor inlet guide vanes being typically closed at the lower ambient temperatures to compensate for the larger mass flow rate of the syngas and the diluent.
79
1.2.2 Implications of CO2 Sequestration for Gas Turbines 1.2.2-3 Conclusions Due to the high H2 content of the syngas stream, the use of current design pre-mixed gas turbine combustors to limit NOx formation is precluded. Diluent addition is required to the syngas in order to reduce the NOx generation when utilizing diffusion type combustors; the amount of diluent addition required by decarbonized syngas is much higher that that required for the un-decarbonized syngas. The H2O vapor content of the working fluid flowing through the turbine, especially in the case when decarbonized syngas is the fuel and while utilizing water vapor as the diluent, is significantly higher than that in the case when natural gas is the fuel. The implications for the gas turbine in such applications are that the turbine firing temperature is derated due to the different aero-heat transfer characteristics and due to the higher cooling air temperatures caused by operation under a higher pressure ratio, while the life of the thermal barrier coatings and any ceramics that may be utilized in advanced gas turbines in the future may be adversely effected. Use of a trapped vortex combustor holds promise as an alternate option for supressing the NOx emissions in syngas application. Penalty of utilizing a SCR in a decarbonized syngas fired combined cycle can be less severe as compared to its use in an IGCC without upstream CO2 capture.
1.2.2-4 Notes ___________________________ 1. A.D. Rao and R. Stobbs,, “An Evaluation of Coal Gasification with CO2 Capture” (presented at the Combustion Canada Conference, Vancouver, September 2003); EPRI Report No. IE-7365, “Engineering and Economic Evaluation of CO2 Removal from Fossil-Fuel-Fired Power Plants” (prepared by Fluor Daniel, Inc, June 1991). 2. See note 1 above (EPRI Report). 3. A.D. Rao and D. Du Plessis, “Prospects for 200 MW Western Canadian Coal IGCC with CO2 Capture” (presented at the Combustion Canada Conference, Vancouver, September 2003). 4. K.Y. Hsu, L.P. Gross, and D.D. Trump, “Performance of a Trapped Vortex Combustor” (J. of Propulsion and Power, Paper No. 95-0810, AIAA 33rd Aerospace Sciences Meeting and Exhibit, Reno, Nevada, Jan 9-12, 1995). 5. T. Ono, “NPRC Negishi IGCC Startup and Operation” (presented at the Gasification Technologies Conference, San Francisco, California, October, 2003); D. Heaven and B. DeSouza, “Technical Issues with SCR in IGCC Applications” (presented at the 6th European Gasification Conference, Brighton, UK, May 2004). 6. See note 5 above (Heaven & DeSouza).
80
BIOGRAPHY
1.2.2 Implications of CO2 Sequestration for Gas Turbines 1.3.2 Advanced Brayton Cycles
Ashok Rao, Ph.D., P.E. Chief Scientist, Power Systems Advanced Power and Energy Program University of California Irvine, CA 92697-3550 phone: (949) 824-7302 ext 345 email:
[email protected]
Dr. Rao serves as the Chief Scientist, Power Systems at UC Irvine Advanced Power and Energy Program. He worked in industry for more than 30 years in the energy conversion area, and previously worked at Fluor as Director in Process Engineering and Senior Fellow in design / development of gasification for power generation and synthetic fuels coproduction. He received several patent awards in the energy conversion area and authored several papers on advanced power cycles and improved IGCC designs. Dr. Rao has also worked for Allis-Chalmers and McDowell Wellman Engineering in coal conversion; responsibilities included taking ideas from drawing board to demonstration scale plants. He holds a Ph.D. in Mechanical Engineering and a M.S. in Chemical Engineering.
1.2
Integrated Coal Gasification Combined Cycle (IGCC)
Gary J. Stiegel NETL 626 Cochrans Mill Road, P.O. Box 10940 Pittsburgh, PA 15236 email:
[email protected]
1.2-1 Introduction Integrated Coal Gasification Combined Cycle (IGCC) refers to the technology of converting coal to a fuel gas by contacting it with a mixture of oxygen (or air) and steam, burning the fuel gas in a combustion turbine/ generator, using the waste heat from the turbine to raise steam, and sending the steam to a steam turbine for additional power generation. IGCC has a number of technical advantages, but until recently, higher capital costs plus the availability of cheap natural gas have limited its application. However, as pollution limits become more stringent and natural gas prices increase, the superior performance of IGCC will make it increasingly attractive, particularly as technical advances reduce costs. Gasification is a well-proven technology that had its beginnings in the late 1700s. In the 19th century, gasification was used extensively for the production of “town gas” for urban areas. Although this application has all but vanished in the 20th century with the widespread availability of natural gas, gasification has found new applications in the production of fuels and chemical feed stocks and in large-scale power generation. Today, gasification technology is being widely used throughout the world. A study conducted in 2004 indicated that there were 156 gasification projects worldwide. Total capacity of the projects in operation was 45,000 MW (thermal) with another 25,000 MW (thermal) in various stages of development.
1.2-2 The Gasification Process1
Massood Ramezan phone: (412) 386-6451 email:
[email protected]. doe.gov
Howard G. McIlvried phone: (412) 386-4825 email:
[email protected]. gov SAIC P.O. Box 10940 Pittsburgh, PA 15236
59
The major difference between combustion and gasification from the point of view of the chemistry involved is that combustion takes place under oxidizing conditions, while gasification occurs under reducing conditions. In the gasification process, a carbon-based feedstock in the presence of steam and oxygen at high temperature and moderate pressure is converted in a reaction vessel called a gasifier to synthesis gas, a mixture of carbon monoxide and hydrogen, generally referred to as syngas. The chemistry of gasification is quite complex and involves many chemical reactions, some of the more important of which are: C + O2 CO2
ΔHr = -393.4 MJ/kmol
(1)
C + ½ O2 CO
ΔHr = -111.4 MJ/kmol
(2)
C + H2O H2 + CO
ΔHr = 130.5 MJ/kmol
(3)
C + CO2 ↔ 2CO
ΔHr = 170.7 MJ/kmol
(4)
CO + H2O ↔ H2 + CO2
ΔHr = -40.2 MJ/kmol
(5)
C + 2H2 CH4
ΔHr = -74.7 MJ/kmol
(6)
Reactions (1) and (2) are exothermic oxidation reactions and provide most of the energy required by the endothermic gasification Reactions (3) and (4). The oxidation reactions occur very rapidly, completely consuming all of the oxygen present in the gasifier, so that most of the gasifier operates under reducing conditions. Reaction (5) is the water-gas shift reaction, which in essence converts CO into H2. The water-gas shift reaction alters the H2/ CO ratio in the final mixture but does not greatly impact the heating value of the synthesis gas, because the heats of combustion of H2 and CO on a molar basis are almost identical. Methane formation, Reaction (6), is favored by high pressures and low temperatures and is, thus, mainly important in lowertemperature gasification systems. Methane formation is an exothermic reaction that does not consume oxygen and, therefore, increases the efficiency
of gasification and the final heating value of the synthesis gas. Overall, about 70% of the fuel’s heating value is associated with the CO and H2 in the gas, but this can be higher depending upon the gasifier type. Depending on the gasifier technology employed and the operating conditions, significant quantities of H2O, CO2, and CH4 can be present in the synthesis gas, as well as a number of minor and trace components. Under the reducing conditions in the gasifier, most of the fuel’s sulfur converts to hydrogen sulfide (H2S), but 3-10% converts to carbonyl sulfide (COS). Fuel-bound nitrogen generally converts to gaseous nitrogen (N2), but some ammonia (NH3) and a small amount of hydrogen cyanide (HCN) are also formed. Most of the chlorine in the fuel is converted to HCl with some chlorine present in the particulate phase. Trace elements, such as mercury and arsenic, are released during gasification and partition among the different phases, such as fly ash, bottom ash, slag, and product gas. Table 1 shows typical gas compositions for some of the more commonly used gasifiers. Table 1. Composition of Raw Syngas from Coal Fed Gasifiers
Gasifier Technology
Sasol/Lurgi1
Texaco/GE Energy2a
BGL2b
E-Gas/Conoco Phillips
Shell/Uhde2c
Type of Bed Coal Feed Form Coal Type Oxidant Pressure, MPa (psia) Ash Form Composition, vol% H2 CO CO2 CH4 Other HC H2S COS N2 + Ar H2O NH3 + HCN HCl H2S/COS Ratio
moving dry Illinois No. 6 oxygen 0.101 (14.7) slag
entrained slurry Illinois No. 6 oxygen 4.22 (612) slag
moving dry Illinois No. 6 oxygen 2.82 (409) slag
entrained slurry Illinois No. 6 oxygen 2.86 (415) slag
entrained dry Illinois No. 5 oxygen 2.46 (357) slag
52.2 29.5 5.6 4.4 0.3 0.9 0.04 1.5 5.1 0.5 20/1
30.3 39.6 10.8 0.1 1.0 0.02 1.6 16.5 0.1 0.02 42/1
26.4 45.8 2.9 3.8 0.2 1.0 0.1 3.3 16.3 0.2 0.03 11/1
33.5 44.9 16.0 1.8 1.0 0.1 2.5 0.2 0.03 10/1
26.7 63.1 1.5 0.03 1.3 0.1 5.2 2.0 0.02 0.03 9/1
Sources: 1
Rath, “Status of Gasification Demonstration Plants,” Proc. 2nd Annu. Fuel Cells Contract Review Mtg., DOE/METC-9090/6112, p. 91.
2
Coal Gasification Guidebook: Status, Applications, and Technologies, Electric Power Research Institute, EPRI TR-102034, 1993. 2a: p. 5-28; 2b p. 5-58; 2c: p.
5-48.
Many other reactions, besides those listed, occur. In the initial stages of gasification, the rising temperature of the feedstock initiates devolatilization of the feedstock and the breaking of weaker chemical bonds to yield tars, oils, phenols, and hydrocarbon gases. These products generally react further to form H2, CO, and CO2. The fixed carbon that remains after devolatilization reacts with oxygen, steam, CO2, and H2. Gasifier Types: All gasifier technologies generally fall into one of three generic types of reactor: moving-bed (also call fixed-bed), fluidizedbed, and entrained flow. In a moving-bed gasifier, large particles of coal move slowly down through the bed while reacting with gases moving countercurrenly. Reaction “zones” are often used to describe the reactions occurring along the length of the gasifier. In the drying zone at the top of the gasifier, the entering coal is heated and dried by the countercurrent flow of syngas, while simultaneously cooling the syngas before it leaves the gasifier. The moisture content of the coal mainly controls the temperature of the discharge gas from the gasifier. Because of the countercurrent operation of this gasifier, hydrocarbon liquids can be found in the product gas which has been problematic for downstream operations; however, techniques have been devised to capture the hydrocarbons and recycle them to the lower part of the gasifier. As the coal continues down the bed, it enters the carbonization zone where the coal is further heated and devolatilized by higher temperature gas. In the gasification zone, the devolatilized coal in converted to syngas by reactions with steam and CO2. In the combustion zone near the bottom of the reactor, oxygen reacts with the remaining char to consume the remaining carbon and to generate the necessary heat for the gasification zone. Depending upon the operation of the combustion zone, the moving bed gasifier can be made to operate in one of two distinct modes, i.e., dry ash or slagging. In the dry-ash version, the temperature is maintained below the ash slagging temperature by the endothermic reaction of the char with steam in the presence of excess steam. In addition, the ash below the combustion zone is cooled by the entering steam and oxidant. In the slagging version, much less steam is used so that the temperature of the ash in the combustion zone exceeds the ash fusion temperature of the coal and molten slag is formed. Moving-bed gasifiers have the following characteristics:
60
Gary J. Stiegel, Massood Ramezan, Howard G. McIlvried • • • •
Low oxidant requirements; Production of hydrocarbon liquids, such as tars and oils; High “cold-gas” thermal efficiency, when the heating value of the hydrocarbon liquids is included; and, Limited ability to handle fines.
Fluidized-bed gasifiers operate in a highly back-mixed mode, thoroughly mixing the coal feed particles with those particles already undergoing gasification. Coal enters at the side of the reactor, while steam and oxidant enter near the bottom, thereby suspending or fluidizing the reacting bed. Char particles entrained in the raw gas leaving the top of the gasifier are recovered by a cyclone and recycled back to the gasifier. Ash particles removed below the bed give up heat to the incoming steam and oxidant. Because of the highly back-mixed operation, the gasifier operates under isothermal conditions at a temperature below the ash fusion temperature of the coal, thus avoiding clinker formation and possible collapse of the bed. The low temperature operation of this gasifier means that fluidized-bed gasifiers are best suited to relatively reactive feeds, such as low-rank coals and biomass, or to lower quality feedstocks, such as high ash coals. Fluidized-bed gasifiers have the following characteristics: • • • •
Accept a wide range of solid feedstocks, including solid waste, wood, and high ash coals; Uniform, moderate temperature; Moderate oxygen and steam requirements; and, Extensive char recycling.
In entrained-flow gasifiers, fine coal particles react with steam and oxidant, generally pure oxygen, at temperatures well above the fusion temperature of the ash. The residence time of the coal in these gasifiers is very short, and high temperatures are required to achieve high carbon conversion. Because of the high reaction temperatures required compared to the other gasifier types, oxygen consumption is higher because of the need to combust more of the feedstock to generate the required heat. To minimize oxygen consumption, and hence cost, these gasifiers are usually supplied with higher quality feed stocks. Entrained-flow gasifiers can operate either in a down-flow or up-flow mode. Entrained-flow gasifiers have the following characteristics: • • • • • • • •
Ability to gasify all coals, regardless of rank, caking characteristics, or amount of fines, although feedstocks with lower ash content are favored; Uniform temperature; Very short feed residence time in the gasifier; Solid fuel must be very finely divided and homogeneous; Relatively large oxidant requirement; Large amount of sensible heat in the raw gas; High-temperature slagging operation; and, Entrainment of some ash/slag in the raw gas.
Syngas Cleanup Before syngas can be burned as a fuel or converted to chemicals, liquid fuels, or hydrogen, impurities in the gas, as shown in Table 1, must be reduced to levels that depend upon the requirements of the downstream process. To clean the syngas, chemical solvents, such as monoethanolamine (MEA), diethanolamine (DEA), and methyl diethanolamine (MDEA), and physical solvents, such as methanol (Rectisol) and mixtures of dimethyl ethers of polyethylene glycol (Selexol), operating at ambient or lower temperatures are employed. The selection of the technology for gas cleanup is dependent on the purity requirements of downstream operations and whether of not capture of carbon dioxide is required. With all of these technologies, the syngas is contacted with the scrubbing liquid in a packed column. In the amine-based systems (MEA, DEA, MDEA), weak chemical compounds are formed between H2S and the amine. Compounds such as COS are unaffected by the amine and must first be hydrolyzed to H2S if deeper sulfur removal is required. The rich amine is then pumped to a second packed column, operating at a higher temperature, where the H2S is stripped from the solvent and sent to sulfur recovery, typically a Claus unit. The lean amine is cooled and returned to the absorber. The Rectisol process uses chilled methanol, at a temperature of about -40oF to -80oF, as the solvent. In this case, the H2S and other sulfur-containing compounds, such as COS, dissolve in the methanol but do not react with it. The methanol is regenerated by flashing, and the lean solvent is then returned to the absorber. Like the Rectisol process, H2S and other sulfur-containing compound are quite soluble in the Selexol solvent, which operates at about 0oF to 100oF. The rich solution is sent to a regeneration column, where a combination of reduced pressure and stripping at an increased temperature is used to remove the absorbed acid gases. The regenerated solvent is returned to the absorber. In current IGCC systems, absorption processes are used to remove H2S, with a minimum of CO2 removal, since CO2 in the fuel gas improves turbine performance. However, should it become necessary to also recover CO2, these processes can be configured to remove both H2S and CO2.
61
1.2 Integrated Coal Gasification Combined Cycle (IGCC) Once the synthesis gas is sufficiently cleaned, various options exist for its utilization, such as the production of electricity via IGCC or the production of chemicals, hydrogen, and liquid fuels by water-gas shift and Fischer-Tropsch (F-T) technology. In IGCC, the clean synthesis gas is sent to a combustion turbine, where the gas is burned to produce electricity. The energy contained in the exhaust gas from the gas turbine is recovered in a heat recovery steam generator (HRSG). Steam from the HRSG goes to a steam turbine for the production of additional electricity. Approximately, two-thirds of the total electricity generated in the IGCC plant is produced by the gas turbine and one third is produced by the steam turbine. Because of the sulfur removal process discussed above, SO2 emissions are very low. Likewise, eliminating ammonia from the syngas in the gas cleaning system and adding a diluent (nitrogen or moisture) to the fuel gas prior to combustion to lower combustion temperature in the turbine results in very low levels of NOx emissions, even in the absence of selective catalytic reduction (SCR). There is growing concern that the increased concentration of CO2 in the atmosphere from the burning of fossil fuels is contributing to global warming with undetermined consequences. This concern had resulted in the development of the Kyoto Protocol, which sets limits on CO2 emissions for the signatory countries. An obvious target for CO2 reductions is large stationary point sources, such as coal-fired power plants. Studies to date have indicated that recovery of the CO2 from the flue gas from these plants is very expensive and inefficient. Because the flue gas is at about atmospheric pressure and the CO2 concentration is typically less than 15%, large volumes of gas have to be treated and the driving force for CO2 absorption is low. With coal gasification, the situation is different. The CO2 partial pressure in the product gas from the gasifier is much higher due to the higher pressure of the syngas (typically 500-700 psi). The higher pressure and the absence of nitrogen dilution result in a much lower gas volume to be treated (on the order of only 0.5% to 1% the volume of flue gas). Furthermore, by using a water-gas shift unit, CO in the fuel gas can be converted to H2 and CO2 before CO2 capture. With this approach, nearly all the carbon in the gasifier feed can be captured as CO2 for use or sequestration. Major potential uses for the captured CO2 include enhance oil recovery (EOR) and enhanced coal bed methane recovery (ECBM). Smaller uses include feedstock for chemicals manufacture and as a fertilizer in greenhouses, but these uses are much too small to have an impact on CO2 emissions. If it becomes mandatory to reduce greenhouse gas (GHG) emissions, it is likely that CO2 will be sequestered by injection into deep saline aquifers, abandoned oil and gas fields, and unminable coal seams.
1.2-3 IGCC Systems
Stack Gas
Electricity Oxygen from ASU or Air
Scrubber Water
Steam Turbine/ Generator
Heat Recovery Steam Generator
Turbine Exhaust
Coal Raw Syngas
Gasifier
Water
Particulate Scrubber
Recycle Ash
Raw Syngas
Scrubber Blowdown
Gas Cooling
Sour Condensate
Sour Syngas
Acid Gas Removal
Sweet Syngas
Acid Gas
Slag or Ash
(including non-volatile trace elements)
Recycle Water
Process Water Treatment
Water Treatment Residuals
Sour Gas
Treated Waste Water
Sulfur Recovery Unit
Gas Gas Turbine/ Turbine/ Generator Generator
Electricity
Air
Tail Gas Recycled to Gasifier
Byproduct Sulfur or H2SO4
Fig. 1. Schematic of Generic IGCC Power Plant
62
Gary J. Stiegel, Massood Ramezan, Howard G. McIlvried IGCC involves the integration of a number of technologies, as shown by the schematic diagram in figure 1. The technologies involved include air separation, gasification, syngas cleanup (including sulfur recovery), and power generation. Figure 2 presents some of the options for the various technology blocks in the gasification based systems. Improvements in any of these technologies will result in an improvement in IGCC. In a typical IGCC unit, coal, oxygen and steam are fed to the gasifier, where they are converted to raw syngas. The syngas is then cooled and cleaned of particulate matter, ammonia, and sulfur compounds. The cleaned gas is sent to the gas turbine where it is mixed with air and burned. Nitrogen from the air separation unit or steam may be added to the syngas to lower the combustion temperature and reduce NOx formation. The hot exhaust from the combustion turbine goes to a HRSG to raise steam for a steam turbine. The combination of a combustion turbine plus a steam turbine bottoming cycle increases the efficiency of IGCC. If it is desired to produce hydrogen, either as a product or to permit CO2 recovery, then a water-gas shift reactor is included along with an acid gas removal system. The hydrogen can be used as a fuel for fuel cells, for petroleum refining, as a chemical intermediate, or burned in the combustion turbine. In this case, the only combustion product is water, and the only pollutant is a small amount of NOx.
RESOURCES
Air/Oxygen Coal Biomass Petroleum Coke Heavy Oil Refinery Wastes MSW Orimulsion Other Wastes
GASIFIERS
OXYGEN-BLOWN Entrained Fl ow ChevronTexaco, E-Gas, Shell, Prenflo, Noell
Flui dized Bed HT Winkler
Moving Bed British Gas Lurg i (BGL) Lurgi (Dry Ash) Transport Reactor Kellogg -----------------------------------AIR-B LOWN Flui dized Bed HT Winkler, IGT “Ugas” KRW Spouting Bed British Coal, Foster Wheeler Entrained Fl ow Mitsubishi Transport Reactor Kellogg
ENVIRONMENTAL CONTROL Particul ate Removal and Recycle Filt ration, Water Scrubbing Chlori de and Alkali Removal Water Scrubbing Aci d Gas Removal Amine Processes Rectisol, Selexo l COS Hydrolysis Sulfur Recovery Claus Process Scott Process Sulfuric Acid Plant Water Treatment Process Water, BFW Tail Gas Treating Turbine NOx Control Nitrogen/Steam Dilution
ENERGY CONVERSION
PRODUCTS
Gas Turbine
Steam
Heat Recovery Steam Generator (HRS G)
Electric Power
Steam Turbine Boiler Syngas Conversion to Fuels & Chemicals Catalytic Conversion Shift Conversion Fischer-Tropsch Fuel Cell
Li qui d Fuels Chemicals Methanol Hydrogen Ammoni a/ Fertilizers Slag Sulfur/Sulfuric Aci d
H2 Turbine
Syngas Mercury Capture Syngas CO2 Capture
Fig. 2. Gasification-Based Energy Conversion System Options
1.2-4 Gasifier Improvements
63
Reliability and performance of the gasifier are key factors impacting the commercial deployment of IGCC technology. Today, single train IGCC plants, such as the Wabash River and Tampa Electric plants, have typically not achieved availabilities in excess of 80% for any sustained period of time. However, for gasification to be accepted for utility applications, availabilities in excess of 90% are required. For other applications, such as in refineries and chemical plants, the availability of the gasifier must be over 97%. Today, these high availabilities can be accomplished, but only through the addition of a spare gasifier at an increase in capital cost. To achieve gasifier high availability, several areas of gasifier operation need to be improved. Feed injectors are considered to be the weakest link in achieving a high on-stream factor, particularly with slurry-fed systems. A typical injector is reported to last from two to six months; however, a minimum life of 12 months is desired. Computational fluid dynamic (CFD) modeling around the injector may help to elucidate the factors that lead to failure. New materials and/or coatings for existing materials are needed to provide protection from sulfidation and corrosion at high reactor temperatures. New injectors are currently being developed based on rocket engine technology to achieve the target life and improve carbon conversion in the gasifier.
1.2 Integrated Coal Gasification Combined Cycle (IGCC) Injector life also appears to be dependent on whether a dry or wet feed system is used. In a dry feed system, injector life is usually better, possibly due to the absence of a large amount of evaporating water. Although improved life has been reported, operations with dry feed systems at high pressures are problematic because of the use of lock hoppers. To eliminate lock hoppers, a high pressure dry feed pump is under development which could result in a significant cost reduction for dry feed systems. For those gasifiers employing refractories to protect the pressure vessel, such as Texaco (now owned by GE Energy) and E-Gas (now owned by Conoco Phillips), new materials that have a useful life in excess of three years must be developed and demonstrated. Depending upon the severity of the gasifier operation and the feedstock being used, refractory liners typically last from six to 18 months. Rebricking a gasifier typically requires three weeks of downtime and costs $1-2 million. If a gasifier must be rebricked once a year, availability is automatically reduced by 5-6%. New refractory materials under development have shown considerable resistance to slag attack under simulated gasifier conditions and are currently being evaluated in commercial coal gasifiers. Actively cooled gasifiers, such as the Shell gasifier, which has steam tubes imbedded in the refractory liner, mitigate the refractory problem, but this route is usually more expensive. A new actively cooled liner that is potentially less expensive than other approaches is under development. Thermocouples used to measure the temperature inside the gasification zone are reported to last about 30-45 days. Failure of the thermocouples is due to corrosion resulting from slag penetration into the refractory and stresses caused by temperature cycles. When thermocouples are lost, the gasifier is typically controlled based on a prior correlation of gasifier temperature versus the methane content of the exit gas. New instrumentation capable of operating in the gasification environment with an expected lifetime of more than a year is required. Several new temperature measuring devices are being developed and tested with a promise of improved performance.
Gas Cleanup Improvements Current synthesis gas cleaning technologies employ chemical or physical solvents and operate at ambient or lower temperature. In an IGCC plant, these technologies typically account for 12-15% of the total capital cost of the plant. Amine-based systems are suitable for meeting today’s emission requirements, but they are not capable of achieving the limits of future potential regulations nor are they applicable for cleaning syngas for chemicals production. For the latter case, more expensive and energy intensive technologies, such as Rectisol, must be employed. What is needed are technologies capable of achieving the performance of a Rectisol unit but at equal or lower cost than an amine system. Considerable effort is currently underway to develop improved sorbents technologies that operate at moderate process temperatures while reducing acid gas concentrations to desired levels at a reduced cost and improved thermal efficiency. Integrated operation in a coal gasifier will be necessary to demonstrate the impact of trace contaminants in coalderived syngas on the performance, longevity, and regenerability of any new sorbent. Selective catalytic oxidation has the potential for achieving sulfur levels well below 1 ppm while operating at moderate process temperatures. In this approach, a small quantity of oxygen is injected into the synthesis gas stream and reacts with H2S over a catalyst to form elemental sulfur. To achieve the desired performance, either the COS in the raw gas stream must be hydrolyzed to H2S or a new catalyst must be developed to directly convert COS to elemental sulfur. For these approaches to be commercially attractive at a moderate process temperature, technologies are needed that can remove other trace contaminants at similar process conditions. Technologies for NH3, chlorides, and Hg removal are being developed and tested. Although not currently regulated, effort is also being focused on the removal of arsenic (As), selenium (Se), and cadmium (Cd) with emphasis on multi-contaminant removal technologies to achieve near-zero emissions of all contaminants.
1.2-5 Gas Separation Improvements Cost effective and efficient gas separation technologies are vital in the production of hydrogen from coal. Gas separation operations occur in two major areas: the separation of oxygen from air for use in the gasifier and the separation of the shifted synthesis gas into pure H2 and CO2 streams. Cryogenic technologies are currently employed for the production of oxygen; however, these plants are very capital and energy intensive. The cryogenic air separation unit in an IGCC plant typically accounts for 12-15% of the total plant capital cost and can consume upwards of 10% of the gross power output of the plant. Advanced dense ceramic membranes possessing both ionic and electronic conductance are being developed as a high temperature approach for air separation. A preliminary engineering analysis comparing these advanced membranes with conventional cryogenic technologies has been performed, and the results indicate that the advanced membranes have the potential for significantly reducing the capital cost of an IGCC plant with a corresponding 1-2 percentage point gain in thermal efficiency. Although many challenges remain in material composition and processing to produce defect-free, chemically and thermally stable membranes with commercially relevant fluxes, significant progress has been made.
64
Gary J. Stiegel, Massood Ramezan, Howard G. McIlvried Separation of hydrogen from shifted synthesis gas, either derived from coal or natural gas, is a key unit operation of any fossil-energy-based hydrogen production system. Membrane technologies have been, and continue to be, explored quite extensively by many investigators. Engineering studies comparing conventional coal gasification processes for producing hydrogen with advanced membranes and other technologies indicate that there is substantial incentive to develop advanced H2/CO2 separation technologies. Membranes can generally be divided into either organic or inorganic. Organic membranes appear to have limited applications for coal-based hydrogen production routes because of their extreme sensitivity to process conditions and trace contaminants. Instead, the bulk of the work for hydrogen separation is focused on inorganic membranes. Inorganic membranes can be classified as either porous or dense, and the latter can be further subdivided into metallic or solid electrolytes (ceramic). One promising membrane uses a manufacturing process that precisely controls the pore size distribution to allow primarily hydrogen to diffuse through the pores, thereby achieving very high separation factors. Considerable effort has also been devoted to metallic membranes, most of which are based on palladium (Pd). Although initially thought to be promising, these membranes have been found to be susceptible to degradation from the presence of both sulfur and CO. There have been reports of metal alloys that show very high hydrogen fluxes at temperatures around 750oF, but the stability of these membranes in the presence of trace contaminants from coal gasification must be determined.
1.2-6 Conclusions Markets and drivers are changing rapidly. Environmental performance is becoming a greater factor as emission standards tighten and market growth occurs in areas where total allowable emissions are capped. Also, reduction of CO2 emissions is one of the challenges in response to global climate change. There is a need for more environmentally sound processes, more efficient and reliable systems, and higher profitability. Industries need technologies that can match these requirements—a way to remain flexible, reduce risk, decrease emissions, increase stockholder return on investment, and consume fewer resources. Gasification is a technology that can meet these requirements. So far, the majority of existing applications have been geared toward the production of a single product or a constant ratio of two or more products per facility. The potential of gasification in expanding markets is in its ability to use low-cost and blended feedstocks and its multi-product flexibility. With deregulation, rapidly changing market demands, fluctuation in natural-gas prices, and increased environmental concerns, gasification has the potential to become a cornerstone technology in many industries. In particular, IGCC could become a dominant technology in the power industry because of the following advantages: • • • • • • • •
65
Ability to handle almost any carbonaceous feedstock; Ability to efficiently clean up product gas to achieve near-zero emissions of criteria pollutants, particulates, and mercury at substantially lower costs and higher efficiencies; Flexibility to divert some syngas to uses other than turbine fuel for load following applications; High efficiency because of the use of both gas turbine and steam turbine cycles; Ability to cost effectively recover CO2 for sequestration, if required; Ability to produce pure H2, if desired; Greater than 50% reduction in the production of solid by-products; and, Substantial reduction in water usage and consumption.
1.2 Integrated Coal Gasification Combined Cycle (IGCC) 1.2-7 Notes _____________________ 1. For a more complete discussion of gasification, refer to the following reports: “Gasification,” by C. Higman and M. van der Burgt, (Elsevier: Gulf Professional Publishing, 2003); “Major Environmental Aspects of Gasification-Based Power Generation Technologies,” by J. Ratafia-Brown, L. Manfredo, J. Hoffmann, and M. Ramezan, U.S. Department of Energy, Office of Fossil Energy, National Energy Technology Laboratory, December 2002.
66
BIOGRAPHY
1.2 Integrated Coal Gasification Combined Cycle (IGCC)
Gary J. Stiegel
Technology Manager - Gasification NETL 626 Cochrans Mill Road, P.O. Box 10940 Pittsburgh, PA 15236 phone: (412) 386-4499 email:
[email protected]
Mr. Gary J. Stiegel has been with the Department of Energy’s National Energy Technology Laboratory over twenty-nine years and is currently Technology Manager for Gasification. In this capacity, he is responsible for strategic planning, budget formulation, program development and oversight, and outreach activities for DOE’s Office of Fossil Energy’s gasification program. Prior to his present assignment, Mr. Stiegel served as the Program Coordinator for the Department’s Indirect Liquefaction and Gas-to-Liquids programs and spent ten years in R&D focusing on coal hydrogenation and the refining of coal-derived liquids. Mr. Stiegel has a Bachelors and Masters degree in chemical engineering and a Masters in Business Administration from the University of Pittsburgh. Prior to joining the Department of Energy, Mr. Stiegel was a process engineer for Union Carbide Corporation. During his career, Mr. Stiegel has published over fifty technical articles on various aspects of coal conversion and reactor engineering and is a registered Professional Engineer in Pennsylvania.
Massood Ramezan SAIC P.O. Box 10940 Pittsburgh, PA 15236 phone: (412) 386-6451 email:
[email protected]
Dr. Ramezan has over twenty five years of diverse experience in engineering, research & development, program management, marketing, energy technology assessment, process evaluation, personnel management, and technical services support in the areas of advanced energy systems and environmental control technologies. Specific project examples include: an environmental assessment of IGCC power systems, analysis of gasification-based multi-product systems with CO2 recovery, and life cycle assessment of advanced power systems. Dr. Ramezan received his B.S., M.S., and Ph.D. degrees in Mechanical Engineering from West Virginia University. He is a registered professional engineer and a member of ASME. He has authored more than 80 technical papers and has received numerous awards. Dr. Ramezan previously taught courses and conducted research in the areas of thermal-fluid sciences.
Howard G. McIlvried SAIC P.O. Box 10940 Pittsburgh, PA 15236 phone: (412) 386-4825 email:
[email protected]
Over 40 years experience in the areas of petroleum refining, petrochemicals, synthetic fuels, and energy conversion. Actively engaged in the preparation of many topical reports and post project assessments for the DOE’s Clean Coal Technology program. Specific project examples include the Tampa Electric Company and the Wabash IGCC projects. Received his BS, MS, and Ph.D. degrees in Chemical Engineering from Carnegie-Mellon University. He is a member of ACS, and has coauthored numerous technical papers and reports related to energy technology.
1.3.1.1
Graz Cycle – a Zero Emission Power Plant of Highest Efficiency
Franz Heitmeir
Wolfgang Sanz
Herbert Jericha Institute for Thermal Turbomachinery and Machine Dynamics Graz University of Technology, Graz, Austria http://www.ttm.tugraz.at
[email protected] [email protected]
81
1.3.1.1-1 Introduction In the last hundred years the concentration of some greenhouse gases in the atmosphere has markedly increased. There is a wide consensus in the scientific community that this seems to influence the Earth surface temperature and thus the world climate. Therefore, in 1997 the Kyoto conference defined the goal of global greenhouse gas emission reduction of about 5% in the next years compared to the 1990 emission level. CO2 is the main greenhouse gas due to the very high overall amount emitted by human activities, and about one third of the overall human CO2 emissions are produced by the power generation sector. In the European Union (EU) there is a strong pressure on public utilities and industry to reduce the CO2 emissions by power generation1. In 2003 the European Parliament passed a directive on emission trading. In 2005 emission allowances were assigned to about 10,000 companies in 25 countries within the EU which cover about 46% of the overall EU CO2 emissions. Companies which do not need their full amount can sell it to companies which need more than assigned. As emission allowances become scarce, they will have an increasing value. First estimates varied between 10 and 20 €/ton CO2 (12 and 24 $/ton CO2) by 2010, but in June 2005 European Union Allowances (EUA) were already being traded at 23 €/ton CO2 ( $28/ton CO2). So there is a strong driving force to develop commercial solutions for the capture of CO2 from power plants. The main technologies are as follows2: - post combustion CO2 capture, e.g. by washing of exhaust gases using amines; - pre-combustion decarbonization of fossil fuels to produce pure hydrogen or hydrogen-enriched fuels for use in conventional power plants; - chemical looping combustion; and, - oxy-fuel cycles with internal combustion of fossil fuels with pure oxygen. The authors believe that oxy-fuel cycles are a promising technology. The combustion with pure oxygen leads to a working fluid consisting mainly of steam and CO2, which allows an easy and cost-effective CO2 separation by steam condensation. Further advantages are the great variety of fuels which can be used (natural gas, syngas from coal or biomass gasification, etc.) and the low NOx generation, since nitrogen is only introduced by fuel bound nitrogen or as a residue in the oxygen to the combustion chamber. The generated NOx as well as other gases are removed together with CO2, so that no pollutants are emitted to atmosphere. On the other hand oxy-fuel cycles need the development of new turbomachinery components and have to bear the high efforts for oxygen supply. Oxygen needed in a large amount for this kind of cycles can be generated by air separation units (ASU) which are in use worldwide with great outputs in steel making industry and even in enhanced oil recovery. The largest air separation plant already in operation for some years in the Gulf of Mexico produces nitrogen for the injection in the gas dome of a large oil field off-shore3. Fortunately, the new working fluid of steam and CO2 allows new power plant cycles of highest efficiency, so that the additional efforts for oxygen supply can be largely compensated. Among them the Matiant Cycle, the Water Cycle, and the Graz Cycle are the best known4.
History
The authors believe that the so-called Graz Cycle has the potential of highest efficiency. The basic principle was developed and published by Jericha in 19855. He presented a power cycle without any emissions which was based on the internal combustion of hydrogen with oxygen in stoichiometric ratio as obtainable from solar power plants. Thermodynamically this steam cycle was an integration of a top Brayton cycle and a bottom Rankine cycle. In the nineties the hydrogen technology lost its impetus, so that the Graz Cycle was adopted for the firing of fossil fuels6. At this time cooperation with Japanese companies and research organizations led to the name
“Graz Cycle”. The working fluid was a mixture of about three quarters steam and one quarter CO2, the electrical efficiency was about 64%. Improvements and further developments since then were presented at many conferences7. In 2000, a variant of the Graz Cycle was proposed with a change of fuel from methane to oxygen blown coal gas (syngas), striving for minimum compression work8. All water of the cycle medium was condensed before compression, thus a minimum compression work could be obtained. In this cycle CO2 was the main component of the working fluid. In the following years the general layout of all components for a 75 MW prototype plant of this type was presented9. But in 2004 there was a return to the original high steam content Graz Cycle (S-Graz Cycle), because it had become clear that the reduction in compression work of almost pure CO2 has led to a considerable lowering of the inlet temperature to the combustion chamber10. So by increasing the steam content in recompression the compression work is increased, leading to a much higher combustion chamber inlet temperature. The heat input to the combustion chamber was lowered considerably thus raising the efficiency to the highest value that could be reached in the course of this cycle optimization. At the same time it turned out that much more steam for cooling could be made effective for the combustion chamber burners and the high temperature turbine (HTT) first blade rows. The resulting highest thermal efficiency of nearly 70% could be obtained if syngas was used as a fuel. The net efficiency, including the efforts of oxygen supply and compression of captured CO2 for liquefaction, is 56%. The general layout of the components for a 75 MW prototype plant showed the feasibility of all components. In recent discussions with gas turbine industry a scale-up to a 400 MW plant was discussed for the S-Graz Cycle scheme. In 2005 further modifications of the Graz Cycle were discussed and their potential was analyzed11. An economic analysis of the Graz Cycle power plant showed the strong dependence of the economics on the still uncertain investment costs. In this work the name “Graz Cycle” means the original “S-Graz Cycle,” which was the more efficient variant and the one which will be pursued in the future.
1.3.1.1-2 Cycle configuration and thermodynamic layout All thermodynamic simulations were performed using the commercial software IPSEpro by SIMTECH Simulation Technology12. This software allowed implementation of user-defined fluid properties to simulate the real gas properties of the cycle medium. The physical properties of water and steam were calculated using the IAPWS_IF97 formulations13; CO2 was also modeled as real gas based on the correlation of Sievers14. Furthermore, a turbine module was developed for the calculation of cooled turbine stages. A simple stageby-stage approach similar to the one presented by Jordal et al. was assumed. This assumption allowed for the calculation of the amount of cooling steam needed per stage15. Within the module, half of the cooling mass flow was mixed to the main flow at the stage inlet, thus contributing to the stage expansion work. The rest was added at the stage exit. Details of the model were presented in Luckel, 200416. The thermodynamic data presented was for a cycle fired with methane, because it gave similar results as natural gas, the most likely fuel to be used in a first demonstration plant. The lower heating value was 50015 kJ/kg. The thermodynamic simulation was based on the following assumptions on efficiencies and losses: • The isentropic turbine efficiency is 90.3% for the High Temperature Turbine (HTT), 90% for the High Pressure Turbine (HPT) and 88% for the Low Pressure Turbine (LPT); • The isentropic efficiency of CO2 compressors is 78% and of CO2/ H2O compressors 88%; • The isentropic efficiency of pumps is 75%; • The mechanical efficiency of the turbomachinery is 99.6% of net power; • The generator efficiency is 98.5%; • The transformer efficiency is 99.65%; • Auxiliary losses are 0.25% of heat input; • The combustor heat loss is 0.25%, the pressure loss 4%; • The oxygen excess is 3% of the stoichiometric ratio in order to keep CO generation low; • The minimum temperature difference at Heat Recovery Steam Generator (HRSG) economizer is 5 K, at superheater 25 K; • HRSG: cold side pressure loss is 28 bar (including 5 bar for HPT pipe); hot side pressure loss is 4 kPa; • The pressure loss of all other heat exchangers is 3%; • Fuel is supplied at 41.7 bar and 150°C; • The cooling water temperature in the condenser is 10°C; • CO2 is released at 1 bar, efforts of a further compression to 100 bar including the remaining steam content at 1 bar (350 kJ/kg) is considered in the power balance; and • The power consumption of oxygen production is 900 kJ/kg (0.25 kWh/kg) and of oxygen compression from an ASU exit pressure of 2.4 bar to combustor pressure is 325 kJ/kg.
82
Franz Heitmeir, Wolfgang Sanz, and Herbert Jericha Figure 1 shows the principle flow scheme of the S-Graz Cycle with the main components and main cycle data.
Fig. 1. Principle flow scheme of Graz Cycle power plant
Basically, the Graz Cycle consists of a high temperature Brayton cycle (compressors C1 and C2, combustion chamber and High Temperature Turbine HTT) and a low temperature Rankine cycle (Low Pressure Turbine LPT, condenser, Heat Recovery Steam Generator HRSG and High Pressure Turbine HPT). The fuel together with the nearly stoichiometric mass flow of oxygen is fed to the combustion chamber, which is operated at a pressure of 40 bar. Steam as well as a CO2/ H2O mixture is supplied to cool the burners and the liner. A mixture of about 74% steam, 25.3% CO2, 0.5% O2 and 0.2% N2 (mass fractions) leaves the combustion chamber at a mean temperature of 1400°C. The fluid is expanded to a pressure of 1.053 bar and 579°C in the HTT. Cooling is performed with steam coming from the HPT (13.7% of the HTT inlet mass flow), increasing the steam content to 77% at the HTT exit. It is quite clear that a further expansion down to condenser pressure would not end at a reasonable condensation point for the water component, so that the hot exhaust gas is cooled in the following HRSG to vaporize and superheat steam for the HPT, the pinch point of the HRSG is 25°C at the superheater exit. But after the HRSG, only 46% of the cycle mass flow is further expanded in the LPT. The LPT exit and thus condenser pressure is 0.043 bar. For a mixture of a condensable (steam) and a non-condensable gas (CO2) the condensation temperature depends on the partial pressure of steam, which continuously decreases during the condensation. For a given condensate exit temperature the condenser pressure determines the amount of steam condensed. In order to maximize the LPT power, the condenser pressure should be reduced as far as possible, but this is counteracted by an increased effort for compressing the gaseous steam/CO2 mixture to atmospheric pressure. So for a given condensate exit temperature of 18°C (for a cooling water temperature of 10°C) the optimum condenser pressure is 0.043 bar, where about half of the combustion water is condensed (see figure 2).
83
Fig. 2. Influence of the condenser pressure on thermal cycle efficiency
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency Gaseous and liquid phases are separated in the condenser. From there on, the gaseous mass flow, which contains the combustion CO2 and half of the combustion water, is compressed to atmosphere (C3/C4) with intercooling and extraction of condensed water, and supplied for further use or storage. At atmosphere the CO2 purity is 94%, further water extraction is done during further compression for liquefaction. After segregating the remaining combustion H2O, the water from the condenser is preheated, vaporized, and superheated in the HRSG. The steam is then delivered to the HPT at 180 bar and 549°C. After the expansion it is used to cool the burners and the HTT stages. The major part of the cycle medium, which is separated after the HRSG, is compressed using an intercooled compressor (C1/C2) and fed to the combustion chamber with a maximum temperature of 600°C. The detailed flow sheet used for the thermodynamic simulation is included in the appendix and gives mass flow, pressure, temperature, and enthalpy of all streams. In order to achieve a high thermal efficiency the heat extracted from a power cycle should be small compared to the input. The cycle arrangement of the Graz Cycle achieves this on the one hand by a very high peak temperature enabling large heat input, and on the other hand by feeding only the smallest possible mass flow of working fluid to the condenser (main heat extractor from the cycle) which has to contain the CO2 generated in the combustor. The major part of the working fluid is compressed in the gaseous phase and so takes its high heat content back to the combustion chamber. Graz Cycle for syngas from coal gasification The Graz Cycle is suited for all kinds of fossil fuels. Best results regarding net cycle efficiency can be obtained for syngas firing from coal gasification. For this investigation it was assumed that syngas is bought from an external gasification plant at elevated costs, so that the production effort was not considered in the thermodynamic balance. Syngas was provided at 500°C, because coal gasification takes place at very high temperatures. This temperature was chosen due to material restrictions. The syngas composition was typical for an oxygen blown coal gasification plant (syngas mole fractions: 0.1 CO2, 0.4 CO, 0.5 H2). Due to the higher carbon content of the fuel, the composition of the working fluid at HTT exit was 69% steam and 31% CO2 (mass fractions). Then, half of the cycle mass flow was expanded in the LPT and fed to the condenser, where the lower steam content led to a slightly higher optimum pressure of 0.05 bar. But in general the main cycle parameters did not change considerably. Power balance Table 1 gives the power balance of the Graz Cycle plant for methane and syngas firing for 143.8 MW heat input. For syngas two variants with syngas at 150°C and 500°C are given in order to better understand the differences to the methane fired version. The net cycle efficiency shown in the last row was calculated according to Equation 1. Table 1 Graz Cycle power balance
(1) 84
Franz Heitmeir, Wolfgang Sanz, and Herbert Jericha Looking at the methane-fired version of the Graz Cycle, we see that the HTT was the major turbomachinery component in the cycle. The thermal cycle efficiency was 66.3%, and accounting for the electrical, mechanical and auxiliary losses, the net electrical cycle efficiency was 64.8%, a value far higher than that which is typical of state-of-the-art combined cycle plants. If considering the efforts for oxygen production and compression to combustion pressure, a net efficiency of 54.6% could be evaluated. If the cycle were to be penalized for the CO2 compression to 100 bar needed for liquefaction, the net efficiency would further reduce to 52.5%, a value still higher than that of most alternative technologies. In comparison of the methane-fired version with the Graz Cycle fired with syngas provided at the same temperature as methane, i.e. 150°C, the turbomachinery power reduces due to the less steam content with a lower heat capacity. The cycle efficiency was slightly reduced by 0.2 %-points due to higher condenser losses. Syngas demands less oxygen per heat input, so that the penalty of oxygen supply decreases considerably. But this gain is partly offset by a larger amount of CO2 generated by syngas firing which then has to be compressed for liquefaction. Finally, the syngas fired version has a net cycle efficiency of 53.1%, 0.6 %-points higher than the methanefired version. If the heat of the syngas production can be used in the Graz Cycle plant for free (it is considered only in the fuel price), the net cycle efficiency would increase by 3 %-points up to 56.1%. Sensitivity study of HTT performance The significance of the thermodynamic simulation was based on the choice of reasonable data for component efficiency and losses. The two key parameters for the Graz Cycle were the HTT efficiency and HTT cooling mass flow because of the very high contribution of this turbine to the overall power generation. Figure 3 shows the influence of the HTT isentropic efficiency. The effect of an improved HTT efficiency was counteracted by the decreased HTT outlet temperature resulting in a decrease of the HPT power output. If we assume an HTT isentropic efficiency of 92% instead of 90.3%, the net cycle efficiency would reach only 53% instead of 53.8%, the value expected if we do not account for the above mentioned effect of the reduced HTT exit temperature on the overall cycle. On the other hand, the HTT cooling mass flow had a more significant influence on the cycle efficiency. It was estimated to be 13.7% using a model evaluated by comparison with conventional gas turbines, but a percentage-point increase in cooling mass flow decreased the net efficiency by 0.22 %-points. These considerations showed that the HTT performance had a decisive influence on the overall cycle efficiency.
Fig. 3. Influence of HTT isentropic efficiency on net cycle efficiency
Modifications of cycle configuration In order to improve the efficiency of the Graz Cycle, several modifications were investigated. The following cycle variants will be discussed in this work: condensation of the cycle working fluid at 1 bar and re-vaporization of the separated water; and, heat supply to the deaerator by the cooling heat of the CO2 compression intercoolers.
85
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency 2.4.1. Condensation at 1 bar and water re-vaporization The working fluid containing the mass flow of CO2 generated in the combustor was expanded in the LPT to a condenser pressure of 0.043 bar. There the steam was condensed, allowing separation of the gaseous CO2. But this configuration had some deficiencies: 1) The condenser was very large and thus expensive, because of the high volume flow and the anticipated reduced heat transfer due to the high inert gas content; and 2) The CO2 mass flow was expanded in the LPT together with steam and afterwards recompressed to atmosphere, so that due to the higher compression effort and additional losses a net loss was generated by this mass flow. Therefore it was suggested in the Austrian patent of the Graz Cycle to condense this mass flow at atmospheric pressure, separate the combustion CO2 and re-vaporize the water at a reduced pressure level using the condensation heat (see figure 4)17. The steam was then fed to the LPT and could be expanded to a condenser pressure lower than that for the working fluid mixture18. Advantages of this configuration are the avoidance of the difficult condenser for the working fluid at vacuum conditions and the saving of the relatively large CO2 compressors C3 and C4 needed for compression to atmosphere. Instead of using a standard condenser, an additional condensation/ re-vaporization unit was needed. This condensation/revaporization unit worked at atmospheric conditions and was similar to distillers used for conversion of sea and brackish water into high purity water by vacuum vapor compression.
Fig. 4. Scheme of condensation/re-vaporization
If the saving of CO2 compression power and the advantage of a lower condenser pressure exceeded the power loss of the LPT due to the reduced mass flow, a net gain in efficiency could be achieved. A thermodynamic study found an optimum for a dual pressure vaporization at the pressure levels 0.55 bar and 0.3 bar. The losses assumed for vaporization were 0.18 bar for the higher pressure level and only 0.07 bar for the lower pressure level. If these relatively low losses could be met, the efficiency for this new configuration would remain the same at 52.5%. So this configuration could lead to reduced plant costs and an even greater efficiency, if the original low condenser pressure cannot be kept for the working fluid condensation. As a second alternative currently investigated, the condensation heat could be utilized in a bottoming steam cycle. It has the advantage of more flexibility, of an easier start-up of the plant and has an easier water make-up. 2.4.2. Deaerator heating by CO2 compression intercoolers In order to remove dissolved gases (N2, O2 and CO2) in the HRSG feed water, a deaerator was arranged in front of the feed pump. Since there was no pure steam at an appropriate pressure available for heating, the feed water was heated close to saturation temperature in a surface heat exchanger that utilized the working fluid extracted in front of the LPT. This fluid passing by the LPT caused a reduction in its power output. To avoid this configuration and the resulting power reduction, it was investigated to supply the necessary heat for the deaerator from the CO2 compression coolers instead of the working fluid by passing the LPT (figure 5).
86
Franz Heitmeir, Wolfgang Sanz, and Herbert Jericha
Fig. 5. Scheme of deaerator supplied with heat from CO2 compression intercoolers
The thermodynamic simulation showed that the heat from the CO2 intercoolers can completely replace the extraction in front of the LPT. So the mass flow and thus the power output of the LPT increased by 8.5%, resulting in an increase of net cycle efficiency by 0.8 %-points up to 53.3%. This improvement showed that there was still room for efficiency improvement of the Graz Cycle, but often in trade-off with higher complexity.
1.3.1.1-3 Turbomachinery Design
87
Prototype plant of 75 MW output Compression and expansion in large power cycles can only be affected with modern turbomachinery. The gases we have to deal with in our case, CO2 and H2O steam, are very compressible at the given high enthalpy heads or pressure ratios. The resulting high changes in volumetric flow in the individual compressors and turbines require a multi-shaft arrangement connected by gears. The design decision to have the high temperature flow channel with minimum surface area and minimum heat loss and also with minimum cooling flow supply leads to the arrangement of turbomachinery as given in figure 6, which also includes the compressors C3 and C4 used to compress the separated CO2 to atmospheric pressure. There it is delivered to a final compression up to 100 bar for liquefaction. The HTT turbine needs 4 stages due to the high heat capacity of the steam-rich cycle medium. The HTT is split into two shafts, where the first stage runs at 23 000 rpm, the other three stages at 12 000 rpm. The two overhang disks of different speed provide the shortest possible high temperature annular flow channel. A bearing is arranged between the second and third stages. In order to reduce the number of generators, the power of all four compressors is balanced with the HTT first stage and the HPT. Both turbines drive the cycle medium compressors C1 and C2 and in normal operation they also drive the CO2 delivering compressors C3 and C4. These compressors are connected via a self-synchronizing clutch and are disconnected from the main high-speed shaft during start-up. Then they are driven by a separate electric motor in a mode similar to the vacuum pump in a steam plant. This arrangement needs two gear boxes, because the compressors C1 and C4 run at 12 000 rpm and the compressor C3 at 3 000 rpm. The stages 2, 3, and 4 of the HTT run at 12 000 rpm and deliver their power via the main gear to the generator, which is driven on the other side by the LPT in a way that is similar to very large steam turbines. The main turbomachinery data and their dimensions for a prototype plant are given in19. Due to the small volumetric flow of the HPT it is designed in the form of a 4-stage partial admission impulse steam turbine. Its arrangement immediately ahead of the HTT allows for cooling of the HTT first stage disk in an effective way. Exhaust steam is fed via labyrinth seals to the front side of the disk thus holding the shaft and the disk at a temperature of around 300°C. The disk is bell shaped with broad width in the center leading to a strong fir-tree root blade attachment which contains the cooling steam inlet ports to the hollow blades. On the other side, the space between the HTT first and second stage disks is again filled with cooling steam from outside, cooling both disks and providing in a form of a stationary steam bearing additional damping to both shafts. Again from here cooling steam is fed into the second disk and its blades. The compressors C1 and C2 have to act on a medium consisting of CO2 and steam. The high volume change requires a change of speed (C1 at 12 000 rpm, C2 at 23 000 rpm) with relatively high Mach numbers at the tip of their respective first blades. But relatively long lasting blades result in low clearance loss and low deterioration of the meridional flow profile. In order to keep the high-speed shaft short and in order to reduce the number of stages in C2 a radial final stage is proposed which can replace 3 or 4 axial stages due to its higher diameter and at the same time can deliver the medium radially outwards, making the inflow to the combustion chamber easier.
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency
Fig. 6. Schematic arrangement of turbomachinery for a 100 MW S-Graz Cycle power plant
Graz Cycle plant of 400 MW output For a Graz Cycle power plant of 400 MW net power output, a different but also very feasible design is suggested which works without gear boxes in the main power shafts. The turbomachinery arrangement consists of three independent shafts. The first shaft is a free running high speed shaft of 7600 or even 8500 rpm. It consists of the working fluid compressors C1 and C2 and the first two stages of the HTT turbine working as a compressor turbine. The turbine power is balanced with the power demand of the two compressors. The second shaft consists of the HTT power turbine and the low pressure turbine LPT. Both turbines run at 3000 rpm and drive the electrical generator. The third shaft consists of the CO2 compressors C3 and C4 and the HPT turbine. It runs at 3000 and preferably 7,600 rpm connected via a gear box. An additional motor/generator delivers power for start-up and is driven in full load by the HPT power which exceeds the demand of the compressors. Development work needed for a Graz Cycle plant Most components of the Graz Cycle are well known but they have to work with an unusual working fluid of steam and CO2. More critical components are as follows: • The combustor for a nearly stoichiometric combustion with oxygen and use of steam and CO2 as cooling medium; • The HTT with a working fluid of about three quarts of steam and one quart of CO2 and steam cooling; • The condenser, condensing steam in the presence of a high content of inert gas. All other components (LPT, HPT, all compressors, HRSG and heat exchangers) can be considered as standard components and do not pose any difficult design problems. Combustion chambers for firing oxygen with methane in a steam environment have already been tested in the USA, in Japan, and Europe20. Recent tests at the National Energy Technology Laboratory (NETL) were performed for a 1 MW combustor, working at 10 bar and an exit temperature of 1200°C21. The combustor was no more difficult to operate than a regular combustor and performed well in terms of CO generation, if a small oxygen surplus of 3% was provided. In summary, all investigations showed that the concept of oxy-fuel combustion using steam dilution is viable. The design of the HTT was studied carefully at Graz University of Technology and discussed at several conferences22. Recent studies by a main gas turbine manufacturer also confirmed the technical feasibility of the HTT, but experience has to be gathered for the behavior of the high-temperature alloys in the steam/ CO2 environment of HTT hot sections. The condenser has to deal with a large volume flow due to its very low pressure and the difficulty of reduced heat transfer in the presence of inert gas. But little experience has yet been gathered for steam condensation at a 20% CO2 content, so that very little data on heat transfer in this environment is available. Further research work is necessary and condensation at 1 bar as discussed above is a reasonable option if very large heat transfer surfaces are required.
1.3.1.1-4 Economic Evaluation Despite the high efficiency and the positive impact on the environment by a Graz Cycle power plant, a future application of this technology and an erection of a power plant mainly depend on the economic balance. The main indicator characterizing the economic performance of a power plant for CO2 capture is the mitigation costs. They represent the increased capital and operational costs incurred by new and additional equipment and lower cycle efficiencies in relation to the CO2 mass flow avoided. The CO2 captured has an economic value of about $10/ton, if it can be used for enhanced oil recovery (EOR) or of about $30/ton in the future CO2 emission trading scenario. These prices show the momentary threshold for the economic operation of zero emission power plants, although it is very difficult to foresee future trends.
88
Franz Heitmeir, Wolfgang Sanz, and Herbert Jericha In order to estimate the mitigation costs for a Graz Cycle plant, an economic comparison with a state-of-the-art combined cycle power plant of 58% efficiency is performed. The economic balance is based on the following assumptions: • The yearly operating hours is assumed at 8500 hrs/yr; • The capital charge rate is 12%/yr; • Methane fuel costs are 1.3 ¢/kWhth; • Syngas is supplied by a syngas producer at 3.5 ¢/kWhth, so no efficiency penalty for the production or additional investment costs are considered; • The investment costs per kW are the same for the reference plant of about 400 MW net power output and the Graz Cycle plant (see below); • Additional investment costs are assumed for the air separation unit (ASU), for additional equipment and CO2 compression to 100 bar (see Table 223); and, • The costs of CO2 transport and storage are not considered because they depend largely on the site of a power plant. Table 2 Estimated investment costs
The assumption of similar investment costs for a conventional and a Graz Cycle power plant is based on a comparison with typical turbomachinery sizes for a 400 MW combined cycle plant as given in table 3. It shows that the turbine power and the HRSG are of similar sizes, whereas the compressor power is remarkably smaller. On the other hand the Graz Cycle needs a larger generator due to the additional power consumption for ASU and CO2 compression. Developmental efforts are needed especially since the HTT and combustor were not considered in the investment costs. Table 3 Comparison of equipment size for a 400 MW plant in terms of power
89
Three indicators characterizing the economic performance of a power plant for CO2 capture are estimated: • The costs of electricity (COE) for both plants; • The differential COE representing the additional costs of electricity due to CO2 capture; • The mitigation or capture costs representing the additional costs incurred by CO2 capture per ton CO2 . Table 4 shows the result of the economic evaluation for methane and syngas firing, respectively. For syngas firing, the reference plant is also syngas-fired without considering an efficiency decrease. The syngas plant has slightly smaller additional investment costs because of the smaller ASU needed. Compared to the reference plant, the capital costs are about 60% – 70% higher by considering only the additional components for O2 generation and CO2 compression. So they contribute mostly to the difference in COE. The fuel costs have the major influence on the COE, especially for syngas firing, but they do not differ largely between reference and Graz Cycle plant. The O&M costs are assumed 15% higher for a Graz Cycle plant due to the operation of additional equipment.
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency Table 4 Economic data for methane and syngas fired Graz Cycle
Due to the more expensive fuel, the COE for syngas firing is by far larger than for methane firing (COE due to fuel). But regarding the differential COE, the difference is 0.74 ¢/kWhel for the methane-fired Graz Cycle and 0.68 ¢/kWhel for the syngas-fired version compared to the respective reference plant. But due to the higher carbon content in syngas, the mitigation costs are only $11.7/ton CO2 for the syngas plant compared to $21.6 /ton CO2 for the methane-fired plant. These values are clearly below the threshold value of $30/ ton showing the economic potential of the Graz Cycle. The results of the economic study depended mainly on the assumptions about investment costs, fuel costs and capital charge rate as well as on the choice of the reference plant. A cost sensitivity analysis was performed and showed that a variation of the capital costs had the main influence on the economics, since they contributed most to the mitigation costs24. Unfortunately, there was a large uncertainty of these costs. A survey of the ASU costs vary in the range of $230 to $400/kWel (the same price as for a complete power plant). Considering this variation solely, the mitigation costs varied between $21.6 and $29.0/ton CO2 for the methane-fired plant (see figure 7).
Fig. 7. Influence of capital costs on the mitigation costs (methane-fired Graz Cycle)
This high sensitivity to the capital costs showed the dilemma in performing an exact economic evaluation, since their estimation for a Graz Cycle power plant was very difficult because of the new turbomachinery components. But the authors claimed that their design of high-speed transonic stages with innovative steam cooling allowed a cost-effective manufacture. In these considerations about the height of additional investment costs, a further advantage of the Graz Cycle, the almost NOx-free combustion was not evaluated. According to exhaust flow NOx and CO catalytic reduction to achieve single-digit emissions (in strict attainment areas) can increase gas turbine genset plant costs by 40 to 50 percent25.
90
Franz Heitmeir, Wolfgang Sanz, and Herbert Jericha 1.3.1.1-5 Conclusions The Graz Cycle is an oxy-fuel power cycle with the capability of retaining all the combustion generated CO2 for further use. Its cycle configuration aims at highest efficiency by reducing the heat extraction in the condenser to a minimum. A thermodynamic investigation of the Graz Cycle fired with methane shows a net efficiency of 52.5%, if the efforts for oxygen supply and CO2 compression to liquefaction are considered. If syngas can be used from an external syngas plant at 500°C, efficiencies can rise up to 56%. Studies show that further efficiency improvements and simplification of the cycle are possible. A layout of all turbomachinery components for a 75 MW prototype plant as well as a 400 MW plant showed the technical feasibility of the Graz Cycle, although some development work is needed for the main components. But the authors claim that their proposed design of high-speed transonic stages with innovative steam cooling allows a cost-effective manufacture. In an economic analysis the Graz Cycle power plant is compared with a state-of-the-art combined cycle plant. The resulting mitigation costs of 22 $/ton CO2 are below a threshold value of 30 $/ton CO2 (assumed for CO2 emission trading), but this value mainly depends on the investment costs assumed. If syngas is used as fuel, the mitigation costs are only about 12 $/ton CO2 due to the higher carbon content of syngas. All investigations done up to now confirm the high efficiency and technical feasibility of the Graz Cycle. In the scenario of increasing costs of CO2 emissions, the investment in such a zero-emission power plant seems very reasonable in the near future.
1.3.1.1-6 Abbreviations and Appendix ASU COE HPT HRSG LPT
- air separation unit - cost of electricity - high pressure turbine - heat recovery steam generator - low pressure turbine
Appendix: Detailed thermodynamic cycle data of a Graz Cycle power plant fired with methane. 41.7 150
41.7 15 41.7 150
299.5 2.867
-17.8 0.002 Efficiencies
116.6 11.58
Cycle (e_m=1)
16 47.14
41.7 15 0
20.37 11.58
116.3 11.58
CO2 in %
0.2418
H2O in %
0.7528
O2 Verdichtung 40 1400
41.7 329.9
2.379 15
15 1091
3028 6.45
4038 77.5
-9.17 11.58 3029 19.32
41.7 329.9
3005 43.73
+O2-Pr./V +CO2
0.5254
180 549 195.3 373.0
15 210.2
46.4543
Net T hermal Power[MW]
95.3643
Cooling mass flow [kg/s]
9.7895
mk/m [%]
13.7788
dt_out [K]
185 554
C2
195.3 363.8
2439 22.66
195.3 363.8
2439 22.66
LPT
24.4998
1.053 578.5
2624 43.73
2120 43.73
dt_out [K]
203.6 367.2
1848 22.66
203.6 367.2
1848 22.66
206.8 159.7
686.2 22.66
0.2127
H2O in %
0.7824 1.013 191.1
0.043 28.21
1881 34.11
0.043 18
261.8 34.11
0.043 18
75.61 23.86
26 66 06 22.
37.11 2270
80.84 2760
1.013 191.1
2270 3
2595 80.84
5.0000
2270 34.11
23414 2. 66
195.7 373
461.2 1.053
1.053 372.2
CO2 in %
2987 80.84
191.1 1.013
13.3 396
2714 43.73
3332 80.84
28 16 3.3 39
3405 22.66
2612 22.66
C1
1.013 191.1
2269 43.73
0.043 18 1.013 191.1
2269 80.84
1 18.0
1 87.6 91.1 213
366.9 22.66
1 94.0
1 18.0
393.8 20.32
1 31.7
h[kJ/kgK] mass[kg/s]
0.25 176.3
132.8 2.338
1 31.7
132.7 2.338
884.1 10.25
0.2425 25
C3
148.8 7.881 137.1 1
75.78 20.32
0.2425 25 0.043 18.0
115.7 3
695.6 10.25
7 .541 5. 78 3
398.2 22. 66
1 31.7
91
141.8186
3028 22.66
HPT
p[bar] t[°C]
T urbine Power [MW] Compressor Power [MW]
3028 3.339 3 749.6
41.7 329.9
1.013 106.1
0.5462
HTT 41.7 329.9 41.7 599
13.7 443
0.6483
+O2-Prod./Verd.
4670 71.05
3028 12.87
41.7 329.9
0.6632
Electr. Efficiency
1 18.0
138.6 10.25
75.55 23.86
1 129.4
75.69 23.86
31.7 1
256.7 7.881
C4
0.6671 55.45
0.2425 25.0
104.9 2.373
2448 1 8 .5
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency 1.3.1.1-7 Notes
____________________________ 1. L. Strömberg, “Overview of CO2 Capture and Storage – Technology and Economics for Coal-Based Power Generation,” (VGB Congress 2003, Copenhagen). 2. R. Gabbrielli and R. Singh, “Thermodynamic Performance Analysis of New Gas Turbine Combined Cycles with no Emissions of Carbon Dioxide,” ASME Paper GT-2002-30117, (ASME Turbo Expo 2002, Amsterdam, The Netherlands). 3. L. Turanskyj and B.A. Keenan, “Turbomachinery for the World’s Largest Nitrogen Plant: Enhanced Oil Recovery to Increase the Output in the Cantarell Oil Field, Mexico,” (Paper at the Exposición Latinoamericana del Petróleo, Maracaibo, Venezuela [2001]). 4. P. Mathieu and E. Iantovski, “Highly Efficient Zero Emission CO2-based Power Plants,” Energy Conversion and Management 38 no. 1 (1997): S141-146; R. A. Anderson, S.E. Doyle, and K.L. Pronske, “Demonstration and Commercialisation of ZeroEmission Power Plant,” (29th Int. Techn. Conference on Coal Utilization & Fuel Systems, Clearwater [2004]). 5. See note 1 above. 6. H. Jericha et al., “ CO2 - Retention Capability of CH4/O2 – Fired Graz Cycle,” (CIMAC Conference Paper, Interlaken, Switzerland [1995]). 7. H. Jericha and M. Fesharaki, “The Graz Cycle – 1500°C Max Temperature Potential H2 – O2 Fired CO2 Capture with CH4 – O2 Firing,” ASME Paper 95-CTP-79, (ASME Cogen-Turbo Power Conference, Vienna, Austria [1995]). 8. H. Jericha, A. Lukasser, and W. Gatterbauer, “Der “Graz Cycle” für Industriekraftwerke gefeuert mit Brenngasen aus Kohleund Schwerölvergasung” (in German, VDI Berichte 1566, VDI Conference Essen, Germany [2000]). 9. Jericha, H., Göttlich, E., 2002, „Conceptual Design for an Industrial Prototype Graz Cycle Power Plant“, ASME Paper 2002 GT-30118, ASME Turbo Expo 2002, Amsterdam, The Netherlands; H. Jericha et al. „Design Optimisation of the Graz Cycle Prototype Plant,“ ASME Paper 2003-GT-38120, (ASME Turbo Expo 2003, Atlanta, USA and ASME Journal of Engineering for Gas Turbines and Power, October 2004, 126: 733-740); F. Heitmeir et al., “The Graz Cycle – A Zero Emission Power Plant of Highest Efficiency,” (XXXV Kraftwerkstechnisches Kolloquium, Dresden, Germany [2003]); F. Heitmeir and H. Jericha, 2005, “Turbomachinery Design for the Graz Cycle: An Optimized Power Plant Concept for CO2 Retention,” Proc. ImechE 219, Part A: J.Power and Energy (2005): 147-155. 10. W. Sanz et al., “Thermodynamic and Economic Investigation of an Improved Graz Cycle Power Plant for CO2 Capture,” ASME Paper GT2004-53722, (ASME Turbo Expo 2004, Vienna, Austria, accepted for publication in ASME Journal of Engineering for Gas Turbines and Power). 11. W. Sanz, et al., “A Further Step Towards a Graz Cycle Power Plant for CO2 Capture,” ASME Paper GT2005-68456, (ASME Turbo Expo 2005, Reno, USA). 12. SimTech Simulation Technology, “IpsePro Overview,” http://www.simtechnology.com/IPSEpro (accessed 2003). 13. W. Wagner and A. Kruse, Properties of Water and Steam (Springer-Verlag Berlin Heidelberg New York, 1998). 14. U. Sievers, „Die thermodynamischen Eigen-schaften von Kohlendioxid (in German),“ VDI Verlag, Düsseldorf. FortschrittBerichte VDI, Reihe 6, Energietechnik, Nr. 155 (1984). 15. K. Jordal, O. Bolland, and A. Klang, „Effects of Cooled Gas Turbine Modelling for the Semi-Closed O2/ CO2 Cycle with CO2 Capture,” ASME Paper 2003-GT-38067 (ASME Turbo Expo 2003, Atlanta, USA). 16. F.Luckel, „Weiterentwicklung des Graz Cycle und der Vergleich mit anderen CO2-Rückhaltekonzepten“ (in German, Diploma Thesis, Graz University of Technology [2004]). 17. H. Jericha and W. Sanz, “Wärmekraftanlagen mit Verbrennung von Kohlenwasserstoffen mit reinem Sauerstoff zur Stromerzeugung bei Rückhaltung von Kohlendioxyd” (in German, Austrian Patent No. AT 409 162 B [2001]). 18. See note 12 above. 19. See note 11 above. 20. B. T. Chorpening, K. H. Casleton, and G. A. Richards, „Stoichiometric Oxy-Fuel Combustion for Power Cycles With CO2 Sequestration,” (Proceedings of the Third Joint Meeting of the U.S. Sections of The Combustion Institute, Chicago, IL [2003]); B. T. Chorpening et al., “Demonstration of a Reheat Combustor for Power Productin With CO2 Sequestration,” ASME Paper 2003-GT-38511 (ASME Turbo Expo 2003, Atlanta, USA); H. Inoue, N. Kobayashi, and T. Koganezawa, “Research and Development of Methane-Oxygen Combustor for Carbon Dioxide Recovery Closed-Cycle Gas Turbine,” CIMAC, Hamburg, Germany (2001); T. Griffin et al., “Staged Catalytic Combustion Method for the AZEP,” ASME paper GT2004-54101 (ASME Turbo Expo, Vienna, Austria [2004]). 21. See note 21 (first two references). 22. H. Jericha et al., “Konstruktion der ersten Stufe der HTT-Gasturbine für den Graz Cycle” (in German), VDI Berichte 1857, VDI Tagung “Stationäre Gasturbinen: Fortschritte und Betriebserfahrungen”, Leverkusen (2004); also see note 10 (2nd and 3rd reference). 23. G. Göttlicher, G., „Energetik der Kohlendioxidrückhaltung in Kraftwerken“ (in German), Fortschritt-Berichte VDI, Reihe 6, Energietechnik, Nr. 421. Düsseldorf: VDI Verlag (1999). 24. See note 11 above. 25. Gas Turbine World, “2003 Handbook for Project Planning, Design and Construction,” Pequot Publishing Inc. (2003).
92
BIOGRAPHY
1.3.1.1 Graz Cycle – a Zero Emission Power Plant of Highest Efficiency
Franz Heitmeir Institute for Thermal Turbomachinery and Machine Dynamics Graz University of Technology, Graz, Austria email:
[email protected] http://www.ttm.tugraz.at
Professor Heitmeir studied Automotive and Airplane Engineering at the Munich College of Applied Engineering and Sciences (with excellence) and Mechanical Engineering and Aerospace Science at the Technical University Munich, Germany, (with excellence). In 1987 he got his PhD in Mechanical Engineering at the University of the Armed Forces in Neubiberg, Germany, (with excellence). The Doctoral thesis was about the burning rates of graphite in high enthalpy flows. From 1987 until 2001 he worked with MTU Munich, a leading gas turbine manufacturer in Germany. At MTU he had a long career in different positions in the divisions research, development and testing as well as in the marketing and sales division. In his position he was head of the two engine programs RB 199 (Tornado fighter airplane) and MTR 390 (Tiger helicopter) and at the same time head of development departments for these two engine programs. Since 2001 he has been head of the Institute for Thermal Turbomachinery and Machine Dynamics at the Graz University for Technology.
Wolfgang Sanz Institute for Thermal Turbomachinery and Machine Dynamics Graz University of Technology, Graz, Austria http://www.ttm.tugraz.at
Professor Sanz studied Mechanical Engineering and Economics in Graz. In 1989 he got his Diploma in Mechanical Engineering (with excellence), and in 1993 his PhD in Mechanical Engineering (with excellence), both at Graz University of Technology. Since 1990 he started working as an assistant at the Institute for Thermal Turbomachinery and Machine Dynamics. From 1994-1995, he was a visiting scientist at Naval Postgraduate School, Monterey, CA, where he worked with Max Platzer on unsteady aerodynamics. In 1998 he made his habilitation for “Thermal Turbomachinery” and became associate professor at Graz University of Technology. He is in charge of national funded projects on CFD and has published over 50 scientific papers, mainly on CFD and CO2 retention. He is a member of the ASME and the Cycle Innovations Committee of the International Gas Turbine Institute IGTI.
Herbert Jericha Institute for Thermal Turbomachinery and Machine Dynamics Graz University of Technology, Graz, Austria http://www.ttm.tugraz.at
Professor Jericha can look back over a working period of 51 years. His start as university assistant in 1954 allowed advanced studies in gas turbine technology at Farnborough, UK, leading to cooperation with the World Power Conference 1956 in Vienna. After his PhD in 1957 he worked in the US with Ingersoll Rand, where he, at that time, already worked on computer programs. Later at Elin Weiz, Austria, he became the leading designer of steam and gas turbines manufactured there. In 1970 Graz University of Technology called him to lead the new founded Institute for Thermal Turbomachinery and Machine Dynamics. By invention, theoretical work, and the establishment of a unique laboratory, he made it known world wide. Repeated authorship in ASME conferences and ASME IGTI contributions as European coordinator and liaison chairman were the path to this success. The most important design proposal is the so called Graz Cycle - a gas turbine system without any emissions.
Clean Energy Systems, Inc.
DOE Turbine Handbook
1.3 TURBINE - BASED ZERO EMISSIONS PLANTS 1.3.1 Oxy-Fuel 1.3.1.2 Clean Energy Systems 1.3.1.2.1 Introduction Clean Energy Systems, Inc. (CES) of Sacramento, CA and DOE’s National Energy Technology Laboratories (NETL) have developed and demonstrated unique technologies that will enable construction and operation of efficient zero-emission power plants (ZEPP). The enabling technologies are an oxy-fueled combustor developed under a DOE/NETL Vision 21 program, an oxy-fueled reheater (RH) designed by NETL and tested at a NASA test facility, and oxy-syngas combustor being developed under a DOE/NETL program. The CES process involves burning high purity oxygen with a hydrocarbon fuel, e.g., natural gas (NG), coal syngas, gasified biomass, etc., in the presence of water to generate a high pressure, high temperature gas comprising approximately 90 % steam, 10 % carbon dioxide (CO2), and a small amount of oxygen (O2). This gas is used to drive steam turbo-generators. CES power plants use cryogenic air separation units (ASU) to provide oxygen. These ASU plants can be made more efficient by the use of axial-flow-type compressors, typical of those found in gas turbines. This section discusses the integration of oxy-fueled combustors and reheaters with steam and gas turbines, gas turbine air compressors, a steam/CO2 condenser, and CO2 compressors/intercoolers. The resulting integrated ZEPPs produces power; generate high quality water, and conditioned CO2, ready for beneficial uses or sequestration. Key issues include ASU/gas turbine compressor flow matching, gas turbine blade cooling using steam rather than air, turbine material compatibility, and gas turbine temperature differences between steam/CO2 and air combustion. Various CEStype ZEPP concepts are illustrated and their performance characteristics defined for a range of operating conditions that are achievable with present day steam and gas turbines.
1.3.1.2.2 The CES Zero Emissions Power Plant Recent test programs by CES and DOE/NETL have successfully demonstrated the enabling combustion technologies of an oxy-fueled (NG) combustor 816 ºC, 104 bar (1500 ºF, 1500 psia)[ 1 , 2 , 3 , 4 , 5 , 6 ] and an oxy-fueled (NG) reheater 1200 ºC, 10 bar (2200 ºF, 145 psia)[ 7 , 8 , 9 ]. A simplified schematic diagram of a CES plant that incorporates these new components is shown in Figure 1.
The following discussion explores the integration of oxy-fueled combustion technologies with gas turbine and steam turbine technologies. CES power plants use the basic Rankine power cycle and consist of four basic systems as described by Martinez-Frias, et al.[ 10 ]. However, there are other alternative cycles that use CES technology and these are discussed in Section 1.3.1.2.8.
N2 Air
Air Separation Plant
* CH4, CO, H2, etc.
O
2
RH
Gas Generator
HP Crude Fuel
Fuel Processing Plant
Coal, Refinery Residues, or Biomass
NG, Oil or Landfill Gas
el * Fu Direct Sales
IP
LP
Elect Gen.
Multi-stage Turbines
HX
Recycle Water
CO2 Recovery
CO2
EOR, ECBM, or Sequestration
Cond. C.W.
Excess Water
Fig.1. The Basic CES Zero-Emissions Power Plant
1. Fuel Processing and Gas Compression: Gaseous fuel, derived from virtually any organic source, e.g., natural gas, gasified coal, biomass or refinery residue, is processed by cleansing any undesirable substances (e.g. sulfur, nitrogen, etc.) and compressed to the combustor injection pressures. 2. Air Separation and Oxygen Compression: Nearly pure oxygen is derived from large cryogenic air separation unit (ASU) and compressed to the combustor injection pressure.
A NG-fired CES plant typically comprises four subsystems: 1. Fuel processing and gas compression 2. Air separation and oxygen compression 3. Power generation (power-train) 4. Carbon dioxide separation and conditioning 3. Power Generation (Power-Train): The power generation system includes three turbines in series driven by high temperature gases consisting of approximately 90 %v steam, 10 %v CO2 and a small amount of oxygen. The excess oxygen suppresses CO2 dissociation and drives the combustion reactions to completion. High temperature gases are generated by an oxy-combustor at approximately 540 °C.– 760 ºC (1000 °F – 1400 °F) and with one or two reheaters operating at 1240 °C – 1760 ºC (2240 °F - 3200 °F). These gases drive multi-stage turbines. The turbines, in turn, drive an electric generator through a common or multiple shaft system, depending upon the selected plant configuration. 4. Carbon Dioxide Separation and Conditioning: This subsystem cools the turbine exhaust in a condenser at atmospheric or sub-atmospheric pressures to condense the steam and separate the CO2. Most of the condensed water is preheated in a feed water heater, located at the turbine exhaust, to recover any residual heat before recirculation back to the combustor. The separated humid CO2 exiting the condenser is extracted and compressed to approximately 145 bar (2100 psia) with multi-stage compressors. Intercoolers between stages remove the remaining water vapor and condition the CO2 for ultimate sequestration. A coal-fired ZEPP is similar to a NG-fired plant except it includes an oxygen-blown coal gasification and syngas cleanup and compression system in place of the NG processing and compression system. Such a power plant is described and shown diagrammatically by MartinezFrias, et. al.[ 11 ].
1.3.1.2.3 ASU and Turbine Compressor Matching CES and integrated coal gasification combined cycle (IGCC) power plants require large dedicated air separation units (ASU) to provide the oxygen required for combustion of the fuel or gasification of the coal. The capacities of ASUs required to support three types of 400 MWe power plants are in the range of 2600 - 6700 metric tons O2/day as shown in Table I. Table I. Typical ASU Plant Sizes Required for Three Types of 400 MWe Power Plants
Plant Type NG-fired CES ZEPP Coal Syngas-fired CES ZEPP IGCC with O2-blown coal gasifier
ASU Size, metric tons O2/day 6500 6700 2600
Cryogenic air separation is currently the most efficient and cost effective technology for producing large quantities of oxygen[ 12 ]. NG and coal syngas fired CES power plants larger than about 200 MWe require ASUs with air compressor capacities exceeding that of existing conventional industrial centrifugal and axial-centrifugal compressor equipment. Only large gas turbine compressors can provide the necessary air from a single unit. The capacity of present-day conventional ASUs is limited to about 3600 metric tons O2/day[ 13 ] and closely matches the
compressed air supply capabilities of a 6F class gas turbine. The very large 9F class gas turbines can meet the compressed air requirements of an ASU that produces about 12,000 metric O2/day[13] and could support a 700 MWe CES power plant. Also, the cost of the air compression system increases from 25% of the total ASU plant cost at 1800 metric tons O2/day to 35% at 3200 metric tons O2/day [12]. Therefore, the air compression system is a key cost element of an ASU. Because large ASU’s require large air compressors available in gas turbines, the economics of large power plants favor the integration of relatively lower cost gas turbine/air compressor unit into CES/ASU systems. Integration eliminates the need for large gearboxes, electric motors, or steam drive motors, and other associated equipment, including the electric motor starting apparatus which can approach the cost and size of the motor itself[ 14 ]. Also, integration simplifies the controls and the control system for improved plant reliability. Gas turbines are available in a wide range of capacities[ 15 ]. Typical gas turbines ranging in compressed airflow rates of 40 - 600 kg/sec (88 -1320 lb/sec) are listed in Table II along with the approximate sizes of equivalent ASU and CES ZEPP power plants such turbines could support. Table II. Typical Gas Turbine Compressor Capacities versus Equivalent ASU and ZEPP Sizes
Gas Turbine M fg Alstom/Cyclone GE/LM 1600 GE/LM 2500 P&W /FT-8 GE/LM 6000 RR/Trent SW /V64.3A GE/7EA GE/7FA GE&SW /H, Al.GT26
Comp. Cap., kg air/sec 40 50 86 86 130 175 200 295 455 600
Equiv. ASU & ZEPP Sizes Plant Size, M W e ASU M etricTons/Day at 42% Eff 800 50 1000 60 1,800 100 1,800 100 2,600 150 3,500 200 4,000 230 6,000 350 9,100 530 12,000 700
From Table II it can be seen that current commercial gas turbine compressors can potentially accommodate integration of ASUs and CES power plants with capacities ranging from approximately 790-12,000 metric tons O2/day and about 50 to 700 MWe, respectively
1.3.1.2.4 Effect of Drive Gases on Gas Turbine Operating Parameters Analysis was performed to define the comparative nominal operating parameters of existing aero-derivative and industrial gas turbines driven by normal air-breathing combustion gases and by CES combustion gases. The assessment involved one-dimensional analyses and the assumptions listed below. A more thorough assessment was conducted by Fern Engineering[ 16 ] , however, further work should be conducted by turbine manufacturing teams that possess precise turbine design information (e.g., exact blade angles, stresses, materials and design limits). Turbine manufacturers are most qualified to accurately determine the design and off-design performance when substituting CES drive gases for air-breathing combustion drive gases.
Assumptions: • Inlet pressure unchanged • Pressure ratio unchanged • Design based on 50% reaction • Blade cooling flow rate equal to or less than design • Inlet temperature equal to or less than design value • Turbine efficiency unchanged and approximately equal to 90% • Blade cooling effectiveness equal to 100% (transpiration cooling) • Blade coolant temperature equal to compressor discharge temperature for air cooling ~510 ºC (~ 950 ºF) aero-derivative and ~230 ºC (~ 450 ºF) industrial) and 230-270 ºC (450-500 ºF) slightly superheated for steam cooling. Criteria for validation: • Aerodynamic similarity is nearly maintained (blade angles and Mach No.) • Operating speed is within the recommended range • Heat transfer and blade temperatures are equal to or less than estimated design limits • Blade root stresses remain approximately the same [proportional to (speed)2 and torque] Because both aero-derivative (high-pressure-ratio) and industrial (low-pressure-ratio) turbines are widely used, the following two designs were considered representative of units for medium size 150 MWe and large 700 MWe ZEPP plants. Baseline Aero-derivative Turbine (150 MWe ZEPP): • • • • • • •
Two-stage design Inlet pressure = 29.93 bar (434 psia) Design Speed = 9586 rpm Design flow rate = 126.0 kg/sec (277 lb/sec) Compressor pressure ratio = 29.4 Cooling air flow rate 9.31 % of main gas flow
• • • • • •
Inlet temperature = 1245 ºC(2273 ºF) Exit pressure = 7.03 bar (102 psia) Speed range = 9500 to 10,800 rpm Turbine mean diameter 77.47 cm (30.5 inches) Cooling air temperature 504 ºC(940 ºF)
Inlet Temperature = 1427 ºC (2600 ºF) Exit Pressure = 1.10 bar (16 psia) Design flow rate = 583 kg/sec (1282 lb/sec) Turbine mean dia. = 211 cm (83 inches)
Baseline Industrial Gas Turbine (700 MWe ZEPP): • • •
Four stage design Inlet Pressure = 19.31 bar (280 psia) Compressor pressure ratio = 19.1
• • •
• •
Design speed = 3600 rpm Cooling air 232 ºC (450 ºF) flow rate 4.8 % of turbine design flow rate.
•
The resulting analyses, comparing nominal operating parameters using air-breathing combustion gases or CES gases at baseline and at a lower temperature, are shown in Table III for both an
aero-derivative and an industrial gas turbine. Various typical operating parameters for the first stage of the turbines and the exit temperature from the last stage are illustrated in the table. The operating parameters for the typical aero-derivative turbine given in Table III shows that parameter matching with the different drive gases is favored by increasing turbine speed by about 12-14 % when using CES drive gases. This speed increase permits close fluid flow angle matching and, when the gas inlet temperature is also decreased slightly, fluid flow angles coincide with the baseline case and last stage exit gas temperature closely approximates the baseline case. It can also be seen that replacing air-breathing combustion gases with CES gases provide 11-16 % higher 1st-stage power output and lowers turbine coolant flow rate by 70-75%. This reduction in flow rate is due to changing the blade coolant from air to steam.
Table III. Operating Parameters of a Typical Aero-derivative Turbine and a Typical Industrial Gas Turbine with Air-Breathing and CES Drive Gases
Parameter Turbine Stage Inlet Gas Temp., ºC (ºF) Inlet Pressure, bar (psia) Exit Pressure, bar ( psia) Exit Gas Temp., ºC (°F) Weight Flow, kg/sec (lb/sec) Speed, rpm Power/Stage, MW Coolant Temp., ºC (ºF) Coolant Cp, kJ/kg-ºC (Btu/lb-°F) Coolant Flow, % gas flow Blade Temp., ºC (ºF) Nozzle Exit Vel.,m/sec (ft/sec) Rotor Exit Vel., m/sec (ft/sec) Mean Blade Speed, m/se (ft/sec) Main Gas Cp, kJ/kg-ºC (Btu/lb-°F) Specific Heat Ratio Nozzle Incid. Angle, Deg. Rotor Incid. Angle, Deg. Last Stage Exit Temp., ºC (°F)
Aero-derivative Turbine Industrial Gas Turbine Air-Breath CES Gases Air-Breath CES Gases st st st st 1 1 1 1 1st 1245 (2273) 1245 (2273) 1121 (2050) 1427 (2600) 1427 (2600) 1427 (2100) 29.93 (434) 29.93 (434) 29.93 (434) 19.31 (280) 19.31 (280) 19.31 (280) 14.55 (211) 14.55 (211) 14.55 (211) 9.24 (134) 9.24 (134) 9.24 (134) 1026 (1878) 1084 (1983) 973(1784) 1176(2149) 1248(2279) 999(1831) 126.0 (277.1) 104.7 (230.3) 109.2 (240.3) 582.7 (1282) 485.0 (1067) 531.4 (1169) 9,586 10,717 10,858 3600 3600 3600 27.64 32.09 30.75 145.8 167.6 153.7 504 (940) 260 (500) 232 (450) 232 (450) 232 (450) 232 (450) 1.033 (0.247) 2.395 (0.551) 2.305 (0.551) 1.033 (0.247) 2.305 (0.551) 2.305 (0.551) 9.31 2.78 2.23 4.80 3.90 1.90 816 (1500) 816 (1500) 816 (1500) 816 (1500) 816 (1500) 816 (1500) 699 (2293) 824 (2703) 792 2597) 744 (2442) 907 (2975) 802 (2632) 435 (1423) 455 (1494) 491 (1611) 460 (1510) 597 (1939) 498 (1633) 389 (1276) 435 (1426) 440 (1445) 396 (1300) 396 (1300) 396 (1300) 1.230 (0.294) 2.385 (0.570) 2.343 (0.560) 1.230 (0.294) 2.385 (0.570) 2.343 (0.560) 1.32 1.21 1.21 1.32 1.20 1.20 0 -4.7 0 0 -9.0 -3.8 0 -2.3 0 0 -6.9 -2.8 833 (1532) 937 (1719) 834 (1534) 637 (1178) 829 (1524) 649 (1200)
The operating parameters for the typical industrial turbine given in Table III shows that when turbine speed is set by the generator speed, the change in fluid angles is somewhat larger when switching to CES drive gases, but is considered to be within the range of capabilities of a highefficiency, reaction-type turbine. Alternatively, reducing the turbine inlet temperature at constant speed provides closer matching of fluid flow angles at the nozzle and rotor exits and last-stage gas exit temperature. At similar turbine inlet temperatures and speeds, replacing air-breathing combustion gases with CES gases provide 15 % higher 1st-stage power output and lowers blade coolant flow about 20% while maintaining constant turbine blade operating temperature. When the turbine inlet temperature is reduced 260 °C (500 °F) and speed is maintained constant, 1ststage power output is still increased about 5 % over the baseline and turbine blade coolant flow is reduced by 60%. A subsequent study by Fern Engineering[16] on a similar, but slightly smaller aeroderivative gas turbine, resulted in the following conclusions: At the design “firing temperature” of 1280 ° (2336 ºF), the CES cycle yields: • Slightly higher power output (~6%) • A lower overall turbine pressure ratio (17.6 vs. 20.3) • Much lower mass flow of working fluid • Significantly cooler HP turbine nozzle metal temperature • Slightly lower power turbine inlet pressure • Slightly hotter power turbine inlet temperature, but a cooler power turbine nozzle metal temperature due to the use of steam cooling • Smaller turbine jet velocity ratios => slightly lower turbine efficiencies • Much hotter power turbine exhaust temp
1.3.1.2.5 Effect of Coolant on Gas Turbine Blade Temperatures Heat transfer study results based on models similar to those of DePaepe and Dick[ 17 , 18 ] are presented in Table IV and Figure 2. Table IV shows the 1st-stage blade baseline temperature of 816 °C (1500 ºF) in a high-pressure-ratio aero-derivative type turbine can be reduced by 246 °C (475 ºF)(1) using steam at 260 °C (500 ºF) for cooling rather than air at 504 °C (940 ºF) at constant volumetric flow rates. Alternatively, the inlet temperature of CES gases to the turbine could be increased approximately 166 °C (330 ºF) when using 260 °C (500 ºF) steam at constant volumetric coolant flow rates and still decrease turbine blade temperature nearly 204 °C (400 ºF) compared to baseline air-breathing case. The lower blade operating temperature, using steam as (1) The study by Reference[16] calculated the temperature reduction as 143 °C (290 ºF) and felt the 246 °C (475 ºF) prediction by CES was too optimistic. However, CES assumed a transpiration cooling effectiveness of 1.0 while Reference[16] assumed a factor 0.58. Measured blade surface temperatures by reference[ 19 ] for 1/2 scale model tests indicated the cooling effectiveness ranged from 0 .60 to 0.80 , for a steam cooled stator and rotor blade operating with 1700 °C (3092 ºF) hot gases(steam) at (355 psia). The measured blade metal temperatures were in the range of 595 to 816 °C (1100 to 1500 ° F).
Table IV. Comparison of Turbine Blade Temperature at Constant Coolant Flow Rates
Coolant Temp., Coolant Flow Rate, Type of Turbine Gas Temp., Blade Temp., Tb, °C (ºF) Coolant Tc,°C (ºF) and Drive Gas Tg, °C (ºF) m3/min (ft3/min) Aeroder.-Air 1245 (2273) 816,(1500) Air 504 (940) 0.872 (30.8) Aeroder.-CES 1245 (2273) 552 (1025) Steam 260 (500) 0.872 (30.8) Aeroder.-CES 1427 (2600) 599 (1110) Steam 260 (500) 0.872 (30.8) 1300 Baseline Air Breathing
1200 Temperature, °C
CES Drive Gas
1100
1000
900
CES Drive Gas at Reduced Inlet Temp. CES Drive Gas + 232 °C Steam Inj.
800 Entropy (no scale) Fig. 2. Gas Temperature Across Two Stages of a Typical Aero-Derivative Turbine
coolant, could result in increased blade stress safety margins and increased life or permit higher gas inlet temperatures, the use of simpler coolant passage designs, or possibly lower-cost blades. The Japanese[ 20 ] investigated various cooling methods for an intermediate pressure steam turbine operating with1700 °C (3100 ºF) steam while using open loop and closed loop cooling circuits with water and 300 °C (572 ºF) steam. These studies indicated the following cooling losses for a 500MW steam plant: (1) closed-circuit water cooling of the combustor, nozzle and stators vanes, and steam cooling of the rotor blades (CCWCN-SCR) has the lowest cycle loss of 10MW; (2) closed-circuit steam cooling of the nozzles and rotor blades (CCSCN-R) has a 21 MW loss; and (3) open-circuit steam cooling for nozzle and rotor blades (OCSCN-R) has a 52 MW loss.
1.3.1.2.6 Gas Turbine Operation with CES Gases versus Air-Breathing Gases The temperature drops across turbine stages change when the nature of the drive gas changes from air-breathing combustion gases to the gases produced by CES gas generators or reheaters because the specific heat ratios of these gases varies from 1.32 to 1.20, Table III. This effect is shown in Figure 2. Where temperature drops are compared for a typical two-stage aeroderivative turbine. Comparing the baseline air-cooled air-breathing case with the CES drive gas case, each with an assumed turbine inlet temperature of 1245 °C (2273 °F), the turbine exit gas temperature for the CES drive gas is higher than for the baseline case by 88 °C (158 °F). However, this difference is negated or markedly reduced when open-loop steam cooling is considered. Assuming steam cooling with 232 °C (418 °F) steam, the temperature drop can be made to coincide with the baseline air-cooled air-breathing case by additional steam injection. Alternatively, the turbine inlet temperature of the CES drive gas can be reduced and made to coincide with the baseline air-cooled air-breathing case. From the preceding examples it can be seen that stage-wise temperature drops or exit temperatures of air-cooled air-breathing turbines can be replicated when the turbines are switched to CES gases with open-loop steam cooling operating practice requires the use of more expensive materials and/or fabrication techniques, the technology base is well established. The major hindrance to increasing steam turbine operating temperatures has been with boiler limitations rather than with turbine limitations. The CES gas generator in CES ZEPP’s removes the boiler and its temperature constraints. Intermediate pressure turbines power plants would not operate at temperatures beyond current commercial gas turbine practice and may operate at even lower blade temperatures because very effective open-loop steam cooling becomes practical.
1.3.1.2.7 Turbine Materials Issues The turbine materials issues addressed in this section concern steam and gas turbine materials operating in high temperature, high-pressure steam environments. The issues primarily involve: 1) matching materials mechanical properties with turbine operating temperatures and stresses and (2) defining materials that are compatible with CES gases (~90% steam, ~10% CO2, and a slight amount of oxygen) at high temperatures and pressures and with weak carbonic acid condensate. Typical materials for various steam and gas turbine components are listed in Table V. Increasing steam turbine operating temperatures beyond ~ 540 to 565 °C (~1000 to1050 °F) requires the use of materials such as high chromium-ferritic steels and austenitic stainless steels[ 21 ]. The temperature limits for this type of materials is near 649 °C (1200 °F) but may extend to slightly higher values. For even higher metal operating temperatures, i.e., 816 °C (1500 °F) and above, high nickel and cobalt alloys become necessary and turbine blades may require even more exotic single-crystal materials. Although increasing steam turbine operating temperatures above current operating practice requires the use of more expensive materials and/or fabrication techniques, the technology base is well established. The major hindrance to increasing steam turbine operating temperatures has been with boiler limitations rather than with turbine limitations. The CES gas generator in CES
Table V. Typical Materials of Construction for Steam and Gas Turbine Components
Component Casings Comb. Liner Transition Duct Discs
Eddystone Steam Turbine 593 - 649°C (1100-1200 °F)[ 22 ] St. 316 (inner), 2 ¼ % Cr-steel (outer) N.A. St. 316 Discalloy
Vanes and Blades K42B, St. 422 (blades) St. 316 (nozzles)
Typical Alloys Solar 816°C, 103 bar (1500 °F, 1500 psia) Steam Turbine[ 23 ] Inconel 939 (inner), 2 ¼ % Cr-steel (outer) N.A Inconel 617 Inconel 718 Inconel 718 (blades), Inconel 939 (nozzles)
Gas Turbines[ 24 , 25 , 26 , 27 ] Inconel 718, Rene’41, Hastelloy X, Haynes 188 Hastelloy X, Haynes 188 Inconel 617, Inconel 939 Inconel 718, Inconel 738, Waspaloy, Udimet 700 Inconel 713C, 718, & 738, Rene’80, Udimet 500 & 700, CMSX-4, FSX-414, M-252, Multimet (N-155)
ZEPP’s removes the boiler and its temperature constraints. Intermediate pressure turbines in CES power plants would not operate at temperatures beyond current commercial gas turbine practice and may operate at even lower blade temperatures because very effective open-loop steam cooling becomes practical. The materials used in both current and advanced steam turbines have demonstrated good compatibility with pure steam but the effects of CES gases, which also contain a minor amount of CO2 and a small amount of oxygen, are not well established. Preliminary compatibility studies with simulated CES gases and typical turbine materials are underway. No significant problems have been encountered in the absence of an aqueous liquid phase but more comprehensive work is required, including the effects of carbonic acid formation in regions subject to condensation. The compatibility of a number of nickel and/or cobalt based alloys with steam at 816 °C, 103 bar (1500 °F and 1500 psia) has been studied by Solar Turbines, Inc. and compared with the behavior of the same alloys in air at 816 °C, 1.01 bar (1500 °F, 14.7 psia). The results of 1000hour tests are summarized in Table VI. The data in Table VI show that the high temperature, high-pressure steam attacks the alloys to a relatively minor extent and with few exceptions similar to the attack of air at a lower pressure. The tests in the steam environment were extended to 4000-hour exposures to provide a firmer basis for selecting materials for a steam turbine (see Table V) that operated successfully at 816 °C, 103 bar (1500 °F, 1500 psia). The results of these latter tests are summarized in Table VII. The data in Tables VI and VII, along with the experience gained with gas turbines operating in an oxidizing environment at moderate pressures and very high temperatures, suggest that CES
Table VI. Metallographic Measurements on Alloys Exposed for 1000 Hours to Steam at 816 °C , 103 bar (1500 °F, 1500 psia) Compared to Air at 816 °C, 1.01 bar (1500 °F, 14.7 psia) [23]
Alloy Inconel 718 Inconel 625 Inconel 800 Hastelloy X Hastelloy S Waspaloy
Outer Scale Thickness, microns Steam Air 3.8-7.6 7.6-17.8 2.5-10.2 2.5-6.4 4.1-6.1 5.1-10.2 0.0-2.5 6.4-22.9 2.5-7.6 2.5-5.1 2.5-5.1 5.1-12.7
Depth of Internal Oxidation, microns Steam Air 10.2-20.3 5.1-15.2 0.0-20.3 0.0–10.2 10.2-20.3 3.8-25.4 0.0-16,5 0.0- 5.1 0.0-6.4 3.8-12.7 7.6-20.3 12.7-33.0
Depth of Alloy Depletion, microns Steam Air 10.2-22.9 0.5-20.3 30.5-40.6 10.2 15.2 0.0 20.3-50.8 0.0 6.1-20.3 0.0 5.1-10.2 15.2-30.5 12.7-25.4
Table VII. Metallographic Measurements on Alloys Exposed for 4000 Hours to Steam At 816 °C, 103 bar (1500 °F, 1500 psia)[ 28 ]
Alloy Inconel 718 Inconel 713 Inconel 625 Inconel 617 Incoloy 800 Hastelloy X Waspaloy
Outer Scale Thickness, microns Nil Nil Nil 0.0-2.5 1.3-3.8 1.0-2.0 9.4-15.7
Depth of Internal Oxidation, microns 5.1-55.9 12.7-24.4 6.4-12.7 12.7-30.5 0.0-6.4 12.7-16.5 31.8-44.5
Depth of Alloy Depletion, microns 15.7-50.8 25.4-38.1 19.1-20.3 30.5-40.6 2.5-3.8 6.1-10.2 63.5-81.3
gases will not pose major compatibility problems when used to drive gas turbines. This outlook remains, however, to be demonstrated.
1.3.1.2.8 Integrated Plant Concepts Integration of CES’ technology with ASU’s, gas turbines, steam turbines and CO2 conditioning equipment to build environmentally friendly zero emissions power plants appears feasible and highly beneficial[ 29 ]. Since CES ZEPP plants involve a number of subsystems, there are many possible concepts. The following two configurations are representative of only a few of the many possible combinations. ZEPP #1, shown in Figure 3, consists of a high pressure 80-100 bar (1200-1500 psia) oxycombustor feeding a high pressure steam/CO2 turbine (HPT) at 600-760 °C (1100-1400 ºF); and an intermediate pressure reheat combustor at 30-40 bar (430-600 psia) feeding an intermediate pressure steam/CO2 turbine (IPT) at 1240-1760 °C (2240-3200 ºF). The IPT exhausts to subatmospheric pressures in the range of 0.15-0.4 bar (2.2-5.8 psia). Residual heat in the LPT exhaust is used to raise cooling steam for the IPT, and to preheat combustor feedwater.
Fuel
Combustor
Nitrogen
ASU
HPT Oxygen Reheat Combustor
Air
IPT Feedwater
FWH
CO2
Condensate
Cond
Fig. 3. Process Flow Diagram of ZEPP #1
ZEPP #2, shown in Figure 4, consists of an intermediate pressure 30-40 bar (430-600 psia) oxycombustor feeding an intermediate pressure steam/CO2 turbine (IPT) at 1240-1760 °C (22403200 ºF) that exhausts to approximately 1 atm. (14.7 psia). The IPT exhaust stream enters a HRSG (heat recovery steam generator) that raises high-pressure steam for a back-pressure HP steam turbine (HPT). Most of the HPT steam exhaust is delivered to the combustor as diluent, and some is used as cooling steam for the IPT. If desired, the HRSG may be fired with an oxyfuel burner to generate additional HP steam. Finally, some of the latent heat in the IPT exhaust is recovered by raising sub-atmospheric steam for a low-pressure steam turbine (LPT).
1.3.1.2.9 Performance Earlier cycle analyses were made by CES and other organizations, using a variety of modeling tools. These include: (1) CES’s in-house code; (2) the commercially available AspenPlus® soft
Feedwater
Nitrogen
ASU
Fuel
Combustor IPT
Oxygen Oxy-Fuel Burner (optional)
Air
IP Steam HP HPT Steam
HP Feedwater
HRSG LP Steam
LP Feedwater
LPT Condensate
Cond
CO2 To condenser
Fig. 4. Process Flow Diagram of ZEPP #2
ware; (3) the Lawrence Livermore National Laboratory’s (LLNL) program developed by Martinez-Frias[10,11] using Engineering Equation Solver (EES) software; and (4) Gates used by Fern Engineering[16]. All four codes were checked against each other for verification when applied to a CES power plant test case and all showed close agreement, provided the same process assumptions were made. Recent analyses have been made by CES personnel for both natural gas and coal-based plants, using the AspenPlus® software combined with updated process information for the ASU, steam/CO2 turbines, steam turbines, CO2 compression system, and gasifier (for coal-based systems). These analyses revealed that the efficiencies of the ZEPP #1 and ZEPP #2 cycles are very similar, provided the ZEPP #2 cycles include supplemental HRSG firing to boost the HP steam flowrate to the corresponding HP steam/CO2 flowrates in the ZEPP #1 cycles. Table VIII lists the key assumptions made in these analyses.
Table VIII. Key Assumptions for Natural Gas and Coal Cycle Analyses
ASU auxiliary load IP turbine isentropic efficiency Temperature of IPT cooling steam Steam turbine isentropic efficiency Compressor isentropic efficiency Compressor per-stage pressure ratio Compressor inter-cooler temperature Condenser temperature HRSG pinch-point Turbine shaft losses Turbine generator losses Compressor motor losses
0.20 kWh/kg O2 91% 380 ºC 88% 82% 2.5 31 ºC 31 ºC 20K 1% 1% 5%
Natural Gas Systems
With this set of assumptions, CES personnel calculated the expected efficiencies of various nearand long-term natural gas-fired CES plants. The key variables were (1) the HP and IP turbine inlet temperatures, and (2) the quantity of steam required for IP turbine cooling. Table IX lists the expected LHV cycle efficiencies for HPT/IPT inlet temperatures of 620/1240ºC, 620/1450ºC, and 760/1760ºC; and open-loop cooling steam flows of 10, 20 and 30%. Table IX. LHV Cycle Efficiencies for CES Natural Gas-Fired Plants
HPT/IPT Inlet Temperatures 620ºC/1240ºC 620ºC/1450ºC 760ºC/1760ºC
IPT Cooling Steam Flow (% of turbine inlet flow) 0% 10% 20% 30% 43.0% 41.7% 40.8% 40.0% 46.0% 44.8% 43.6% 51.0% 49.3% 48.0%
The various HPT/IPT inlet temperatures, which were provided by Siemens, represent sequential advancements that may be made through 2015. Since these cycle studies did not include a detailed analysis of the IPT cooling requirements for each case, a wide range of IPT cooling steam flowrates were considered. Also, all IPT cooling was assumed to be by open-loop steam cooling, where the steam is injected into the drive gas. As shown in the table, the flowrate of cooling steam has a significant impact on the cycle efficiency since it lowers the effective inlet temperature of the drive gas, particularly at the higher turbine inlet temperatures. This highlights the importance of optimizing the turbine cooling methodology to minimize its impact on the cycle performance.
Coal-Based Systems
Under award DE-FC26-05NT42645 (“Coal-Based Oxy-Fuel System Evaluation and Combustor Development”), CES personnel performed detailed cycle analyses on coal-based plants where an Illinois #6 coal is gasified and the clean syngas used as fuel in the CES oxy-combustors. These are referred to as IGCES (integrated gasification CES) plants. The studies incorporated input and interface information from Siemens (turbines), Air Products (ASU), Future-Energy (gasifier), and MAN-Turbo (compressors). Near-term (2010) and long-term (2015) cases were considered, each of which had its own set of assumptions. These assumptions, also provided by Siemens, appear in Table X. Table X. Key Turbine Assumptions for Near- and Long-Term Coal-Based IGCES Plants
Parameter HPT inlet temperature IPT inlet temperature IPT cooling steam flowrate IPT exhaust pressure
Near-term (2010) 620ºC 1450ºC 25% 1.0 bar
Long-term (2015) 760ºC 1760ºC 15% 0.15 bar
This analysis was performed in more detail than the previous natural gas analyses as the study included an extensive information exchange with key equipment suppliers, particularly Siemens. For this reason, the analysis focused on a fewer number of cases than the natural gas study. The cycle analyses revealed that the performance of an IGCES plant is sensitive to the gasifier configuration, particularly the mode of heat recovery from the hot syngas stream. Most commercial gasifiers provide the option of cooling the hot syngas in a syngas cooler which raises saturated HP steam at pressures in the range of 100 bar (1,500 psia). In the ZEPP #1 cycle, the steam from the syngas cooler can be injected into the HP oxy-combustor, heated and expanded through the HPT, reheated in the reheat combustor, and expanded through the IPT. This represents efficient use of the steam. An alternate approach is to quench the hot syngas via water injection to produce a cooler, saturated syngas stream. Some of the latent heat in this stream may be recovered by raising steam at a lower pressure than the partial pressure of moisture in the syngas, and expanding this steam through an LP turbine. Although this is a simpler technique with lower capital costs, the energy losses associated with syngas quenching have a negative impact on the overall cycle performance. Table XI lists the expected HHV cycle efficiencies for the near-term and long-term cases, for plants with either syngas cooler or syngas quench systems.
Table XI. HHV cycle efficiencies for CES coal-based plants
Availability Near-term cycle (2010) Long-term cycle (2015)
Syngas Heat Recovery Method Quench System Syngas Cooler 27.2% 30.0% 34.2% 37.0%
As shown in the table, technical advancements that may be made by 2015 will have a significant impact on the cycle performance. Also, a syngas cooler is preferable to a syngas quench system to maximize cycle efficiency.
1.3.1.2.10 Conclusions The use of modified IP (gas) turbines along with steam turbines in CES power plants enables high-efficiency, near-zero power generation. In some applications, the gas turbine compressor could be used as the air supply source for an ASU. This integration of systems eliminates the need for large electric drive motors, gearboxes, etc. to drive the compressors and, therefore, could significantly reduces plant capital cost and plant operational and maintenance costs. The high capacities of gas turbine compressors also permits construction of larger single train ASU’s that could support integrated CES-type ZEPP’s. Gas turbines operating with CES drive gases have lower temperature drops per stage and this results in higher temperatures for the later stages. To alleviate these higher temperatures, additional steam at 204-260 °C (400–500 ºF) could be injected at each stage to reduce the temperature to that compatible with the stage. Alternatively as a temporary solution, turbine inlet temperature could be reduced 93 °C (167 ºF) for aero-derivative and 260 °C (470 ºF) for industrial turbines with consequent small reductions in plant efficiencies. Matching aero-derivative gas turbines with CES drive gas is more easily accomplished when the turbine speed can be increased by about 12% and the inlet temperature is reduced by 93 °C (167 ºF). These changes permit almost exact fluid angle matching. Matching industrial gas turbines with CES drive gas where the turbine speed is set by the generator speed, i.e., 3600 rpm, causes modest changes in fluid angles. Reducing the inlet temperature by 260 °C (470 ºF) minimizes this mismatch and allows the fluid angles to remain in the range of high efficiency turbine operation. Ultimately, gas turbines with CES drive gases should be able to operate at 1450 °C (2640 ºF), and higher temperatures, using water and warm steam for stationary components such as GG/RH, transition sections, nozzles, stators, etc., (this could be a separate closed circuit that operates before start-up and after shut-down) and the more effective CES open-loop transpiration steam cooling for rotating components such as rotor disks and blades. Blade temperatures of both high-pressure-ratio aero-derivative type turbines and low-pressureratio industrial gas turbines can be reduced appreciably using open-loop steam cooling rather than air. The lower blade operating temperatures increase blade stress margins of safety and life and could permit the use of lower cost blades or increase the turbine inlet temperature to achieve higher efficiencies. Increasing steam turbine operating temperatures above current operating practice requires the use of more expensive materials and/or fabrication techniques but the technology base is well
established. Initially, intermediate pressure turbines in CES power plants would not operate at temperatures beyond current commercial gas turbine practice. Materials compatibility data and experience gained with gas turbines operating in an oxidizing environment at moderate pressures and very high temperatures, suggest that CES gases should not pose major compatibility problems when used to drive turbines or CES ZEPP’s, however, this requires further confirmation, including full-scale testing. The use of existing low-pressure steam turbines with CES drive gases may require design modifications to eliminate condensation and the compatibility issues presented by carbonic acid that would otherwise form.
1.3.1.2.11 References [1]
Anderson, R. E., Baxter, E. and Doyle, S.E. (Clean Energy Systems, Inc.), Final Report: Fabricate and Test Advanced Non-Polluting Turbine Drive Gas Generator, prepared for the United States Department of Energy National Energy Technology Laboratory; Cooperative Agreement No. DE-FC26-00NT 40804, September 1, 2000 to June 1, 2003.
[2]
Anderson, R., Brandt, H., Doyle, S., and Viteri, F., “A Demonstrated 20 MWt Gas Generator for a Clean Steam Power Plant”, presented at 28th International Technical Conference on Coal Utilization & Fuel Systems, Clearwater, FL, March 10-13, 2003
[3]
Anderson, R., Brandt, H., Doyle, S., Pronske, K., and Viteri, F., “Power Generation with 100% Carbon Capture and Sequestration”, presented at the 2nd Annual Conference on Carbon Sequestration, Alexandria, VA, May 5-8, 2003
[4]
Anderson, R.E., Doyle, S.E., and Pronske, K.L., “Demonstration and Commercialization of Zero-Emission Power Plants”, 29th International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida. April 18-22, 2004
[5]
Anderson, R.E. and Pronske, K.L., “Kimberlina - A Zero-Emission Multi-Fuel Power Plant and Demonstration Facility”, 30th International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida. April 17-21, 2005
[6]
Anderson, R.E. and Bischoff, R.W., “Durability and Reliability Demonstration of a Near-Zero-Emission, GasFired Power Plant, California Energy Commission, Publication #CEC 500-2006-074, July 2006
[7 ] Chorpening, B., Richards, G.A., and Casleton, K.H., U.S. Dept. of Energy, National Energy Technology Laboratory (NETL); Woike, M., Willis, B., NASA Glen Research Center, Plum Brook Station, Hoffman, L. Clean Energy Systems, Inc.; “ Demonstration of a Reheat Combustor for Power Production With CO2 Sequestration”; ASME Turbo Expo Land, Sea, Air, June 16-19,2003, Atlanta, GA. USA [8]
Chorpening, B., Casleton, K.H., and Richards, G.A., U.S. Dept. of Energy, National Energy Technology Laboratory (NETL), Woike, M., Willis, B., NASA Glen Research Center, Plum Brook Station, “ Stoichiometric Oxy-Fuel Combustion for Power Cycles with CO2 Sequestration”, Proceedings of the Third Joint Meeting of the U.S. Sections of The Combustion Institute, March 16-19, 2003, Chicago, IL
[9]
Richards, G.A., Casleton, K.H., and Chorpening, B., U.S. Dept. of Energy, National Energy Technology Laboratory (NETL); Morgantown, WV 26505, USA, “Dilute Oxy-Fuel Combustion Technology for Zero – Emission Power”, 1st International Conference on Industrial Gas Turbine Technologies, Brussels, July 10-11, 2003 [10] Martinez-Frias, J., Aceves, S., Smith, J.R. (Lawrence Livermore National Laboratory), and Brandt, H. (Clean Energy Systems, Inc.), “Thermodynamic Analysis of Zero-Atmospheric Emissions Power Plant”, presented at ASME International Conference, New Orleans, LA, November, 2002 [11] Martinez-Frias, J., Aceves, S., Smith, J.R. (Lawrence Livermore National Laboratory), and Brandt, H. (Clean Energy Systems, Inc.), “A Coal-Fired Power Plant with Zero-Atmospheric Emissions”, IMECE2003-43923, presented at 2003 ASME Mechanical Engineering Congress & Exposition, Washington, D.C., November 1521, 2003 [12] Smith, A.R., Klosek, J., Sorensen, J.C., and Woodward, D.W.; “ Air Separation Unit Integration for Alternative Fuel Projects”, Air Products and Chemicals, Inc. Allentown, Pa 18195, ASME Paper No. 98-GT63 [13] Smith, A.R. and Dillon, J.L. IV; “Gas Turbine Applications for Large Air Separation Units”, Air Products and Chemicals, Inc., Allentown, PA 18195, ASME Paper No. 99-GT-321, presented at the Int. Gas Turbine & Aerospace Congress & Exposition, Indianapolis, IN, June 7-10, 1999 [14] Scharle, W., J., Wilson, Air Products and Chemicals, “Oxygen Facilities for Synthetic Fuel Projects”, ASME Journal of Engineering for Industry, November,1981, Vol. 103 pp 409-417 [15] 2003 GTW Handbook, Gas Turbine World, Vol. 23, (2003), Pequot Publishing, Inc., Southport, CT [16] Phillips, J. N. “Integration of Commercial Gas Turbine Technology into a Clean Energy Systems Zero Emission Power Plant”, Report No. 5909-08-3 To Clean Energy Systems, Inc., June 7. 2004, by Fern Engineering, Inc., Pocasset, MA.
BIOGRAPHY
1.3.1.2 Clean Energy Systems
Fermin ‘Vic’ Viteri Clean Energy Systems, Inc. 11330 Sunco Drive, Suite A Rancho Cordova, CA 95742 Bus. (916) 379-9143 Fax (916) 379-9146 email:
[email protected]
Mr. Viteri worked 36 years at Aerojet’s Liquid Rocket Co., Sacramento, CA. and became manager of Rotating Machinery where he directed the analysis, design, and testing of high speed turbomachinery for pump fed rocket engines, waterjets for Navy patrol boats and submarine torpedo ejector pumps. All of these pumps were driven with specially built turbines or by gas turbines that were commercially purchased. Mr. Viteri is one of the original founders of Clean Energy Systems, Inc., a company involved in the research and development of oxy-fuel Zero Emissions Power Plants (ZEPP). He served as President of the company from 1996 to 1999. Currently, Vic is involved with convertng a recently aquired 5 MW biomass plant in Bakersfield, CA. to an oxy-fuel ZEPP and supporting studies of similar plants for Norway and the Netherlands.
1.3.1.3 Hydrogen-Fueled Power Systems
1.3.1.3-1 Introduction The concept of a hydrogen economy was introduced in the 1960s as a vision for future energy requirements to replace the inevitable exhaustion of fossil fuels. In the hydrogen economy, the storable and transportable hydrogen is envisioned to be a dominant energy carrier. The hydrogen can also be exploited as a clean, renewable, and nonpolluting fuel. The use of hydrogen as a fuel is attractive for a number of reasons: • Hydrogen burns with 15-22% higher thermal efficiency than that of gasoline; • From an environmental standpoint, hydrogen combustion with pure oxygen results in no emissions of the greenhouse gases, CO, CO2, SOx, and NOx; and • It generates only steam and water. Serious hydrogen-fueled turbine development program primarily comes from the initiatives of the Japanese government in 1992, through its New Energy and Industrial Technology Development (NEDO). It created the World Energy Network (WE-NET) Program, a 28-year effort from 1993 to 2020, directed at research and development of the technologies needed to develop a hydrogen-based energy conversion system1. Part of this effort is directed toward research and development of a hydrogen-fueled combustion turbine system2 which can efficiently convert the chemical energy stored in hydrogen to electricity via a heat engine in which the hydrogen is combusted with pure oxygen. Turbine manufacturers developing hydrogen-fueled power generation cycles under the WE-NET program include Westinghouse, Toshiba and Mitsubishi Heavy Industries3. The hydrogen-fuel power systems resulting from the WE-NET program and others reported in the literature are summarized below.
1.3.1.3-2 The High Temperature Steam Cycle (HTSC) Power System
Wen-Ching Yang Department of Chemical & Petroleum Engineering University of Pittsburgh Pittsburgh, PA 15261 (724) 327-3011
[email protected]
107
The High Temperature Steam Cycle (HTSC) power system, shown in figure 1 and reported by Kizuka et al.,, was based on that suggested by Jericha, et al.4. It has two closed cycles: the topping and the bottoming cycles. The topping cycle consists of a compressor, combustor, intermediate pressure (IP) turbine, steam cooler 1, and steam cooler 2. The bottoming cycle includes a low pressure (LP) turbine, condenser, preheater, high pressure (HP) turbine, combustor, and IP turbine. The key design considerations are to increase the outlet temperature of the IP turbine to generate more steam in the bottoming cycle and to increase the inlet temperature of the LP turbine to obtain more power. The HTSC power system makes use of the closed-loop cooling systems. Closed-loop cooling techniques will be favored in advanced combustion turbines because they (1) eliminate the disruption of the turbine flowfield caused by coolant ejection, (2) eliminate mixing losses caused by coolant ejection, (3) reduce the decrease in gas path temperature caused by turbine cooling, and (4) return the turbine coolant to the primary cycle5. The conceptual designs of the cooling systems were studied by Kizuka et al.,6. The three systems studied are (1) closed-circuit water cooling system for nozzle blades and steam cooling system for rotor blade; (2) closed-circuit steam cooling system for nozzle and rotor blades; and (3) open-circuit steam cooling system for nozzle and rotor blades. The main component parameters employed to evaluate the cycle performance is summarized in table 1. The reported performance of three different types of cooling systems for a 1700°C-class, hydrogen-fueled combustion gas turbine varies from 54.9 % to 61.3 % efficiency based on HHV.
Fig. 1. Process Diagram for the HTSC Power System Table 1 Main Component Parameters for HTSC Plant
IP Turbine (Topping Cycle Gas Turbine) Compressor Inlet Pressure (MPa) Inlet Temperature (oC) Outlet Pressure (MPa) Rotational Speed (rpm) Combustor Fuel Outlet Temperature (oC) Turbine Stage Rotational Speed (rpm) Inlet Gas Flow (kg/s) Outlet Pressure (MPa) Bottoming Cycle HP Turbine Inlet Pressure (MPa) Outlet Pressure (MPa) LP Turbine Inlet Pressure (MPa) Outlet Pressure (MPa)
0.14 114 4.90 6,500 H2 + O2 1,700 2(IHP) + 6(ILP) 6,500(IHP), 3,000(ILP) 222.3 0.16 19.00 5.00 0.15 0.05
1.3.1.3-3 The New Rankine Cycle Figure 2 depicts the so-called new Rankine cycle configuration involving direct steam expansion proposed and investigated by Funatsu et al.7. The combustion gas in the cycle consists of only pure steam by combusting hydrogen and oxygen stoichiometrically in the high pressure combustor (HPCOMB) and low pressure combustor (LPCOMB). The feed water from the condenser (COND) is pressurized by the boiler feed water pump (BFP) to the super critical pressure, and then heated in the heat recovery boiler (HRBL). The steam generated in the HRBL is first expanded in the high pressure turbine (HT). The exhaust from HT is then reheated in the high pressure combustor (HPCOMB) and expanded in the intermediate pressure gas turbine (IHT) again. Finally, the reheat and expansion is repeated in the low pressure combustor (LPCOMB) and the low pressure gas turbine (ILT). The steam from ILT exchanges heat with the feed water in HRBL, expands through the low pressure steam turbine (LT), and condenses in the condenser (COND). The operating parameters at different stations shown in figure 2 are reported in table 2. The reported thermal efficiency is over 60 percent HHV or 71.89 percent LHV.
108
Wen-Ching Yang
Fig. 2. Process Diagram for the New Rankine Cycle Power System
1.3.1.3-4 The Rankine Cycle With Reheat And Recuperation In the studies by the Westinghouse team on an optimum hydrogen-fueled combustion turbine system, alternative turbine systems were evaluated8. (The Power Generation Division of Westinghouse Electric Corporation was acquired by Siemens AG in 1998 and became Siemens Westinghouse Power Corporation, a fully-own subsidiary of Siemens). The Rankine cycle with regeneration and reheat was identified as the best system with potential to reach the highest efficiency. A Rankine cycle with reheat and recuperation was thus selected as the general reference system for the study of hydrogen-fueled power systems. Westinghouse has assessed both a near-term reference plant and a long-term reference plant. The near-term plant requires moderate development based on extrapolation of current steam turbine technology for the HP turbine (650°C inlet temperature), and extrapolation of current combustion turbine technology for the IP turbine (1700°C inlet temperature). In contrast, the long-term plant requires more extensive development for an additional intermediate high pressure (IHP) reheat turbine (1700°C inlet temperature), and is more complex than the near-term plant¸ with closedloop steam cooling9 of the IHP and IP turbines, and extractive feedwater heating. A single reheat stage is used in the near-term plant and two reheats are used in the long-term plant. The HRSG operates at atmospheric pressure, producing supercritical steam. The Westinghouse studies identified the trade-offs between the efficiency benefits and the developmental challenges of the near-term and long-term reference plants. The near-term plant achieves 65.2 percent net plant efficiency, and the long-term plants achieve 71.4 percent net plant efficiency. Even though the near-term plant does not achieve the goal of 70.9 percent net plant efficiency, its relative simplicity and low cost make it attractive as a next step. Process flow diagrams for the near-term and long-term plants are shown in figures 3 and 4. Representative temperatures and pressures are listed at various locations on the flow diagrams. Table 3 provides a summary of the essential technical assumptions for the component performance factors, and the reference plant requirements and boundary conditions. Table 2 The New Rankine Cycle Performance Parameters
Location
Pressure Flow MPa kg/s 1 38.0 104.2 2 34.3 104.2 3 7.35 104.2 4 6.99 115.8 5 0.98 133.1 6 0.93 148.0 7 0.14 156.1 8 0.13 156.1 9 0.005 145.1 10 25.7 3.2 11 12 20.4 13 2.6 Generating Power 500MW 60.57 % (HHV) Thermal Efficiency 109
Temperature C 100 750 464 1700 1146 1700 1129 148 33 o
71.89 % (LHV)
Enthalpy kJ/kg 447.9 3851.1 3315.9 6506.3 4798.1 6510.5 4965.5 2771.7 2315.3 -
1.3.1.3 Hydrogen-Fueled Power Systems
Fig. 3. Process Diagram for the Near-Term Rankine Cycle with Reheat and Recuperation
Fig. 4. Process Diagram for the Long-Term Rankine Cycle with Reheat and Recuperation
110
Wen-Ching Yang Table 3 Estimated Component Performance Factors and Reference Plant Conditions
PLANT Capacity (MWe) Site Type Ambient Air Temperature (oC) Ambient Pressure (bar) Relative Humidity (%) Cooling Sea Water Temperature (oC) Hydrogen and Oxygen Supply Temperature (oC) Hydrogen and Oxygen Supply Pressure Hydrogen Purity (%) Oxygen Purity (%) Hydrogen HHV (kJ/kg) CONDENSER Type Shell Pressure (bar) PUMPS Adiabatic Efficiency (%) Motor Efficiency (%) Mechanical Efficiency (%) HRSG Tube-Side Pressure Drop (%) Shell-Side Pressure Drop (%) Heat Loss (% of heat transferred) COMBUSTORS Pressure Drop (%) Combustion Efficiency (%) Heat Loss (% of heat input) TURBINES Rotation Speed (rpm) Bearing Losses (% of shaft power) Adiabatic Efficiency (%; HP, IHP, IP, LP) Exhaust Losses Steam Leaks (% of inlet flow) Shaft Leakages Windage and Pumping for Steam Leaks Steam Cooling (% of inlet flow for IHP, IP) Coolant Pressure Drop (%) LP Turbine Maximum Moisture Content (%) LP Turbine Minimum Inlet Temperature (oC) OTHERS Steam Piping Pressure Drop (%) Generator Efficiency (%) House Load (% of plant shaft power)
111
500 Greenfield, Sea-Side 15 1.01325 60 21 15 As Required 100 100 141742 Vacuum 0.0508 85 98 98 3 3 0.2 3 99.9 0.1 3,600 0.6 93, 93, 93, 93 Neglect for HP Turbine 2 Neglect Neglect 15, 15 10 15 110 3 99.2 1.5 (near-term); 1.0 (long-term)
In the near-term plant, figure 3, hydrogen and excess oxygen are combined in the HP combustor and are mixed with recycle steam to produce a 650°C combustion product at a nominal pressure of 350 bar. This is then expanded through the uncooled HP turbine, producing an exhaust stream having nominal conditions of 290°C and 40 bar. This exhaust stream, containing excess oxygen, is combined with stoichiometric hydrogen to generate the IP combustion products at a temperature of 1700°C. The IP turbine expands this combustion stream to exhaust conditions of about 740°C and 1.15 bar. The IP turbine exhaust stream is cooled in the HRSG down to about 110°C before being expanded through the LP turbine to about 33°C and the condenser pressure. A bleed stream of water is taken from the water condensate. Feedwater pumps provide high-pressure water for the HRSG to produce high-pressure steam for recycle, and intermediate-pressure water for the HRSG to produce intermediate pressure steam for open-loop cooling of the IP turbine. The conceptual design characteristics for HP turbine and IP turbine for the near-term plant are tabulated in tables 4 and 5. Table 6 shows the conceptual design characteristics of the HRSG.
1.3.1.3 Hydrogen-Fueled Power Systems Table 4 Near-Term HP Turbine Design Characteristics
Cooling Needs Inlet Temperature (oC) Exhaust Temperature (oC) Inlet Pressure (bar) Expansion Ratio Gas Flow (kg/s) Number of Stages Number of Cooled Stages Blade Heights (first stage/last stage) (cm) Turbine Casing Diameter (m) Length of Flow Passage (cm) Total Turbine Length (m)
None 650 288 350 8.75 150 15 0 2.3/6.4 1.7 270 4.9
Table 5 Near-Term IP Turbine Conceptual Characteristics
Cooling Needs Inlet Temperature (oC) Exhaust Temperature (oC) Inlet Pressure (bar) Expansion Ratio Gas Flow (kg/s) Number of Stages Number of Cooled Rows Blade Heights (first stage/last stage) (cm) Turbine Casing Diameter (inlet/exhaust) (m) Length of Flow Passage (cm) Total Turbine Length (m)
Open-Loop Steam 1700 740 38.8 33.7 154 6 11 7.4/36.4 2.3/3.7 200 5.6
Table 6 HRSG Conceptual Characteristics
FLUE SIDE Inlet Gas Inlet Gas Flow (kg/s) Inlet Gas Temperature (oC) Inlet Gas Pressure (bar) Outlet Temperature (oC) Pressure Drop (bar) TUBE SIDE Inlet Fluid Inlet Temperature (oC) IP Inlet Pressure (bar) Inlet Flow (kg/s) HP Inlet Pressure (bar) HP Outlet Flow (kg/s) IP Outlet Temperature (oC) HP Outlet Temperature (oC) HRSG DIMENSIONS HRSG Length (m) HRSG Width (m) HRSG Height (m)
Steam with trace oxygen and noncondensibles 177 740 1.15 110 0.09 0.15 µS cation conductivity water 33 40 121 385 98 269 397 24 12 14
Figure 4 depicts the long-term reference plant process diagram. The differences between the long-term plant and the near-term plant diagrams are primarily related to the insertion of an additional turbine stage, the IHP turbine, and the use of extractive feedwater heating. The long-term plant HP turbine expands recycle steam at 650°C and 350 bar. The HP turbine exhaust stream has nominal conditions of 387°C and 75 bar. A portion of the HP exhaust stream is used for closed-loop cooling of the IHP turbine and the IP turbine. The remainder of the HP exhaust stream is combined with hydrogen fuel, excess oxygen, and the exiting IHP and IP turbine closed-loop cooling streams. The IHP combustion produces a 1700°C, 70 bar stream to be expanded in the IHP turbine. The IHP turbine exhaust stream is at about 1100°C and 10 bar pressure. The IHP exhaust stream is combined with a stoichiometric amount of hydrogen to generate IP combustion products at 1700°C and 9.6 bar for expansion in the IP turbine. The IP turbine exhaust is about 1100°C and 1.2 bar pressure. Both the IHP turbine and the IP turbine produce very high temperature exhaust stream requiring high temperature piping designs. The IP expansion stream is cooled in the HRSG to about 240°C before expansion in the LP turbine. Steam is extracted from the IP turbine for feedwater heating. In contrast to the near-term plant, the long-term plant HRSG produces recycle steam at a single pressure level to be expanded in the HP turbine.
112
Wen-Ching Yang The HP turbine in both the near-term and long-term plants is close to, or within the range of current steam turbine technology10. In contrast, the long-term plant IHP turbine is a large technology step above both current steam turbine practices and advanced combustion turbine developments. The IP turbine has conditions close to the conditions of advanced combustion turbines being developed for natural gas fuels 11. The conceptual design of the reference plant was developed by using several design tools and sources of engineering experience. TM A commercial process simulator (ASPEN PLUS ) was applied to develop the reference plant process flow diagrams and thermal performance estimates. Westinghouse proprietary design codes were used to design and size turbines, combustors, and the HRSG. Advanced combustion turbine engineering experience resulting from development programs in the United States12 and advance steam turbine engineering experience resulting from past studies and testing at Westinghouse13 were used to extrapolate current technology to the demands of the reference plant. Plant heat and materials balances and cycle calculations were generated for the reference plant at its rated load of 500 MWe using the component performance assumptions described. No hydrogen or oxygen preheat is used in the plants, and extractive feedwater heating is used in the long-term plant only, providing a simple, compact plant configurations. Table 7 lists the reference plant performance and power generation breakdown for the long-term and the near-term plants. The plant net efficiency for the near-term plant is estimated to be 65.2 % (LHV) and that for the long-term plant, 71.4% (LHV). Table 7 Reference Plant Thermal Performance
HP Turbine (MWe) IHP Turbine (MWe) IP Turbine (MWe) LP Turbine (MWe) Gross Power (MWe) Gross Efficiency (%) Generator Losses (MWe) Pumping Power (MWe) BOP Losses (MWe) Net Power (MWe) Net Efficiency (%)
Long-Term Plant
Near-Term Plant
42.3 189.9 215.4 67.9 515.5 73.5 4.1 6.2 5.2 500.0 71.4
81.6 362.5 74.1 518.2 67.6 4.1 6.5 7.8 500.0 65.2
The reference plant environmental performance is expected to be superior to that of other power generation concepts using other fuels, by having very low nitrogen oxides, sulfur oxide, particulate, toxic species, and green-house gas emissions. The generation of solid waste and liquid/sludge wastes would also be negligible. The only significant emissions could result from fuel or oxygen contaminants, or from noise.
1.3.1.3-5 Developmental Requirements In a final report14 published in 1999, Westinghouse discussed the developmental requirements for realization of hydrogenfueled power systems to be developed in the general areas of 1) Materials, 2) Closed-loop turbine cooling, and 3) Hydrogen combustion for the major equipment components: combustors, compressors, expanders, and heat exchangers. The nature of these developmental requirements depend on the specific component operating conditions. The feasibility of manufacturing the equipment components with current technology must also be considered. The near-term plant HP turbine development should center on combustion phenomena and design features of the combustor and materials selection. The near-term HP turbine design needs to be an adaptation of conventional HP steam turbine designs. The IP turbine development needs should center on the component cooling needs- airfoils, cylinder, blade rings, and rotor bearings. The cooling design should follow from closed-loop steam cooling features currently under development for advanced combustion turbines operating under similar temperature and pressure conditions. HP and IP combustion development should focus on the nature of the hydrogen combustion products and the control of its chemistry, as well as the combustor operability: flame stability, ignition, flame detection, and combustion-induced oscillations. Combustor liner and transition cooling should also be part of the combustor development testing. Materials testing should focus on the durability of candidate materials during exposure to the hydrogen combustion products at simulated commercial operating duty conditions. This includes hydrogen embrittlement, oxidation, stress corrosion cracking, corrosion deposits, and creep rupture. The development of appropriate corrosion coatings and TBCs, and their associated bond coating, should be a major part of the materials program. These may include advanced combustion turbine Ni-based super-alloys (single crystal and directionally-solidified), and advanced combustion turbine TBCs (e.g., yttia-stabilized zirconia, alumina, TiN). The materials testing associated with the IP turbine should also consider the limited use of ceramic components in ring segments, combustor liners and transitions, and, vane and blade leading edges. The stability of candidate ceramic materials and alternative ceramic forms in simulated combustion-steam environments (e.g. zirconia in composite forms, tile forms, and fibrous insulating forms) should also be considered.
113
1.3.1.3 Hydrogen-Fueled Power Systems 1.3.1.3-6 Conclusions The development of a hydrogen-fueled power plant with an efficiency higher than 70 percent (LHV) can be accomplished. Four conceptual reference cycle power systems, the high-temperature steam cycle, the new Rankine cycle and the near-term and long-term Rankine cycle with reheat and recuperation, were reviewed and reported. The reference plants environmental performance is expected to be superior to that of other power generation concepts using other fuels, by having very low nitrogen oxide, sulfur oxide, particulate, toxic species, and green-house gas emissions. The generation of solid waste and liquid/sludge wastes would also be negligible. The only significant emissions could result from fuel or oxygen contaminants, or from noise. To reach the ultimate reality, the hydrogen-combustion turbine cycles described require development in the general areas of materials, closed-loop turbine cooling, and hydrogen combustion for the major equipment components: combustors, compressors, expanders, and heat exchangers.
1.3.1.3-7 Notes _______________________________ 1. MITI, “Comprehensive Approach to the New Sunshine Program which Supports the 21st Century,” Sunshine Journal (Agency of Industrial Science and Technology in Ministry of International Trade and Industry [MITI]) 4 (1993): 1-6. 2. NEDO, “International Clean Energy Network Using Hydrogen Conversion (WE-NET), 1993 Annual Summary Report on Results (New Energy and Industrial Technology Development Organization [NEDO]) (1994). 3. NEDO, “Subtask 8 – Development of Hydrogen-Combustion Turbine, Study for an Optimum System for HydrogenCombustion Turbine,” 1995 Annual Technical Results Report (New Energy and Industrial Technology Development Organization [NEDO]) (1996). 4. N. Kizuka et.al., “Conceptual Design of the Cooling System for 1700°C-Class, Hydrogen-Fueled Combustion Gas Turbines,” Trans. ASME 121 (1999): 108-115; H. Jericha, O. Starzer, and M. Theissing, “Towards a Solar-Hydrogen System,” ASME CogenTurbo, IGTI, 6 (1991): 435-442. 5. E. D. Alderson, G. W. Scheper, and A. Cohn, “Closed Circuit Steam Cooling in Gas Turbines,” ASME Paper 87-JPGC-GT-1 (1987); T. Ikeguchi and K. Kawaike, “Effect of Closed-Circuit Gas Turbine Cooling Systems on Combined Cycle Performance,” ASME Paper 94-JPGC-GT-8 (1994). 6. Kizuka (see note 4 above). 7. T. Funatsu, M. Fukuda, and Y. Dohzono, “Start Up Analysis of a H2-O2 Fired Gas Turbine Cycle,” ASME Paper 97-GT-491 (1997). 8. R. L. Bannister, R. A. Newby, and W. C. Yang, “Development of a Hydrogen-Fueled Combustion Turbine Cycle for Power Generation,” ASME Paper 97-GT-14 (1997). 9. D. A. Little, R. L. Bannister, and B. C. Wiant, “Development of Advanced Turbine Systems,” Proceedings, ASME Cogen Turbo Power ’93 (New York: ASME, 1993). 10. R. L. Bannister, and G. J. Silvestri, “The Evolution of Central Station Steam Turbines,” Mechanical Engineering 111, no. 2 (1989): 70-78; R. L. Bannister et al., “High-Temperature Supercritical Steam Turbines,” Mechanical Engineering 109 no. 2 (1987): 60-65. 11. R. L. Bannister et al., “Turbines for the Turn of the Century,” Mechanical Engineering 116 no. 6 (1994): 68-75. 12. D. J. Amos et al., “Update on Westinghouse’s Advanced Turbine System Program,”, ASME Paper 97-GT-369 (1997); I. S. Diakunchak et al., “Technology Development Programs for the Advanced Turbine Systems Engine,” ASME Paper 96-GT-5 (1996). 13. G. J. Silvestri, R. L. Bannister, and A. Hizume, “Optimization of Advanced Steam Condition Power Plants,” Journal of Engineering Gas Turbines and Power 114 (1992): 612-620. 14. R. L. Bannister, R. A. Newby, and W. C. Yang, “Final Report on the Development of a Hydrogen-Fuelde Combustion Turbine Cycle for Power Generation,” J. Eng. Gas Turbines and Power 121(1999): 38-45.
114
BIOGRAPHY
1.3.1.3 Hydrogen-Fueled Power Systems
Wen-Ching Yang Department of Chemical & Petroleum Engineering University of Pittsburgh Pittsburgh, PA 15261 phone: (724) 327-3011 email:
[email protected]
Dr. Yang worked for Westinghouse Electric Corporation and Siemens Westinghouse Power Corporation for more than 36 years primarily in the area of advanced fossil fuel power generation systems for IGCC applications. In the past 6 years, he concentrated on the detailed design of commercial gas turbines through cold flow simulation testing, experimentation at commercial sites, theoretical modeling, and development of advance sensors to monitor performance of gas turbines. He was elected a Fellow of American Institute of Chemical Engineers in 1992 and awarded a honorary Guest Professorship at the Thermal Engineering Department of Tsinghua University in Beijing, China, in 1996. He holds several patents relating to turbine applications in the area of partial oxidation, thermal chemical recuperation, and hydrogen-fueled power plants. He retired at the end of July 2004 and is now an adjunct Professor at the Department of Chemical and Petroleum Engineering, Univeristy of Pittsburgh.
1.3.2 Advanced Brayton Cycles
1.3.2-1 Introduction Gas turbines could play a key role in the future power generation market addressing issues of producing clean, efficient, affordable, and fuel-flexible electric power. Numerous projections estimate that gas turbines will comprise a significant portion of the required generation capacity in the 21st century. Novel advanced gas turbine cycle modifications intended to improve the basic Brayton cycle performance and reduce pollutant emissions are currently under development or being investigated by gas turbine manufacturers and Research and Development (R&D) organizations. Preliminary conceptual analyses of advanced cycles indicate that it may be possible to achieve an improved combination of efficiency, emissions, and specific power output which in turn should reduce the power generation equipment cost on a $/kW basis. Developing turbine technology to operate on coal-derived synthesis gas and hydrogen is critical to the development of advanced power generation technologies and the deployment of FutureGen plants. The FutureGen plant concept may also be deployed in natural gas-based plants with respect to generating power with near-zero emissions while utilizing these advanced Brayton cycle machines and securing fuel diversity.
1.3.2-2 Gas Turbine Technology A conventional gas turbine cycle consists of pressurizing a working fluid (air) by compression, followed by combustion of the fuel; the energy thus released from the fuel is absorbed into the working fluid as heat (see figure 1). The working fluid with the absorbed energy is then expanded in a turbine to produce mechanical energy, which may in turn be used to drive a generator to produce electrical power. Unconverted energy is exhausted in the form of heat which may be recovered for producing additional power. The efficiency of the engine is at a maximum when the temperature of the working fluid entering the expansion step is also at a maximum. This occurs when the fuel is burned in the presence of the pressurized air under stoichiometric conditions.
Ashok Rao, Ph.D., P.E. Chief Scientist, Power Systems Advanced Power and Energy Program University of California Irvine, CA 92697-3550 phone: (949) 824-7302 ext 345 email:
[email protected]
115
Fig. 1. Gas Turbine and the Ideal Brayton Cycle P-V Diagram
When natural gas is burned with air under stoichiometric conditions, however, the resulting temperature is greater than 1940ºC (3500ºF) depending on the temperature of the combustion air. It is therefore necessary to utilize a large excess of air in the combustion step, which acts as a thermal diluent and reduces the temperature of the combustion products, this temperature being dependent on the gas turbine firing temperature which in turn is set by the materials used in the turbine parts exposed to the hot gas and the cooling medium (its temperature and physical properties) as well as the heat transfer method employed for cooling the hot parts. A fraction of the air from the
60%
compressor is bled off as cooling air when air is utilized for cooling, the air being extracted from the compressor at appropriate pressures depending upon where it is utilized in the turbine. From a cycle efficiency and engine specific power output (kW per kg/s of suction air flow) standpoint, it is important to minimize the amount of cooling air as well as the excess combustion air. 50% The necessity to use a large excess of pressurized air in the combustor as well as for turbine cooling when air cooling is employed creates a large parasitic load on the cycle, since compression of the air requires mechanical energy and this reduces the net power produced from the system, as well as reducing the overall efficiency of the system. Some of the technological advances being made or being investigated to improve the Brayton cycle include the following, in addition to the changes in the basic cycle configuration such as the inclusion of reheat combustion, intercooling (which is justified for very high pressure ratio cycles), recuperation and humidification: • • • • • • •
Rotor inlet temperature of 1700ºC (3100ºF) or higher which would require the development and use of advanced materials including advanced thermal barrier coatings and turbine cooling techniques including closed loop steam cooling Advanced combustor liner (combustion air and combustion products being hotter) required due to increases in rotor inlet temperatures High blade metal temperature in the neighborhood of ~1040ºC (1900ºF) while limiting coolant amount would again require the development and use of the advanced materials including advanced thermal barrier coatings Pressure gain combustor Cavity or trapped vortex combustor High pressure ratio compressor (greater than 30 to take full advantage of higher firing temperature) Integration capability with high temperature ion transport membrane air separation in IGCC applications. 45%
Addition of novel bottoming cycles is yet another approach to improving the overall plant (combined cycle) performance. Overall cycle efficiencies utilizing advanced technology gas turbines approaching 65% on natural gas on an LHV basis may be expected (see figure 2). Some of these developments are described in the following.
1900 1800 1700 1600
O
Temperature C
1500 1400
Film cooling
1300 1200
Rotor Inlet Firing Temperature
1100
TBC
1000
IN738
900
IN939 IN792 DS
800
S-816 M252 U500 U700 N 80 A 700 1950
1955
1960 1965
1970
1975
1980
1985
1990
Single Crystal Material Improvement 1995
2000
2005
2010
2015
2020
2025
Year
Fig. 2. Impact of Firing / Metal Temperature on Efficiency
Gas Turbine Firing Temperature Current-state-of-the-art gas turbines have firing temperatures (rotor inlet temperatures) that are limited to about 2600ºF. This increase in firing temperature has been made possible by being able to operate the turbine components (that come into contact with the hot gasses) at higher temperatures while at the same time utilizing closed circuit steam cooling. In a state-of-the-art air-cooled gas turbine with firing temperature close to 1320ºC (2400ºF), as much as 25% of the compressor air may be used for turbine cooling, which results in a large parasitic load of air compression. In air-cooled gas turbines, as the firing temperature is increased, the demand for cooling air is further increased. Closed circuit steam cooling of the gas turbine provides an efficient way of increasing the firing temperature without having to use a large amount of cooling air. Furthermore, steam with its very large heat capacity is an excellent coolant. Closed circuit cooling also minimizes momentum and dilution losses in the turbine while the turbine operates as a partial reheater for the steam cycle. Another major advantage with closed circuit cooling is that the combustor exit temperature and thus the NOx emissions are reduced for a given firing temperature; the temperature drop between the combustor exit gas and the turbine rotor inlet gas is reduced since the coolant used in the first stage nozzles of the turbine does not mix with the gasses flowing over the stationary vanes. Note that control of NOx
116
Ashok Rao, Ph.D., P.E. emissions at such high firing temperatures becomes a major challenge. The General Electric (GE) H series gas turbines as well as the Siemens and Mitsubishi G series gas turbines incorporate steam cooling although the GE turbine includes closed circuit steam cooling for the rotors of the high pressure stages. Taking the firing temperature beyond 1430ºC (2600ºF) poses challenges for the materials in the turbine hot gas path. Single crystal blading has been utilized successfully in advanced turbines but in addition to this, development of advanced thermal barrier coatings would be required. Extensive use of ceramics may be predicated. Reheat or sequential combustion is an alternate approach to decreasing the amount of excess combustion air without increasing the firing temperature. Gas Turbine Pressure Ratio The optimum pressure ratio for a given cycle configuration increases with the firing temperature of the gas turbine. Thus to take full advantage of the higher firing temperature of the gas turbine with firing temperature in the neighborhood of 1700ºC (3100ºF) the required pressure ratio may be in excess of 30. Another constraint to consider is the temperature of the last stage buckets in the turbine. This temperature may have to be limited to about 650ºC (1200ºF) from a strength of materials standpoint since the last stage buckets in large scale gas turbines tend to be very long and a certain minimum pressure ratio would be required to limit this temperature. Combustor Developments Pressure Gain Combustor A pressure gain combustor produces an end-state stagnation pressure that is greater than the initial state stagnation pressure. An example of such a system is the constant volume combustion in an ideal spark ignited engine. Such systems produce a greater available energy in the end state than constant pressure systems. It has been shown that the heat rate of a simple cycle gas turbine with a pressure ratio of 10 and a turbine inlet temperature of ~1200ºC (2200ºF) can be decreased by more than 10% utilizing such a constant volume combustion system1. Pulse combustion which relies on the inherent unsteadiness of resonant chambers can be utilized as a pressure gain combustor. Research continues at the U.S. DOE and at NASA for the development of pressure gain combustors. Trapped Vortex Combustor The Trapped Vortex Combustor (TVC) has the potential for numerous operational advantages over current gas turbine engine combustors. These include lower weight, lower pollutant emissions, effective flame stabilization, high combustion efficiency, and operation in the lean burn modes of combustion. The TVC concept grew out of fundamental studies of flame stabilization and is a radical departure in combustor design using swirl cups to stabilize the flame. Swirl-stabilized combustors have somewhat limited combustion stability and can blow out under certain operating conditions. On the other hand, the TVC maintains a high degree of flame stability because the vortex trapped in a cavity provides a stable recirculation zone that is protected from the main flow in the combustor. The second part of a TVC is a bluff body dome which distributes and mixes the hot products from the cavity with the main air flow. Fuel and air are injected into the cavity in a way that it reinforces the vortex that is naturally formed within it. The TVC may be considered a staged combustor with two pilot zones and a single main zone, the pilot zones being formed by cavities incorporated into the liners of the combustor2. The cavities operate at low power as rich pilot flame zones achieving low CO and unburned hydrocarbon emissions, as well as providing good ignition and the lean blowout margins. At higher power conditions (above 30% power) the additional required fuel is staged from the cavities into the main stream while the cavities are operated at below stoichiometric conditions. Experiments have demonstrated an operating range that is 40% wider than conventional combustors with combustion efficiencies of 99%+. Use of the TVC combustor holds special promise as an alternate option for suppressing the NOx emissions in syngas applications where pre-mixed burners may not be employed. More details on this type of combustor may be found in Section 3.2.1.4.1. Catalytic Combustor Lean stable combustion can be obtained by catalytically reacting the fuel-air mixture with a potential for simultaneous low NOx, CO and unburned hydrocarbons. It also has the potential for improving lean combustion stability and reducing combustioninduced pressure oscillations. The catalytic combustor can play a special role in IGCC applications to reduce NOx emissions. More details on this type of combustor may be found in Sections 3.2.2 and 3.2.2.1. IGCC Applications
117
The H2O vapor content of the working fluid flowing through the turbine when firing syngas while utilizing water vapor as the diluent, is significantly higher than that in the case when natural gas is the fuel (i.e., compared to the case when natural gas is fired in dry low NOx combustors). The following implications exist for the gas turbine in such applications: 1. Derating of the turbine firing temperature due the different aero-heat transfer characteristics and 2. Life of the thermal barrier coatings, and any ceramics that may be utilized in advanced gas turbines in the future.
1.3.2 Advanced Brayton Cycles Additionally, a gas turbine designed for a certain firing temperature on natural gas would see derating of the firing temperature not only due to the increased concentration of H2O vapor in the working fluid but also due to the increase in the pressure ratio since the temperature of the cooling air increases as the pressure ratio is increased. In the case of a steam-cooled gas turbine, however, derating of the firing temperature may be less significant (since the cooling steam temperature may be maintained independently of the gas turbine pressure ratio), unless the low pressure air-cooled stages of the gas turbine become the bottleneck. Furthermore, if dual fuel capability, i.e., operating capability on natural gas and on syngas is required, a large surge margin would be necessary for the compressor with a pressure ratio in excess of 30 and may require a twin-spool aero-compressor for high pressure ratios. Air extraction from the engine to supply the air separation unit may alleviate some of these challenges. Integration capability with high temperature membrane air separation in IGCC applications may be a requirement in the future when these advanced gas turbines are deployed. Capabilities for extraction of ~ 50% of the compressor discharge air for the membrane unit while introducing hot (~800ºC or 1500ºF) depleted air from it into the gas turbine combustor would be required. Within the combustor, its liner design and materials would be impacted. Novel Cycles Humid Air Turbine (HAT) Cycle The mechanical energy required for air compression in the Brayton cycle can be reduced by utilizing interstage cooling. However, from an overall cycle efficiency standpoint, interstage cooling can be utilized advantageously if the heat removed from the compressed air in the intercooler can be efficiently recovered for conversion to power. If the entire heat is simply rejected to the atmosphere, the overall cycle efficiency may actually decrease depending upon the cycle pressure ratio, since it results in the consumption of more fuel to compensate for the energy lost through the intercooler. Only at very high pressure ratios can intercooling be justified in most cycles.
Fig. 3. HAT Cycle
In the HAT cycle a significant portion of the excess air that is required as thermal diluent in a gas turbine, is replaced with water vapor (see figure 3)3. The water vapor is introduced into the system in an efficient manner, by pumping of a liquid followed by low temperature evaporation. Pumping a liquid requires less mechanical energy compared to gas (air) compression. Evaporation of the water into the compressed air stream is accomplished using low temperature heat, in a counter-current multistage humidification column, rather than generating steam in a boiler. This method of humidification permits the use of low temperature heat for accomplishing the evaporation of water. For example, water which boils at 100oC or 212oF at atmospheric pressure may be made to evaporate at room temperatures when exposed to a stream of relatively dry air. The process also reduces the parasitic load of compressing the combustion air by intercooling the compressor, while recovering most of the heat removed in the intercooler for the humidification operation. Thus, a more thermally efficient power cycle is achieved. Humidification of the compressed air also leads to a reduction of NOx emissions. The humid air is preheated by heat exchange with the turbine exhaust in a recuperator to recycle the exhaust energy to the combustor, thereby eliminating the expensive steam bottoming cycle required in a combined cycle.
118
Ashok Rao, Ph.D., P.E. The advantages of the HAT cycle are: • Less than 5 ppmV NOx without post-combustion treatment • High efficiency without a steam bottoming cycle • Applicable to micro- and mini-turbines for distributed generation • Excellent part-load performance, efficiency essentially constant down to 60% of full load • Performance quite insensitive to ambient temperature • Water usage less than that for a combined cycle employing wet cooling tower and if desired, water may be recovered from HAT exhaust • High specific power • Integrates synergistically with reliable low-cost “Total Quench” gasifier • In coal based Zero Emission plants, the “Total Quench Gasifier” option is of choice • In natural gas Zero Emission based plants where CO2 is recovered from exhaust, CO2 concentration is higher (dry basis). The disadvantages of the HAT cycle are: • Requires intercooled-regenerative gas turbine for optimum performance • Compressor / turbine flow mismatch deviates from normal gas path design • Development cost could be high although possible for compressor / turbine “mix and match” • Does not take advantage of steam-cooled blade technology (water cooled technology could prove better in advanced machines) • Use of ceramics in hot gas path may be a challenge due to high moisture content of the working fluid. Inlet Air Fogging Another approach to reducing the parasitic load of air compression in a gas turbine is to introduce liquid water into the suction air4. The water droplets will have to be extremely small in size and be in the form of a fog to avoid impingement on the blades of the compressor causing erosion. As the water evaporates within the compressor from the heat of compression, the air being compressed is cooled which in turn causes a reduction in the compressor work. Note that the compression work is directly proportional to the absolute temperature of the fluid being compressed. A benefit in addition to increasing the specific power output of the engine is the reduction in the NOx due to the presence of the additional water vapor in the combustion air. A number of gas turbines have been equipped with such a fogging system. Care should be taken, however, in specifying the water treatment equipment since high quality demineralized water is required as well as in the design of the fogging system to avoid impingement of the compressor blades with water droplets. Fuel Cell Hybrids A fuel cell, as an electrochemical device is similar to a battery that converts chemically bound energy directly into electricity but unlike conventional batteries, the chemical energy to the cell is supplied on a continuous basis in the form of a fuel such as natural gas or synthesis gas while the oxidant (air) is also supplied continuously. Higher conversion efficiencies are achievable with a fuel cell when compared to heat engines; the chemical energy is directly converted into electricity, the intermediate step of conversion into heat as in a heat engine is eliminated and thus without being constrained by temperature limitations of the materials as is the case with heat engines. A fuel cell-based hybrid cycle consists of combining a fuel cell with a heat engine to maximize the overall system efficiency. Overall system efficiencies greater than 60% on natural gas on an LHV basis may be achieved. High temperature fuel cells such as solid oxide and molten carbonate fuel cells are most suitable for such applications. In the case of a high pressure fuel cell based hybrid (see figure 4), the combustor of the gas turbine is replaced by the fuel cell system while in the case of a low pressure fuel cell based hybrid, the heat rejected by the fuel cell may be transferred to the working fluid of the gas turbine through a heat exchanger (indirect cycle)5. The characteristics for gas turbines needed in these hybrid applications are: • Recuperation (currently only small gas turbines, i.e., less than 15 MW are offered as recuperated engines for generator drives) • Low firing temperature of less than 1000ºC (1800ºF) and pressure ratio in the range of 4 to 12 • Combustors accepting hot and depleted fuel and air when gas turbine combustors are utilized for oxidation of the fuel cell anode exhaust gas • Oil free bearings to avoid carbon deposition in the anode section of the fuel cell.
119
Fig. 4. A Pressurized SOFC Hybrid
1.3.2 Advanced Brayton Cycles 1.3.2-3 Conclusions Gas turbines could play a key role in the future power generation market including coal based FutureGen plants. Potential exists to take the overall cycle efficiencies to 65% on natural gas on an LHV basis, 60% being the state-of-the-art combined cycle efficiency with the technological advances being made or being investigated which include higher rotor inlet temperature of 1700ºC (3100ºF) or higher and higher blade metal temperature ~1040ºC (1900ºF) made possible with the use of advanced materials including advanced thermal barrier coatings and turbine cooling techniques including closed loop steam cooling, advanced combustor liners to handle the higher temperatures within the combustor, pressure gain and cavity combustors, high pressure ratio compressors (greater than 30 to take full advantage of higher firing temperature) and integration capability with high temperature ion transport membrane air separation in IGCC applications. In tandem, changes to the basic cycle configuration such as the inclusion of reheat combustion and intercooling which is advantageous in very high pressure ratio cycles would be complementary in achieving the goals of higher thermal efficiency and higher engine specific power output. These desirable attributes could also be further enhanced by the use of advanced combustor concepts such as the pressure gain combustor while the TVC holds the promise of an alternate option for suppressing the NOx emissions, especially in syngas applications.
1.3.2-4 Notes _________________ 1. R.S. Gemmen, G. A. Richards and M. C. Janus, “Development of a Pressure Gain Combustor for Improved Cycle Efficiency, Proceedings of the ASME Cogen Turbo Power Congress and Exposition (1994). 2. D. L. Burrus, A. W. Johnson and W. M. Roquemore, and D. T. Shouse, “Performance Assessment of a Prototype Trapped Vortex Combustor for Gas Turbine Application,” Proceedings of the ASME IGTI Turbo-Expo Conference (New Orleans, June 2001). 3. A. D. Rao, “Process for Producing Power,” U.S. Patent No. 4,289,763 dated May 16, 1989. 4. R. Bhargava and C. B. Meher-Homji, “Parametric Analysis of Existing Gas Turbines with Inlet Evaporative and Overspray Fogging,” Proceedings of the ASME IGTI Turbo-Expo Conference, (Amsterdam, June 2002). 5. K. Litzinger, et. al., “Comparative Evaluation of SOFC Gas Turbine Hybrid System Options,” Proceedings of the ASME Turbo Expo Conference (Reno-Nevada, June 2005); G. Agnew, et. al., “The Design and Integration of the Rolls-Royce Fuel Cell Systems 1 MW SOFC, “Proceedings of the ASME Turbo Expo Conference (Reno-Nevada, June 2005); R. Schonewald, “TurboMachinery Requirements for Practical SOFC-Gas Turbine Hybrid Systems,” Proceedings of the ASME Turbo Expo Conference (Reno-Nevada, June 2005); H. Ghezel-Ayagh, “Hybrid Controls,” Presented at the ICEPAG Conference (Irvine, California, September 2004).
120
BIOGRAPHY
1.2.2 Implications of CO2 Sequestration for Gas Turbines 1.3.2 Advanced Brayton Cycles
Ashok Rao, Ph.D., P.E. Chief Scientist, Power Systems Advanced Power and Energy Program University of California Irvine, CA 92697-3550 phone: (949) 824-7302 ext 345 email:
[email protected]
Dr. Rao serves as the Chief Scientist, Power Systems at UC Irvine Advanced Power and Energy Program. He worked in industry for more than 30 years in the energy conversion area, and previously worked at Fluor as Director in Process Engineering and Senior Fellow in design / development of gasification for power generation and synthetic fuels coproduction. He received several patent awards in the energy conversion area and authored several papers on advanced power cycles and improved IGCC designs. Dr. Rao has also worked for Allis-Chalmers and McDowell Wellman Engineering in coal conversion; responsibilities included taking ideas from drawing board to demonstration scale plants. He holds a Ph.D. in Mechanical Engineering and a M.S. in Chemical Engineering.
1.3.3
Partial Oxidation Gas Turbine (POGT) Cycles
Joseph K. Rabovitser, Ph.D.
Serguei Nester, Ph.D.,
Gar Technology Institute, Energy Utilization center 1700 S. Mount Prospect Road Des Plaines, IL 60018 Phone: 847-768-0548 847-768-0541
[email protected],
[email protected]
David James White TRITEK Consulting 3633 Millikin Avenue San Diego, CA 92122-2413 Phone: 858-453-8653 Email:
[email protected]
121
1.3.3-1 Introduction There are two main features that distinguish a Partial Oxidatation Gas Turbine from a conventional gas turbine. These are associated with the design arrangement and the thermodynamic processes used in operation. A primary design differentiating feature of the POGT when compared to a conventional gas turbine is that POGT utilizes a non-catalytic partial oxidation reactor (POR) in place of a normal combustor. An important secondary distinction is that a much smaller compressor is required, one that typically supplies less than half of the air flow required in a conventional gas turbine. From an operational and thermodynamic point of view the key distinguishing feature is that the working fluid provided by the POR (a secondary fuel gas) has a much higher specific heat than lean complete combustion products and more energy per unit mass of fluid can be extracted by the POGT expander than is the conventional case. (This is why the POGT uses a smaller compressor than a conventional gas turbine.) A POR operates at fuel rich conditions typically at equivalence ratios on the order of 2.5, and virtually any hydrocarbon fuel can be combusted. Because of these fuel rich conditions, incomplete combustion products are used as the hot section working fluid. A POGT thus produces two products: power and a secondary fuel that usually is a hydrogen rich gas. This specific feature creates a great opportunity to provide high efficiencies and ultra-low emissions (single digit NOx and CO levels) when the secondary fuel is burned in a bottoming cycle. When compared to the equivalent standard gas turbine bottoming cycle combination, the POGT provides an increase of about 10 percent points in system efficiency. The overall efficiency of a POGT two-staged power system is typically high and can approach 70% depending on the POGT operating conditions and the chosen bottoming cycle. In figure 1 a generic arrangement of a two-stage or air-staged reheat power system with a POGT as a topping cycle is shown. The bottoming cycle can be either a low pressure (or vacuum) combustion turbine, or an internal combustion engine, or a solid oxide fuel cell, or any combination of them. In addition, the POGT can be used as the driver for cogeneration systems. In such cogeneration systems the bottoming cycle can be a fuel-fired boiler, an absorption chiller, or an industrial furnace. The POGT is ideally suited for the coproduction of power and either hydrogen, or synthesis gas (syngas), or chemicals. Some of the important applications are described below.
Fig. 1. Generic Schematic of POGT System
1.3.3-2 Background Research and development (R&D) into the application of POGT concepts for power generation was first performed by the Institute of High Temperature (IVTAN) in the former Soviet Union in the late 1950s1. The result of this R&D was the demonstration of a working POGT. In one published application by IVTAN2, residual fuel oil is partially combusted to produce high-pressure steam and fuel gas, which is then cooled and cleaned to remove ash and sulfur compounds. The steam and purified fuel gas are then used for power generation. A 1970 patent for a POGT by Jacques Ribesse of the JARIX company in Brussels, Belgium,
was followed by a technical paper in 19713, and a second paper describing further improvements in 19914, which described the gas turbine, air compressor, catalytic partial oxidation reactor (POR), and expansion turbine. Partial or total combustion of the combustible gas (leaving the POR) and passing through the expansion turbine was accomplished by injecting air into the turbine vanes. This simultaneously accomplished both the needed cooling and, through local combustion, an isothermal expansion5. In 1992, IVTAN published a paper describing an innovative combined cycle utilizing a POGT for the repowering of existing natural-gas-fired steam turbine power plants. The retrofit modifications were estimated to improve fuel efficiencies to between 70-80% and reduce NOx emissions by a factor of 10 or more6. Efficiencies are increased mainly because of (1) complete use of the thermal energy of the hot pressurized gasifier product gas supplied by the POGT; (2) reduced air flow requirements typically about 65% of that used for a conventional expansion turbine, (3) larger volumetric gas flow in the turbine (15-20%), taking into account the lower specific mass of the partial oxidation products, (4) higher specific heat of the turbine working fluid, and (5) close to isothermal expansion, allowing a better utilization potential of the heat7. Hodrien and Fairbairn in 1993 evaluated the POGT in a report prepared for British Gas as a highly promising cycle with a potential efficiency above 60%8. Further study at the University of Leige (Belgium) in collaboration with other European partners, which included preliminary analysis and testing, concluded POGT has good potential for power generation applications and Combined Heat and Power (CHP) applications as well9. The Gas Technology Institute (GTI) has been actively working on the POGT concept since 1995. With support from the U.S. Department of Energy (DOE) and Gas Research Institute (GRI), GTI (formerly IGT) teamed up with SWPC (formerly Westinghouse) to perform a system study of POGT applications10. The cycles studied included (1) a conventional natural-gas-fired gas turbine with a POGT utilized as a topping cycle, (2) a combined cycle plant joining a POGT with a steam turbine, and (3) a repowering system for coalfired power plants using a POGT as a topping cycle. In a continuation of this work Westinghouse performed technical feasibility studies and cost analyses of the PO power cycle11 and concluded that there was potential for significant plant heat rate and cost-of-electricity improvements. In a recent development effort to demonstrate a POGT for on-site CHP generation, GTI with support from the California Energy Commission (CEC) and GRI, has teamed with Solar Turbines Incorporated (Solar), Tritek Consulting, Alturdyne Incorporated , and the Belcan Corporation to develop, build, and install at GTI a 10-MWth (34 MMBtu/hr) pressurized research non-catalytic POR, intended to replace the combustor of the Solar Spartan T-350 conventional gas turbine modified to operate in a POGT mode.
1.3.3-3 Overview The POGT has great potential as a driver for a wide range of bottoming cycles for power generation. A POGT can effectively co-produce both power and syngas from which hydrogen can be extracted. It can also be used in a cogeneration mode where the bottoming cycle systems are industrial furnaces, boilers, or absorption chillers. Depending on the fuel type and if normal ambient air is used (rather than oxygen enriched air) the exit fuel gases from the POGT are essentially low to medium heating value secondary fuels with variable but high hydrogen contents. In general the lower the hydrogen content of the fuel molecule the lower the exhaust gas hydrogen concentrations will be. An increase in hydrogen content can be obtained by adding steam to the POR which through reforming reactions will increase the hydrogen content of the POR exhaust. Typically the POR will operate at temperatures on the order of 2000 to 2400°F thus keeping the maximum turbine inlet temperature to this level and allowing usage of existing and proven fleet of turbine expanders for POGT application. The simultaneous endothermic reforming and exothermic oxidation reactions that occur within the POR tend to thermally balance each other at a particular temperature (depending on the equivalence ratio) thus eliminating any destructive run-away reactions. The POR exhaust gases, which Fig. 2. POGT Cycle Comparisons have very high specific heats, provide a significant improvement over air as the working fluid. Expansion of the POR gases over a turbine provides a much greater power extraction per unit mass of working fluid than is possible for the products of lean conventional combustion systems. This working fluid improvement results in the specific power of the POGT proper, being almost twice that of conventional gas turbines. Generally improved specific power provides in turn improved profitability for the manufacturer and lower unit costs for the customer. Higher efficiencies typically improve the customer’s profitability. The POGT exhibits both improved specific power and increased efficiencies when compared to conventional gas turbines. This is clearly shown in figure 2 in which a comparison of the POGT with a number of gas turbine cycles is provided.
122
Joseph K. Rabovitser, Ph.D., Serguei Nester, Ph.D., 1.3.3-4 POGT Applications This section covers the key applications that are perfectly suited to a POGT. These applications have been divided into power generation, co-production of power and synthesis gas /hydrogen, and cogeneration applications. Each application is covered in details in the following sections. POGT for Power Generation The benefits of POGT in power generation systems when compared to those based on a conventional gas turbine are that it provides fundamentally higher energy conversion efficiencies and inherently lower nitrogen oxide (NOx) emissions without any catalytic combustion or post combustion catalytic treatment. Typically in the POGT power generation approaches more of the oxygen in the air is consumed in the staged combustion arrangement than in conventional gas turbine power systems; O2 in the stack is about 3% for POGT systems vs. 14-16% for conventional systems. This leads to the generally higher conversion efficiencies. The overall combustion system can be regarded as being similar to a richlean combustor with an expansion turbine located between the rich and lean sections. The expansion process cools the gases that enter the lean combustion section creating easier premixing and lower flame temperatures than are encountered in a standard rich-lean combustor. It is this critical difference that allows the POGT system to provide NOx emission levels less than 3 parts-per-million, by volume (ppmv), dry (corrected to 15% oxygen). The POGT does require a more advanced control system than a conventional gas turbine primarily for starting and shut-down. In particular if the exhaust fuel gas composition is to be varied then additional control features have to be added to the basic system. The integration of the POGT with bottoming systems such as a steam based combined cycle is straightforward as is the integration with a fuel cell. Using a gas turbine as a bottoming cycle is more difficult and usually results in either a sub-atmospheric pressure system being employed or forward integration of the gas turbine with the POGT. A steam bottoming cycle involving the return of all of the steam generated (in a fuel fired boiler) to the POGT is shown in figure 3. This is considered to be a steam injected simple cycle. This general approach is similar to steam injected gas turbines and produces a significant increase in power over the non-steam injected case. A combined cycle version of this steam injected POGT is shown in figure 4. In this arrangement the steam produced in the bottoming cycle is split into two streams. One of the streams is injected into the POGT proper and the other is used to drive a steam turbine. This particular approach using steam injection not only increases the power output of the POGT but also increases the hydrogen content of the exhaust secondary fuel gases as mentioned above. The higher hydrogen content provides improved stability for the lean combustion systems that are used to fire the boilers. This improved stability allows the boiler combustion systems to operate at very lean (low NOx emission) conditions. The injected steam not only enters into the POR reforming reactions but is also be used for cooling of the POR and hot-section structures where needed.
123
Fig. 3. Steam Injected Simple Cycle POGT
Fig. 4. Steam Injected Combined Cycle POGT
1.3.3 Partial Oxidation Gas Turbine (POGT) Cycles The use of a gas turbine as the bottoming cycle typically involves a much closer integration than say with a fuel cell or boiler. This forward integration results in a configuration that has been termed an Air Staged Reheat (ASR) system. Perhaps the simplest version of an ASR in an all electric power generating system is shown as a schematic in figure 5. In gas turbine terminology this could be a two shaft, single-spool arrangement with a reheat combustor. The POGT turbine provides power to both the low pressure and high pressure compressors while a fired (power) turbine produces the power for export. This configuration has high overall efficiencies because of the higher levels of oxygen consumed in the two stage combustion process when compared to a single combustor arrangement. The fuel cell integrated as a bottoming cycle with a POGT is shown in figure 6. In this particular version the fuel cell air is provided by an auxiliary fan while the fuel is supplied as the POGT exhaust. This approach of using the POGT exhaust which is a high hydrogen content fuel gas for the fuel cell lends itself well to both the solid oxide and molten carbonate fuel cells. Figure 6 shows an application involving a solid oxide fuel cell. In general the POGT is a better candidate topping cycle to the fuel cell than a conventional gas turbine. The POGT eliminates the need for a reformer and the integration problems that reformers create for fuel cell thermal management, as well as improves start-up and shut-down operation.
Fig. 5 Integrated Air Staged Reheat POGT
POGT for Co-Production of Hydrogen and Power One of the more promising applications of the POGT is the co-production of hydrogen and electrical power. In figure 7 a POGT system intended for the production of both synthesis gas or hydrogen (from natural gas) together with electrical power is shown. It shares many features with the steam injected systems shown in figures 3 and 4. Instead of reinjecting all of the steam generated into the POR part of the steam produced is reacted with part of the POGT exhaust in a shift reactor to increase the hydrogen content. Part of the exhaust is burned in a boiler to produce the required steam. This particular POGT arrangement is a very effective approach to producing synthesis gas or hydrogen and can be viewed as a “self-powered” reformer when this kind of application is needed.
Fig. 6. Hybrid Fuel Cell POGT System
POGT for Cogeneration with Industrial Furnaces, Boilers or Chillers The use of POGT systems as the drivers for cogeneration is likely to be very effective primarily because of the use of the POGT exhaust as a fuel. Not only is the exhaust thermal energy made available as in a conventional gas turbine cogeneration system but the fuel when combusted boosts the application
Fig. 7. POGT Co-Production of Power and Hydrogen
124
Joseph K. Rabovitser, Ph.D., Serguei Nester, Ph.D., temperatures to higher levels. These higher temperatures increase the efficiency of the bottoming cycle. The exhaust which is a low to medium energy gas typically has low flame temperature and thus produces minimal levels of NOx. An arrangement of the POGT with a high temperature furnace (HTF) or an industrial boiler (IB) is shown in figure 8. High temperature steam suitable for process uses or even power generation is typically produced. Furnaces such as those used in steel annealing and reheat, glass melting, or aluminum reclamation are candidates for this approach because the associated processes typically require both electrical power and high temperatures.
1.3.3-5 Conclusions POGT is a highly flexible device that when integrated with a bottoming cycle can provide significant improvements over conventional gas turbines in both efficiency and gaseous emissions particularly in small megawatt-size power generation systems. The core POGT because of its very high specific power (kW/ (lb/s)) should have a lower specific cost ($/kW) than a conventional gas turbine. Thus the POGT represents a promising type of gas turbine which could be widely used for power generation, cogeneration, and co-production of power and hydrogen, syngas or chemicals. POGT systems efficiency is in the lower fifties for simple cycle systems and the upper sixties for combined cycles. Typically the NOx emission levels are below 3-ppmv without post combustion catalytic treatment. A conventional gas turbine could be converted to a POGT by replacement of the conventional combustor with a POR, and by downsizing the compressor. Modifications to the turbine and the hot section cooling systems could also be needed.
1.3.3-6 Acronyms and Abbreviations AC FFB HPAC HTF IB LPAC LPT LRC POR POGT
125
Air Compressor Fuel Fired Boiler High-Pressure Air Compressor High Temperature Furnace Industrial Boiler Low -Pressure Air Compressor Low-Pressure Turbine Lean Reheat Combustor Partial Oxidation Reactor Partial Oxidation Gas Turbine
Fig. 8. High Temperature Furnace or Industrial Boiler Cogeneration
1.3.3 Partial Oxidation Gas Turbine (POGT) Cycles 1.3.3-7 Notes _________________________ 1. J.K. Rabovitser, M. J. Khinkis, R. L. Bannister, and F. Q. Miao, “Evaluation of Thermochemical Recuperation and Partial Oxidation Concepts for Natural Gas-Fired Advanced Turbine Systems,” presented at International Gas Turbine and Aeroengine Congress & Exhibition, Birmingham UK, June 1996. 2. S.A. Christianovich, V. M. Maslennikov, and V. L. Sterenberg, “Steam-Gas Power Stations with Multi-Stage Residual Oil Combustion,” Applied Energy, Great Britain, 2 (1995):175-187. 3. J. Ribesse, “Gas Turbine with Catalytic Reactor for the Partial Oxidation of Natural Gas and Its Application in Power Stations,” Gas Wärme International,July-August, 1971. 4. J.J. Ribesse, “Isothermal Gas Turbine Using Catalytic Partial Oxidation,” International Patent WO 91/05946, May 2, 1991. 5. J.J. Ribesse, “The Isotherm Partial Oxidation Gas Turbine,” Eur. J. Mech. Eng. 36 no. 1 (1991): 27-32. 6. V.M. Maslennikov and V. J. Sterenberg, “Steam-Gas Units for the Modification of Existing Steam Power Plants,” IVTAN, Moscow, 1992. 7. See note 1 above. 8. R.C. Hodrien and G.W. Fairbairn, “Power Into the 21st Century,” Gas Engineering & Management (March, 1994). 9. B. Kalitventzeff, M. N. Dumont, and F. Marechal, “Process Integration Techniques in the Development of New Energy Technologies: Application to the Isothermal Gas Turbine,” Univ. Liege, Belgium; M.A. Korobitsyn, P.W. Kers, and G.G. Hirs, “Analysis of a Gas Turbine Cycle with Partial Oxidation,” American Society of Mechanical Engineers (Paper), 98-GT-33 (1998): 7;G. Heyen and B. Kalitventzeff, “A Comparison of Advanced Thermal Cycles Suitable for Upgrading Existing Power Plants,” Applied Thermal Engineering 19 (1998): 227-237; F. Desmaré and G. Heyen, “High-Temperature Heat and Power Generation using a partial Oxidation Gas Turbine : Application to an annealing Furnace,” IcheaP-5 Symposium, Florence, May 20-23, 2001; M. Korobitsyn, “Enhancing Direct-Fired Power Plants Performance by Use of Gas Turbine Technology,” Journal of Propulsion and Power 16 no. 4 (2000): 568-571; B. Kalitventzeff, M.N. Dumont, and F. Maréchal, “Process Integration Techniques in the Development of New Energy Technologies : Application to the Isothermal Gas Turbine, proceedings of Chisa 1998, August 24-28, Prague; B. Albrecht, “Reactor Modeling and Process Analysis for Partial Oxidation of Natural Gas,” (PhD diss. University of Twente, Oct. 14, 2004). 10. See note 1 above. 11. Newby et. al, “An Evaluation of a Partial Oxidation Concept for Combustion Turbine Power Systems,” ASME paper 97AA-24, 1997
126
BIOGRAPHY
1.3.3 Partial Oxidation Gas Turbine (POGT) Cycles
Joseph K. Rabovitser, Ph.D. Gar Technology Institute, Energy Utilization center 1700 S. Mount Prospect Road Des Plaines, IL 60018 phone: 847-768-0548 email:
[email protected]
Dr. Joseph Rabovitser is a director of power generation, at the Gas Technology Institute. Since 1994, he has been involved in the development of the partial oxidation gas turbine (POGT) technology, and currently he is the project manager / principal investigator of the ongoing project “Development of a POGT for Combined Electricity and Hydrogen Enriched Fuel Gas Generation,” and he directs several other research programs including development and deployment of high efficiency and ultra-low NOx boilers, burners for gaseous and solid fuels, and novel partial oxidation gas turbine for CHP and multi-stream cogeneration system. Dr. Rabovitser has over 30 years of extensive experience in R&D, engineering, and computer modeling of various power plants equipment. He has over 145 publications, including three books (with co-authors), 44 articles in technical journals and proceeding, and 32 patents.
Serguei Nester, Ph.D. Gas Technology Institute, Energy Utilization center 1700 S. Mount Prospect Rd. Des Plaines, IL 60018 phone: 847-768-0541 email:
[email protected]
Dr. Serguei Nester is a senior engineer at Gas Technology Institute, Des Plaines, Illinois. He conducts combustion research and development for industrial applications. His responsibilities include CFD modeling, design, development, and testing of novel combustion equipment. Currently Dr. Nester is involved in the development of the Partial Oxidation Gas Turbine (POGT) technology, including Partial Oxidation Reactors, and small and midsize partial oxidation gas turbines, POGT cycle analysis, combinations of POGT with boilers, furnaces and fuel cells, gas turbine/fuel cell hybrids, fundamental studies of partial oxidation of natural gas. Also, he is involved in the development of downstream supplemental firing combustion equipment.
David James White TRITEK Consulting 3633 Millikin Avenue San Diego, CA 92122-2413 phone: 858-453-8653 email:
[email protected]
Mr. David J. White is the president of TRITEK Consulting whose specialty is future gas turbine technologies. Mr. David J. White is a chemical engineer and combustion specialist with degrees from Manchester University (BSc.) and Royal College of Aeronautics (MSc.). He has worked in a research capacity for a number of companies including Rolls-Royce, Garrett AiResearch, and Solar Turbines Incorporated. He retired early from Solar Turbines Incorporated and started TRITEK Consulting. Mr. David J. White has provided valuable contributions to a number of programs including: • Variable power afterburners for the Rolls-Royce Spey-Engined Phantom. • Hypersonic Ramjet Engine (HRE) for the X-15 (Garrett AiResearch) • Advanced Turbine Systems (Solar Turbines Incorporated) • Low NOx Combustion Systems (Several Companies) • Partial Oxidation Gas Turbine Design (GRI)
1.4
Hybrid Gas Turbine Fuel Cell Systems
Professor Jack Brouwer, Ph.D. Associate Director National Fuel Cell Research Center University of California Irvine, CA 92697-3550 email:
[email protected] http://www.nfcrc.uci.edu
1.4-1 Introduction With increasing energy demands, dwindling fossil energy resources, and environmental concerns associated with criteria pollutants and greenhouse gases, significant attention in the gas turbine community has been focused on increasing efficiency and reducing emissions. A highly efficient and low emitting concept that has been considered for the future is the hybrid gas turbine high temperature fuel cell concept. Hybrid fuel cell technologies may enable the U.S. to meet its future energy demands while enhancing energy efficiency, reliability and security, and reducing environmental impact. Hybrid systems are comprised of integrated gas turbines and fuel cells with other technologies. A myriad of potential configurations exists with hundreds of cycles proposed and investigated. In each case these hybrid cycles exhibit a synergistic energy and environmental performance enhancement through novel individual technology components, unique systems integration, advanced energy conversion devices, innovative pollutant mitigation approaches, and/or increased fuel flexibility and applicability. These types of hybrid systems have been developed and proposed for operation on natural gas, coal, biomass and other fossil fuels. Both experimental and theoretical analyses of such hybrid gas turbine fuel cell systems have indicated that such hybrid systems can achieve very high fuel-to-end-use efficiency and very low emissions. The environmental and energy efficient performance of these hybrid systems could allow them to make major contributions to new and secure fossil-fueled energy infrastructure and could assist in the provision of fuels, value added products, and introduction of the hydrogen economy. Integrated hybrid cycles exhibit synergies not present in typical combined cycles with fuel-to-electricity efficiencies higher than either the fuel cell or gas turbine alone and costs for a given efficiency that may become lower than either alone. Significant improvement of high temperature fuel cell technology robustness and cost is required for the development of hybrid gas turbine fuel cell systems. The advancement of high temperature fuel cell technology in the last decade has been significant and expectations are that it will become commercially viable in coming decades. Once high temperature fuel cells become commercially viable, stand-alone fuel cell systems may compete with gas turbine technology in the electricity production sector. However, this will not occur without a natural evolution toward significant use of hybrid systems that use both gas turbine and fuel cell technology. This natural evolution will be driven by the superior efficiency and emissions performance of hybrid systems. Economies, industry, citizens and the environment could all benefit from the advancement and deployment of gas turbine fuel cell hybrid systems due to high energy efficiency, and reduced environmental impact. No fossil-fuel based technology can compete with the high efficiency and environmental performance of gas turbine fuel cell hybrid systems. In addition, the market applications for hybrid gas turbine fuel cell technologies are myriad. They include the future potential application to large central station power plants operated on a variety of fuel resources, distributed generation support of traditionally energy intensive industries, local commercial applications and various distributed generation scenarios. In addition, hybrid fuel cell technologies can be used to support the auxiliary power and propulsion power needs of aircraft, spacecraft, satellites, ships, and trains. Although the potential for gas turbine fuel cell hybrid systems is significant, the front-end risk associated with developing this technology is considerable. Broad investment in industry, at national laboratories, and in university research and development is required to advance hybrid gas turbine fuel cell technology.
1.4-2 Background
127
Hybrid gas turbine fuel cell systems are comprised of two major components, a high temperature fuel cell and a gas turbine engine. Since this handbook provides sufficient background information on gas turbine technology, background information on fuel cell technology for use in integrated hybrid cycles is the focus of this section. Brief background information regarding gas turbine technology for hybrid applications is included.
1.4-3 Fuel Cell Technology Fuel Cell History and Background In the late 1830s, Sir William Grove and Professor Christian Friedrich Schoenbein discovered fuel cells. The discovery was accomplished by postulating and proving that the process of water hydrolysis could be reversed through the provision of hydrogen and oxygen to the electrode surfaces of an electrolytic cell to produce water and electricity. Unfortunately, the fuel cells of Sir Grove and his contemporaries did not garner serious interest in an era where high-powered steam engines were proving to produce power levels of interest. In the early 1960s, however, fuel cells began to be developed as a power generating technology for space applications that required strict environmental and efficiency performance. The successful demonstration of efficient and environmentally sensitive fuel cells in space led to their serious consideration for terrestrial applications in the 1970s. From 1970 through the 1980s, fuel cell research and development was confined to a small number of companies and research institutions. Due to the emergence of several new fuel cell types (e.g. solid oxide and molten carbonate fuel cells), the last decades have produced a tremendous expansion and diversification of developers and manufacturers, which has expanded the list of potential products and applications of fuel cells. Almost all research and development efforts have pointed to and successfully demonstrated the environmentally sensitive features of fuel cell technology, principally regarding ultra-low to zero criteria pollutant emissions. In addition, many efforts have proven that fuel cells can produce electricity with fuel-to-electricity conversion efficiencies that are higher than similarly sized traditional electricity production technology (e.g., gas turbine). The advancement of fuel cell technology has been accomplished with significant investments from the U.S. Department of Energy, European Union, Japanese New Energy Technology Development Organization (NEDO) and similar agencies around the world together with hundreds of companies. Significantly, the largest annual investments in fuel cell technology today are coming from the private sector and major industries (e.g., automobile, power generation). These investments are often focused on overcoming the principal barrier to widespread use of fuel cells in today’s market, fuel cell capital cost. There are many specific technical challenges and technical hurdles that contribute to the high capital cost of fuel cell technology that the fuel cell community is currently addressing to make fuel cells commercially viable. While the details of these technical challenges are not discussed herein, note that until and unless fuel cell technology itself becomes commercially viable the potential for hybrid gas turbine fuel cell systems will be limited. Note additionally, however, that government agencies, national laboratories, researchers and industry agree that successful commercialization and market viability of fuel cells is likely to occur within the decade.
General Fuel Cell Characteristics Fuel cells are electrochemical devices that convert the chemical energy of a fuel directly to usable energy - electricity and heat - without combustion. This is quite different from most electric generating devices (e.g., steam turbines, gas turbines, reciprocating engines), which first convert the chemical energy of a fuel to thermal energy, then to mechanical energy and finally to electricity. Fuel cells are similar to batteries containing electrodes and electrolytic materials to accomplish the electrochemical production of electricity. Load 2e Batteries store chemical energy in an electrolyte and convert it to electricity on demand, until the chemical energy has been depleted. Applying Oxidant In Fuel In an external power source can recharge depleted secondary batteries, but primary batteries must be 1/2 O2 replaced. Fuel cells do not store chemical energy, H2 Positive Ion but rather, convert the chemical energy of a fuel to or electricity. Thus fuel cells do not need recharging Negative Ion and can continuously produce electricity as long as H 2O fuel and oxidant are supplied. Figure 1 presents the basic components of a H 2O fuel cell, which include a positive electrode (anode), negative electrode (cathode) and an electrolyte. Fuel is supplied to the anode (positive electrode) while oxidant is supplied to the cathode (negative electrode). Fuel is electrochemically oxidized on Depleted Fuel and Depleted Oxidant and the anode surface and oxidant is electrochemically Product Gases Out Product Gases Out reduced on the cathode surface. Ions created by the Cathode Anode Electrolyte electrochemical reactions flow between anode and (Ion Conductor) cathode through the electrolyte. Electrons produced at the anode flow through an external load to the Fig. 1. General schematic of a fuel cell cathode completing an electric circuit.
128
Jack Brouwer A typical fuel cell requires gaseous fuel and oxidant flows. Hydrogen is the preferred fuel because of its high reactivity, which minimizes the need for expensive catalysts, and because electro-oxidation of hydrogen leads only to water emission. Hydrocarbon fuels can be supplied but typically require conversion to hydrogen or a hydrogen-rich mixture before electrochemical reaction can occur. This fuel processing step can be accomplished prior to entering the fuel cell (for lower temperature fuel cells) or within the fuel cell (for higher temperature fuel cells). Oxygen in air is the preferred oxidant because of its availability in the atmosphere. As indicated in Figure 1, the electrolyte serves as an ion conductor. The direction of ion transport depends upon the fuel cell type, which determines the type of ion that is produced and transported across the electrolyte between the electrodes. The various fuel cell types are described in a subsequent section. A single fuel cell is only capable of producing about 1 volt, so typical fuel cell designs link together many individual cells to form a “stack” that produces a more useful voltage. A fuel cell stack can be configured with many groups of cells in series and parallel connections to further tailor the voltage, current and power produced. The number of individual cells contained within one stack is typically greater than 50 and varies significantly with stack design. Figure 2 presents the basic components that comprise the fuel cell stack. These components include the electrodes and electrolyte of Figure 1 with additional components required for electrical connections and to provide for the flow of fuel and oxidant to each cell in the stack. These key components include current collectors, separators, and gas flow channels, which are often integrated into one design as in the “interconnect” design pictured in Figure 2. This interconnect serves as current collector and Fig. 2. Basic components of a fuel cell stack gas separator and provides the flow channels for both fuel and oxidant. The interconnect provides the electrical connections between cells and physically separates the oxidant flow of one cell from the fuel flow of the adjacent cell. The channels serve as the distribution pathways for the fuel and oxidant. The preferred fuel for most fuel cell types is hydrogen. Hydrogen is not readily available, but, and the infrastructure for provision of hydrocarbon fuels is well established in our society. Thus, fuel cell systems that have been developed for practical power generation applications to-date have been designed to operate on hydrocarbon fuels. This typically requires the use of a fuel processing system or “reformer” as shown in Figure 3. The fuel processor typically accomplishes the conversion of hydrocarbon fuels to a mixture of hydrogen rich gases and, depending upon the requirements of the fuel cell, subsequent removal of contaminants or other species to provide pure hydrogen to the fuel cell. In addition to the fuel cell system requirement of a fuel processor for operation on hydrocarbon fuels, Figure 3 presents the need for a power conditioning or inverter system component as well. This is required for the use of current end-use technologies that are designed for consuming alternating current (AC) electricity, and for grid connectivity in distributed power applications. Since the fuel cell produces direct current (DC) electricity, the power conditioning section is a requirement for fuel cell systems that are designed for distributed generation today. In the future, systems and technologies may be amenable to the use of DC electricity, which would allow significant cost savings. EXHAUST EXHAUST HEAT
WATER
USEFUL HEAT
FUEL
FUEL PROCESSOR
HYDROGEN RICH GAS
POWER GENERATOR (FUEL CELL)
DC POWER
POWER CONDITIONER
AIR
POWER GENERATOR (FUEL CELL STACK)
Fig. 3. Basic schematic of the major components in a fuel cell system
129
AC POWER
1.4 Hybrid Gas Turbine Fuel Cell Systems Fuel Cell Types There are five principle types of fuel cells that are currently in various stages of commercial availability, or undergoing research, development and demonstration. These five fuel cell types are significantly different from each other in many respects; however, the key distinguishing feature is the electrolyte material. The type of electrolyte material is generally used to describe each fuel cell type. Thus the five types of fuel cells are (in alphabetical order): (1) Alkaline Fuel Cell (AFC), (2) Molten Carbonate Fuel Cell (MCFC), (3) Phosphoric Acid Fuel Cell (PAFC), (4) Proton Exchange Membrane Fuel Cell (PEMFC), and (5) Solid Oxide Fuel Cell (SOFC). Alkaline fuel cells were the first type of fuel cell to be widely used for space applications. AFCs contain a potassium hydroxide (KOH) solution as the electrolyte. AFCs operate at temperatures between 100oC and 250oC (211oF and 482oF). Molten carbonate fuel cells are now being tested in full-scale demonstration plants from 250 kW to 1.5 MW output. The electrolyte in an MCFC is an alkali carbonate (sodium, potassium, or lithium salts, Na2CO3, K2CO2, or Li2CO3) or a combination of alkali carbonates that is retained in a ceramic matrix of lithium aluminum oxide (LiAlO2). An MCFC operates at 600 to 700oC where the alkali carbonates form a highly conductive molten salt with carbonate ions (CO3=) providing ionic conduction through the electrolyte matrix. Phosphoric acid fuel cells have been commercialized for stationary power applications and remain the most mature fuel cell technology. PAFCs use a concentrated 100% phosphoric acid (H3PO4) electrolyte retained on a silicon carbide matrix and operate at temperatures between 150 and 220oC. The proton exchange membrane fuel cell is also known as the solid polymer or polymer electrolyte fuel cell. A PEMFC contains an electrolyte that is a layer of solid polymer (usually a sulfonic acid polymer, whose commercial name is NafionTM) that allows protons to be transmitted from one face to the other.1 PEMFCs require relatively pure hydrogen that typically must be humidified. PEMFCs operate at a temperature of below 90oC because of limitations imposed by the thermal properties of the membrane itself.2 The development of PEMFC technology is primarily sponsored by the transportation and portable power market sectors. Solid oxide fuel cells are currently being demonstrated in various sizes from 1kW up to 250-kilowatt plants. SOFCs utilize a non-porous metal oxide (usually yttria-stabilized zirconia, Y2O3-stabilized ZrO2) electrolyte material. SOFCs operate between 650 and 1000oC, where ionic conduction is accomplished by oxygen ions (O=). Table 1 presents a summary comparison of the four primary fuel cell types that are under serious consideration for power generation applications. Notice that the higher temperature fuel cells do not require an external reformer. The PAFC and PEMFC units tend to use precious metal catalysts, while those of the MCFC and SOFC units are typically nickel-based. These differences lead to many differences in design and function, which will be described in more detail in the next section. Table 1. Key features of the four fuel cell types used in power generation applications after Hirschenhofer et al. (1998).3 High temperature fuel cells highlighted.
Feature
MCFC Immobilized Molten Carbonate
PAFC
PEMFC
SOFC
Immobilized Phosphoric Acid
Ion Exchange Membrane
Ceramic
600-650oC
200oC
80oC
600-1000oC
Charge Carrier
CO3=
H+
H+
O=
External Reformer for natural gas
No
Yes
Yes
No
Prime Cell Components
Stainless Steel, nickel, carbonate salts
Graphite, Teflon, phosphoric acid
Carbon, plastics, special polymers
Catalyst
Nickel
Platinum
Platinum
Product Water Management
Gaseous Product
Evaporative
Evaporative
Product Heat Management
Internal Reforming + Process Gas
Process Gas + Independent Cooling Medium
Process Gas + Independent Cooling Medium
Electrolyte Typical Operating Temperature
Ceramics, high temperature metals Nickel, Perovskites Gaseous Product Internal Reforming + Process Gas 130
Jack Brouwer While any one of the above fuel cell types can be integrated into a hybrid gas turbine fuel cell cycle, the advantages of integration are most prominent with the high temperature fuel cells (i.e., MCFC and SOFC). This is due to the fact that a gas turbine engine can more effectively utilize the heat produced at the higher operating temperatures of MCFC and SOFC technology than it can that produced by other fuel cell types. In a complementary fashion, the MCFC and SOFC technologies can directly benefit from the pressure and temperature conditions (higher pressure and preheating of reactants) that a gas turbine engine can produce in an integrated hybrid cycle. As a result of this complementary operation, the focus of this chapter and all remaining discussion will be on hybrid systems that use high temperature fuel cells (MCFC and SOFC) only.
High Temperature Fuel Cells High temperature fuel cells (HTFCs), such as solid oxide fuel cells and molten carbonate fuel cells are especially well suited for hybrid operation. The high operating temperature of these fuel cell systems allows integration with a gas turbine engine at a temperature that is mutually beneficial. Depending upon the design of the integrated hybrid system, waste heat from the fuel cell could be converted in the gas turbine engine to electricity or waste heat from the gas turbine could be put to good use preheating the reactants for the fuel cell or providing heat for fuel processing.
MCFC The MCFC, also called a carbonate fuel cell, is one of the fuel cell technologies that has proven efficiency and environmental performance. In addition, significant reductions in carbonate fuel cell capital cost are expected in the near future. In particular, the use of carbonate fuel cells in the distributed power market is already significant and could offer an ideal solution to increased energy demands with concurrent expectations for reliability and environmental sensitivity. The carbonate fuel cell concept involves conduction of carbonate ions (CO3=) within an immobilized mixture of molten carbonate salts. Other cell components are based on nickel and stainless steels, which contribute to initial capital cost, but, are significantly less expensive than the precious metal catalysts used in lower temperature fuel cells. Relatively inexpensive nickel (Ni) and nickel oxide (NiO) are adequate to promote reaction on the anode and cathode respectively at the high operating temperatures of an MCFC.4 Since the charge carrier is an oxidant, several fuel species can be oxidized within the anode compartment leading to inherently greater fuel flexibility. To-date, carbonate fuel cells have been operated on hydrogen, carbon monoxide, natural gas, propane, landfill gas, marine diesel, and simulated coal gasification products. The typical operating temperature of a carbonate fuel cell is around 650oC. This temperature is almost ideal from the system perspective, since it allows higher Nernst potential (ideal Nernst potential increases with decreasing temperature) while still providing high temperature thermal energy sufficient to sustain and support reformation chemistry. Thus carbonate fuel cell system designs typically contain an internal reformer. The carbonate fuel cell demonstrations to-date, have therefore been able to show the highest fuelto-electricity conversion efficiencies of any stand-alone fuel cell type. The primary developer of carbonate fuel cell technology is FuelCell Energy Corporation, the developer and manufacturer of the Direct Fuel CellTM concept. FuelCell Energy has demonstrated carbonate fuel cells from 10kW to 2MW of electrical output on a variety of fuels. Hitachi and IHI are also developing carbonate fuel cell technology for stationary power applications and have recently, successfully demonstrated carbonate fuel cell technology in Kawagoe. Japan. Ansaldo Ricerche has also demonstrated a 100kW carbonate fuel cell in Milan, Italy. Carbonate fuel cell systems have the highest fuel-to-electricity conversion efficiency (>50%) of any fuel cell type. In addition, carbonate fuel cell technology is expected to experience dramatic initial capital cost reductions in upcoming years. Carbonate fuel cell technology is more fuel flexible than lower temperature fuel cell technologies and is well suited to marine, military, and traction applications. The high temperature thermal effluent of a carbonate fuel cell allows significant co-generation and/or integration with a heat engine cycle in hybrid applications. Several carbonate fuel cell hybrid systems with fuel-to-electricity efficiencies greater than 70% have been conceptualized with some under development today. Hybrid MCFC systems have been developed and tested by FuelCell Energy and Capstone Turbine in Danbury, Connecticut and in Japan.
SOFC
131
A SOFC is a solid state fuel cell constructed of ceramic materials (metal oxides) and metals. SOFCs share the solid state electrolyte feature with the proton exchange membrane fuel cell (PEMFC). Solid state construction offers the potential for increased reliability and durability with less corrosion and no need to manage electrolyte evaporation or circulation. Typically the anode of an SOFC is nickel zirconia (Ni-ZrO2) and the cathode is strontium-doped lanthanum manganite (Sr-doped LaMnO3).5 SOFCs offer the stability and reliability of all-solid-state ceramic construction. High-temperature operation, up to 1,000oC, allows more flexibility in the choice of fuels and can produce very good performance in combined-cycle and hybrid applications. SOFCs approach 50 percent electrical efficiency in the simple cycle systems operated on natural gas, and 85 percent total thermal efficiency in co-generation applications.6 The SOFC concept involves conduction of oxygen ions (O=) within the electrolyte at high temperatures (650-1000oC) making it inherently more fuel flexible than other fuel cell types. Whereas most other fuel cells are susceptible to carbon monoxide (CO) poisoning, SOFCs can use CO as a fuel to produce electricity. To-date, SOFCs have been operated on hydrogen, carbon monoxide, natural gas, propane, landfill gas, diesel and JP-8. The high temperature operation of a SOFC has advantages and disadvantages. The advantages include the use of high temperature heat to reform hydrocarbon fuels to hydrogen(H2)/carbon monoxide(CO) mixtures for direct use in the fuel cell. This reformation process
1.4 Hybrid Gas Turbine Fuel Cell Systems requires heat to proceed. The high temperature heat also allows significant co-generation and/or integration with a heat engine cycle. The disadvantages of high temperature operation include the need to insulate the technology to protect from injury and the requirement of more costly materials of construction. SOFCs have higher overall fuel-to-electricity efficiency than lower temperature fuel cells (e.g., PEMFC) operated on available hydrocarbon fuels (e.g., natural gas). When integrated with a heat engine cycle, efficiency can be increased even further. The hybrid SOFC cycle, which integrates a SOFC into a gas turbine cycle, offers the potential of fuel-to-electricity efficiencies in the 75-80% range. This remarkably high efficiency is unmatched by any other technology. Although in the early stages of development, hybrid designs and systems are now emerging with the first demonstration being accomplished by Southern California Edison and Siemens Westinghouse Power Corporation at the National Fuel Cell Research Center. SOFC systems are being advanced by a number of companies and organizations with three major fuel cell stack designs emerging. The major design types are tubular, planar, and monolithic. Tubular SOFC designs are closer to commercialization and are being produced by Siemens Power Corporation, Mitsubishi Heavy Industries, Acumentrics, among others. The planar and the monolithic designs are at an earlier stage of development typified by sub-scale, single cell and short stack development (kW scale). More than 100 companies are advancing and commercializing SOFC technology around the world and especially in the U.S., Europe and Japan. Primary U.S. SOFC companies include GE, Acumentrics, FuelCell Energy, Versa Power, Ceramatec, Inc., Technology Management, Inc., SOFCo, Cummins, and Ztek, Inc., among others. SOFC systems have been operated all over the world, proving SOFC performance and features. Examples include the tubular SOFC design of Siemens Westinghouse Power Corporation that has demonstrated over 85,000 hours of operation with low cell degradation, and the SOFCo planar SOFC design exhibiting power densities up to 1000W/l. Because of the high potential of SOFC technology to produce robust (long lasting), high power density, fuel flexible, and low cost fuel cell systems, significant industry and agency investment is currently focused on SOFC technology, especially in Europe, Japan and the United States. Notably, the Solid State Energy Conversion Alliance of the U.S. Department of Energy includes six industryled teams (General Electric, Siemens, Cummins, FuelCell Energy, Acumentrics, and Delphi) and a core technology research program including national laboratory and university researchers that is focused on developing low cost, high power density and robust SOFC technology.
Gas Turbine Technology for Hybrid Applications A typical hybrid system recovers the thermal energy in the fuel cell exhaust and converts it into additional electrical energy through a heat engine. Several heat engines have been considered for this type of system including gas turbines, steam turbines and reciprocating engines. The only conversion device that has been tested in this role to-date is a micro-gas turbine (or micro-turbine generator, MTG). An MTG is a type of gas turbine engine that is particularly amenable to integration with a high temperature fuel cell in a hybrid system. This is due to several features of an MTG that are well matched to the requirements of the high temperature fuel cell in the hybrid system such as: • MTGs require relatively low turbine inlet temperature, which can be supplied by the exhaust of a high temperature fuel cell, • MTGs operate at relatively low pressure ratios that are amenable to either direct use in the high temperature fuel cell or in other components of the hybrid system, • MTGs often use recuperation to improve efficiency, which introduces components and features (e.g., heat exchangers, large gas volume between compressor and turbine) that make the gas turbine engine design more amenable to a hybrid cycle, • The fuel cell can be operated under pressurized conditions improving it’s output and efficiency, • Sufficient thermal energy is contained in the fuel cell exhaust to power an MTG compressor (for fuel cell pressurization) and an electric generator (to produce additional electricity, • The current size of most fuel cell systems is relatively small (between 250kW and 1.5 MW), which matches well with the smaller output MTG, • The power density of the system can be increased, and • Overall system cost is potentially lower on a $/kW basis (primarily due to the increased output from the fuel cell). Note that the gas turbine engine characteristics noted above as desirable for hybrid applications are not necessarily those that are desired for stand-alone gas turbine engines. Usually one desires higher turbine inlet temperatures and higher pressure ratios to improve the performance of a gas turbine engine. In hybrid cycles with a high temperature fuel cell, the gas turbine engine is not required to operate at either high pressure ratios or with high turbine inlet temperature making the performance characteristics of the gas turbine relatively simpler to achieve. That is, less sophisticated gas turbine technology may be all that is required for a hybrid system, although improvements in compressor and turbine efficiency etc. are still desirable. Although the MTG is currently well-suited for integration into hybrid gas turbine fuel cell systems, future high temperature fuel cell technologies may become large and able to withstand significantly higher pressures. Analyses have shown that synergistic effects of the combined gas turbine fuel cell system lead to electrical conversion efficiencies of 72-74 percent (LHV) for systems under 10 MW, whereas efficiencies greater than 75% could be achieved with larger systems. As fuel cells advance and scale-up and pressurization of MCFC and/or SOFC technology becomes viable, larger and more sophisticated gas turbine engines (e.g., axial compressors and turbines, higher pressure ratios, high turbine inlet temperature) will be required.
132
Jack Brouwer 1.4-4 Hybrid Gas Turbine Fuel Cell Concept The gas turbine fuel cell hybrid system was first conceived in the mid-1970s. By 1998 over ten hybrid concepts had been patented, offering variations in fuel cell type, in the position of the components in the integrated system, and in system operating pressure. The basic concept of a hybrid gas turbine fuel cell system is illustrated in Figure 4 where a fuel cell replaces the combustor of a typical Brayton (gas turbine) cycle. This leads to direct fuel-to-electricity production from the fuel cell (in the place of chemical-to-thermal energy conversion of a combustor) with the waste heat of the fuel cell being used to provide all compression power and additional electricity through a turbo-generator. Note that electrochemical production of electricity lowers emissions and increases efficiency and as a result around 80% of the electricity is produced via electrochemistry in the fuel cell with the remainder being produced in the turbo-generator.
Fig. 4. Basic design concept of a hybrid gas turbine fuel cell system
System studies have been carried out for the U.S. DOE and others for hybrid systems up to 300 MWe capacity (using 40 MWe power blocks). In 2000 the first tests and demonstrations of hybrid gas turbine fuel cell systems began with efforts in the U.S. and Japan. Both MCFC and SOFC hybrid systems have been built and tested each proving the potential for such systems to achieve high efficiency and low emissions production of electricity from natural gas. To-date five hybrid gas turbine fuel cell systems have been tested, each using a different design concept.
Hybrid Design Concepts The basic design goal of a hybrid system is to integrate two or more energy conversion devices—or two or more fuels for the same device—into a single system that provides benefits in terms of fuel flexibility, efficiency, availability, economics or sustainability that either of the devices alone could not provide. Synergy is the term often used to describe hybrid systems since the components when integrated are complementary and lead to combined performance characteristics that are greater than the sum of the individual elements. In the particular case of a gas turbine fuel cell hybrid system the primary design concepts are as follows: (1) convert most of the fuel by electro-oxidation in the fuel cell leading to low emissions of criteria pollutants and relatively high fuel-to-electricity conversion efficiency, (2) use fuel cell heat and turbine exhaust heat elsewhere in the system (e.g., fuel processing, reactant preheating, provide compression power) in a manner in which overall efficiency is enhanced, (3) use the high pressure produced by the gas turbine in a manner that improves fuel cell output and efficiency (if possible), and (4) use the separated fuel and oxidant streams of the fuel cell to enhance other features (e.g., CO2 sequestration, fuel production) of the hybrid cycle (if possible). Synergy can be realized in hybrid cycles in many ways, as indicated by the last two concepts introduced above. The operation of a fuel cell at the high pressure conditions between the compressor and turbine of a gas turbine engine leads both to an increase in the fuel cell power output (for the same fuel cell stack) and a reduction in some of the electrochemical losses leading to higher efficiency (most notably improving electrochemical kinetics). In addition, the fact that a fuel cell operates with separated fuel and oxidant streams provides multiple opportunities for enhanced hybrid cycle performance from easier sequestration of a higher concentration CO2 stream to significant potential for hydrogen and/or synthetic hydrocarbon fuel production. These synergistic aspects of hybrid gas turbine fuel cell systems are of particular interest and are part of the reason why analyses and testing to-date has determined that such hybrid systems have great promise.
133
1.4 Hybrid Gas Turbine Fuel Cell Systems Hybrid System Configuration The fuel cell and gas turbine of a hybrid system can be configured in several different fashions with a myriad of potential cycle configurations that are possible. Four basic parameters can be defined that provide a means of characterizing the hybrid system configurations. The parameters are: (1) Fuel cell topping cycle, (2) Fuel cell bottoming cycle, (3) Direct hybrid cycle, and (4) Indirect hybrid cycle. Most hybrid cycles that have been conceived and studied to-date can be characterized using these four parameters. A fuel cell “topping” cycle is one in which the gas turbine is considered the balance of plant (BOP) with the turbo-machinery placed downstream of the fuel cell in the cycle. The basic design concept presented in Figure 4 represents this type of fuel cell topping cycle. Essentially fuel cell topping cycles use a fuel cell in the place of a combustor in the typical Brayton cycle, and the turbine is placed downstream of fuel cell. The turbine uses the fuel cell exhaust to produce compressive power and additional electricity while the fuel cell is the primary electricity generator. A fuel cell “bottoming” cycle is one in which the gas turbine turbo-machinery resides upstream of the fuel cell. The fuel cell is placed downstream of turbine and uses the gas turbine exhaust as its air supply stream. Typically the fuel cell remains the primary generator. This type of bottoming cycle is particularly well-suited to the MCFC since this type of fuel cell requires carbon dioxide in the oxidant stream (to make the carbonate ions), which can be provided by an upstream gas turbine engine combustor. In a “direct” hybrid cycle flow from upstream elements is directly used in downstream elements of the cycle. Heat exchangers to de-couple to two cycles are not required, but may be used for other purposes. The fuel cell of a direct hybrid cycle is typically operated at pressurized conditions extant between the compressor and turbine of the gas turbine. This presents more significant challenges with control and with fuel cell operation and degradation. However, direct hybrid cycles typically have higher efficiency that indirect hybrid cycles. An “indirect” hybrid cycle uses devices (usually heat exchangers) to de-couple the gas turbine and fuel cell components of the system so that flow from upstream components does not enter downstream components. Thermal integration of the cycle involves more losses (e.g., in the additional heat exchangers) and the fuel cell is typically operated at atmospheric pressure. This leads to a less challenging system to control and operate and lesser challenges for fuel cell operation and degradation. However, indirect hybrid cycles tend to be less efficient than direct cycles and the cost and size of heat exchanger components can be considerable. Figure 5 presents an example of a direct hybrid gas turbine fuel cell topping system configuration in which the fuel cell is operated in-between the compressor and turbine of the gas turbine engine. Note that a recuperator (heat exchanger) is still used in this cycle configuration to preheat the fuel and air before it enters the fuel cell.
Fig. 5. Schematic of a direct hybrid gas turbine fuel cell topping cycle Source: See note 7.
134
Jack Brouwer
Fig. 6. Schematic of an indirect gas turbine fuel cell bottoming cycle Source: See note 7.
Figure 6 presents a schematic of an indirect fuel cell bottoming cycle in which the turbine operates on air that does not come into contact with the fuel cell exhaust, but rather receives heat through a heat exchanger. Note that the fuel cell operates at atmospheric pressure and uses the pure air exhaust from the turbine. From this brief introduction of the hybrid concept it should be apparent that myriad cycle configurations are possible. Any one of the cycles presented above could use additional heat exchangers, boilers, fuel or oxidant separations technologies, fuel production and purification equipment, a steam turbine bottoming cycle, and/or other devices. Depending on the size of the system and the desired products, the number of components could be large and the cycle could become quite complex. Most of these systems can be expected to have complex control issues that need to be resolved.
1.4-5 Early Hybrid Gas Turbine Fuel Cell Developments The first U.S. patents identifying hybrid fuel cell technologies were issued in the mid 1970s.7 More recently various fuel cell hybrid cycles have been identified in U.S. patents.8 The most noteworthy distinction between these patented arrangements is the placement of the fuel cell with respect to the turbine as well as the pressure under which the cell operates. Each of these patented hybrid cycle concepts can be characterized by the four parameters defined above.
Initial Analyses Significantly, it was not until the late 1990s, once high temperature fuel cell technology had progressed sufficiently to consider manufacturing large SOFC and MCFC stacks, that detailed analyses and experimental investigation of hybrid gas turbine fuel cell systems began in earnest. In 1998 the U.S. Department of Energy Office of Fossil Energy initiated five studies to conceptualize and assess variations on the fuel cell turbine hybrid concept. These studies, funded by the turbines program, included molten carbonate fuel cells (MCFC), solid oxide fuel cells (SOFC), off-the-shelf turbines, and conceptual turbines. Four of these studies examined cycle configurations in the 20-MW class power system. The fifth study, by McDermott, assessed a sub-MW cycle. Table 2 summarizes the results of these studies. Table 2. Overall results from the U.S. Department of Energy hybrid gas turbine fuel cell system studies initiated in 1998.
135
Company
FuelCell Energy
Siemens Westinghouse
Cycle configuration
MCFC Indirect
SOFC Turbine Bottoming
Nominal Size
20 MW
20 MW
Nominal Efficiency *
71 %
60 %
* These are nominal efficiencies and should not be directly compared
Siemens Westinghouse Staged SOFC Turbine Bottoming
M-C Power
McDermott
MCFC Turbine Bottoming
SOFC Indirect
20 MW
20 MW
750 kW
67 - 70 %
66 - 70 %
71 %
1.4 Hybrid Gas Turbine Fuel Cell Systems In 1999 the turbines program funded a study by Rolls Royce with the goal to produce a turbo-generator, which would cost approximately $400/kW. When coupled with fuel cells, the turbine would produce approximately 25% of the power for a hybrid in the 1 MW to 5 MW class. The gas turbine would be capable of providing pressurization from 5 pressure ratio (PR), to approximately 15 PR and higher, all from the same special purpose gas turbine system design. As a stand-alone device, the turbine would produce 1.5 MW of electric power in a simple cycle mode, without the need of a recuperator (recuperators are not needed in mini-turbines to achieve 30% efficiency, which reduces costs by 25-30%, reduces space requirements, and contributes to more reliable operation). In the stand-alone mode, its efficiency would be approximately 33% (comparable to larger 5 MW class gas turbines). The exhaust energy could be used to operate a combined heat and power cycle.
Detailed and Experimental Studies After these studies were completed several tests of the hybrid concept were initiated by the U.S. Department of Energy with industry cost-sharing from FuelCell Energy, General Electric (formerly Honeywell), and Siemens Power Corporation (formerly Siemens Westinghouse). Each of the projects focused on the sub-1MW class hybrid system. FuelCell Energy’s hybrid system is comprised of a 250 kW fuel cell stack and a 30 kW Capstone micro-turbine in an indirect fuel cell bottoming configuration. The sub-MW system tests have provided valuable data that has proven the high efficiency and low emissions performance of such systems. FuelCell Energy also designed a 40 MW fuel cell turbine hybrid power system in this effort. The General Electric project includes the sub-MW design and test of a Solid State Energy Conversion Alliance (SECA) solid oxide fuel cell and a micro-turbine. The project evaluated several turbine cycle configurations, including topping, bottoming, direct and indirect, and allowed for the evaluation of integration and scale-up issues for SECA-based hybrid systems. The Siemens Power Corporation hybrid design included a 100 kW tubular SOFC integrated with a 60 kW Ingersoll Rand microturbine generator. This system was built and tested at the National Fuel Cell Research Center, in Irvine, California. In this test the hybrid direct gas turbine fuel cell topping cycle configuration was demonstrated. This test included pressurization of the fuel cell to provide a total of 220 kW of power from the hybrid system. Testing proved that high efficiency and ultra-low emissions was achievable with these types of hybrid cycles, but, that integration and operation is considerably difficult with such complex hybrid systems. Early optimism regarding the ease of integrating fuel cells with off-the-shelf micro-turbines has been tempered by technical issues encountered in the test program at the National Fuel Cell Research Center. It is now recognized that integrating fuel cells and turbines is challenging. Existing gas turbines do not match the pressure ratios, mass flows, and other critical operating and performance parameters of the small high temperature fuel cells that are currently available. Nonetheless, the early tests have proven the high efficiency and ultra-low emissions performance characteristics of hybrid gas turbine fuel cell technology so that optimism regarding the potential of these types of cycles to significantly contribute to future energy demands remains high.
Detailed Paper Studies for Future Hybrid Systems Additional studies on hybrid systems and the results of recent tests identified significant potential benefits from a combined fuel cell/turbine power system. These benefits include the ability to achieve net electrical efficiencies in the 70 % + range, to configure systems in the 20 MW to 40 MW and larger size range, and to significantly surpass emission standards for criteria pollutants while reducing the emissions of CO2 / kW-hr. Studies also predict lower cost of electricity and lower capital costs than alternative power generation systems. For example, a market study by Research Dynamics Corporation suggested that such products could compete on a cost-of-electricity basis with other DG technologies and capture 8.2 GW of market share by 2005. The historical record of the evolution of hybrid gas turbine fuel cell technology has been documented by White9, and the various technical elements of the hybrid technology have been presented in a series of sessions sponsored by the American Society of Mechanical Engineers (ASME) International Gas Turbine Institute (IGTI).10 Under the sponsorship of the U.S. DOE, a multi-disciplinary team led by the Advanced Power and Energy Program (APEP) of the University of California, Irvine is defining the system engineering issues associated with the integration of key components and subsystems into central power plant systems that meet stretch performance and emission goals for both natural gas and coal fuel operation. The myriad of fuel processing, power generation, and emission control technologies are narrowed down to selected scenarios in order to identify those combinations that have the potential to achieve high efficiency and minimized environmental impact while using fossil fuels. The technology levels considered are based on projected technical and manufacturing advances being made in industry and on advances identified in current and future government supported research. Examples of systems included in these advanced cycles are high-temperature fuel cells, advanced gas turbines, ion transport membrane separation and hydrogen-oxygen combustion. The overall objectives of DOE study were to (1) produce electricity and transportation fuels at competitive costs, (2) minimize environmental impacts associated with fossil fuel use, and (3) attain high efficiency. The efficiency target for natural gas fueled plants was 75% on a LHV basis while that for coal fueled plants was 60% on an HHV basis. All cycles were to include producing electricity with the potential for CO2 capture and sequestration and co-production of transportation fuels. This study determined that the only technology that could meet these goals is hybrid gas turbine fuel cell technology.11
136
Jack Brouwer 1.4-6 Dynamic Simulation of Hybrid Systems In both thermodynamic simulation and experiment hybrid gas turbine fuel cell systems have demonstrated lower environmental impact and higher efficiency compared to conventional combustion driven power plants. Lower carbon dioxide emissions can be achieved through higher fuel-to-electrical efficiencies, while NOx and other criteria pollutant emissions are greatly reduced by primary electrochemical conversion of the fuel versus the combustion process of conventional plants. Understanding of the dynamic performance of hybrid systems is important to the advancement of the technology and the development of controls for future systems. In this section, a dynamic model of a hybrid system is described and applied to analyze a specific hybrid cycle that is applicable to distributed generation. More complex cycles have been considered for larger scale power plants that may utilize a combined cycle to drive the efficiency up and the environmental impact down.12 Today much work is being done to reduce the cost and increase the reliability of SOFC systems. Several cell geometries are being advanced by fuel cell manufacturers including tubular and planer SOFC designs, and even cell geometries that combine planer and tubular features. Each geometry has its advantages and disadvantages with regard to thermal expansion compliance, power density, potential cost, manufacturability, and internal resistivity.13 Many companies are advancing these different types of SOFCs, but no commercial products exist today. Only demonstration and prototype systems have been built and tested to-date. Mathematical models provide a cost effective and efficient tool in aiding the development of SOFCs and SOFC/GT systems. Several entities around the world have developed steady state simulation capabilities for FC/GT systems. These research groups include efforts at the Georgia Institute of Technology, University of Genova, NFCRC, Nanyang Technical University and others.14 Dynamic gas turbine fuel cell simulation capabilities are less common, but increasingly being developed as the demand for dynamic understanding and controls development grows. Examples of previous dynamic simulation efforts include work at the National Energy Technology Laboratory, and FuelCell Energy among others.15 Model evaluation is very important and there remains a great need to produce experimental hybrid system data. To-date there have been two hybrid systems built and successfully demonstrated. An indirect bottoming cycle (with respect to the FC) has been built and demonstrated by FuelCell Energy that integrated a molten carbonate fuel cell and a Capstone C30 gas turbine. This system successfully ran for 2900 hours in grid-connected mode at 51.7% fuel-to-electrical efficiency. See Ghezel-Ayagh16 for more information on this system. The second system was a direct topping cycle (with respect to the FC), which is the system of direct interest to the current work.
Experiment Description Siemens Power Corporation developed the very first pressurized SOFC/GT hybrid system using their tubular SOFC stack design. This system, presented in Figure 7, was tested at the NFCRC with support from Southern California Edison, the U.S. Department of Energy and others. The system was designed, constructed and tested to demonstrate and prove the hybrid concept. The system operated for over 2900 hours and produced up to 220 kW at fuel-to-electricity conversion efficiencies of up to 53%. In parallel, NFCRC developed dynamic simulation capabilities for each of the system components together with a simulation framework for modeling and developing control strategies for integrated SOFC/GT systems.
Fig. 7. Pressurized 220 kW SOFC/Gas Turbine Hybrid
137
1.4 Hybrid Gas Turbine Fuel Cell Systems A diagram of the integrated SOFC/GT system is presented in Figure 8. This system is comprised of a tubular SOFC with integrated internal reformer and anode off-gas oxidizer as illustrated in Figure 9. These components (stack, reformer) are placed between the compressor and turbine so that they operate under pressurized conditions. The gas turbine is a dual shaft Ingersoll-Rand 75 kW gas turbine. The integrated cycle also includes a recuperative heat exchanger and a separate turbine generator set (see Figure 8). Note that there are also two bypass valves that can divert flow around the heat exchanger and around the SOFC. Dynamic and steady state data were gathered during operation. Nominally the SOFC produced 180 kWe while the GT produced 40kWe of the total power. The dynamic data produced by the SOFC/GT system was primarily gathered during start-up and shutdown. The primary goal of the experimental effort was to demonstrate the hybrid concept for 3000 hours of steady state operation without detailed investigation of dynamic responses to perturbations. Under nominal operating conditions the SOFC stack was pressurized to three atmospheres, resulting in improved performance (through better electrode kinetics) and increased output (through increased Nernst potential of higher reactant partial pressures). The SOFC stack produces 100 kWe at atmospheric pressure, whereas in the hybrid configuration it produced as much as 180 kWe. A more detailed description of the system is presented in other works.17
Air
Exit
Turbine 2
Turbine 1
Compressor
Generator/ Motor
Heat Exch.
Natural Gas
Comb
Comb
Natural Gas
Cathode
SOFC
Natural Gas
Anode
Fig. 8. Diagram of the pressurized 220 kW SOFC/gas turbine hybrid system
AIR PLENUM
AIR IN (500 °C)
EXHAUST GAS OUT (800 °C)
DEPLETED FUEL RECIRCULATION
DESULFERIZED NATURAL GAS IN 1
COMBUSTION ZONE AIR INJECTION TUBE TUBULAR FUEL CELL (1000 °C) PROCESS AIR FUEL
EXHAUST
REFORMERS
Fig. 9. Tubular SOFC stack design.
2
138
Jack Brouwer Dynamic Model Description The equations that govern the dynamic performance of or each of the system components are solved in a modular fashion for each of the components of the 220 kW hybrid system in a Matlab/SimulinkTM format. The models were designed and constructed to be reliable and robust. All of the models are based on the fundamental mass, momentum, and energy conservation equations plus detailed solutions of electrochemical, chemical, and heat transfer processes.
SOFC Model The SOFC model developed for the current application is a simplified bulk model that simulates the overall performance of a pressurized tubular SOFC. The current model does not capture the spatial variations of operating parameters throughout the SOFC stack. This simplified model is deemed sufficient for simulating a complete hybrid system. However, spatially resolved models may be required to more accurately simulate the performance of specific SOFC stack designs and to garner more insights into stack behavior. Such models have been developed previously at the NFCRC and will be considered for future integration in a full hybrid system model.18 The governing equations of the SOFC model are introduced, starting with the Nernst potential EQ(1), which provides the reversible cell potential for a given fuel and oxidant composition.
R T X H 2 X O22 12 E = E 0 + u ln PCATHODE 2 F X H 2O 1
(1)
While EQ(1) solves for ideal cell potential, the actual cell potential for any fuel cell under real operating conditions will be reduced due to irreversibilities referred to as polarizations or overpotential losses. The modeling of realized cell voltage can be achieved by calculating each of the three primary overpotentials (activation, ohmic, and concentration) in bulk fashion and subtracting them from the ideal Nernst potential as in EQ(2)
VCell = E − η A − η C − η R
,
(2)
where Vcell is the actual cell voltage for a given current, ηA is the activation polarization loss, ηC is the concentration polarization loss, and ηR is the ohmic polarization loss. Calculation of these polarizations is based on a first principles understanding of the overall performance of a fuel cell. For a given temperature and pressure, all three polarizations are typically only a function of current demand. The loss associated with sluggish kinetics due to low temperatures and/or lack of availability of active catalytic cell sites is modeled using a relationship for activation polarization. This polarization is more dominant at low current densities. The activation polarization is calculated as
ηA =
i Ru T . ln i F αn 0
(3)
The key parameter that determines activation polarization for a specific fuel cell is io, which is the exchange current density. Exchange current density is associated with the catalytic activity of a particular cell and corresponds to the rate at which the electrodes exchange ions with the electrolyte under equilibrium conditions (no net current flow). α represents the distribution of intermediate species at the triple phase boundary, indicating whether these species more closely resemble reactants or products. α has a value between zero and one (usually taken to be 0.5). The irreversibility associated with concentration gradients near the active cell surface is modeled by EQ(4)
ηC = −
139
Ru T i ln 1 − n F iL
.
(4)
The new term here is iL, which is the limiting current density. Limiting current density corresponds to the maximum current that the fuel cell can produce to equal the maximum supply speed of reactants. To avoid this polarization, the fuel cell is usually operated at lower current densities or at higher pressures (if power density is a concern). Since activation polarization is reduced at high temperature, and since high temperature fuel cells are typically operated at relatively low current density, ohmic polarization is usually the most significant electrochemical loss. At normal operating conditions, this ohmic loss is primarily due to low ionic conductivity of the electrolyte and/or low electrical conductivity of associated interconnect materials. Resistance can also be high, if the cell is operating at a temperature below the optimum due to the strong temperature dependence of electrolyte ionic resistivity. The potential loss associated with cell resistance is
1.4 Hybrid Gas Turbine Fuel Cell Systems ,
(5)
where i is the current density and Reff is the effective overall cell resistance. Several fuel cell parameters affect the cell resistance including inherent electrolyte ionic conductivity, electrolyte thickness, electrode and interconnect electronic conductivities and geometry of the electrolyte affects the internal resistance. Thinner electrolyte layers can be designed to reduce ionic ohmic polarization, but the thickness is bound by the requirements of the cell to endure structural stresses produced by different thermal expansion of the materials that are sandwiched together. The effective resistance used in the current model includes consideration of the cell materials and geometry as well as a temperature dependence that is based on empirical data gathered from test cell and laboratory experiments on the tubular SOFC design of Siemens Westinghouse. The SOFC model incorporates the dynamic equations that solve for conservation of mass or species, momentum, and energy. For species conservation the equation assuming a well-stirred reactor approach is used. Vcv C
dX = N in (X in − X ) − X∑ R + R d t
(6)
There are seven species considered: methane, carbon monoxide, carbon dioxide, hydrogen, water, nitrogen, and oxygen. Using Faraday’s law of electrolysis EQ(7) the electrochemistry vectors for the reaction rates in the SOFC anode and cathode become equations EQ (8) and EQ (9) for the anode and cathode respectively.
rj =
a *i nF
(7)
i R anode ,e = ACell * 0 0 0 − 2F
+
i 2F
0 0
i R cathode ,e = ACell * 0 0 0 0 0 0 − 4 F
(8) (9)
Reformation and water-gas-shift chemical reactions occur simultaneously with the electrochemical reactions in the anode compartment of the SOFC. The reaction vector for the internal reformation chemical reactions is added to the electrochemistry reaction vector and inserted into EQ(6) to solve for dynamic species conservation. The internal reformation model considers the chemical kinetics of three concurrent chemical reactions, steam reformation of methane and water-gas shift as follows:
CH4 + H2O CO + 3 H2
(10)
CO + H2O CO2 + H2
(11)
CH4 + 2H2O CO2 + 4 H2
(12)
The forward rates of these steam reformation and water-gas shift are determined by Arrhenius rate expressions. The reformation model uses rates that are consistent with the use of typical nickel-based catalysts.19 This should be reasonable considering the nickel-YSZ composition of the cathode and nickel felt electrical connection materials in the anode compartment. The rate equation of reaction EQ(10) is
PCH 4 PH 2O PCO PH0.25 R1 = k 1 − PH2.5 K p1 2
/ DEN 2 .
(13)
The rate of reaction EQ(11) is
PCO PH 2O PCO 2 R2 = k 2 − PH K p2 2
/ DEN 2 .
(14)
140
Jack Brouwer The rate of reaction EQ(12) is
PCH PH2 O PCO PH0.5 4 2 2 2 R3 = k 3 − P 3.5 K p 3 H2
/ DEN 2 .
(15)
The denominator used in each of the reaction rate expressions above is:
DEN = 1 + K CO PCO + K H 2 PH 2 + K CH 4 PCH 4 +
K H 2O PH 2O
(16)
PH 2
According to the Arrhenius equation and van’t Hoff equation, the reaction constants ki (i =1-3) and Kj (j =CO, CH4, H2O, or H2) in the above equations can be calculated from the pre-exponential factors Ai and Aj, and the absorption parameters E i and ∆H j from the following equations
E k i = Ai exp − i , T R
∆H j K j = A j exp − T R
(17)
.
(18)
The constants used in the current model are presented in Table 3 and Table 4. CO is assumed to be consumed/created only by water-gas shift and steam reformation. Direct electrochemical oxidation of CO and hydrocarbons is possible under current anodic conditions, but it occurs at a sufficiently slow rate that this assumption has been shown to be reasonable in previous studies.20 Table 3. Reformation constants
Rate Constant k1
Activation Heat of energy Pre-exponential factor Rate Constant adsorption (kJ/mol) (kJ/mol) 240.1
15
1.336 x 10 (0.5)
(kmol·MPa k2
k3
67.13
243.9
7
14
3.22 x 10 (0.5)
-4
8.23 x 10 -1
/kgcat·h)
1.955 x 10 (kmol/kgcat·h·MPa)
(kmol·MPa
-70.65
K CO
Preexponential factor (MPa )
-38.28
K CH4
-3
6.65 x 10 -1
(MPa ) 5
K H2O
88.68
1.77 x 10 (unitless)
K H2
-82.9
6.12 x 10
/kgcat·h)
-8
-1
(MPa ) Table 4. Equilibrium constants
Equilibrium constant
141
11
Dimensions (MPa)
2
= 1.77 x 10 *exp(4400/T)
(MPa)
0
= K p1 *K p2
(MPa)
2
K p1
= 1.198 x 10 *exp(-26830/T)
K p2 K p3
-2
1.4 Hybrid Gas Turbine Fuel Cell Systems The SOFC model solves for the energy balance between the anode and cathode gas streams and the fuel cell materials. The cell materials (electrode-electrolyte assembly) the energy balance is solved using EQ(19). There is heat generated within the porous fuel cell electrode-electrolyte assembly were the hydrogen is being electrochemically oxidized. Based on the lower heating value of hydrogen, the energy that is not being converted to electrical energy produces heat in the SOFC stack as in EQ(20).
dρCmassT V dt
(19)
= E in − E out + QGEN
cv
∆H f , H 2O ( g ) QGEN = − VCELL * i nF
(20)
As for the anode and cathode gases, EQ(21) solves the energy balance for each of these control volumes.
dCCv ,molar T V
cv
(21)
= E in − E out
dt
The gas stream flows are assumed to be fully developed laminar flow. This assumption permits the use of an altered form of the Darcy equation EQ(22) for the solution of momentum conservation (calculating the fuel cell pressure drop) as follows
∆P = f where ∆P is the pressure drop, the hydraulic diameter.
L ρv 2 Dh 2
(22)
f is the friction factor, L is the characteristic length, ρ is density, v is average velocity, and Dh is
Heat Exchanger and Combustor Model The recuperator heat exchanger, SOFC heat exchanger, and combustor models are one dimensional in the streamwise direction. The heat exchangers solve the conservation equations for mass, momentum, and energy. The same equations EQ(19), EQ(21), and EQ(22) are used for the heat exchangers and combustor models except that there is no heat generation in the heat exchanger models.
Gas Turbine Model A transient mathematical model of a gas turbine system has been developed using in the same Matlab/SimulinkTM framework. The model predicts the behavior of a lumped parameter compressor and expander attached via rotating shaft. The dynamic expressions that account for gas compressibility and mass storage are solved in separate diffuser volumes as depicted in Figure 10. A generator load can be applied to the shaft or the system can operate as a turbo-charger. The current one-dimensional lumped parameter approach is flexible to allow the incorporation of semi-empirical data from a specific production or prototype gas turbine. The semi-empirical data that is used in the current dynamic gas turbine modeling approach is in the form of non-dimensionalized compressor and turbine “maps.” These compressor and turbine maps provide steady state mass flow, pressure ratio, and efficiency as a function of rotational speed. Usually two maps each are required for the compressor and turbine. The first map plots the pressure ratio versus dimensionless mass flow for a series of fixed (sometimes non-dimensionalized) rotational speeds. The second map gives the normalized isentropic efficiency versus dimensionless mass flow for a series of fixed (sometimes non-dimensionalized) rotational speeds. Exhaust Out
Diffuser Volume
Compressed Air Out Power = Torque*RPM
Diffuser Volume
Generator Load
Air In
Compressor Control Volume
Hot Compressed Air In
Expander Control Volume
Fig. 10. Schematic approach to dynamic simulation of a gas turbine engine
142
Jack Brouwer Typical non-dimensionalization of the mass flow is as follows:
T m R γR
01
(23)
D 2 P01
is the fluid mass flow, R is the gas constant, γ is the ratio of specific heats, T01 is the stagnation temperature at the inlet, D is where m a characteristic length, and P01 is the stagnation pressure at the inlet. Dimensionless rotor speed can be given by
N D γR T
,
(24)
01
where N is the rotational speed. Using the mapped compressor and turbine performance, the mass flow through such can be determined for any given speed, discharge pressure, and inlet condition. The solution strategy for both the turbine and compressor dynamics involves iterative determination of mass flow. For a given rotational speed and pressure ratio, a mass flow is guessed and iteratively converged upon until a pressure ratio matches the ratio of discharge pressure to inlet pressure. An iterative approach is necessary because the discharge pressure is determined by the swallowing capacity of components downstream of the compressor (or turbine). Once the mass flow is determined, a compressor (or turbine) efficiency can be determined from the efficiency map. Knowing the isentropic efficiency, the compressor (or turbine) exit temperature can be determined from the isentropic relations described in the following paragraphs. The inlet temperature of the compressor is known and once the compressor isentropic efficiency is extracted from the performance maps the compressor stagnation exit temperature, T02, can be calculated by the using EQ(25).
1 T02 = T01 1 + η comp
(
P02 P01
γ −1 γ
− 1 )
(25)
The specific heat, CP, is calculated next as a function of temperature based on third-order curve fits for a gas mixture containing up to seven molecular species (CH4, CO, CO2, H2, H2O, N2, O2). Using Cp and the temperature of each state the enthalpies can be calculated by EQ (26) and used to calculate the compressor work using EQ (27). T01
h01 − h02 = ∫ C P (T )dT
(26)
T02
PC = m Comp (h01 − h02
)
(27)
After the compressor exit state is determined a dynamic expression that accounts for gas compressibility and mass storage in a separate compressor diffuser volume is solved as follows
γR dP T = (m in − m out ) dt V
(28)
Thus, for a given moment in time, all the parameters necessary to assess the dynamic compressor performance are calculated. As for the gasifier turbine work or the turbine supplying work to compressor the turbine inlet temperature (T03) is known. Using performance maps the isentropic turbine efficiency can be extracted from the turbine efficiency map and used in EQ(29) to calculate the turbine exit temperature.
P T04 = T03 1 + η T 04 P03
γ −1 γ
− 1
(29)
Once the temperatures are known then turbine mass storage can be assessed by solution of EQ (28) for the turbine. Then the enthalpies at each state (EQ(30)) are calculated in order to calculate the turbine power using EQ (31). T03
h03 − h04 = ∫ C P (T )dT
(30)
T04
143
PT = m Comp (h03 − h04
)
(31)
1.4 Hybrid Gas Turbine Fuel Cell Systems The above calculations are performed at every time step in the gas turbine transient model. To capture the dynamics associated with the rotational inertia of the GT, the summation of torques is used to calculate the angular acceleration, which is integrated over time to calculate the shaft speed of the gasifier turbine. Equation (32) is solved with the known turbine and compressor powers and rotational inertia, J, and rotational speed, ω , of the turbo machinery as follows:
dω 1 = ΡT − ΡC . dt Jω 1
(
)
(32)
For the second turbine (power turbine) the same equations are used for to calculate the state (5 and 6) temperatures and enthalpies. As for the sum of the torques, the second shaft has the generator load instead of the compressor load as in EQ (33).
dω 1 = ΡT − ΡLOAD dt Jω 2
(
)
(33)
The generator operates at 3600 RPM for 60 Hz AC electricity production; therefore the load from the generator is dynamically adjusted to maintain the RPM of the power turbine at 3600 RPM. Do to alterations made to the nozzle of the gas turbine to accommodate the over sizing of the gas turbine with the rest of the system the power turbine was operated at a lower RPM of 3000. This produced 50 Hz AC power from the generator.
Data-Model Comparison Approach The Matlab/SimulinkTM modules described above were integrated into a system model that could simulate the 220 kW pressurized tubular SOFC/gas turbine hybrid system of Figure 7. The system configuration was identical to that presented in Figure 8, with a fuel cell that represented to performance of an integrated SOFC/reformer module depicted in Figure 9. Experimental start-up data is presented for model verification. A series of control moves must occur during start-up in order to heat up the fuel cell, control temperatures and temperature ramp rates throughout the system, and maintain operation of the gas turbine. Two combustors were used to supply heat during start-up and SOFC stack warm-up. In the simulation results, the control moves that are identical to those recorded during the experiment were implemented in the simulation.
Dynamic Simulation Results and Comparison The data, as stated earlier, that were acquired from the system and selected for simulation and comparison are system startup data. During this period of operation the SOFC/GT hybrid system was slowly ramped up in power to minimize the mechanical stresses from thermal shock. Figure 11 presents the control moves made by the operator during start-up of the hybrid system. The controlled parameters were the SOFC load, recuperator and SOFC bypass and the fuel flow to the system (SOFC load not shown in Figure 11). The bypass valves were used to control the temperature of the SOFC stack. The recuperator bypass controlled the inlet temperature of the air entering the stack. The SOFC bypass was used to control the mass flow through the SOFC stack. The hybrid system utilized a dual shaft turbine. As a result the total mass flow through the system could not be controlled independent of SOFC power. With a single shaft gas turbine one can adjust turbo machinery speed (and thus compressor mass flow) by manipulation of generator load. The free-spinning turbine and compressor of the dual shaft machine thus required an SOFC bypass flow to control speed (and mass flow). Model Inputs for the 220 kW SOFC/GT Hybrid 6
Recuperator Bypass Valve
45
5.8
SOFC Bypass Valve
40
5.6
SOFC Fuel Flow
35
5.4
30
5.2
25
5
20
4.8
15
4.6
10
4.4
5
4.2
0
Natural Gas Flow (SLPM)
Bypass Valve Percent Open
50
4 0
25000
50000
75000
100000 125000 150000
Time (sec)
Fig. 11. Fuel flow and bypass valve positions used in the experiment and simulations
144
Jack Brouwer
145
165 SOFC Power (kW)
Model
Experiment
160 155 150 145 140 0
25000
50000
75000
100000 125000 150000
Time (sec) Fig. 12. SOFC power experimental and simulation comparison Gas Turbine Power Experimental and Model Comparison for the 220 kW SOFC/GT Hybrid Model Experimental
29
GT Power (kW)
27 25 23 21 19 17 15 0
20000
40000
60000
80000
100000 120000 140000 160000
Time (sec)
Fig. 13. Gas turbine power simulation and experimental data comparison
760
Model Experiment
750
TIT (K)
In Figure 12, the SOFC power is ramped up from 147 kW to 158 kW over a period of 100,000 seconds. The model simulation follows the SOFC power closely. The model input for the SOFC is the current demand and fuel flow rate. The cathode inlet temperature, operating fuel cell voltage, overall SOFC temperature, internal reformer temperature, combustor temperature, pressure and other operating parameters are calculated and dependent on the solution of the integrated hybrid system dynamic performance as calculated using the simulation modules described herein. Sudden drops in SOFC power were observed in the experiment as the SOFC bypass valve was opened to allow more air to bypass the SOFC. At low load (time = 10,000 seconds) the model does not capture this dynamic. However, a similar dynamic that occurs when the fuel cell is producing 157 kW (around t = 90,000 seconds) is slightly captured by the model. It is believed these sudden drops in power are due to the changes in the airflows through and around the SOFC stack. The discrepancies of the experiment and model data during changes in the SOFC bypass valve position are due to uncertainties in the exact flow dynamics and flow amounts altered by the SOFC bypass valve Measurement data for the bypass valve position is not very accurate since the valve type used was a pneumatically actuated butterfly valve. Pressure, mass flow, and temperature deviations would lead to different mass flows being bypassed for the same valve position. Also the first degrees of movement of the valve dramatically change the amount of mass flow being bypassed. One could estimate the bypass mass flow using an enthalpy balance if accurate data for mass flow and temperatures around each bypass valves were known. Since this information was not available bypass mass flow rates were estimated by valve position only with rates averaged over a 5 minute time period. The gas turbine performance, power output, system airflow rate and SOFC pressure are each totally dependent on the SOFC performance (in model and experiment). The power output of the gas turbine (Turbine 2 of Figure 8) is left to float depending on the SOFC stack exhaust conditions and the percentage of air that bypasses the SOFC. As more air bypasses the SOFC stack, the cooling of the SOFC stack decreases and the turbine inlet temperature (TIT) is reduced resulting in
740 730 720 710 0
25000
50000
75000 100000 125000 150000 Time (sec)
Fig. 14. Model and experiment turbine inlet temperature (TIT) comparison
1.4 Hybrid Gas Turbine Fuel Cell Systems lower gas turbine power output. Figure 13 presents the experimental data and model results for the gas turbine power during the SOFC stack power ramp up. The model follows the power output of the gas turbine quite well with a few deviations during the SOFC stack ramp up. The model does not predict turbine power as accurately when the SOFC bypass valve is being adjusted. Some of the errors again are associated with the limited experimental data on actual bypassed mass flows. The model well captures the change in the turbine inlet temperature (TIT) that corresponds to the SOFC ramp-up conditions as shown in Figure 14. There is a slight error in the TIT that peaks at about 6 degrees, but the overall trend is captured throughout the entire dynamic response to SOFC ramp-up perturbations. It can be seen that the change in TIT is the dominant parameter that affects gas turbine power. Temperatures predicted by the dynamic model and observed in the experiment for several of the system states are presented in Figure 15 for the starting and end-point conditions presented in previous figures. The temperatures throughout the system are fairly close, but there are some differences. There is a 5% difference in compressor mass flow, which could cause the model to predict lower temperatures, but instead the model predicts higher temperatures. The reason the model predicts higher temperatures is due to inadequate accounting of the heat losses through out the system. The only heat losses currently considered in the system occur in combustor 1 and 2 (where there is significant heat loss). There are not any heat losses accounted for in the current SOFC and recuperator models, around which the largest temperature discrepancies are presented. Additional work is required in order to accurately quantify the heat losses associated with the SOFC stack and the recuperator. Nonetheless, the dynamic and steady state performance predictions are impressive, given the system complexities.
T1 20 C 14.7 psia 635 g/s 38 C T2 14.7 psia 625 g/s
Air
20 C 14.7 psia 660 g/s 20 C 14.7 psia 659 g/s
GENERATOR RECUPERATOR
Stack Gas T1 216 C
246 C
T2 202 C
236 C
T1 578 C
576 C
T2 509 C
516 C
BY-PASS
Cathode
Fuel
Anode
SOFC
T1 760 C
752 C
T2 727 C
735 C
T1 769 C
788 C
T2 822 C
803 C
T1 575 C
575 C
T2 552 C
566 C
OXIDIZER OBSERVED SIMULATION RESULTS RESULTS
Fig. 15. Comparison of temperature states in hybrid system for initial and final conditions
Hybrid Dynamic Simulation Conclusions A modular approach for dynamic simulation of the major system components of a hybrid FC/GT system is presented. The dynamic models were developed in a Matlab/SimulinkTM environment. Using the dynamic simulation modules, a detailed hybrid system model was constructed to simulate a 220kW tubular SOFC/GT hybrid system. The dynamic model well captures the dynamic performance of the integrated experimental system for transient operating conditions observed during a system start-up. Some system dynamics are not well captured by the model especially those associated with the bypass valve dynamics, which were not adequately understood at the time of model application. Overall, the dynamic model quite accurately captures the particular set of hybrid SOFC/GT performance data during a start-up transient. This comparison shows that the model, built from first principles, can reasonably predict the dynamic performance of a complex hybrid FC/GT system. Thus verified, the dynamic model can be used to provide operational insights and guidance for design and controls development. Comparisons of system simulation results to experimental data are rare in the literature, due to the dearth of available experimental results. Although the current paper does not fully validate the current system model, it provides confidence that users can apply the dynamic models in developing control algorithms and proper procedures for start-up and shutdown of these types of complex and integrated hybrid fuel cell systems. Future investigations will be performed to further validate the systems simulation tools and test the limits of component and system dynamic responses to load demands and other possible perturbations. Being able to test these scenarios with an accurate system model provides an insightful, economical, and safe means for system research and technology advancement.
1.4-7 Hybrid System Control
Hybrid cycles comprised of high temperature fuel cells, such as the molten carbonate fuel cells (MCFC) or solid oxide fuel cells (SOFC) are very promising for generating electric power in the future, initially at the small to medium scale (250 kW to 20 MW), and later in large scale central plants (>100 MW). However, hybrid gas turbine fuel cell systems are in need of
146
Jack Brouwer significant advancement before they are introduced as commercial products. Some progress is needed to address the specific challenges that are introduced by coupling a fuel cell with a gas turbine given their disparate dynamic response characteristics. Thus a significant need for developing and testing control methods and strategies for hybrid gas turbine fuel cell power plants is required. As an example, hybrid systems are sensitive to ambient conditions due the sensitivity of compressors to air density. At higher temperatures the air becomes less dense requiring a compressor to do more work to pressurize and move the air through the system. As for a hybrid system, it is challenging to maintain sufficient compressor mass flow for extreme conditions since the fuel cell is operated at a fixed temperature. If the gas turbine operates at a fixed speed there are no options for controlling the mass flow. The total power output of the system may have to be sacrificed in order to maintain appropriate fuel cell operating temperature by lowering the load demand on the fuel cell. For the purposes of better understanding the dynamics of hybrid gas turbine fuel cell hybrid systems and for development of controls, NFCRC has developed dynamic modeling tools for FC/GT hybrid systems. In previous work21, transient performance and controls analyses of atmospheric hybrid systems with MCFCs were presented. Load perturbations were implemented to analyze the MCFC/GT hybrid response. In these investigations it was discovered that additional control loops are necessary to control the MCFC operating temperature. For example, varying fuel utilization across the MCFC provided some means for control but was limited. Variable speed operation of the gas turbine was tested and showed more promise, but still was limited in the particular system at lower power demands. For a larger turn-down in system power a bypass or auxiliary combustor is needed in parallel.22 For part-load operation of a FC/GT hybrid it has been shown that a variable speed gas turbine is a required feature for both pressurized23 and atmospheric systems.24 The variable speed gas turbine provides better control of the compressor mass flow. In the previous section a dynamic system model was described and results were compared to experimental data from the Siemens SOFC/GT system. A dual shaft turbine was used in that particular SOFC/GT system. The dual shaft turbine prevented the direct control of the compressor mass flow, which limited the operation flexibility. The system had to be operated at the maximum power safely allowed. In the current section, a 1.15 MW pressurized SOFC/GT hybrid model is developed. A diagram of the system is presented in Figure 16 and a schematic of the SOFC module is presented in Figure 17. The system was designed around the Capstone C200 microturbine generator. Design parameters for the C20025 and the hybrid plant are presented in Table 5.
Fig. 16. Pressurized SOFC/GT hybrid cycle
Fig. 17. SOFC module
147
1.4 Hybrid Gas Turbine Fuel Cell Systems Table 5. Design parameters SOFC/GT system
Design Parameter
Value
Unit
System System Power
1150
Combustor Efficiency
1
Recuperator Effectiveness
1
Heat Exchanger Effectiveness
kW
0.4
System Efficiency
0.73 Gas Turbine
Shaft Speed*
60000
RPM
950
C
Turbine Inlet Temperature* Turbine Efficiency
1
Mass Flow*
1.3
Compressor Inlet Temperature Compressor Discharge Pressure* Compressor Efficiency
kg/sec
1500.0%
C
43569.8%
kPa
75.0%
Gas Turbine Power Mechanical Loss (Shaft) Gas Turbine Power Electronics Efficiency
2
kW
RPM *8.33E-10 98% and 14 kW load
Compressor Leakage
0.02
Compressor Filter Loss
0.02 SOFC Module
SOFC Stack Power
960
kW
SOFC Active Area
320
m
4,000
A/m
Current Density SOFC Operating Voltage
0.75
SOFC Power Electronics
100.0%
Anode Recircuation
80.0%
SOFC Stack Fuel Utilization
85.0%
SOFC Average Operating Temperature
900
2 2
V
C
*Willis, 2005
Dynamic Model and Operating Conditions The dynamic model used in this section is identical to that described in the previous section except that the components are arranged to simulate the cycle presented in Figure 16. The electrochemical performance for the SOFC is based on the results presented by Kim et al.26 The work presented in this section is on the controller design and dynamic analysis of a SOFC/GT hybrid system. Two different cases are presented: (1) a base-load system is exposed to changing ambient temperature; (2) a load-following system is exposed to the same ambient conditions while following a load demand curve. The design electrical power production of the SOFC/GT hybrid system is 1.15 MW. For the base-load case the system maintains 1.15 MW (1150 kW) of net electrical power production. The SOCF/GT hybrid system is operated in an extreme environment with a vast fluctuating ambient temperature. The ambient temperature is varied from -5°to 35°C in a sinusoidal form to emulate the daily temperature fluctuation. This range of temperatures accounts for colder or frigid regions and hot regions where the system may be operated. The system is tested in load-following mode with a varying load demand. The system is subjected to the same daily ambient conditions in all cases, while meeting a sinusoidal demand of power from 1150 kW at the peak of the day to 950 kW at the minimum power production of the day.
Controller Design A decentralized controller design is used to control the hybrid system. The objective of the system controllers is to maintain constant power production while maintaining the SOFC operating temperature close to its design operation temperature of 900°C. Figure 18 presents the controller design. The controller design consists of a gas turbine shaft speed controller, system power controller,
148
Jack Brouwer and a SOFC temperature and fuel flow controller. The shaft speed controller is a cascade controller with the outer loop consisting of a feed forward and a feedback flow controller for the RPM set point. The inner loop manipulates the gas turbine power to achieve the a set point provided by the outer loop. The feed forward aspect of the outer loop uses a look-up table to determine the RPM setting for a given system power. The feedback loop corrects the RPM setting for any SOFC temperature deviations. The feedback portion is very important when the compressor is operating at an off design setting. For example, extreme ambient conditions would require RPM correction. The system power controller manipulates the SOFC current in order to meet the power demand. The gas turbine power is treated as a disturbance for this particular controller. Therefore, the SOFC power is altered continuously by manipulating the current to meet the power demand that has not been met by the gas turbine. There is additional control of the SOFC temperature via the bypass valve located between the turbine exhaust and the recuperator. The bypass, when used, lowers the inlet temperature to the SOFC module. The fuel flow is manipulated to achieve fuel utilization of 85%. The fuel flow controller is a feed forward controller based on the current of the SOFC. The fuel utilization after one pass through the anode section is approximately 53%. The design electrical power production of the SOFC/GT hybrid system is 1.15 MW. For the base-load case the system maintains 1.15 MW (1150 kW) of net electrical power production. The SOCF/GT hybrid system is operated in an extreme environment with a vast fluctuating ambient temperature. The temperature changes account for colder or frigid regions and hot regions where the system may be operated. The system is tested in load-following mode with a varying load demand. The same daily ambient conditions are applied to the system while demanding a sinusoidal power profile that varies from 1150 kW at the peak of the day to 950 kW at the minimum power production time of the day.
Hybrid System Control Results Base-Load Case The SOFC hybrid system is simulated in base-load mode. The system is to produce its design power while operating in varying ambient conditions. As stated before the ambient temperature is varied in the range of + 20 ْC. A sinusoidal temperature profile with a period of one day is used. The peak temperature is at 12 noon. Figure 19 presents the total power produced by the hybrid plant along with the SOFC and the gas turbine power. The total power produced by the hybrid plant is constant with very small deviations. The gas turbine power changes dramatically to control the shaft speed. The SOFC power changes in order to compensate for the changes in the gas turbine power.
Fig. 18. Controller design
149
1.4 Hybrid Gas Turbine Fuel Cell Systems The SOFC temperature is presented in Figure 20 along with ambient temperature and percent bypass mass flow. The SOFC temperature is maintained within 25 ْC of the design operating temperature of 900 ْC. The effects of the ambient temperature are seen when plotted with the SOFC temperature. The high ambient temperature increases the compressor outlet temperature and also decreases the compressor mass flow by reducing the air density. The reduction of the compressor mass flow can be seen in Figure 21. The dip in SOFC temperature just before 7 hours is a result of the slight increase in mass flow from the compressor presented in Figure 21 just before the mass flow sharply decreases. The mass flow from the compressor increases with the sudden increase of the shaft speed also presented in Figure 21. Two things promote this increase in shaft speed: (1) the ambient temperature is at the design inlet temperature of the compressor resulting in a more efficient compressor and (2) the TIT in Figure 22 increases providing more power to the shaft. The gas turbine power increases in Figure 19 at 6 hours to overcome this surge of net power being produced by the gas turbine. The TIT eventually lowers as the bypass valve opens and the ambient temperature continues to rise. This reduces the power produced by the turbine and thus increases the compressor work. The gas turbine power is dramatically decreased by the controllers at 7 hours, as shown in Figure 19, to allow the shaft speed to increase so that the SOFC can be provided sufficient air for cooling. Even though the gas turbine power is dramatically reduced, the shaft speed does not increase sufficiently. The extra work by the compressor prevents the shaft from speeding up and supplying more mass flow. The reduction of mass flow in the system reduces the operating pressure of the system as shown in Figure 21. The bypass valve prevents the SOFC from overheating when the mass flow from the compressor does not fully recover. The bypass valve opens to reduce the temperature of the air entering the SOFC module. The effects of the bypass valve on the SOFC operating temperature are shown in Figure 20. The SOFC inlet temperature (state #1 of Figure 16 and Figure 17) is reduced as shown in Figure 22. This decreases the cathode inlet temperature (see state #4 of Figure 17), which helps prevent the SOFC stack from overheating. The cathode and turbine inlet temperature along with catalytic oxidizer temperature are presented in Figure 22. The SOFC operating temperature rises at around 20-21 hours. The bypass valve closed rapidly at this time triggering this sudden rise in SOFC temperature. The bypass valve partially opens again when the SOFC temperature exceeds 900ْC.
Fig. 19. Total power, SOFC power and gas turbine power produced
Fig. 20. SOFC average temperature, ambient temperature and bypassed mass flow
150
Jack Brouwer
Fig. 21. Compressor mass flow, pressure and normalized shaft speed
Fig. 22. SOFC, cathode, turbine inlet, and catalytic oxidizer temperatures
The system efficiency, SOFC fuel and oxygen utilization and fuel flow are presented in Figure 23. The system efficiency fluctuates between 65% and 72%. At the peak ambient temperature, the gas turbine net power is reduced to sustain sufficient mass flow from the compressor. The SOFC power is increased to offset the power drop from the gas turbine. The increased power from the SOFC increases the fuel flow which decreases the system efficiency when more fuel is required for the same net power produced by the system. The SOFC fuel utilization presented in Figure 23 is the fuel utilization after one pass through the anode section of the SOFC stack. After recirculation, the overall SOFC module electrochemical fuel utilization is 85%.
Fig. 23. System efficiency, SOFC oxygen and fuel utilization and fuel flow
151
1.4 Hybrid Gas Turbine Fuel Cell Systems Load-following Case The same ambient temperature perturbation as presented in the previous case is applied to the hybrid system in the case presented in this section. In addition, the hybrid system must follow a load demand. The load demand varies form 950 kW to 1150 kW. A sinusoidal power demand with a period of one day is used. The peak demand is at 12 noon. The total power, SOFC power and gas turbine power are presented in Figure 24. The system was excellent in following the power demand. The fluctuations the gas turbine power can be seen in Figure 24. The gas turbine remains around 140 kW during the entire day. Unlike the case before, the gas turbine does not reach 180 kW during the colder parts of the day since the system is operating at a lower power demand at that time of day. The SOFC power has a sinusoidal profile with only fluctuations due to the gas turbine power. If the ambient temperature had not been so extreme at 12 hours, the gas turbine would have produced more net power.
Fig. 24. Total power, SOFC power and gas turbine power produced
The SOFC temperature in Figure 25 is kept within 25ْC of the design temperature as in the case presented earlier. The impact of the ambient temperature on the system can be seen in Figure 25. The same spikes and dips occur in the SOFC temperature as did in the previous case, but the logic behind them is more obvious in these results. The dip in SOFC temperature at 8 hours is a result of the sudden increase in mass flow from the compressor presented in Figure 26. The bulge at 7 hours is more apparent in this case. The mass flow from the compressor increases with the sudden increase of the shaft speed also presented in Figure 26. The same two sources as described in the previous section triggered this sudden change in shaft speed (1) more efficient compressor and (2) increase in TIT. The gas turbine power increases in Figure 24 at the same time to overcome this surge of net power being produced by the gas turbine. The TIT eventually lowers as the bypass valve opens and the ambient temperature continues to rise. This reduces the power produced by the turbine and the increases the compressor work as before. The gas turbine power is decreased by the controllers in Figure 24 to allow the shaft speed to continually increase so that the SOFC can have sufficient cooling. In this case the gas turbine power does not have to change as much since it is already at the right power range for 1150 kW system power production with 35ْC ambient temperature.
Fig.25. SOFC average temperature, ambient temperature and bypassed mass flow
152
Jack Brouwer
Fig. 26. Compressor mass flow, pressure and normalized shaft speed
Figure 27 presents the cathode and SOFC inlet temperature. The drop in both of these temperatures from the opening of the bypass valve can be seen between 7-17 hours. There is a more dramatic change in the SOFC inlet than the cathode inlet temperature because the heat exchanger becomes more effective due the increase in temperature differences between the SOFC inlet and catalytic oxidizer temperature. The catalytic oxidizer increase from the increase in SOFC power (more anode of gas), SOFC temperature and the reduction of mass flow (higher oxygen utilization, Figure 28). The TIT increases because of the catalytic oxidizer temperature increase, but less since the heat exchanger is more effective in transferring the heat from one flow to the other.
Fig. 27. SOFC, cathode, turbine inlet, and the catalytic oxidizer temperatures
The system efficiency in Figure 28 has the same profile as in the earlier case, but is higher when the power production is lower due to the higher operating voltage or more efficient operation of the SOFC. The oxygen utilization and the fuel utilization are similar to the case presented in the previous section. The oxygen utilization does reach higher levels of 47%, which indicates that the mass flow would be desired to be increased for better performance.
153
Fig. 28. System efficiency, SOFC oxygen and fuel utilization and fuel flow
1.4 Hybrid Gas Turbine Fuel Cell Systems Summary of Hybrid System Control Study A SOFC/GT hybrid system was developed with controllers that allow load-following capabilities. A base-load case with varying ambient temperature for two days was simulated and presented. The system maintains constant power (100% design power) while being exposed to an ambient temperature that varies significantly from -5°to +35°C. The system controllers responded to changes in the ambient temperature and successfully maintained the SOFC operating temperature within 25°C of the design operating temperature. The gas turbine power had to be continuously manipulated in order maintain the correct shaft speed and in turn the adequate amount of compressor mass flow. A sinusoidal load profile was demanded of the hybrid system with peak power demand at 12 hours at the same time of the peak ambient temperature. The system followed the load very well. In this case, the gas turbine remained closer to 140 kW during the entire load perturbation. The oxygen utilization increased during the load perturbation. Ideally for a SOFC/GT hybrid system, much like a gas turbine system, the oxygen and fuel utilization would remain constant over the entire range of load demand. This fixes the air-to-fuel ratio in the SOFC module. A controller that enforces constant oxygen utilization is needed to maintain consistent SOFC temperature and operation. Precisely controlling the compressor mass flow during large fluctuations in ambient temperature is challenging. If not carefully executed, the gas turbine can become unstable. The constant RPM approach with temperature error correction attempts to control the mass flow, but there are deviations in the SOFC temperature and oxygen utilization. These deviations are acceptable and the RPM control of the gas turbine provides a more stable means of controlling the system. In future work a combination of RPM control and direct mass flow control will be investigated. This type of control approach will provide stability and accurate control of the system mass flow.
1.4-8 Research & Development Needs for Hybrid Gas Turbine Fuel Cell Systems Over a period of more than 10 years the U.S. Department of Energy has sponsored workshops and conferences on the topic of gas turbine fuel cell hybrid systems. In many of these venues stakeholders from industry, agencies, national laboratories and universities have gathered to discuss the latest findings and results from hybrid projects and work together to identify the remaining research and development topics that should be addressed to advance hybrid systems. This section presents a summary of research and development needs for hybrid gas turbine fuel cell systems that is developed in part on the basis of input from these workshops and conferences.
System Component Design & Development Fuel Cells Both solid oxide fuel cells (SOFC) and molten carbonate fuel cells (MCFC) are well suited for hybrid fuel cell heat engine designs and application. General advancement of SOFC and MCFC technology will be very important to the hybrid fuel cell program. In addition, there are specific research and development needs for fuel cells that should be addressed in order to facilitate their integration with heat engines in hybrid systems. These R&D issues include: • • • • • • •
understanding of pressurized operation design for pressurized operation o significant pressure differentials o significant pressure fluctuations integration with oxidizers for increased thermal output to heat engine when needed increased fuel cell power density (ease of integration) nickel replacement or additive sulfur tolerant anode redox resistant anode
Research is required to enable fuel cells to meet the demands that hybrid cycles might place on them. Some of the particular needs that new fuel cell technology may need to provide to reach the expected hybrid system performance targets include the following: 1) Advanced materials a) Increased current densities (to reduce the size and cost of fuel cells, improved materials for electrodes and electrolytes are required) b) Improved mechanical properties (to withstand thermal stresses induced by successive starts and stops, and mechanical vibrations induced by turbomachinery and/or by motion in mobile applications) 2) Decrease air to fuel ratio (to decrease size of fuel cell itself, the equipment upstream of the fuel cell supplying the air and downstream of the fuel cell handling the exhaust gas, as well as increase the efficiency of the hybrid by being able to operate the gas turbine at a higher firing temperature) 3) Improved heat transfer to remove heat generated by cell a) For example, use more internal reforming to absorb heat generated by cell
154
Jack Brouwer 4) High speed solid state inverter technologies a) New materials (semiconductor compounds and/or layered structures of different materials) b) New switching techniques 5) Fuel flexibility a) Tolerance to fuel contaminants (sulfur and chlorine compounds)
Combustors Combustors will be required for all fuel cell systems for startup and possibly shut-down (e.g., to keep the gas turbine operating and supplying cooling air to the stack), as well as to accommodate dynamic load variations through increased heat engine output. Combustor advancement that would be valuable to a hybrid program includes those with the following possible features and/or research requirements: • • • • •
modulatable combustor (0-100% load), relight capability, can withstand constant flux of high temperature fuel cell products through inactive combustor (during steady state operation) without cooling air, high reliability, availability, maintainability, durability, low cost
Inverters and Power Electronics Inverters and power electronics must be designed and manufactured specifically for fuel cell hybrids with the understanding that accepting input from both the heat engine and fuel cell would be preferred. The integration of the inverter and power electronics with hybrid power plants is not well understood and has not been well investigated. The following inverter component studies would be useful in a hybrid fuel cell program: • • • • • • • •
low cost inverters simplified inverters inverter and power electronics systems analysis identifying cost, complexity, trade-offs of various inverter architectures inverter and power electronics systems analysis identifying cost, complexity, trade-offs of using separate inverters for fuel cell and heat engine versus a single inverter, integration of inverter with system effects of inverter and system architecture on power quality and reliability inverter innovation to reduce cost for lower power inverters inverter and power electronics robustness to match expected low maintenance of hybrid system
Sensors and Controls New sensors and/or the application of reliable and robust sensors and diagnostics as well as well understood control algorithms and strategies based upon sensitive measurements and manipulated inputs. Research and development is required in several sensors and controls areas: •
• • • • • •
identification and development of appropriate in-situ sensors for measurement of critical parameters o reformer composition o temperatures o pressures o flow rates (including various mixtures at high temperature and pressure) identification and fundamental understanding of system response to appropriate manipulated variables identification and understanding of controlled parameter response to manipulated variables high temperature (high pressure) sensors, valves, measurement and control technologies understanding of dynamic and steady state response to manipulated variables development of control strategies and methodologies based upon this understanding intelligent components (automatically sense failure before it occurs)
Gas Turbines or other Heat Engines
155
It is well known that one cannot simply replace the combustion-provided heat input to a heat engine with that available from a fuel cell and expect the heat engine to perform very well. One must, in order to take full advantage of hybrid cycles, develop and design heat engines that can both handle the flow and thermal input features that a fuel cell can provide, as well as perform well under these conditions. This requires a considerable shift in the focus of research and development for heat engine technologies. Whereas for standalone heat engines, increases in temperature and pressure are almost always desired to increase efficiency, application of heat engines to hybrid cycles may require movement toward designs with lower temperatures and pressures to maximize efficiency.
1.4 Hybrid Gas Turbine Fuel Cell Systems The development of heat engines with the following general features are desired for a hybrid system: • • • • • •
ability to perform well on lower quality thermal input (e.g., lower turbine inlet temperature (TIT) for the case of a gas turbine) ability to withstand long-duration thermal cycling (due to thermal mass of the fuel cell) ability to perform well with lower pressure ratios, larger window of operation to allow for system turndown and avoid shut-down of integrated system (e.g., movement of surge line away from typical operating conditions of a compressor – increase surge margin) controllability with slow time-response output of fuel cell (due to thermal mass of fuel cell) robust heat engines (to match maintenance cycle of fuel cell)
Research is required to enable gas turbine engines to meet the demands and features of various hybrid cycle designs. Some of the particular needs that new gas turbine technology may need to provide to reach the expected hybrid system performance targets include the following: 1) 2) 3) 4) 5) 6) 7) 8) 9)
Recuperative cycle configurations Advanced cycle configurations (intercooling, humid air turbine) Combustor (capable of accepting hot vitiated air and hot depleted fuel) Reduced emissions combustor (reduced NOx, CO and hydrocarbons) Catalytic combustor (capable of accepting reduced excess air, possibly approaching stoichiometric conditions) Reduced turbine cooling penalty (advanced turbine materials including ceramics and cooling technologies) Increased pressure ratio Increased compressor and turbine aerodynamic efficiencies Fuel flexibility a) Tolerance to fuel contaminants (chlorine compounds and alkaline earth compounds)
Simplify/Optimize System Configuration Several aspects of fuel cell hybrid systems could be simplified and/or optimized to advance hybrid technology, lower its cost and make it more reliable. Challenges in this general area include a fundamental understanding of fuel cell and fuel cell hybrid steady state and dynamic performance, increasing system and component RAMD, and increasing the power density of the fuel cell for better system integration.
Increase Fuel Cell Power Density or Thermal Output • • • • •
Increase power density Higher operating temperature fuel cells Increase TIT More stack electrical and thermal output Develop FCs that operated at higher & lower temperatures to facilitate FC staging
Hybrid Steady State and Dynamic Performance Optimization • • • •
Mathematical Models for S.S./dynamic response System Configuration Studies Thermoeconomics Subsystem consolidation
RAMD • • • •
System failure modes & criticality affects (FMCA) RAMD Tests on systems and/or subsystems Component Tests--accelerated or otherwise Power electronics RAMD study
Packaging
156
Jack Brouwer Optimize/Customize Turbine Subsystem The gas turbine subsystem of a gas turbine fuel cell hybrid system could be optimized or customized in many different ways to make the subsystem amenable in some cases, and more suitable in most cases to the high reliability and high efficiency characteristics that are desireable in a hybrid system. This section identifies specific advances in gas turbine technology that would be helpful to the advancement of hybrid systems. Features of hybrid-optimized gas turbine systems that need advancement in a hybrid program include: • • • • • • • • • • • • • • • • • • • •
At least 8700 hours of continuous, maintenance-free operation Match the fuel cell operating and maintenance cycles (e.g., major and minor overhauls) Develop never-sieze bearings Increase surge margin in design for fuel cell Hot section (e.g., combustor, heat exchanger) endurance (increased life) Develop control strategies consistent with “new” thermal input Develop interface requirements, and standard interface strategies Maximize commonality amongst hybrid components with stand alone line of gas turbines that can meet these needs Material selection (near term) Materials development (long term) Optimize control strategies for efficiency taking into account protection of both GT and fuel cell On-board intelligent diagnostics--specific sensor development Variable geometry compressor or turbine blades to achieve a suitable design for surge margin Package alternator with power electronics Advanced bearings--maintenance free or “on the fly” (magnetic, air, lube free)--never seize Control for “graceful” depressurization High temperature heat exchangers and recuperators Design GT for proper size range Design turbine for lower turbine inlet temperature (TIT) Design GT for lower pressure ratio
Analysis Tools for Combined System Several advances are required in the areas of steady-state and dynamic modeling tools to both garner insight into the fundamental operation of systems and components, and to use for cycle optimization, design of control systems and strategies, and use for developing next generation hybrid systems. Neither steady-state nor dynamic modeling tools are readily available for simulating state-of-the-art fuel cells that are to be used in hybrid fuel cell systems. These detailed models must be developed to gain insight into both fuel cell component operation and design, as well as system and sub-system level understanding of hybrid systems. In addition, system level modeling tools must be developed that include fundamentally sound, but simplified models for fuel cells and other system components for use in developing, analyzing and optimizing system configurations as well as designing control strategies for hybrid systems. The following developments and features are to be advanced under the current hybrid systems development plan: • • • • • • • • • • • • • •
Computer model for each component and system under consideration Validation of computer models by comparison to literature data for accurate predictions Acquisition of quality data from systems, subsystems and simulators for model validation Development and/or use of user friendly interfaces for the modeling tools to facilitate widespread use throughout the community Provide guidance and assist in the design of control algorithms Develop mechanisms for sharing empirical data previously held as proprietary Accurately determination of time scales of simulation required for components and entire hybrid systems Develop capabilities for concurrent dynamic simulation and materials / stress analyses Use super-computers and parallel processing to enhance computing power Assess existing software packages for amenability to new models and use as user interface Coordinate with previous efforts in related fields (e.g., Advanced Gas Turbine Systems Research (AGTSR) Program) Use models to identify potential control (manipulated) variables and measured outputs Use models to discover troublesome system or component behavior before systems are built Use Siemens Westinghouse Power Corporation, Southern California Edison 220 kW hybrid system, and FuelCell Energy, Capstone Turbines hybrid system as the first two data sources
Integration and Optimization of Fuel Cell/Turbine Combined Cycles 157
The integration and optimization of fuel cell gas turbine cycles depends upon the identification of a hybrid market and the development of system designs that address market needs. The first market that appears ripe for the use of hybrid systems is the distributed generation market. This is a logical first application for hybrid systems since the major fuel cell and gas turbine manufacturers
1.4 Hybrid Gas Turbine Fuel Cell Systems are already developing systems and components in the size range that is applicable to distributed generation, and since natural gas fired systems are amenable to this market. In addition, hybrid systems have high efficiency and ultra-low emissions features that make them attractive in a wide variety of applications and markets. The same type of hybrid design, however, will not work in each market application. Thus, the current hybrid plan must address the issues of system integration and optimization for a variety of applications. The current plan includes supporting advances in integration and optimization to address the following: • • • • • • • • • • • • • •
Service of the Distributed Generation / Industrial Market Service of larger central plant systems Service of SECA portable power , APU, and mobile applications Definition of appropriate system configurations for each of the above markets Matching of the transient response of fuel cells and turbines for each application Integration of fuel cell and turbine power output for each application Optimization of size and performance to include stored energy, fuel cell, and turbine management and control Understanding and control of transients over entire duty cycle, which differs for each application Establish criteria for installed cost, operating cost, reliability, and Life-Cycle Analysis (LCA) for each application Optimization of components and systems for improved performance in each application Perform trade-off analyses for each application Establish and demonstrate performance to the marketplace Study current operating flexibility and amenability to desired designs Test prototype systems, even to the point of destructive testing to investigate applicability
Research is required to enable integrated hybrid systems to meet performance expectations. Some of the particular needs that the integration technology may need to provide to reach the expected hybrid system performance targets include the following: 1) 2) 3) 4)
5) 6) 7) 8) 9) 10)
Systems Analysis (to identify more efficient and cost effective hybrids) Off-design Performance Analysis (part-load and sensitivity to ambient conditions) Dynamic and Transient Analysis (load following capability, rapid start-up and shut-down) Fuel Processing a) Compact Reformers b) Membrane Reformers for Natural Gas (reactors that separate one of the products of reaction such as hydrogen or carbon dioxide as it is formed from the reaction mixture) c) Resilient Reformer Catalysts for Natural Gas (less susceptible to poisons such as sulfur and chlorine compounds, and carbon deposition which will allow use of lower steam to carbon ratios) d) Cost Effective Partial Oxidation Plants for “Dirty Fuels” such as coal, biomass, refinery residues (compact reactor system designs and operating at lower temperatures to increase cold gas efficiency, and reduce oxygen consumption) e) Cost Effective Air Blown Catalytic Partial Oxidation and/or Reforming of Distillate for Mobile Applications such as Ships and Locomotives (compact reactor system designs and catalysts less susceptible to poisons such as sulfur and chlorine compounds, and carbon deposition which will allow use of lower steam to carbon ratios, operating at lower temperatures to increase cold gas efficiency, and reduce air usage in case of PoX) f) Cost Effective Air Blown Catalytic Partial Oxidation and/or Reforming of Diesel and Gasoline for Automotive Applications (compact reactor system designs and catalysts less susceptible to poisons such as sulfur and chlorine compounds, and carbon deposition which will allow use of lower steam to carbon ratios, operating at lower temperatures to increase cold gas efficiency, and reduce air usage in case of PoX) g) Cost Effective Air Blown Partial Oxidation and/or Reforming of Diesel/Gasoline Substitutes such as Alcohols and Dimethyl Either (compact reactor system designs and catalysts less susceptible to carbon deposition which will allow use of lower steam to carbon ratios, operating at lower temperatures to increase cold gas efficiency, and reduce air usage in case of PoX) Cost Effective and Efficient Oxygen Production (e.g., ion transport membranes) Cost Effective and Efficient Hydrogen Separation from Syn Gas (e.g., ion/proton transport membranes) Fuel Cleanup and Desulfurization a) Regenerable Desulfurization of Natural Gas b) Hot Gas Cleanup of Syn-Gas for Particulate, Sulfur and Chlorine Compounds Removal High Temperature Heat Exchangers (transferring heat from atmospheric or low pressure fuel cell combustors to working fluid of high pressure ratio gas turbines) – next 15 years Compact Mobile Unit Sub-system Designs (e.g., to be able to operate with unstable liquid levels caused by motion) Hydrogen Storage
Specific Functionality and Specification of System Components Since hybrid systems are new and evolving in both their design and the design of individual components one part of the hybrid program must focus on the specific functionality and specification of system components as they are to be applied in hybrid systems.
158
Jack Brouwer Market and Design Analyses Developing an understanding of the market for fuel cell hybrid systems will be integrated into the development efforts. These efforts will endeavor to develop: • • • • • • • • • • • • • • • •
Market understanding Impacts of the market(s) on HPS designs Sensitivity of parameters between markets and systems Equipment parameters and costs for each market segment Definition and identification of competing technologies Understanding of the impacts of regulatory policy Definitions of potential markets and market segmentation Interaction of systems with one another and utility grids Accurate models of equipment and integration (e.g., with the grid) Strategies for sharing of market information amongst competitive entities Develop steady state equipment and integrated systems analysis tools Develop dynamic equipment and integrated systems analysis tools Sources for neutral, objective and reliable information gathering and dissemination Collaboration amongst equipment user groups Collaboration amongst manufacturers Centers for multi-disciplinary research (Business, Economics, Engineering)
Integration of Fuel Cells and Engines Integration of the disparate technologies that comprise a hybrid system is perhaps the most significant challenge of the hybrid program. Hybrid systems are most definitely not comprised of the simple linking of two or more technologies through an easily configurable interface. A hybrid system is an entity distinct from and superior to the sum of its parts. In this context the integration of fuel cells and engines to form an integrated whole hybrid is one of the most important tasks of the hybrid program. The goals of the integration aspects of the program include developing hybrids and hybrid concepts with: • • • •
Competitive System Cost Modularity Operational Flexibility Simplicity (O&M)
It is anticipated that the integration elements of the hybrid program can be addressed in stages with the following overall staged goals: • • • • •
Adapt Existing Components (2005) Optimize configurations (2005-2010) Configure and test optimized hardware (2005-2010) Learn lessons from first generation optimized systems (2010) Develop and demonstrate next generation hybrids (2015)
Some of the research and development challenges that are faced in this integration aspect of the program include the following: • • • • • • • •
Interface of Major Components Reconfiguration of Major Components System Operational Features Control Flow Matching Thermal Management System Balance Safety
Key tools and strategies that must be developed and used in the integration elements of the hybrid program include:
159
• • • •
Performance Modeling Design Systems Controls Research and Development Performance Analysis
1.4 Hybrid Gas Turbine Fuel Cell Systems • • • • • • •
Packaging Re-Engineer Hardware based upon Requirements Simplification of System Flow Modeling Dynamic System Modeling Eliminate Balance of Plant Items Combine Balance of Plant Items
1.4-9 Acknowledgments: Special contributions to this work have been made by Dr. Rory A. Roberts, who accomplished most of the dynamic modeling efforts presented herein, and Mr. Yaofan Yi, who conducted most of the thermodynamic analyses.
1.4-10 Notes ___________________________ 1. Gottesfeld, S., and Zawadinski, T., Advances in Electrochemical Science and Engineering, Volume 5, Alkire, R., Gerischer, H., Kolb, D., and Tobias, C., editors, 1998. 2. Appleby, A.J., and Yeager, E.B., Energy, vol. 137, p. 11, 1986. 3. Hirschenhofer, J.H. Stauffer, D.B., Engleman, R.R. and Klett, M.G., Fuel Cell Handbook, Fourth Edition, DOE Contract No. DE-AC21-94MC31166, Reading, PA, 1998. 4. Baker, B.S. “Carbonate Fuel Cells – A Decade of Progress,” 191st Meeting, Electrochemical Society, May, 1997. 5. Singhal, S.C., “Recent Progress in Tubular Solid Oxide Fuel Technology,” Proceedings of the Fifth International Symposium on Solid Oxide Fuel Cells (SOFC-V), The Electrochemical Society, Inc., Pennington, NJ, 1997; Minh, N.Q., “Ceramic Fuel Cells,” J. American Ceramic Society, vol. 76, issue 3, pp. 563-588, 1993. 6. Ibid. 7. Dennis, R., U.S. Department of Energy, “Hybrid Fuel Cell Systems,” International Colloquium on Environmentally Preferred Advanced Generation, 2003; Bloomfield et al., United States Patent 3,973,993; 8/10/76; Landau, United States Patent 3,976,506; 8/24/76; Bloomfield, United States Patent 3,976,507; 8/24/76; 8. Hendriks et al., United States Patent 5,319,925; 1/14/94; Pietrogrande et al., United States Patent 5,314,761; 5/24/94; Domeracki et al., United States Patent 5,413,879; 5/9/95; Micheli et al., United States Patent 5,449,568; 9/12/95; Shingai et al., United States Patent 5,482,791; 1/9/96; Hsu et al., United States Patent 5,693,201; 12/2/97; Wolfe et al., United States Patent 5,678,647; 10/21/97; Skowronski, United States Patent 5,811,201; 9/22/98; 9. White, D.J., Solar Turbines Incorporated; TTS85/492; “Energy Conversion: A Vision of the Future;” 1997. 10. IGTI (1999). The Hybrid Cycle: Integration of the Gas Turbine with a Fuel Cell Session, The International Gas Turbine Institute Turbo-Expo, Indianapolis, June. a) Developmental Status of Hybrids (1999). ASME 99-GT-400 (Abbie Layne, Mark Williams, Scott Samuelsen, Patricia Hoffman); b) Hybrid Gas Turbine and Fuel Cell Systems in Perspective Review (1999). ASME 99-GT-419 (David White); c) Solid Oxide Fuel Cell Power System Cycles (1999). ASME 99-GT-356 (Stephen E. Veyo, Wayne L. Lundberg); d) The Hybrid Cycle: Integration of a Fuel Cell with a Gas Turbine (1999). ASME 99-GT-430 (John D. Leeper); e) The Hybrid Cycle: Integration of Turbomachinery with a Fuel Cell (1999). ASME 99-GT-361 (Sy Ali, Robert R. Moritz); f) Technical Development Issues and Dynamic Modeling of Gas Turbine and Fuel Cell Hybrid Systems (1999). ASME 99-GT-360 (Eric Liese, Randall Gemmen, Faryar Jabbari, Jacob Brouwer); IGTI (2000). The Hybrid Cycle: Integration of the Gas Turbine with a Fuel Cell Session, The International Gas Turbine Institute Turbo-Expo, Munich, May; a) Hybrid Heat Engines: The Power Generation Systems of the Future (2002). ASME 2000-GT-0549 (Abbie Layne, Mark Williams, Scott Samuelsen, Patricia Hoffman);
160
Jack Brouwer
161
b) Tubular Solid Oxide Fuel Cell/Gas Turbine Hybrid Cycle Power Systems Status (2002). ASME 2000GT-0550 (Stephen Veyo, Larry Shockling, Jeffrey Dederer, James Gillett, Wayne Lundberg); c) A Prototype for the First Commercial Pressurized Fuel Cell System (2002). ASME 2000-GT-0551 (Sy Ali, Robert Moritz); d) Ultra High Efficiency Hybrid Direct Fuel Cell/Turbine Power Plan (2002). ASME 2000-GT-0552 (Anthony J. Leo, Hossein Ghezel-Ayagh, Robert Sanderson); e) Analysis Strategies for Tubular SOFC Based Hybrid Systems (2002). ASME 2000-GT-0553 (Ashok Rao, Scott Samuelsen); f) Development of Dynamic Modeling Tools for Solid Oxide and Molten Carbonate Hybrid Fuel Cell Gas Turbine Systems (2002). ASME 2000-GT-0554 (Randall Gemmen, Eric Liese, Jose Rivera, Faryar Jabbari, Jacob Brouwer); IGTI (2001). The Hybrid Cycle: Integration of the Gas Turbine with a Fuel Cell Session, The International Gas Turbine Institute Turbo-Expo, June, New Orleans; a) Hybrid Fuel Cell Heat Engines: Recent Efforts (2001). ASME 2001-GT-0588 (Abbie Layne, Mark Williams, Norman Holcombe, Scott Samuelsen); b) A Turbogenerator for Fuel Cell/Gas Turbine Hybrid Power Plant (2001). ASME 2001-GT-0524 Sy Ali, Robert Moritz); c) A Thermodynamic Analysis of Tubular SOFC Based Hybrid Systems (2001). ASME 2001-GT-0522 (Ashok Rao, G.S. Samuelsen); d) A High-Efficiency SOFC Hybrid Power System Using the Mercury 50 ATS Gas Turbine (2001). ASME 2001-GT-0521 (Wayne Lundberg, Stephen Veyo, Mark D. Moeckel); IGTI (2002). The Hybrid Cycle: Integration of the Gas Turbine with a Fuel Cell Session, The International Gas Turbine Institute Turbo-Expo, June, Amsterdam; a) The National Energy Technology Laboratory’s Hybrid Power Systems Program (2002). ASME GT2002-30668 (Richard Dennis, Mark Williams, Abbie Layne, Scott Samuelsen, Norm Holcombe); b) Status of Pressurized SOFC/Gas Turbine Power System Development at Siemens Westinghouse (2002). ASME GT-2002-30670 (Stephen Veyo, Kevin Litzinger, Shailesh Vora, Wayne Lundberg); c) Power Plant System Configurations for the 21st Century (2002). ASME GT-2002-30671 (Ashok Rao, Scott Samuelsen, Fred Robson, Rodney Geisbrecht); d) Rao, Ashok; Samuelsen, Scott; Robson, Fred; Geisbrecht, Rodney (2002). Power Plant System Configurations for the 21st Century, ASME GT-2002-30671; e) Veyo, Stephen; Litzinger, Kevin; Vora, Shailesh, and Lundberg, Wayne (2002). Status of Pressurized SOFC/Gas Turbine Power System Development at Siemens Westinghouse, ASME GT-2002-30670. 11. Rao, A. D., Samuelsen, G. S., 2002, “Analysis Strategies for Tubular Solid Oxide Fuel Cell Based Hybrid,” Journal of Engineering for Gas Turbines and Power. 124(July 2002), pp. 503-509. 12. Massardo, A. F., Lubelli, F., 2000, “Internal Reforming Solid Oxide Fuel Cell-Gas Turbine Combined Cycles (IRSOFC-GT): Part A- Cell Model and Cycle Thermodynamic Analysis,” Journal of Engineering for Gas Turbines and Power. 122, pp. 27-35. 13. Costamagna, P., et al., 2004, “Electrochemical model of the integrated planar solid oxide fuel cell (IP-SOFC),” Chemical Engineering Journal. 102(1), pp. 61-69. 14. Bessette, N.F.,1994, “Modeling and Simulation for SOFC Power Systems,” in Mechanical Engineering, Georgia Institute of Technology, Atlanta, p. 209; Yi, Y., Smith, Thomas P., Brouwer, Jacob, Rao, Ashok D., 2003, “Simulation of a 220 kW Hybrid SOFC Gas Turbine System and Data Comparison.” Journal of Power Sources; Chan, S.H., Ho, H. K., Tian, Y., Modeling of Simple Hybrid Solid Oxide Fuel Cell and Gas Turbine Power Plant. Journal of Power Sources, 2002. 109: p.111-120; See Notes 11, 12, and 13. 15. Gemmen, R. S., Liese, Eric, Rivera, Jose G., Jabbari, Faryar, and Brouwer, Jacob. 2000, “Development of Dynamic Modeling Tools for Solid Oxide and Molten Carbonate Hybrid Fuel Cell Gas Turbine Systems.” 2000-GT-0552, in ASME Turbo Expo. Munich, Germany: ASME; Liese, E. A., Gemmen, Randall S., Jabbari, Faryar, Brouwer, Jacob, 1999, “Technical Development Issues and Dynamic Modeling of Gas Turbine and Fuel Cell Hybrid Systems,” Journal of Engineering for Gas Turbines and Power; Lukas, M. D., Lee, Kwang Y., Ghezel-Ayagh, Hossein, 1999, “Development of a Stack Simulation Model for Control Study on Direct Reforming Molten Carbonate Fuel Cell power Plant,” IEEE Transactions on Energy Conversion. PE-468-EC-0-01-1999; Lukas, M. D., Lee, Kwang Y., Ghezel-Ayagh, Hossein, 2000, “Operation and Control of Direct Reforming Fuel Cell Power Plant,” IEEE Power Engineering Society. 16. Ghezel-Ayagh, H., Daly, Joseph M., Wang, Zhao-Hui, Advances in Direct Fuel Cell / Gas Turbine Power Plants. 2003 ASME Turbo Expo, Atlanta, Georgia, 2003. GT2003-38941. 17. Veyo, S.E., Lundberg, Wayne L., Vora, Shailesh D., Litzinger, Kevin P., Tubular SOFC Hybrid Power System Status. Preceedings of ASME Turbo Expo 2003, 2003. GT2003-38943; See Note 14. 18. Roberts, R. A., Brouwer, J., Gemmen, R.S., and Liese, E.A., 2003, “Inter-Laboratory Dynamic Modeling of a Carbonate Fuel Cell for Hybrid Application.” GT2003-38774, in 2003 ASME Turbo Expo, Atlanta, Georgia;
1.4 Hybrid Gas Turbine Fuel Cell Systems Roberts, R. A., Brouwer, J., Gemmen, R.S., and Liese, E.A., 2004, “Dynamic Simulation of Carbonate Fuel CellGas Turbine Hybrid Systems.” GT2004-53653, in ASME Turbo Expo. Vienna, Austria. 19. Xu, J., Froment, Gilbert F., Methane Steam Reforming, Methanation and Water-Gas Shift: I. Intrinsic Kinetics. AIChE Journal, 1989. 35(1): p. 88-96. 20. Weber, A., Bastain, Sauer, Muller, Axel C., Herbstriitt, Dirk, Ivers-Tiffee,Ellen, Oxidation of H2, CO and Methane in SOFCs with Ni/YSZ-Cermet Anodes. Solid State Ionics, 2002. 152-153: p. 543-550. 21. Roberts, R. A., et al. 2005, “Development of Controls for Dynamic Operation of Carbonate Fuel Cell-Gas Turbine Hybrid Systems.” GT2005-68774, in ASME Turbo Expo. Reno, NV, USA: ASME; See Note 18. 22. Roberts, R. A., 2005, “A Dynamic Fuel Cell-Gas Turbine Hybrid Simulation Methodology to Establish Control Strategies and an Improved Balance of Plant,” Ph.D. thesis, in Mechanical and Aerospace Engineering, University of California, Irvine, Irvine, p. 316. 23. See Note 13. 24. Roberts, R. A. and Brouwer, J., 2005, “Dynamic Simulation of a Pressurized 220 kW Solid Oxide Fuel Cell-Gas Turbine Hybrid System: Modeled Performance Compared to Measured Results,” Journal of Fuel Cell Science and Technology, (Accepted August, 2005). 25. Willis, J. 2005, “Capstone Microturbines.” in ICEPAG. Irvine, California: APEP. 26. Kim, J.-W., et al., 1999, “Polarization Effects in Intermediate Temperature, Anode-Supported Solid Oxide Fuel Cells,” Journal of The Electrochemical Society. 146(1), pp. 69-78.
162
BIOGRAPHY
1.4 Hybrid Gas Turbine Fuel Cell Systems
Professor Jack Brouwer, Ph.D. Associate Director National Fuel Cell Research Center University of California Irvine, CA 92697-3550 email:
[email protected] http://www.nfcrc.uci.edu
Dr. Jack Brouwer is Adjunct Assistant Professor of Mechanical and Aerospace Engineering at U.C. Irvine (UCI) and the Associate Director of the National Fuel Cell Research Center (NFCRC). Dr. Brouwer completed doctoral studies in Mechanical Engineering and Chemical Engineering at the Massachusetts Institute of Technology (MIT). Prior to joining the NFCRC, Dr. Brouwer was a research faculty at the University of Utah, a Senior Engineer at Reaction Engineering International, and a Staff Scientist at Sandia National Laboratories. Dr. Brouwer has expertise in energy systems, fuel cell technology, turbulent reacting flows, computational fluid dynamics, chemical kinetics, and electrochemical reactions with concurrent heat, mass and momentum transfer in electrochemical systems. Dr. Brouwer is leading research and development efforts including projects on hydrogen refueling, the world’s first independent fuel cell vehicle testing, the world’s first testing and evaluation of a hybrid fuel cell gas turbine system, the development and application of dynamic fuel cell and hybrid fuel cell systems simulations, and the advancement of reformation technologies for gaseous, liquid, and solid hydrocarbon fuels. Dr. Brouwer is a regular instructor at UCI in the areas of fuel cells, thermodynamics, heat transfer, and combustion. He developed and introduced the first graduate level fuel cell course to UCI in 2002, and is a regular instructor in fuel cell short courses around the world.
2.0
Axial-Flow Compressors
2.0-1 Introduction The compressors in most gas turbine applications, especially units over 5MW, use axial flow compressors. An axial flow compressor is one in which the flow enters the compressor in an axial direction (parallel with the axis of rotation), and exits from the gas turbine, also in an axial direction. The axial-flow compressor compresses its working fluid by first accelerating the fluid and then diffusing it to obtain a pressure increase. The fluid is accelerated by a row of rotating airfoils (blades) called the rotor, and then diffused in a row of stationary blades (the stator). The diffusion in the stator converts the velocity increase gained in the rotor to a pressure increase. A compressor consists of several stages: 1) A combination of a rotor followed by a stator make-up a stage in a compressor; 2) An additional row of stationary blades are frequently used at the compressor inlet and are known as Inlet Guide Vanes (IGV) to ensue that air enters the first-stage rotors at the desired flow angle, these vanes are also pitch variable thus can be adjusted to the varying flow requirements of the engine; and 3) In addition to the stators, another diffuser at the exit of the compressor consisting of another set of vanes further diffuses the fluid and controls its velocity entering the combustors and is often known as the Exit Guide Vanes (EGV). In an axial flow compressor, air passes from one stage to the next, each stage raising the pressure slightly. By producing lowpressure increases on the order of 1.1:1 to 1.4:1, very high efficiencies can be obtained as seen in table 1. The use of multiple stages permits overall pressure increases of up to 40:1 in some aerospace applications and a pressure ratio of 30:1 in some Industrial applications. The last twenty years has seen a large growth in gas turbine technology. The growth is spear headed by the increase in compressor pressure ratio, advanced combustion techniques, the growth of materials technology, new coatings and new cooling schemes. The increase in gas turbine efficiency is dependent on two basic parameters: 1. 2.
Increase in Pressure Ratio Increase in Firing Temperature
It also should be remembered that the Gas Turbine Axial Flow Compressor consumes between 55%-65% of the power produced by the Turbine section of the gas turbine. Table 1 Axial Flow Compressor Characteristics
Meherwan P. Boyce 2121 Kirby Drive, Number 28N Houston, TX 77019 713-807--0888 713-807-0088 Fax boycepower.com
[email protected]
163
Type of Application
Type of Flow
Inlet Relative Velocity Mach Number
Pressure Ratio per Stage
Efficiency per Stage
Industrial
Subsonic
0.4-0.8
1.05-1.2
88%-92%
Aerospace
Transonic
0.7-1.1
1.15-1.6
80%-85%
Research
Supersonic
1.05-2.5
1.8-2.2
75%-85%
The aerospace engines have been the leaders in most of the technology in the gas turbine. The design criteria for these engines was high reliability, high performance, with many starts and flexible operation throughout the flight envelope. The engine life of about 3500 hours between major overhauls was considered good. The aerospace engine performance has always been rated primarily on its Thrust/Weight ratio. Increase in engine Thrust / Weight Ratio is achieved by the development of high aspect ratio blades in the compressor as well as optimizing the pressure ratio and firing temperature of the turbine for maximum work output per unit flow. The Industrial Gas Turbine has always emphasized long life and this conservative approach has resulted in the Industrial Gas Turbine in many
aspects giving up high performance for rugged operation. The Industrial Gas Turbine has been conservative in the pressure ratio and the firing temperatures. This has all changed in the last ten years; spurred on by the introduction of the “Aero-Derivative Gas Turbine” the Industrial Gas Turbine has dramatically improved its performance in all operational aspects. This has resulted in dramatically reducing the performance gap between these two types of gas turbines. Figure 1 indicates the growth of the Pressure Ratio in a gas turbine over the past 50 years. The growth of both the Pressure Ratio and Firing Temperature parallel each other, as both growths are necessary to achieving the increase in thermal efficiency in Gas Turbines. The Axial flow compressor in most of the advanced gas turbine is a multistage compressor consisting of 17-22 stages with an exceedingly high pressure ratio. It is not uncommon to have pressure ratios in industrial gas turbines in the 17 to 20:1 range with some units having pressure ratios in the 30:1 range. Figure 2 shows a multistage high-pressure axial flow compressor rotor. The low-pressure increase per stage also simplifies calculations in the preliminary design of the compressor by justifying the air as incompressible in its flow through the stage.
Engine Pressure Ratio Development
45 40
Pressure ratio
35 30 25 20 15
Pressure Ratio Aircraft
10
Pressure Ratio Industrial
5 0 1940
1950
1960
1970
1980
1990
2000
2010
Year
Fig.1.Development of pressure ratio over the past 50 years
Fig. 2. Axial Flow Compressor Rotor
164
Meherwan P. Boyce Figure 3 shows the stators, the stationary blades which are in between each rotor blade and causes the flow to be diffused (increase in the static pressure, reduction of the absolute velocity). The early stages of the stators in figure 3 are adjustable, as can be noted by their circular base. The adjustable stators allow the stator to be positioned to the correct flow angle leaving the blades as the air mass flow varies with load and inlet temperature.
Fig. 3. Axial Flow Compressor Stators located in the casing
As with other types of rotating machinery, an axial compressor can be described in a cylindrical coordinate system. The z axis is along the axis of rotation which is along the running length of the compressor shaft, the radius r is measured outward from the shaft, and the angle of rotation θ is the angle turned by the blades in figure 4. This coordinate system will be used throughout this discussion of axial-flow compressors.
Fig. 4. Coordinate System for Axial-Flow Compressor
Fig. 5 Variation of Temperature Velocity, and Pressure through an Axial-Flow Compressor
Figure 5 shows the pressure, velocity, and total temperature variation for flow through several stages of an axial compressor. As indicated earlier in figure 3, the length of the blades, and the annulus area, this is the area between the shaft and shroud, decreases throughout the length of the compressor. This reduction in flow area compensates for the increase in fluid density as it is compressed, permitting a constant axial velocity. In most preliminary calculations used in the design of a compressor, the average blade height is used as the blade height for the stage.
165
2.0 Axial-Flow Compressors 2.0-2 Blade and Cascade Nomenclature Since airfoils are employed in accelerating and diffusing the air in a compressor, much of the theory and research concerning the flow in axial compressors are based on studies of isolated airfoils. The nomenclature and methods of describing compressor blade shapes are almost identical to that of aircraft wings. Research in axial compressors involves the inter effect of one blade on the other; thus, several blades are placed in a row to simulate a compressor rotor or stator. Such a row is called a cascade. When discussing blades, all angles which describe the blade and its orientation are measured with respect to the shaft (Z axis) of the compressor. The airfoils are curved, convex on one side and concave on the other, with the rotor rotating toward the concave side. The concave side is called the pressure side of the blade, and the convex side is called the suction side of the blade. The chordline of an airfoil is a straight line drawn from the leading edge to the trailing edge of the airfoil, and the chord is the length of the chordline as seen in figure 6. The camberline is a line drawn halfway between the two surfaces, and the distance between the camberline and the chordline is the camber of the blade. The camber angle θ is the turning angle of the camber line. The blade shape is described by specifying the ratio of the chord to the camber at some particular length on the chordline, measured from the leading edge. The aspect ratio AR is the ratio of the blade length to the chord length. The term “hub-to-tip ratio” is frequently used instead of aspect ratio. The aspect ratio becomes important when three-dimensional flow characteristics are discussed. The aspect ratio is established when the mass flow characteristics are discussed. The aspect ratio is established when the mass flow and axial velocity have been determined.
Fig. 6. Blade profile nomenclature
The pitch Sb of a cascade is the distance between blades, usually measured between the camberlines at the leading or trailing edges of the blades. The ratio of the chord length to the pitch is the solidity σ of the cascade. The solidity measures the relative interference effects of one blade with another. If the solidity is on the order of 0.5-0.7, the single or isolated airfoil test data, from which there are a profusion of shapes to choose, can be applied with considerable accuracy. The same methods can be applied up to a solidity of about 1.0 but with reduced accuracy. When the solidity is on the order of 1.0-1.5, cascade data are necessary. For solidity in excess of 1.5, the channel theory can be employed. The majority of present designs are in the cascade region. The blade inlet angle β1 is the angle formed by a line drawn tangent to the forward end of the camber line and the axis of the compressor. The blade outlet angle β2 is the angle of a line drawn tangent to the rear of the camberline. Subtracting β2 from β1 gives the blade camber angle. The angle that the chordline makes with the axis of the compressor is γ, the setting or stagger angle of the blade. High-aspect ratio blades are often pretwisted so that at full operational speed the centrifugal forces acting on the blades will untwist the blades to the designed aerodynamic angle. The pretwist angle at the tip for blades with AR ratios of about four is between two and four degrees. The air inlet angle α1, the angle at which incoming air approaches the blade, is different from β1. The difference between these two angles is the incidence angle i. The angle of attack α is the angle between the inlet air direction and the blade chord. As the air is turned by the blade, it offers resistance to turning and leaves the blade at an angle greater than β2. The angle at which the air does leave the blade is the air outlet angle α2. The difference between β2 and α2 is the deviation angle δ. The air turning angle is the difference between α1 and α2 and is sometimes called the deflection angle. The original work by NACA and NASA is the basis on which most modern axial-flow compressors are designed. Under NACA, a large number of blade profiles were tested. The test data on these blade profiles is published. The cascade data conducted by NACA is the most extensive work on its kind. In most commercial axial-flow compressors in Gas Turbines built before 1990, NACA 65 series blades are used. These blades are usually specified by notation similar to the following: 65-(18) 10. This notation means that the blade has a lift coefficient of 1.8, a profile shape 65, and a thickness/chord ratio of ten percent (10%). The lift coefficient can be directly related to the blade camber angle by the following relationship for 65 series blades: Θ ≈ 25 CL.
(1)
166
Meherwan P. Boyce The new advanced compressor rotors have fewer blades with higher loadings, and the blades are thinner, larger, and are designed using advanced radial equilibrium theory, which create Three Dimensional and Controlled Diffusion shaped airfoils (3D/CDA), with smaller clearances and higher loading per stage.
2.0-3 Elementary Airfoil Theory When a single airfoil is parallel to the velocity of a flowing gas, the air flows over the airfoil as shown in figure 7a. The air divides around the body, separates at the leading edge, and joins again at the trailing edge of the body. The main stream itself suffers no permanent deflection from the presence of the airfoil. Forces are applied to the foil by the local distribution of the steam and the friction of the fluid on the surface. If the airfoil is well designed, the flow is streamlined with little or no turbulence. If the airfoil is set at the angle of attack to the air stream, as in figure 7b, a greater disturbance is created by its presence, and the streamline pattern will change. The air undergoes a local deflection, though at some distance ahead of and behind the body the flow is still parallel and uniform. The upstream disturbance is minor compared to the downstream disturbance. The local deflection of the air stream can, by Newton’s laws, be created only if the blade exerts a force on the air; thus, the reaction of the air must produce an equal and opposite force on the airfoil. The presence of the airfoil has changed the local pressure distribution and, by the Bernoulli equation, the local velocities. Examination of the streamlines about the body shows that over the top of the airfoil, the lines approach each other, indicating an increase of velocity and a reduction in static pressure. On the underside of the airfoil, the action separates the streamlines, resulting in a static pressure increase.
Fig. 7. Flow around an airfoil at various angles of attack: a, parallel to the velocity of a flowing gas; b, set at the angle of attack to the airstream; c, pressure measurement at various points on airfoil’s surface.
Measurement of the pressure at various points on the surface of the airfoil will reveal a pressure distribution as shown in figure 7c. The vectorial sum of these pressures will produce some resultant force acting on the blade. This resultant force can be resolved into a lift component L at right angles to the undisturbed air stream, and a drag component D, moving the airfoil in the direction of flow motion. This resultant force is assumed to act through a definite point located in the airfoil so that the behavior will be the same as if all the individual components were acting simultaneously. By experimentation, it is possible to measure the lift and drag forces for all values of airflow velocity, angles of incidence, various airfoil shapes. Thus, for any one airfoil the acting forces can be represented as shown in figure 8a. Using such observed values, it is possible to define relations between the forces
where:
167
D L
= CD Aρ V2/2 = CL Aρ V2/2
L A D ρ CL V CD
= lift force = surface area = drag force = fluid density = lift coefficient = fluid velocity = drag coefficient
(2) (3)
2.0 Axial-Flow Compressors Two coefficients have been defined, CL and CD, relating velocity, density, area, and lift or drag forces. These coefficients can be calculated from wind-tunnel tests and plotted as shown in figure 8b versus the angle of attack for any desired section. These curves can then be employed in all future predictions involving this particular foil shape.
Fig. 8. Characteristics of the lift and drag forces on an airfoil
Examination of figure 8 reveals that there is an angle of attack which produces the highest lift force and lift coefficient. If this angle is exceeded, the airfoil “stalls” and the drag force increases rapidly. As this maximum angle is approached, a great percentage of the energy available is lost in overcoming friction, and a reduction in efficiency occurs. Thus, there is a point, usually before the maximum lift coefficient is reached, at which the most economical operation occurs as measured by effective lift for a given energy supply.
2.0-4 Laminar-Flow Airfoils Just before and during World War II, much attention was given to laminar-flow airfoils. These airfoils are designed so that the lowest pressure on the surface occurs as far back as possible. The reason for this design is that the stability of the laminar boundary layer increases when the external flow is accelerated (in the flow with a pressure drop), and the stability decreases when the flow is directed against increasing pressure. A considerable reduction in skin friction is obtained by extending the laminar region in this way, provided that the surface is sufficiently smooth. A disadvantage of this type of airfoil is that the transition from laminar to turbulent flow moves forward suddenly at small angles of attack. This sudden movement results in a narrow low-drag bucket, which means that the drag at moderate-to-large attack angles is much greater than an ordinary airfoil for the same attack angle as seen in figure 9. This phenomenon can be attributed to the minimum pressure point moving forward; therefore, the point of transition between laminar and turbulent flow is also advanced toward the nose as shown in figure 10. The more an airfoil is surrounded by turbulent airflow, the greater its skin friction will be.
Fig. 9. NACA measurements of drag coefficients for two laminar airfoils
168
Meherwan P. Boyce
Fig 10. Laminar Flow Airfoils
2.0-5 Energy Increase In an axial flow compressor, air passes from one stage to the next with each stage raising the pressure and temperature slightly. By producing low-pressure increases on the order of 1.1:1-1.4:1, very high efficiencies can be obtained. The use of multiple stages permits overall pressure increases up to 40:1. Figure 5 shows the pressure, velocity, and total temperature variation for flow through several stages of an axial flow compressor. It is important to note here that the changes in the total conditions for pressure, temperature, and enthalpy occur only in the rotating component where energy is inputted into the system. As seen also in figure 5, the length of the blades, and the annulus area, which is the area between the shaft and shroud, decrease throughout the length of the compressor. This reduction in flow area compensates for the increase in fluid density as it is compressed, permitting a constant axial velocity. In most preliminary calculations used in the design of a compressor, the average blade height is used as the blade height for the stage. A heuristic approach for a multiple stage gas turbine compressor would be that the energy rise per stage would be constant, rather than the commonly held perception that the pressure rise per stage is constant. The energy rise per stage can be written as:
∆H =
[H 2 − H1 ] NS
(4)
where: H1, H2 = Total Inlet and Exit Enthalpy Btu/lbm (kJ/kg) and Ns = number of stages. Assuming that the gas is thermally and caloricaly perfect (cp, and γ are constant) equation 4 can be rewritten as:
(5)
where: Tin = Total Inlet Temperature (ºF, ºC) and P1, P2 = Total Inlet and Exit Pressure (psia, bar).
2.0-6 Velocity Triangles
169
As stated earlier, an axial-flow compressor operates on the principle of putting work into the incoming air by acceleration and diffusion. Air enters the rotor as shown in figure 11 with an absolute velocity (V) and an angle α1, which combines vectorially with the tangential velocity of the blade (U) to produce the resultant relative velocity W1 at an angle α2. Air flowing through the passages formed by the rotor blades is given a relative velocity W2 at an angle α4, which is less than α2 because of the camber of the blades. Note that W2 is less than W1, resulting from an increase in the passage width as the blades become thinner toward the trailing edges. Therefore, some
2.0 Axial-Flow Compressors diffusion will take place in the rotor section of the stage. The combination of the relative exit velocity and blade velocity produce an absolute velocity V2 at the exit of the rotor. The air then passes through the stator, where it is turned through an angle so that the air is directed into the rotor of the next stage with a minimum incidence angle. The air entering the rotor has an axial component at an absolute velocity VZ1 and a tangential component Vθ1.
Fig. 11. Typical velocity triangles for an axial-flow compressor
Applying the Euler turbine equation
H=
1 [U1Vθ 1 − U 2Vθ 2 ] gc
and assuming that the blade speeds at the inlet and exit of the compressor are the same and noting the relationships,
(6)
Vθ 1 = Vz1 tanα1
(7)
Vθ 2 = Vz 2 tanα 3
(8)
Equation (1) can be written
H=
U1 (Vz1 tan α 2 − Vz 2 tan α 3 ) gc
Assuming that the axial component (VZ) remains unchanged,
(9)
(10) The previous relationship is in terms of the absolute inlet and outlet velocities. By rewriting the previous equation in terms of the blade angles or the relative air angles, the following relationship is obtained: U1 - U2 = VZ1 tan α1 = VZ1 tan α2 = VZ2 tan α3 + VZ2 tan α4. Therefore, (11) The previous relationship can be written to calculate the pressure rise in the stage: (12) which can be rewritten .
(13)
170
Meherwan P. Boyce The velocity triangles can be joined together in several different ways to help visualize the changes in velocity. One of the methods is to simply join these triangles into a connected series. The two triangles can also be joined and superimposed using the sides formed by either the axial velocity, which is assumed to remain constant as shown in figure 12a, or the blade speed as a common side, assuming that the inlet and exit blade speed are the same as shown in figure 12b.
Fig. 12. Velocity triangles
2.0-7 Degree of Reaction The degree of reaction in an axial-flow compressor is defined as the ratio of the change of static head in the rotor to the head generated in the stage:
R=
H rotor H stage
(14)
The change in static head in the rotor is equal to the change in relative kinetic energy:
Hr = and
(
1 2 2 W1 − W1 2gc
)
(15)
W1 = Vz1 + (Vz1 tanα 2 ) 2
2
2
(16)
W2 = Vz 2 + (Vz 2 tanα 4 ) 2
2
2
(17)
Therefore, 2
V H r = z (tan 2 α 2 − tan 2 α 4 ) 2gc Thus, the reaction of the stage can be written
R=
Vz tan 2 α 2 − tan 2 α 4 2U tan α 2 − tan α 4
(18)
Simplifying the previous equation,
171
R=
Vz (tan α 2 + tan α 4 ) 2U
(19)
2.0 Axial-Flow Compressors In the symmetrical axial-flow stage, the blades and their orientation in the rotor and stator are reflected images of each other. Thus, in a symmetrical axial-flow stage where V1 = W2 and V2 = W1 as seen in figure 13, the head delivered in velocity as given by the Euler turbine equation can be expressed by the following relationships:
Fig. 13. Symmetrical velocity triangle for 50% reaction stage
[
]
1 ( U 12 − U 22 )+ (V12 − V22 )+ (W22 − W12 ) 2gc 1 ( H= W22 − W12 ) . 2gc H=
(20) (21)
The reaction for a symmetrical stage is fifty percent (50%). The fifty percent (50%) reaction stage is widely used, since an adverse pressure rise on either the rotor or stator blade surfaces is minimized for a given stage pressure rise. When designing a compressor with this type of blading, the first stage must be preceded by inlet guide vanes to provide prewhirl, and the correct velocity entrance angle to the first-stage rotor. With a high tangential velocity component maintained by each succeeding stationary row, the magnitude of W1 is decreased. Thus, higher blade speeds and axial-velocity components are possible without exceeding the limiting value of 0.70-0.75 for the inlet Mach number. Higher blade speeds result in compressors of smaller diameter and less weight. Another advantage of the symmetrical stage comes from the equality of static pressure rises in the stationary and moving blades, resulting in a maximum static pressure rise for the stage. Therefore, a given pressure ratio can be achieved with a minimum number of stages, a factor in the lightness of this type of compressor. The serious disadvantage of the symmetrical stage is the high exit loss resulting from the high axial velocity component. However, the advantages are of such importance in aircraft applications that the symmetrical compressor is normally used. In stationary applications, the symmetrical compressor is normally not used. In stationary applications, where weight and frontal area are of lesser importance, one of the other stage types is used. The term “asymmetrical stage” is applied to stages with reaction other than 50%. The axial-inflow stage is a special case of an asymmetrical stage where the entering absolute velocity is in the axial direction. The moving blades impart whirl to the velocity of the leaving flow which is removed by the following stator. From this whirl and the velocity diagram as seen in figure 14, the major part of the stage pressure rise occurs in the moving row of blades with the degree of reaction varying from 60% to 90%. The stage is designed for constant energy transfer and axial velocity at all radii so that the vortex flow condition is maintained in the space between blade rows.
Fig. 14. Axial-entry stage velocity diagram
The advantage of a stage with greater than 50% reaction is the low exit loss resulting from lower axial velocity and blade speeds. Because of the small static pressure rise in the stationary blades, certain simplifications can be introduced such as constantsection stationary blades and the elimination of interstage seals. Higher actual efficiencies have been achieved in this stage type than with the symmetrical stage - primarily because of the reduced exit loss. The disadvantages result from a low static pressure rise in the stationary blades that necessitates a greater number of stages to achieve a given pressure ratio and thus creates a heavy compressor. The lower axial velocities and blade speed, necessary to keep within inlet Mach number limitations, result in large diameters. In stationary applications where the increased weight and frontal area are not of great importance, this type is frequently used to take advantage of the higher efficiency.
172
Meherwan P. Boyce The axial-outflow stage diagram in figure 15 shows another special case of the asymmetrical stage with reaction greater than 50%. With this type of design, the absolute exit velocity is in an axial direction, and all the static pressure rise occurs in the rotor. A static pressure decrease occurs in the stator so that the degree of reaction is in excess of 100%. The advantages of this stage type are low axial velocity and blade speeds, resulting in the lowest possible exit loss. This design produces a heavy machine of many stages and of large diameter. To keep within the allowable limit of the inlet Mach number, extremely low values must be accepted for the blade velocity and axial velocity. The axial-outflow stage is capable of the highest actual efficiency because of the extremely low exit loss and the beneficial effects of designing for free vortex flow. This compressor type is particularly well-suited for closed-cycle plants where smaller quantities of air are introduced to the compressor at an elevated static pressure.
Fig. 15. Axial-outflow stage velocity diagram
While a reaction of less than 50% is possible, such a design results in high inlet Mach numbers to the stator row, causing high losses. The maximum total divergence of the stators should be limited to approximately 20o to avoid excessive turbulence. Combining the high inlet for the limiting divergence angles produces a long stator, thereby producing a longer compressor. Radial Equilibrium The flow in an axial-flow compressor is defined by the continuity, momentum, and energy equations. A complete solution to these equations is not possible because of the complexity of the flow in an axial-flow compressor. Considerable work has been done on the effects of radial flow in an axial-flow compressor. The first simplification used considers the flow axisymmetric. This simplification implies that the flow at each radial and axial station within the blade row can be represented by an average circumferential condition. Another simplification considers the radial component of the velocity as much smaller than the axial component velocity, so it can be neglected. For the low-pressure compressor with a low-aspect ratio, and where the effect of streamline curvature is not significant, the simple radial equilibrium change of the radial velocity component along the axial direction is zero (∂Vrad/∂Z = 0) and the change of entropy in the radial direction is zero negligible (a s/∂r = 0). The Meridional velocity (Vm) is equal to the axial velocity (VZ), since the effect of steamline curvature is not significant. The radial gradient of the static pressure can be given
Vθ2 ∂P =ρ ∂r r
.
(22)
Using the simple radial equilibrium equation, the computation of the axial velocity distribution can be calculated. The accuracy of the techniques depends on how linear V2θ/r is with the radius. The assumption is valid for low-performance compressors, but it does not hold well for the high-aspect ratio, highly loaded stages where the effects of streamline curvature become significant. The radial acceleration of the Meridional velocity and the pressure gradient in the radial direction must be considered. The radial gradient of static pressure for the highly curved streamline can be written (23) where ∈ is the angle of the streamline curvature with respect to the axial direction and rc is the radius of curvature.
173
To determine the radius of curvature and the streamline slope accurately, the configuration of the streamline through the blade row must be known. The streamline configuration is a function of the annular passage area, the camber and thickness distribution of the blade, and the flow angles at the inlet and outlet of the blade. Since there is no simple way to calculate the effects of all the parameters, the techniques used to evaluate these radial accelerations are empirical. By using iterative solutions, a relationship can be obtained. The effect of high radial acceleration with high-aspect ratios can be negated by tapering the tip of the compressor inward so that the hub curvature is reduced.
2.0 Axial-Flow Compressors Diffusion Factor The diffusion factor first defined by Lieblien is a blade-loading criterion:
W V − Vθ 2 . D = 1 − 2 + θ 1 2σW1 W1
(24)
The diffusion factor should be less than 0.4 for the rotor tip and less than 0.6 for the rotor hub and the stator. The distribution of the diffusion factor throughout the compressor is not properly defined. However, the efficiency is less in the later stages due to distortions of the radial velocity distributions in the blade rows. Experimental results indicate that even though efficiency is less in the later stages, as long as the diffusion loading limits are not exceeded, the stage efficiencies remain relatively high. The Incidence Rule For low-speed airfoil design, the region of low-loss operation is generally flat, and it is difficult to establish the precise value of the incidence angle that corresponds to the minimum loss as seen in figure 16. Since the curves are generally symmetrical, the minimum loss location was established at the middle of the low-loss range. The range is defined as the change in incidence angle corresponding to a rise in the loss coefficient equal to the minimum value. The following method for calculation of the incidence angle is applicable to cambered airfoils. Work by NASA on the various cascades is the basis for the technique. The incidence angle is a function of the blade camber, which is an indirect function of the airturning angle. (25) where i0 is the incidence angle for zero camber, and m is the slope of the incidence angle variation with the air-turning angle (ξ). The zero-camber incidence angle is defined as a function of inlet air angle and solidity as seen in figure 17 and the value of m is given as a function of the inlet air angle and the solidity as seen in figure 18. The incidence angle io is for a 10% blade thickness. For blades of other than 10% thickness, a correction factor K is used, which is obtained from figure 19.
Fig. 16. Loss as a function of incidence angle
Fig 18. Slope of incidence angle variation with air angle
Fig. 17. Incidence angle for zero-camber airfoil
Fig. 19. Correction factor for blade thickness and incidence angle calculation
174
Meherwan P. Boyce The incidence angle now must be corrected for the Mach number effect (δm). The effect of the Mach number on incidence angle is shown in figure 20. The incidence angle is not affected until a Mach in number of 0.7 is reached. The incidence angle is now fully defined. Thus, when the inlet and outlet air angles and the inlet Mach number are known, the inlet blade angle can be computed in this manner.
Fig. 20. Mach-number correction for incidence angle
2.0-8 The Deviation Rule Carter’s rule, which shows that the deviation angle is directly a function of the camber angle and is inversely proportional to
(
)
the solidity δ = mθ 1 / σ has been modified (Boyce) to take into account the effect of stagger, solidity, Mach number, and blade shape as shown in the following relationship: (26) where mƒ is a function of the stagger angle, maximum thickness, and the position of maximum thickness as seen in figure 21. The second term of the equation should only be used for camber angles 0 < θ > 8. The third term must be used only when the mach number is between 0.75 < M > 1.3.
Fig. 21. Position of maximum thickness effect on deviation
The use of NACA cascade data for calculating the exit air angle is also widely used. Mellor has replotted some of the low-speed NACA 65 series cascade data in convenient graphs of inlet air angle against exit air angle for blade sections of given lift and solidity set at various staggers. Figure 22 shows the NACA 65 series of airfoils.
175
2.0 Axial-Flow Compressors
Fig. 22. The NACA 65 series of cascade airfoils
The 65 series blades are specified by an airfoil notation similar to 65-(18)10. This specification means that an airfoil has the profile shape 65 with a camber line corresponding to a life coefficient (CL) = 1.8 and approximate thickness of 10% of the chord length. The relationship between the camber angle and the lift coefficient for the 65 series blades is shown in figure 23.
Fig. 23. Approximate relation between camber (θ) and CL0 of NACA 65 series
The low-speed cascade data have been replotted by Mellor in the form of graphs of α2 against α1 for blade sections of given camber and space-chord ratio but set at varying stagger γ, and tested at varying incidence (i = αi - β1) or angle of attack (α1 - γ) as seen in figure 24. The range on each block of results is indicated with heavy black lines, which show the attack angle at which the drag coefficient increases by 50% over the mean unstalled drag coefficient.
Fig. 24. The NACA 65 series cascade data (Reprinted, by permission of G. Mellor, Massachusetts Institute of Technology, Gas Turbine Laboratory Publication.)
176
Meherwan P. Boyce NACA has given “design points” for each cascade tested. Each design point is chosen on the basis of the smoothest pressure distribution observed on the blade surfaces: if the pressure distribution is smooth at one particular incidence at low speed, it is probable that the section will operate efficiently at a higher Mach number at the same incidence, and that this same incidence should be selected as a design point. Although such a definition appears somewhat arbitrary at first, the plots of such design points against solidity and camber give consistent curves. These design points are replotted in figure 25, showing the angle of attack (α1 - γ) plotted against space-chord ratio and camber is independent of stagger. If the designer has complete freedom to choose space-chord ratio, camber, and stagger, then a "design point" choice may be made by trial and error from the plots of figure 24 and 25. For example, if an outlet angle (α2) of 15 is required from an inlet angle of 35, a reference to the curves of the figures will show that a space-chord ratio of 1.0, camber 1.2, and stagger 23 will give a cascade operating at its design point. There is a limited variety of cascades of different space-chord ratios, but only one cascade that will operate at “design point” at the specified air angles. For example, if the space-chord ratio were required to be 1.0 in the previous example, then the only cascade that will produce design point operation is that of camber 1.2, stagger 23.
Fig 25. Design angles of attack (α1 - γ) for NACA 65 series
Such a design procedure may not always be followed, for the designer may choose to design the stage to operate closer to the positive stalling limit or closer to the negative stalling (choking) limit at design operating conditions to obtain more flexibility at offdesign conditions.
2.0-9 Compressor Operation Characteristics A compressor operates over a large range of flow and speed delivering a stable head/pressure ratio. During start up the compressor must be designed to operate in a stable condition at low rotational speeds. There is an unstable limit of operation known as ‘surging’, and it is shown on the performance map as the surge line. The surge point in a compressor occurs when the compressor back pressure is high and the compressor can not pump against this high head causing the flow to separate and reverse its direction. Surge is a reversal of flow and is a complete breakdown of the continuous steady flow through the whole compressor. It results in mechanical damage to the compressor due to the large fluctuations of flow which results in changes in direction of the thrust forces on the rotor creating damage to the blades and the thrust bearings. The phenomenon of surging should not be confused with the stalling of a compressor stage. Stalling is the breakaway of the flow from the suction side of the blade aerofoil thus causing an aerodynamic stall. A multi-stage compressor may operate stably in the unsurged region with one or more of the stages stalled, and the rest of the stages unstalled.
Compressor Surge
177
Compressor surge is a phenomenon of considerable interest; yet, it is not fully understood. It is a form of unstable operation and should be avoided. It is a phenomenon that unfortunately occurs frequently, sometimes with damaging results. Surge has been traditionally defined as the lower limit of stable operation in a compressor, and it involves the reversal of flow. This reversal of flow occurs because of some kind of aerodynamic instability within the system. Usually, a part of the compressor is the cause of the aerodynamic instability, although it is possible for the system arrangement to be capable of augmenting this instability. Compressors are usually operated at a working line, separated by some safety margin from the surge line. Extensive investigations have been conducted on surge. Poor quantitative universality or aerodynamic loading capacities of different blades and stators, and an inexact knowledge of boundary-layer behavior make the exact prediction of flow in the compressor at the off-design stage difficult. A decrease in the mass flow rate, an increase in the rotational speed of the impeller, or both can cause the compressor to surge. Whether surge is caused by a decrease in flow velocity or an increase in rotational speeds, the blades or the stators can stall. One should note that operating at higher efficiency implies operation closer to surge. It should be noted here that total pressure increases occur only in the rotational part of the compressor, the blades. To make the curve general, the concept of aerodynamic speeds and corrected mass flow rates has been used in the performance maps in this chapter.
2.0 Axial-Flow Compressors The surge line slope on multistage compressors can range from a simple single parabolic relationship to a complex curve containing several break-points or even “notches.” The complexity of the surge line shape depends on whether or not the flow limiting stage changes with operating speed from one compression stage to another; in particular, very closely matched stage combinations frequently exhibit complex surge lines. In the case of compressors with variable inlet guide vanes, the surge line tends to bend more at higher flows than with units which are speed controlled. Usually surge is linked with excessive vibration and an audible sound; yet, there have been cases where surge not accompanied by audible sound has caused failures. Usually, operation in surge and, often, near surge is accompanied by several indications, including general and pulsating noise level increases, axial shaft position changes, discharge temperature excursions, compressor differential pressure fluctuations, and lateral vibration amplitude increases. Frequently, with high pressure compressors, operation in the incipient surge range is accompanied by the emergence of a low frequency, asynchronous vibration signal which can reach predominant amplitudes, as well as excitation of various harmonics of blade passing frequencies. Extended operation in surge causes thrust and journal-bearing failures. Failures of blades and stators are also experienced due to axial movement of the shaft causing contact of blades and stators. Due to the large flow instabilities experienced severe aerodynamic stimulation at one of blade natural response frequencies is caused, leading to blade failure. The performance map, of an axial flow compressors displays the variation of total pressure ratio across a compressor, as a function of corrected mass flow (usually expressed as percent of design value), at a series of constant corrected speed lines (Nc). The axial flow compressor adiabatic efficiency (ηc) is shown as islands on the performance map, and can also be depicted versus corrected mass flow and is shown for a representative multi-stage compressor in figure 26.
PR
Surge Line
Fig. 26. Multi-Stage Axial Compressor Maps
On a given corrected speed line, as the corrected mass flow is reduced, the pressure ratio (usually) increases until it reaches a limiting value on the surge line. For an operating point at or near the surge line the “orderly” flow (i.e. nearly axisymmetric) in the compressor tends to “break” down (flow becomes asymmetric with rotating stall) and can become “violently” unsteady. Thus the surge line is a locus of unstable compressor operating points and is to be avoided. To cope with this, one specifies the surge margin SM defines as:
(27)
In Equation (27) PRsurge/working denotes the pressure ratio on the surge/working line at the same corrected mass flow rate; thus the corrected speed would be higher for operating points on the surge line. For operation on a constant corrected speed line an alternative definition for surge margin in terms of corrected mass flow on the working line and on surge line at the same corrected speed would be preferable. For stable operation of a multi-stage compressor a surge margin is specified. Compressors are designed to operate at a condition referred to as the design point. At the design point the various stages mounted on the same shaft are matched aerodynamically i.e. the inlet flow to each stage is such that the stage is at the design point and this occurs for only one combination of corrected speed and mass flow (for this reason the design point is also known as match point). While the design point is one at which the compressor will operate most of the time, there are situations of low-speed operation during the starting of gas turbines where the compressor must also provide adequate pressure rise and efficiency. For compressor operations
178
Meherwan P. Boyce at corrected speed or at the same corrected speed but corrected mass flow different from those at design, difficulties arise due to the requirements of matching the inlet flow to one stage to the outlet flow from those upstream. As an illustration, consider changes along the constant corrected speed line. The effect of reduction in mass flow relative to the working line results in a higher pressure rise and therefore a greater increase in density in the first stage than was predicted at design. The greater increase in density means the second stage has an even lower value of flow coefficient than the first stage, with an even greater increase in density. The effect is cumulative, so that the last stage approaches stall while the front stage is only slightly altered. Conversely increasing the mass flow relative to the working line would result in a lower pressure rise and therefore a smaller increase in density. The smaller increase in density means the second stage has an even higher value of flow coefficient than the first stage, with an even smaller increase in density. The consequence is that the last stage approaches stalling at negative incidence with low efficiency performance. Similarly one can also show that reducing the rotational speed along the working line through the design point can lead to stalling of front stages and windmilling of rear stages. Methods for coping with low-speed difficulties include use of compressor air bleed at the intermediate stage, use of variable geometry compressor, and use of multi-spool compressors or combinations of the above.
Compressor Choke The compressor choke point is when the flow in the compressor reaches Mach 1 at the blade throat, a point where no more flow can pass through the compressor. This phenomenon is often known in the industry as “Stone Walling.” The more stages, the higher the pressure ratio, and the smaller the operational margin between surge and choke regions of the compressor as shown in figure 27.
Fig. 27. A High Pressure Multistage Axial Flow compressor map
Compressor Stall There are three distinct stall phenomena. Rotating stall and individual blade stall are aerodynamic phenomena; stall flutter is an aero elastic phenomenon.
Individual Blade Stall This type of stall occurs when all the blades around the compressor annulus stall simultaneously without the occurrence of a stall propagation mechanism. The circumstances under which individual blade stall is established are unknown at present. It appears that the stalling of a blade row generally manifests itself in some type of propagating stall and that individual blade stall is an exception.
Rotating Stall
179
Rotating, or propagating stall, was first observed by Whittle and his team on the inducer vanes of a centrifugal compressor. Rotating stall (propagating stall) consists of large stall zones covering several blade passages and propagates in the direction of the rotation and at some fraction of rotor speed. The number of stall zones and the propagating rates vary considerably. Rotating stall is the most prevalent type of stall phenomenon. The propagation mechanism can be described by considering the blade row to be a cascade of blades as shown in figure 28. A flow perturbation causes blade 2 to reach a stalled condition before the other blades. This stalled blade does not produce a sufficient pressure rise to maintain the flow around it, and an effective flow blockage or a zone of reduced flow develops. This retarded flow diverts the flow around it so that the angle of attack increases on blade 3 and decreases on blade 1. In this way a stall ‘cell’ may move
2.0 Axial-Flow Compressors along the cascade in the direction of the lift on the blades. The stall propagates downward relative to the blade row at a rate about half the rotational speed; the diverted flow stalls the blades below the retarded-flow zone and unstalls the blades above it. The retarded flow or stall zone moves from the pressure side to the suction side of each blade in the opposite direction of rotor rotation. The stall zone may cover several blade passages. The relative speed of propagation has been observed from compressor tests to be less than the rotor speed. Observed from an absolute frame of reference, the stall zones appear to be moving in the direction of rotor rotation. The radial extent of the stall zone may vary from just the tip to the whole blade length. Table 2 shows the characteristics of rotating stall for single and multistage axial-flow compressors.
Fig. 28. Propagating Stall in a Blade Cascade Table 2 Summary of Rotating Stall Data
Type of Velocity Diagram
Hub-tip Radius Ratio
Symmetrical
0.50
0.90 0.80 0.76
Free vortex Solid body Vortex transonic
0.72 0.60
0.60 0.50 0.50 0.40
Single-Stage Compressors Weight-flow Fluctuation Propagation during stall, Number of Rate, Stall Stall Zones Speed, abs/ Rotor Speed 3 4 5 1 8 1 7 8 5 3 4 3 2 6, 8 1 2 1 1 1 3 2 1 2
0.420 0.475 0.523 0.305 0.87 0.36 0.25 0.25 0.25 0.23 0.48 0.48 0.49 0.245 0.48 0.36 0.10 0.45 0.12 0.816 0.634 0.565
1.39 2.14 1.66 1.2 0.76 1.30 2.14 1.10 1.10 2.02 1.47 2.02 1.71 0.71=1.33 0.60 0.60 0.68 0.60 0.65
Radial Extent of Stall Zone
Type of Stall
Partial
Progressive
Total Partial Total Partial
Abrupt Progressive Abrupt Progressive
Total
Total Partial Partial Total Partial Total Partial Total Total Partial
Progressive Progressive Progressive Abrupt Progressive Abrupt Progressive Progressive Abrupt Progressive
180
Meherwan P. Boyce
Progressive Stall indicates the gradual increase in blocked annulus area due to stall. Abrupt Stall is a single stall zone covering as much as half the annulus area and extending over the entire blade span with discontinuity in the pressure curve. Complete Compressor Stall is applied to multistage compressors to describe a discontinuous performance curve similar to that for abrupt stall, and these points define the stall-limit line. 1 2
Stall Flutter This phenomenon is caused by self-excitation of the blade and is an aero-elastic phenomenon. It must be distinguished from classic flutter, since classic flutter is a coupled torsional-flexural vibration that occurs when the free-stream velocity over a wing or airfoil section reaches a certain critical velocity. Stall flutter, on the other hand, is a phenomenon that occurs due to the stalling of the flow around a blade. Blade stall causes Karman vortices in the airfoil wake. Whenever the frequency of these vortices coincides with the natural frequency of the airfoil, flutter will occur. Stall flutter is a major cause of compressor blade failure. Several types of flutter have been identified and these are indicated as various flutter boundaries on the operating map of a c and N c , additional non-dimensional parameters have to be introduced to high-speed (transonic) compressor in figure 29. Besides m adequately characterize the flutter boundaries. One such parameter is the reduced frequency which is given by the ratio of blade chord to the wavelength of the unsteady disturbance induced by the blade motion. Often the inverse of reduced frequency, the reduced velocity is used instead. More recently Khalak (2002) proposed and developed a framework for flutter operability assessment in which a set of four non-dimensional parameters is used to characterize the flutter boundary. These parameters are the corrected mass flow, the corrected
speed, the compressible reduced frequency (where c denotes blade chord length, ω0 the modal frequency) and the combined mass-damping parameter (ratio of mechanical damping to blade mass). In analogy with the surge margin, a flutter margin FM is specified in equation (28): (28)
181
PRflutter is the pressure ratio on the flutter boundary at the same corrected mass flow corresponding to that for PRworking on the working line. For operation on a constant corrected speed line, it would be preferable to define flutter margin in terms of corrected mass flow on the working line and on the flutter boundary at the same corrected speed.
2.0 Axial-Flow Compressors
Fig. 29. Flutter regions on the operating map of a transonic compressor (after Mikolajczak, et al., 1975)
An example of a typical failure due to flutter in an axial flow compressor fifth stage is discussed in this section. There were three blade failures of the fifth stage blade all within 3-10 hours of operation. The cause of the failure had to be determined. A dynamic pressure transducer with a voltage output was used to obtain the frequency spectra. In the first four stages of the compressor no outstanding vibration amplitudes were recorded. A signal was noted at 48N (N being the running speed), but the amplitude was not high, and it did not fluctuate. A measurement at the low-pressure bleed chamber taken from the fourth stage showed similar characteristics. The compressor high-pressure bleed chamber occurs after the eighth stage. A measurement at this chamber showed a high, fluctuating 48N signal. As there are 48 blades on the fifth-stage wheel, a problem in the fifth-stage was suspected. However, above the fifth-stage are blade rows of 86N (2 x 48N), so further analysis was needed. It was found that the measurement at the high-pressure bleed chamber showed only a very small 86N amplitude compared to the high amplitude of the 48N frequency. Since blade rows of 86 blades were closer to the high-pressure bleed chamber, the expected high signal should have been 86N compared to 48N under normal operating conditions. This high amplitude of 48N indicated that it was the fifth-stage which caused the high, fluctuating signal; thus, a stall condition in that section was probable. Figures 30, 31, 32, and 33 show the spectrum at speeds of 4,100; 5,400; 8,000; and 9,400 rpm. At 9,400 rpm, the second and third harmonics of 48N were also very predominant.
Fig. 30. High-Pressure Bleed Chamber - 4,100 rpm
Fig. 31. High-Pressure Bleed Chamber - 5,400 rpm
182
Meherwan P. Boyce
Fig. 32. High-Pressure Bleed Chamber - 8,000 rpm
Fig. 33. High-Pressure Bleed Chamber - 9,400 rpm
Next, the fifth-stage pressure was measured. Once again, high amplitude at 48N was found. However, a predominant reading was also observed at 1,200 Hz frequencies. Figures 34 and 35 shows the largest amplitudes at speeds of 5,800 and 6,800 rpm, respectively.
Fig. 34. Fifth Stage Bleed Pressure - 5800 rpm
Fig. 35.Fifth-Stage Bleed Pressure - 6,800 rpm
At the compressor exit, predominate frequencies of 48N existed up to speeds of 6,800 rpm. At 8,400 rpm, the 48N and 86N frequencies were of about equal magnitudes - the only signal where the 48N and 86N frequencies were the same. The pressure was measured from a static port in the chamber. All other pressures were measured from the shroud, thus indicating the phenomena occurred at the blade tip. Since the problem was isolated to the fifth stage, the conclusion was that the stall occurred at the fifth-stage rotor tip. The solution to the problem was the redesign of the fifth stage blade with a modified angle so that it would not be as subject to stall flutter.
2.0-10 Compressor Performance Parameters For a gas compressor, the functional dependence of compressor exit total/stagnation pressure Ptexit and the adiabatic compressor efficiency ηc can be expressed as follows:
, P , T , N, ν, R, γ, design D) . (Ptexit, ηc) = F( m tin tin
183
(29)
The gas properties of relevance to the compression process are characterized by the kinematic viscosity ν, specific heat ratio γ, and the gas constant R. The geometry dependence of the machine is set by the design and its characteristic size D such as the tip diameter of compressor. Use of dimensional analysis reduces the complexity of Equation (29) (noting that γ and design D, can be regarded as non-dimensional) to yield
2.0 Axial-Flow Compressors
(30)
For a given compressor and for inlet conditions for which γ does not vary, Equation (30) reduces to (31)
At high enough Reynolds number (> 3 x 105), changes in this number have little effect on compressor performance so that
(
Ptexit ,ηc Ptin
)
m Ttin N , Ptin T tin i.e. can be correlated in terms of
m Ttin Ptexit N ,ηc = F , Ptin Ptin Ttin
(Section. 32a)
(32a)
As no functional dependence is implied if the non-dimensional variables on the right hand side is scaled by a constant, one can thus choose to replace them by the corrected mass flow rate
m θ m c = δ
and corrected speed N c = N so that θ
m θ N Ptexit = F(m c , N c ). , η c = F , Ptin δ θ
In equation (32b),
θ=
(Section. 32b)
(32b)
Ptin Ttin and δ = where the reference temperature Tref and the reference pressure Pref are taken to be the Pref Tref
sea-level value for the standard atmosphere, 59.6ºF (15°C) and 14.7 psia (101 KN/m2) respectively. The advantage of using these corrected variables is that their numerical magnitude is similar to the actual value so that its significance is not obscured. We can also use the Euler Turbine Equation (8) for a compressor stage
c p (Ttexit − Ttin ) = ω [(rVθ )2 − (rV θ )1 ]
(Section. 33)
(33)
to elucidate the functional dependence and to deduce why the performance characteristics look the way they are on a compressor map. Assuming isentropic flow (i.e. no loss) then the stagnation pressure ratio across the (ideal) stage is given by
(34)
In equation (33) and (34) subscript 1 and 2 refer to variable evaluated at rotor inlet and rotor exit respectively, Vθ denotes tangential velocity, V the axial velocity, ω the angular velocity of rotor, αexit the absolute flow angle at stator exit, βexit the relative flow angle at rotor exit, and r the radius. Upon introducing the corrected variables into equation (34) we have
P R
s
{
γ
}
= 1 + k 0 N c2 − k1 N c m c G (M 1 )(tan α exit + tan β exit ) γ −1
(Section. 35)
(35)
184
Meherwan P. Boyce where G(M1) has a weak dependence on the incoming Mach number M1, k0 ∝ r2 and k1 ∝ r. For a given compressor stage (tan α exit + tan β exit ) is fixed and neglecting the variation in G(M1) we have P R s =P R s (m c , N c ) . The general dependence of PRs c and N c is shown in figure 39 as a series of dashed lines of constant corrected speed for the ideal stage; equation (35) can be on m c and N c . The solid lines (of constant corrected speed) used to obtain the trend in the variation of the ideal stage characteristic with m P R v s m in figure 36 are the s c curves with stagnation pressure losses taken into account. Flow angle varies as corrected mass flow rate changes along a given corrected speed line. The point of minimum difference between the dash (ideal) and the solid (actual) curve corresponds to a corrected mass flow that yields an angle of incidence for minimum loss; moving away from this point along a constant c or decreasing the corrected speed line amounts to changing the incidence angle (increasing the angle of incidence for decreasing m m angle of incidence for increasing c ) so as to lead to higher loss. This is reflected in the increasing difference between the two curves (ideal versus actual) at corrected mass flow other than that corresponding to minimum loss. One thus deduces from the above arguments c and N c . The pressure ratio of a complete that the actual pressure rise (and the efficiency) can also be characterized in terms of m compressor consisting of many stages can be obtained by taking the products of the stage performance.
Fig. 36. Performance map of compressor stage
2.0-11 Performance Losses in an Axial-Flow Compressor The calculation of the performance of an axial-flow compressor at both design and off-design conditions requires the knowledge of the various types of losses encountered in an axial-flow compressor. The accurate calculation and proper evaluation of the losses within the axial-flow compressor are as important as the calculation of the blade-loading parameter, since unless the proper parameters are controlled, the efficiency drops. The evaluation of the various losses is a combination of experimental results and theory. The losses are divided into two groups: (1) losses encountered in the rotor, and (2) losses encountered into the stator. The losses are usually expressed as a loss of heat and enthalpy. A convenient way to express the losses is in a nondimensional manner with reference to the blade speed. The theoretical total head available (qtot) is equal to the head available from the energy equation (qth = qtot) plus the head which is loss from disc friction.
qtot = qth + qdf
(36)
The adiabatic head that is actually available at the rotor discharge is equal to the theoretical head minus the heat losses from the shock in the rotor, the incidence loss, the blade loadings and profile losses, the clearance between the rotor and the shroud, and the secondary losses encountered in the flow passage
qia = qth − qin − qsh − qbl − qc − qsf .
(37)
Therefore, the adiabatic efficiency in the impeller is
η imp = 185
qia qtot
.
(38)
2.0 Axial-Flow Compressors The calculation of the overall stage efficiency must also include the losses encountered in the stator. Thus, the overall actual adiabatic head attained would be the actual adiabatic head of the impeller minus the head losses encountered in the stator from wake caused by the impeller blade, the loss of part of the kinetic head at the exit of the stator, and the loss of head from the frictional forces encountered in the stator
qoa = qia − qw − qex − qosf .
(39)
Therefore, the adiabatic efficiency in the stage
ηstage =
qoa . qtot
(40)
The losses as mentioned earlier can be further described: 1. 2. 3. 4. 5. 6. 7. 8.
Disc friction loss. This loss is from skin friction on the discs that house the blades of the compressors. This loss varies with different types of discs. Incidence loss. This loss is caused by the angle of the air and the blade angle not being coincident. The loss is at a minimum to about an angle of ± 4o, after which the loss increases rapidly. Blade loading and profile loss. This loss is due to the negative velocity gradients in the boundary layer, which gives rise to flow separation. Skin friction loss. This loss is from skin friction on the blade surfaces and on the annular walls. Clearance loss. This loss is due to the clearance between the blade tips and the casing. Wake loss. This loss is from the wake produced at the exit of the rotary. Stator profile and skin friction loss. This loss is from skin friction and the attack angle of the flow entering the stator. Exit loss. This loss is due to the kinetic energy head leaving the stator.
Figure 37 shows the various losses as a function of flow. Note that the compressor is more efficient as the flow nears surge conditions.
Fig. 37. Losses in an axial-flow compressor stage
186
Meherwan P. Boyce 2.0-12 New Developments in Axial Flow Compressors The new advanced compressor rotors have fewer blades with higher loadings, and the blades are thinner, larger, and are designed using advanced radial equilibrium theory, which create Three Dimensional and Controlled Diffusion shaped airfoils (3D/CDA), with smaller clearances and higher loading per stage. There are also trends towards water injection at the inlet or between compressor sections which will likely affect airfoil erosion life. The smaller clearances (20-50 mils) and high pressure ratios tend to increase the probability of encountering rubs. These tip rubs usually occur near the bleed flow sections of the turbines where there are inner diameter changes and the compressor casing could be out of round. Figure 38 shows one such blade that encountered tip rub.
Fig. 38. Compressor Blade with Tip Rub
The advanced compressor blades also usually have squealer sections on the blade tips, which are designed to wear in a safe manner if the blades are in contact with the casing. Figure 39 is one such blade. These rubs, if severe can lead to tip fractures and overall destruction of the downstream blades and diffuser vanes due to domestic object damage (DOD).
Fig. 39. Axial Flow Compressor rotor blade with squealer tip
The very high temperature at the exit of the compressor, which in some cases exceeds a 1000ºF, causes a very hot compression section, which also requires the cooling of the bleed flows before they can be used for cooling the turbine section. This requires large heat exchangers and in some combined cycle plants steam is used to cool the compressed air. This also limits the down time between startups of the turbines. Design margins are set by Finite Element Modeling (FEM) at the element level which results in lower safety margins than previous designs. The costs of these larger, thinner, less-rub tolerant, and more twisted-shape airfoils are usually higher. When several of the major characteristics of advanced gas turbines are examined from a risk viewpoint (i.e., probability and consequences of failure), there are no characteristics which reduce the probability of failure and/or decrease the consequence of failure.
187
2.0 Axial-Flow Compressors Table 3 indicates the changes in the compressor blades that are now prevalent on the advanced gas turbines. The first column represents previous gas turbine designs, the second column represents new gas turbine designs, and the last column indicates the change in risk ( represents higher) for the design differences. Most of the comparisons are self-explanatory. Table 3 State of Gas Turbine Technology Compressors
New Designs
Previous Designs
Risk
•
2D double circular arc or NACA 65 profiles
•
3D or Controlled Diffusion Airfoil (CDA) profiles
•
Large number of airfoils
•
Reduced airfoil count
•
Repeating stages/shorter chords
•
Stages unique/longer chords
• •
Low/ modest Aspect ratios Large clearances
• •
High Aspect ratios Smaller clearances
•
Low/modest pressure ratios (Rc)
•
Much higher pressure ratios (Rc)
•
Low/modest blade loading per stage
•
High blade loading per stage
• • • • •
Wider Operating margin Thicker leading edges Dry operation Bulk safety margins Lower costs
• • • • •
Narrow operating margin Thinner leading edges Wet operation Safety margins by FEM Higher costs
2.0-13 Recent Advances and Research Requirements There is considerable research is being carried out on improving the performance of axial flow compressor. This research is being carried out in many different aspects of the axial flow compressor: 1.
Effects of Aspect Ratio on blade loading, blade excitation, and the pre-twist blade angles (centrifugal forces on the blade). Increase in blade loading was carried out by increasing the Aspect ratio of the blade. Blade aspect ratios were increased to AR = 9. At these high aspect ratios the blades had to be designed with mid span shrouds, and tip shrouds. This decreases the efficiency of the stage; however, without the shrouds the pre-twist blade angle had to be increased to about 12º, and the blade excitation resulted in blade failure. Presently most blade designs are limited to an AR=4.
2.
Increasing the operational range (surge – choke) at a given compressor speed, by developing new blade profiles to reduce blade stall in compressors Cascade Tests The data on blades in an axial-flow compressor are from various types of cascades, since theoretical solutions are very complex, and their accuracy is in question because of the many assumptions required to solve the equations. The most thorough and systematic cascade testing has been conducted by NACA staff at the Lewis Research Center. The bulk of the cascade testing was carried out at low mach numbers and at low turbulence levels. The NACA 65 blade profiles were tested in a systematic manner by Herrig, Emergy, and Erwin. The cascade tests were carried out in a cascade wind tunnel with boundary-layer suction at the end walls. Tip effects were studied in a specially designed water cascade tunnel with relative motion between wall and blades. Cascade tests are useful in determining all aspects of secondary flow. For better visualization, tests have been conducted in water cascades. The flow patterns are studied by injecting globules of dibutyl phatalate and kerosene in a mixture equal to the density of water. The mixture is useful in tracing secondary flow, since it does not coagulate. An impeller designed for air can be tested using water if the dimensionless parameters, Reynolds number (Re), and specific speed (Ns) are held constant
Re =
ρ airVair D ρ waterVwater D = µ air µ water
(41)
188
Meherwan P. Boyce Ns = where:
ρ V D
Qair Qwater = 3 N air D N water D 3
= medium density = velocity = impeller diameter
(42)
µ = viscosity N = speed
Using this assumption, one can apply this flow visualization method to any working medium. One designed apparatus consists of two large tanks on two different levels. The lower tank is constructed entirely out of Plexiglas and receives a constant flow from the upper tank. The flow entering the lower tank comes through a large, rectangular opening which houses a number of screens so that no turbulence is created by water entering the lower tank. The center of the lower tank can be fitted with various boxes for the various flow visualization problems to be studied. This modular design enables a rapid interchanging of models and work on more than one concept at a time. Blade Profile To study the effect of laminar flow, the blades were slotted as shown in figure 40. For the blade treatment cascade rig experiment, a Plexiglas cascade was designed and built. Figure 41 shows the cascade. This cascade was then placed in the bottom tank and maintained at a constant head. Figure 42 shows the entire setup, and figure 43 shows the cascade flow. Note the large extent of the laminar-flow regions on the treated center blades as compared to the untreated blades.
Fig. 40. Perspective of compressor blade with treatment
189
Fig. 42. Apparatus for testing axial-flow cascade model
Fig. 41. Cascade model in axial-flow test tank
Fig. 43. Treatments on center cascade blade
2.0 Axial-Flow Compressors 3.
Reduction of flow leakage at the compressor tips The effect of casing treatment in axial-flow compressors was studied in a water cascade tunnel. In this study the same Reynold number and specific speeds were maintained as those experienced in an actual axial-flow compressor. In an actual compressor the blade and the passage are rotating with respect to the stationary shroud. It would be difficult for a stationary observer to obtain data on the rotating blade passage. However, if that observer were rotating with the blade passage, data would be easier to acquire. This was accomplished by holding the blade passage stationary with respect to the observer and rotating the shroud. Furthermore, since casing treatment affects the region around the blade tip, it was sufficient to study only the upper portion of the blade passage. These were the criteria in the design of the apparatus. The modeling of the blade passage required provisions for controlling the flow in and out of the passage. This control was accomplished by placing the blades, which partially form the blade passage, within a Plexiglas tube. The tube had to be of sufficient diameter to accommodate the required flow through the passage without tube wall effect distorting the flow as it entered or left the blade passage. This allowance was accomplished by using a tube three times the diameter of the blade pitch. The entrance to the blades was designed so that the flow entering the blades was a fully developed turbulent flow. The flow in the passage between the blade tip and the rotating shroud was laminar. This laminar flow was expected in the narrow passage. A number of blade shapes could have been chosen; therefore, it was necessary to pick one shape for this study, which would be the most representative for casing treatment considerations. Since casing treatment is most effective from an acoustic standpoint in the initial stages of compression, the maximum point of camber was chosen toward the rear of the blade (Z = .6 chord). This type of blade profile is most commonly used for transonic flow and is usually in the initial stages of compression. The rotating shroud must be in close proximity to the blade tips within the tube. To get this proximity, a shaft-mounted Plexiglas disc was suspended from above the blades. The Plexiglas disc was machined as shown in figure 44. The Plexiglas tube was slotted so that the disc could be centered on the centerline of the tube and its stepped section lowered through the two slots in the tube. Clearances between the slot edges and the disc were minimized. One slot was cut directly above the blade passage emplacement. The other slot was sealed off to prevent leakage. As the disc was lowered into close proximity to the blade tips, the blade passage was completed. The clearance between disc and blade was kept at 0.035 of an inch. The disc, when spun from above, acted as the rotating shroud.
Fig. 44. Details of the various casing treatments. (Each treatment was on a separate disc)
There are only two basic casing treatment designs other than a blank design - which corresponds to no casing treatment at all. The first type of casing treatment consists of radial grooves. A radial groove is a casing treatment design in which the groove is essentially parallel to the chordline of the blade. The second basic type is the circumferential groove. This type of casing treatment has its grooves perpendicular to the blade chordline. Figure 45 is a photograph of two discs showing the two types of casing treatment used. The third disc used is a blank, representing the present type of casing. The results indicate that the radial casing treatment is most effective in reducing leakage and also in increasing the surge-to-stall margin. Figure 46 shows the leakage at the tips for the various casing treatments. Figure 47 shows the velocity patterns observed by the use of various casing treatments. Note that for the treatment along the chord (radial), the flow is maximized at the tip. This flow maximum at the tip indicates that the chance of rotor tip stall is greatly reduced.
190
Meherwan P. Boyce
Fig. 45. Two discs with casing treatment
Fig. 46. Mass flow leakage at tips for various casing treatments
Fig. 47. Velocity patterns observed in the side view of the blade passage for various casing treatments
4.
Enhancement of Numerical Solutions of the Navier-Stokes Equation (viscous compressible flow) The solution of the full Navier Stokes equation requires much enhanced numerical techniques. The old solutions used inviscous flow and quasi three dimensional flow solutions. There are many new enhanced numerical programs underway to solve the equation in its entirety.
191
5.
Supersonic Blade Profiles for higher pressure ratio per stage (>2.1)
2.0 Axial-Flow Compressors Transonic blades have been designed with the point of maximum thickness at about 0.6 of blade chord from the leading edge of the blade. Supersonic blade design has problems with standing shock waves which can occur as the flow enters the stators. The losses with the diffusion process is very high and thus design changes are being tested so that the flow entering the diffuser is easily swallowed, and that if any shock waves exist they are oblique shocks with minimal losses. Cascade testing is being conducted on various profiles to ensure that the stage losses are minimized. 6.
Compressor interstage cooling by water injection between stages In this system the water is injected into the mid-stages of the compressor to cool the air and approach an isothermal compression process as shown in figure 48. The water injected is usually mechanically atomized so that very fine droplets are entered into the air. The water is evaporated as it comes in contact with the high pressure and temperature air stream. As water evaporates, it consumes about 1058 BTU (1117 kJ) (latent heat of vaporization) at the higher pressure and temperature resulting in lowering the temperature of the air stream entering the next stage. This lowers the work required to drive the compressor. The intercooling of the compressed air has been very successful when applied to high-pressure ratio engines.
Fig. 48. Mid-Compressor Cooling showing a schematic as well as an actual application in a GE LM 6000 Engine (courtesy GE Power Systems)
2.0-14 Compressor Blade Material Compressor blading is made by forging, extrusion or machining. All production blades, until the advent of he new Advanced Gas Turbines, have been made from stainless steels, Type 403 or 403 Cb both having about 12 Cr. This family of alloys has properties which include good ductility at high strength levels, uniform properties, and good strength at temperatures up to about 900ºF (482ºC). Because of the new axial flow compressors which have pressure ratio of 30:1 to 40:1, and exit temperatures between 1000ºF – 1150ºF (538ºC - 621ºC), new compressor blade material, a precipitation hardened, martensitic stainless steel such as 15-5 PH nominal, was introduced into production for advanced and uprated machines, as shown in Table 4. This material provides increased tensile strength without sacrificing stress corrosion resistance. Substantial increases in the high-cycle fatigue and corrosion fatigue strength are also achieved with this material, compared with the Type 403 stainless steel with 12Cr. Superior corrosion resistance is also achieved due to the metal’s higher concentration of chromium and molybdenum content. Compressor corrosion results from moisture containing salts and acids collecting on the blading. During operation, moisture can be present because of rain, use of evaporative coolers, fogging systems, or compressor water washes, or condensation resulting from humid air being accelerated at the compressor inlet. Moisture may be present in the compressor during operation up to between stage 5 and stage 8, where it usually becomes warm enough to prevent condensation. When the turbine is not in operation, the compressor can still become wet if metal temperatures are below the local dew point. This can happen to units stored in humid environments. The chemistry of this moisture deposit, especially the salt in the air depositing on the blading, determines the severity of the corrosion phenomenon.
192
Meherwan P. Boyce Table 4 Compressor Blade Material
Compressor Blade Type
Max. Temp
COMPONENTS PERCENT C
S
Mn
P
Si
Cr
Mo
Ni
Cu
Al
Cb
O
Fe
-
-
Bal
AISI 403
900ºF
.11
AISI 403+Cb
900ºF
.15
-
-
-
-
12
-
-
-
-
0.2
-
-
Bal
Martensitic High Temperature Stainless Steel
1250ºF
.08
-
.14
-
.4
15.6
.08
3.86.5
2.9
.9
-
-
-
Bal
<.07
<.03
<1.0
<0.04-
<1.0
1415.5
-
3.55.5
3.2
.9
.15.45
-
-
Bal
15-5 PH, nominal
12
Mg
The high temperature blade alloy is normally produced by vacuum-arc remelting to reduce inclusions, and is advertised to have a balanced chemistry that minimizes the formation of delta-ferrite. Inclusions and the delta-ferrite would provide planes of weakness in that part. It is not uncommon for the mill to supply forging stock that has first been given a 1900ºF heat treatment, just for better forgeability. The forged blanks are then usually reheat-treated at 1900ºF, followed by hardeningtreatments between 1100ºF and 1150ºF depending on the properties sought. There is a general correlation between hardness and strength (tensile/fatigue). A hardness of RC 32 suggests that the tensile strength is around 150000 psi and that the hardening temperature used during manufacture was somewhere around 1100ºF to 1150ºF. Coating of the compressor blades is now very common. Compressor blades suffer a great amount of corrosion pitting from impurities in the air stream. This corrosion pitting has led to blade failures. Compressor blades in many cases have over 100,000 hours but due to pitting can be reduced considerably to between 20,000 – 60,000 hours. It has been a very common practice for over 30 years to coat at least the first 5-8 stages depending on the compressor design. The first stages are considered to be the “wet stages” because many units now use on line water washes, as well as have evaporative cooling and fogging for power augmentation. Coating for these blades is usually consistent of a duplex type coating, which must be at least 3 mils in thickness. This coating as most typical coating has a sacrificial undercoat coating which is placed on the base metal and is covered by a ceramic coating. Ni-Cd coating is also used in selected applications, and new coatings consisting of an aluminum slurry coating which has a protective ceramic top layer that provides improved erosion resistance are also being introduced. This type of coating, as compared to conventional aluminum slurry coatings, is better in corrosion protection and substantially better in erosion resistance. This type of coating also improves the performance of the gas turbine by reducing the amount of power consumed by the compressor. Tests conducted show a reduction of 2%-3% in the power consumed by the compressor which pays back additional cost of coating in 4-6 months of operation. The aspect ratio of axial flow compressors including the IGV vary from about AR = 4, to an aspect ratio of about AR = 0.5. All IGV’s and the first five to eight stages of rotating and stationary airfoils in the compressor are made from Martensitic High Temperature Stainless Steel; or 15-5 PH nominal blade material, the next stages are usually coated AISI 403 or 403 Cb.
2.0-15 Acknowledgements This chapter has been taken liberally from the author’s book Gas Turbine Engineering Handbook. The author would like to express his sincere thanks to Dr. Choon Sooi Tan and Dr. Yifang Gong for their contributions to the sections on Stall Flutter, and Compressor Performance Parameters. Dr. Tan is a senior research engineer, and Dr. Gong is a research engineer at the MIT Gas Turbine Laboratory. Dr. Tan is a leading authority on unsteady and three dimensional flow in multistage Turbomachinery and is an author of 38 publications and a co-author of the book entitled, Internal Flow: Concepts and Applications, Cambridge University Press, 2004. Dr Gong is an authority on compressor aerodynamics and instability in compressor/compression systems; he is presently working on the design and development of a gas turbine power plant using supercritical CO2 as the working fluid.
193
2.0 Axial-Flow Compressors 2.0-16 Bibliography _________________________________ 1. Boyce, M.P,, Gas Turbine Engineering Handbook, Second Edition, Butterworth-Hienemann 2003 2. Herrig, L.J., Emery, J.C., and Erwin, J.R., “Systematic Two Dimensional Cascade Tests of NACA 65 Series Compressor Blades at Low Speed,” NACA R.M. E 55H11 (1955). 3. Boyce, M.P., “Fluid Flow Phenomena in Dusty Air,” (Thesis), University of Oklahoma Graduate College, 1969, p. 18. 4. Boyce M.P., Schiller, R.N., and Desai, A.R., “Study of Casing Treatment Effects in Axial-Flow Compressors,” ASME Paper No. 74-GT-89. 5. Boyce, M.P., “Secondary Flows in Axial-Flow Compressors with Treated Blades,” AGARD-CCP-214 pp. 5-1 to 5-13. 6. Giamati, C.C., and Finger, H.B., “Design Velocity Distribution in Meridional Plane,” NASA SP 36, Chapter VIII (1965), p. 255. 7. Hatch, J.E., Giamatic, C.C., and Jackson, R.J. “Application of Radial Equilibrium Condition to Axial-Flow Turbomachine Design Including Consideration of Change of Enthropy with Radius Downstream of Blade Row,” NACA RM E54A20 (1954) 8. Holmquist, L.O., and Rannie, W.D., “An Approximate Method of Calculating Three-Dimensional Flow in Axial Turbomachines” (Paper) Meeting Inst. Aero. Sci., New York, January 24-28, 1955. 9. Lieblein, S., Schwenk, F.C., and Broderick, R.L., “Diffusion Factor for Estimating Losses and Limiting Blade Loading in Axial-Flow Compressor Blade Elements,” NACA RM #53001 (1953). 10. Stewart, W.L., “Investigation of Compressible Flow Mixing Losses Obtained Downstream of a Blade Row,” NACA RM E54120 (1954). 11. Boyce, M.P., “Transonic Axial-Flow Compressor.” ASME Paper No. 67-GT-47. 12. Carter, A.D.S., “The Low-Speed Performance of Related Aerofoils in Cascade,” Rep. R.55, British NGTE, September, 1949. 13. Mellor, G., “The aerodynamic Performance of Axial Compressor Cascades with Applications to Machine Design,” (Sc. D. Thesis), M.I.T. Gas Turbine Lab, M.I.T. Rep. No. 38 (1957). 14. Graham, R.W. and Guentert, E.C., “Compressor Stall and Blade Vibration,” NASA SP 365, (1956) Chapter XI, p.311. 15. Cumpsty, N. A., 1989, Compressor Aerodynamics, Longman Group UK Ltd., London, England. 16. Cumpsty, N. A., 1998, Jet Propulsion, Cambridge University Press, Cambridge, England. 17. Hill, P. G., Peterson, C. R., 1992, Mechanics and Thermodynamics of Propulsion, Second Edition, Addison-Wesley Publishing Company, Reading MA. 18. Kerrebrock, J. L., 1992, Aircraft Engines and Gas Turbines, MIT Press, Cambridge, MA. 19. Khalak, A., 2002, “A Famework for Futter Clearance of Aeroengine Blades”, Journal of Engineering for Gas Turbine and Power, Vol 124, No. 4. Also ASME 2001-GT-0270, ASME Turbo Expo 2001, New Orleans, LA, 2001. 20. Mikolajczak, A. A., Arnoldi, R. A., Snyder, L.E., Stargardter, 1975, “Advances in Fan and Compressor Blade Flutter Analysis and Prediction,” Journal of Aircraft 12. 21. Caltech Lecture Notes on Jet Propulsion JP121 Graduate Course (Instructor: Zukoski E. E.)
194
BIOGRAPHY
2.0 Axial-Flow Compressors
Meherwan P. Boyce 2121 Kirby Drive, Number 28N Houston, TX 77019 phone: (713) 807--0888 fax: (713) 807-0088 email:
[email protected] boycepower.com Dr. Meherwan P. Boyce, P.E., Fellow ASME & IDGTE; has over 42 years of experience in the field of TurboMachinery in both industry and academia. His industrial experience covers 20 years as Chairman and CEO of Boyce Engineering International, and five years as a designer of compressors and turbines for gas turbines for various gas turbine manufacturers. His academic experience covers a 15 year period, which includes the position of Professor of Mechanical Engineering at Texas A&M University and Founder of the TurboMachinery Laboratories and The TurboMachinery Symposium, which is now in its thirtieth year. He is the author of several books such as the Gas Turbine Engineering Handbook (Butterworth & Heinemann), Cogeneration & Combined Cycle Power Plants (ASME Press), and Centrifugal Compressors, A Basic Guide (PennWell Books). He is a contributor to several Handbooks; his latest contribution is to the Perry’s Chemical Engineering Handbook Seventh Edition (McGraw Hill) in the areas of Transport and Storage of Fluids, and Gas Turbines. Dr. Boyce has taught over 100 short courses around the world attended by over 3000 students representing over 400 Companies. He is a Consultant to the Aerospace, Petrochemical and Utility Industries globally, and is a much-requested speaker at Universities and Conferences throughout the world. Dr. Boyce is Chairman of the Plant Engineering & Maintenance Division of ASME, and Chairman of the Electric Utilities Committee of the of ASME’s International Gas Turbine Institute. He is also a Chairman of the ASME Conferences Committee. In 2002 Dr Boyce was chairman of two major conferences the Advanced Gas Turbine and Condition Monitoring Conference sponsored by DOE and EPRI, and the Gas Turbine Users Associations Conference. Dr. Boyce has authored more than 100 technical papers and reports on Gas Turbines, Compressors Pumps, Fluid mechanics, and TurboMachinery. He is a Fellow of the ASME (USA) and the Institution of Diesel and Gas Turbine Engineers (UK), and member of SAE, NSPE, and several other professional and honorary societies such as Sigma Xi, Pi Tau Sigma, Phi Kappa Phi, and Tau Beta Phi. He is the recipient of the ASME award for Excellence in Aerodynamics and the Ralph Teetor Award of SAE for enhancement in Research and Teaching He is also a Registered Professional Engineer in the State of Texas. Dr. Boyce received a B.S. and M.S. in Mechanical Engineering from the South Dakota School of Mines and Technology and the State University of New York, respectively, and Ph.D. in Aerospace & Mechanical Engineering in 1969 from the University of Oklahoma.
3.1.1
Static and Dynamic Combustion stability
3.1.1-1 Introduction The objective of this article is to provide the reader with some background on blowoff and combustion instability, often referred to as a combustor’s “static stability” and “dynamic stability”. In particular, this chapter will focus upon this phenomenon in lean, premixed combustion systems operating with any of a variety of fuels, such as natural gas or syntheticgas. Blowoff refers to the flame physically leaving the combustor and “blowing out” of the combustor. This issue is often referred to as “static stability”. Blowoff occurs when the flame cannot be anchored in the combustor. Combustion instability, or “dynamic instabilities” refer to damaging oscillations driven by fluctuations in the combustion heat release rate. These oscillations cause wear and damage to combustor components and, in extreme cases, can cause liberation of pieces into the hot gas path and resulting damaging to downstream turbine components.
3.1.1-2 Static Stability
Timothy C. Lieuwen Associate Professor School of Aerospace Engineering Georgia Institute of Technology Atlanta, GA 30332-0150 email:
[email protected]
197
As the propagation speed of essentially all flames is substantially lower than flow velocities in realistic systems, special flame stabilization systems are necessary to anchor the flame. These include rapid expansions or bluff bodies in the flow, so that there is a re-circulating flow field that recirculates hot products back to the incoming reactants. Swirling combustors introduce this recirculation with purely aerodynamic means - the flow actually reverses direction and forms a recirculation bubble when the fluid has a sufficient swirl number, a phenomenon referred to as “vortex breakdown”. Whatever the stabilization method, a flame can only be stabilized in a combustor over a certain range of conditions, even if those conditions lie within its flammability limits. For example, at a fixed stoichiometry, as the flow velocity is increased, at some point the flame will not be able to remain anchored but will blow off. Alternatively, at a fixed flow velocity, as the equivalence ratio is decreased, at some point the flame blows off. Predicting blowout behavior is complicated by a lack of understanding of the flame characteristics at the stabilization point. Nonetheless, empirically anchored phenomenological methods for correlating blowout behavior have been reasonably successful. Most approaches consider the ratio of two time scales: a chemical kinetic time and residence time, τchem/τres. The chemical time characterizes how much time is required for the reaction while the residence time characterizes the time which the reactants reside in the reaction zone1. This ratio is often referred to as a combustor loading parameter. Simply put, if this residence time is shorter than the chemical time, the flame will blow off. It must be emphasized that the detailed flow and chemical processes are much more complex than this simple picture might suggest; nonetheless, more sophisticated approaches generally reduce to a correlation of this form. When applied to blowoff limits of premixed flames, this chemical time can be estimated as:
τ chem = α S L2
(1)
where SL and α denote the laminar flame speed and thermal diffusivity, respectively2. The residence time is generally scaled as d/Uref, where d and Uref denote a characteristic length scale (e.g., a recirculation zone length) and velocity scale, respectively. Putting this together, blowoff limits should scale with the Damköhler number:
τ res S 2d = L . τ chem αU ref
(2)
Determining the correct length and velocity scale is not straightforward. Note that Uref need not directly scale with approach flow velocity, Uu, due to the acceleration of the burned gas3. Since the burned gas velocity scale is given by Ub=(Tb/Tu)Uu, then Uref =f(Uu, Tb/Tu). Similar considerations apply for the recirculation zone scale, d. For this reason, prior workers have often had to measure the recirculation zone length in order to use Eq. (2) (e.g., see Ref 6.). Furthermore, the chemical time calculation is complicated by thermal-diffusive effects (i.e., H2 diffuses much more rapidly than air or other fuels), as the local fuel/air ratio of the mixture may differ from the global average. While clearly there are important issues such as appropriate choice of length and velocity scale, Damköhler number scalings do a reasonable job in scaling blowout data across a wide range of fuel compositions, as shown in several prior publications. As such, the manner in which the blowoff trends of a system are affected by variations in fuel composition can be inferred from the chemical kinetic times of the mixtures. To illustrate, figure 1, plots the dependence of the chemical time, upon fuel composition of H2/CO/CH4 mixtures at a fixed flame temperature of 1900°K. Note the order of magnitude variation in chemical time from the fast H2 mixtures to slow CO mixtures. One clear implication of this result is that higher hydrogen mixtures will blowoff at leaner equivalence ratios, as can be seen by figure 2, which plots the equivalence ratio of the mixture at blowoff. It should be emphasized that fluid mechanics, and not just chemical kinetics, must be accounted for in understanding how blowoff limits will vary with composition. Because the flow field and the flame are coupled, variations of the chemistry do impact the flow.
3.1.1-3 Dynamic Stability Overview Combustion instabilities refer to large amplitude oscillations of pressure, heat release, velocity, and other variables inside the combustion chamber. They often occur at discrete frequencies associated with the natural acoustic modes of the combustor. Such instabilities have been encountered during the development and operation of most high performance propulsion and power generating devices. They are spontaneously excited by feedback between unsteady heat release and, generally, one of the natural acoustic modes of the combustor. Their occurrence is usually problematic because they produce large amplitude pressure and velocity oscillations that result in enhanced heat transfer and thermal stresses to combustor walls, oscillatory mechanical loads that result in low or high cycle fatigue of system components, and flame blowoff or flashback.
Fig. 1. Dependence of chemical time (ms) upon fuel composition at fixed adiabatic flame temperature of 1900°K [pressure is 4.4 atm with 460K reactants temperature] (reproduced with permission from authors). Source: Q. Zhang, D. Noble, and T. Lieuwen, “Blowout Measurements in a Syngas-Fired Gas Turbine Combustor,” Annual Pittsburgh Coal Conference (2005).
0.4
0.35 φ at LBO
Da =
Flame
0.3
0.25
0.2
0
No Flame 0.2
0.6 0.4 % H2
0.8
1
Fig. 2. Dependence of LBO equivalence ratio upon H2 mole fraction at approach flow velocities of 6 m/s and 4.4 atm combustor pressures, 460 K inlet temperature (reproduced with permisson from authors). Source: See fig. 1.
198
Timothy C. Lieuwen A generic feedback loop is shown in figure 3, illustrating the sequence of events responsible for self-excited oscillations in the combustion chamber: (1) Fluctuations in the velocity, pressure, fuel/air ratio, etc. excite a fluctuation in the heat release rate, (2) The heat release fluctuation excites acoustic oscillations, (3) The acoustic oscillations generate the disturbance in Step (1) above, closing the feedback loop. Depending upon the phase between the pressure and heat release (discussed below), the flame may add or remove energy from the acoustic field during each cycle, represented by one complete loop in this diagram. If the energy supplied to the acoustic field by the combustion process exceeds the energy losses of the mode, the acoustic amplitude will grow in time until it saturates, at some limit, cycle amplitude. Generally, combustion instabilities occur at frequencies associated with natural acoustic modes of the combustor. These include, e.g., bulk (i.e., Helmholtz type oscillations), axial, and transverse (i.e., tangential and/or radial) modes (see figure 4). On occasion, however, the oscillations are not associated with a purely acoustic mode and are excited by a coupled “convective-acoustic mode, which occurs at frequencies lower than those of purely acoustic modes. Such oscillations occur when a hot gas packet (due to, e.g., partial flame extinction) or vortex convects through the nozzle, where it excites an acoustic wave that propagates back to the flame4, exciting another convected wave, thus repeating the process. These types of modes are often encountered in systems that are operating at conditions close to flame blowoff.
Heat Release Oscillations
Flow and Mixture Perturbations
Acoustic Oscillations
Fig. 3. Illustration of feedback loop responsible for combustion instability.
Transverse Radial Mode Longitudinal
Why do Instabilities Occur?
199
In order to understand why instabilities occur, we must understand why the flame adds energy to the acoustic field. Rayleigh’s Criterion describes Transverse Azimuthal Mode these conditions. Essentially, it states that heat release disturbances add energy to the acoustic field if the heat is added/removed to or from the gas when its Fig. 4. Longitudinal and transverse acoustic modes in pressure is above/below its mean value. This statement is mathematically cylindrical combustors. described by the integral in Eq. (2). This equation shows that the heat addition process locally adds energy to the acoustic field when the magnitude of the phase between the pressure and heat release oscillations, θpq, is less than ninety degrees (i.e., 0<| θpq|<90). Conversely, when these oscillations are out of phase (i.e., 90<| θpq|<180), the heat addition oscillations damp the acoustic field. Rayleigh’s criterion describes the conditions under which unsteady heat release adds energy to the acoustic field. However, even if energy is transferred from the combustion process to the acoustic field, this does not Fuel/air ratio necessarily imply that the combustor is unstable – this can only happen oscillations Fuel flow rate if the rate of energy supplied by the periodic combustion process to the oscillations Flame area and acoustic field is larger than the rate at which acoustic energy is dissipated reaction rate within the combustor and/or transmitted through its boundaries. oscillations Having established the conditions under which energy is added Flow rate Combustion oscillations to the acoustic field by the flame, we need to consider the mechanisms Products through which these heat release disturbances are generated. A number of mechanisms can produce heat release fluctuations in gas turbines, as Unsteady mixing, vaporization, indicated in figure 5. These include: atomization Vortex/flame 1. Fuel Feed Line-Acoustic Coupling. Pressure oscillations in the interactions combustor modulate the pressure drop across unchoked fuel nozzles. This, in turn, modulates the fuel injection rate into the system, causing an Fig. 5. Potential mechanisms of combustion instability. oscillatory heat release process that drives the acoustic oscillations. 2. Equivalence Ratio Oscillations6. Combustor pressure oscillations propagate into the premixer section where they modulate mixing processes and fuel and/or air supply rates, thus producing a reactive mixture whose equivalence ratio varies periodically in time. The resulting mixture is convected into the flame where it produces heat release oscillations that drive the instability.
3.1.1 Static and Dynamic Combustion stability 4. Oscillatory Flame Area Variation7. Interactions of acoustic velocity oscillations with the flame cause periodic variation of the flame area and, thus, a periodic heat addition process that drives the acoustic field. 5. Vortex Shedding8. Large scale, coherent vortical structures due to flow separation from flameholders and rapid expansions, as well as vortex breakdown in swirling flows, are often present in gas turbine combustors, as shown in figure 6. The vortical structures distort the flame and cause its surface area to oscillate, thus producing heat release oscillations. Although the details are excluded here, one can show that in order for any one of these mechanisms to be self-exciting, the characteristic times related to the physical processes responsible for the heat release disturbance must be of similar magnitude as the acoustic period. For example, if the mechanism is equivalence ratio oscillations or vortex shedding, a combustion instability may occur when the following relationship holds: τ convect + τ chem = kT
(3)
Fig. 6. Computed image of swirling flame distorted by vortical structures (reproduced with permission of Y. Huang and V. Yang). Source: Y Huang and V. Yang, “Effect of Swirl on Combustion Dynamics in a Lean-Premixed SwirlStabilized Combustor,” Proceedings of the Combustion Institute 30 (2004): 1771-1778.
where τconvect refers to the time required for either the equivalence ratio oscillation or vortex to convect from its point of formation to the “center of mass” of the flame, τchem refers to the chemical delay time, T refers to the acoustic period, and k is a series of constants whose value depend upon the combustion chamber acoustics9. Fuel composition variations impact this relationship, Eq. (3), by affecting both characteristic times on the left of the equation. Their impact on the chemical time is clear. Their impact on the convective time delay can be better understood from the following equation: τ convect = ( LFl / n + Lst ) / u
(4)
where u refers to the mean flow velocity, Lst refers to the flame “standoff distance” from wherever the disturbance originates, LFl is the flame length, and n is a constant that determines the location of the flame “center of mass”. For example, an n value of ½ refers to a flame that is effectively concentrated at its midpoint. Variations in fuel composition impact both the flame standoff location, flame length and the constant n (by altering the flame shape). For situations where the flame temperature remains constant, fuel composition impacts upon the flame standoff location can be approximately inferred from the turbulent flame speed. Increases in turbulent flame speed cause the flame to anchor farther upstream and vice-versa. If the flame temperature varies as well, the situation is much more complex, as the recirculating flow structure can be altered as well in a complex manner. Similar considerations apply for the flame length, which also scales with the turbulent flame speed. One point worth emphasizing is that no fuel is intrinsically more “stable” or “unstable” than another. In other words, stability is determined by whether the equality in Eq. (4) is satisfied – depending upon flow velocity, flame location, and a variety of other factors, any particular fuel can be either stable or unstable. This point is emphasized because it is sometimes stated, incorrectly, that the addition of hydrogen has a stabilizing influence upon dynamic stability. While hydrogen certainly does have a stabilizing influence on static stability, due to its high flame speed, hydrogen fueled combustors can (and do!) become quite dynamically unstable. One instance where hydrogen addition can promote dynamic stability in general is under near blowout conditions where low frequency dynamic instabilities occur. By promoting a more statically stable flame, hydrogen addition could potentially make these types of dynamic instabilities less problematic.
Growth and Saturation of Instabilities The amplitude of the instability grows if the rate of energy addition to the oscillations exceeds the rate of energy dissipation by damping processes. As the amplitude of the oscillations increases, the energy addition and dissipation processes become amplitude dependent and the amplitude of the oscillations attains its maximum value when the time average of the energy addition and removal equal one another. The resulting oscillations are referred to as a limit cycle. The objective of this section is to consider the growth and saturation of the instability amplitude. The mechanisms that initiate combustion instabilities are typically grouped into linear and nonlinear categories. A linearly unstable system is one that is unstable with respect to infinitesimally small disturbances; e.g., a ball perfectly balanced at the crest of a hill. To further illustrate the dependence of the stability and limit cycle of a system upon the amplitude of the oscillations, A, consider the hypothetical, amplitude dependent, driving, H(A), and damping, D(A) processes, which are described in figure 8. As shown, the
200
201
H(A )
D(A )
A LC
A
Fig. 7. Hypothetical dependence of the acoustic driving, H(A) and damping, D(A), processes upon the instability amplitude, A.
Driving/Damping
“driving” and “damping” curves intersect at the origin, indicating that a zero amplitude oscillation is a potential equilibrium point. This equilibrium point is, however, unstable, as any small disturbance that moves the system away from the origin produces a condition in which H(A) is larger than D(A), resulting in further growth of the disturbance. Because these two curves diverge near the origin, their difference increases with amplitude, implying that the amplitude growth rate increases with amplitude. Nonlinear combustor processes control the dynamics of the oscillations as the driving and damping processes become amplitude dependent. Figure 8 describes a situation where H(A) saturates and D(A) increases linearly with the amplitude A, thus resulting in an intersection of the two curves at the limit cycle amplitude, ALC. A nonlinearly unstable system differs from a linearly stable one in that it is stable with respect to small amplitude disturbances but is unstable when subjected to disturbances whose magnitude exceeds a certain threshold value, AT. A simple example of a nonlinearly unstable system is a ball in a depression on the top of a hill. When pushed, this ball returns to its equilibrium point as long as it is subjected to disturbances with amplitude that does not get it over the side walls of the depression. However, for sufficiently large disturbance amplitude, the ball rolls out of the depression and down the hill. Similar behavior may be observed in combustors. Although nominally stable, if disturbed hard enough, the combustor may become unstable. A typical manifestation of combustors with this type of behavior is hysteresis, where the parameter values where instability occurs differ depending upon whether the parameter is increasing or decreasing. Figure 9 provides an example of the amplitude dependences of H(A) and D(A) that produces the above discussed behavior. In this case, the system has three equilibrium points where the driving and damping curves intersect. Specifically, the damping exceeds the driving when A
AT grow until their amplitude attains the value A=ALC. Consequently, two stable solutions exist at this operating condition. The one observed at any point in time will depend upon the history of the system. Two other phenomena are often observed in unstable combustors under limit cycle conditions. First, is the generation of harmonics. In other words, an instability at 251 Hz generates harmonic oscillations at 502 Hz, and possibly 753 Hz and higher harmonics as well. Second, the presence of oscillations also changes the mean flame position and flow field. For example, the flame may become either shorter or longer. Unfortunately, the factors that influence the limit cycle instability amplitude are very poorly understood. As such, it is not possible to comment on the influence of fuel composition upon instability amplitudes.
Driving/Damping
Timothy C. Lieuwen
H(A )
D(A ) A
AT
A LC
Fig. 8. Hypothetical dependence of the acoustic driving, H(A) and damping, D(A), processes upon amplitude, A, that produce triggering of instabilities.
3.1.1 Static and Dynamic Combustion stability 3.1.1-4 Notes
_________________________ 1. E.E. Zukoski, “Afterburners,” in Aerothermodynamics of Gas Turbine and Rocket Propulsion, G. Oates, Ed., 1997; D. Spaulding, “Some Fundamentals of Combustion,” Ch. 5, Butterworth Press: London, 1955; J. Longwell, E. Frost, and M. Weiss, “Flame Stability in Bluff-Body Recirculation Zones,” Ind. Eng. Chem. 45 no. 8 : 1629-1633; S. Hoffmann, P. Habisreuther, and B. Lenze, “Development and Assessment of Correlations for Predicting Stability Limits of Swirling Flames,” Chemical Engineering and Processing 33 (1994): 393-400. 2. S.L. Plee, and A.M. Mellor, “Characteristic Time Correlation for Lean Blowoff of Bluff Body Stabilized Flames,” Comb. Flame 35 (1979): 61-80; K. Radhakrishnan, J. Heywood, and R. Tabaczynski, “Premixed Turbulent Flame Blowoff Velocity Correlation Based on Coherent Structures in Turbulent Flows,” Comb. Flame 42 (1981): 19-33. 3. See note 2 (Plee & Mellor). 4. K. Yu, A. Trouve, and J. Daily, “Low-frequency Pressure Oscillations in a Model Ramjet Combustor,” J. Fluid Mech. 232 (1991): 47-72; F. Marble and S. Candel, “Acoustic Disturbance from Gas Non-uniformity Convected Through a Nozzle,” J. Sound Vib. 55 (1977): 225-243. 5. Donald W. Kendrick, Torger J. Anderson, and William A. Sowa, “Acoustic Sensitivities of Lean-Premixed Fuel Injectors in a Single Nozzle Rig,” ASME Paper #98-GT-382. 6. T. Lieuwen, H. Torres, C. Johnson, and B.T. Zinn, “A Mechanism for Combustion Instabilities in Premixed Gas Turbine Combustors,” Journal of Engineering for Gas Turbines and Power 123 no. 1 (2001): 182-190. 7. S. Candel, “Combustion Dynamics and Control: Progress and Challenges,” Proc. Comb. Inst. 29 (2002): 8. U.G. Hegde, D. Reuter, B.R. Daniel and B.T. Zinn, “Flame Driving of Longitudinal Instabilities in Dump Type Ramjet Combustors,” Comb. Sci. and Tech. 55 (1987): 125-138; K. Schadow and E. Gutmark, “Combustion Instability Related to Vortex Shedding in Dump Combustors and Their Passive Control,” Prog. Energy Combust. Sci. 18 (1992): 117-132. 9. See note 6 above.
202
BIOGRAPHY
3.1.1 Static and Dynamic Combustion stability
Timothy C. Lieuwen Associate Professor School of Aerospace Engineering Georgia Institute of Technology Atlanta, GA 30332-0150 email: [email protected]
Dr. Tim. Lieuwen is an Associate Professor at the Georgia Institute of Technology. He is an active researcher in the areas of unsteady combustion phenomenon and acoustics. Dr. Lieuwen is the author of 2 book chapters, and over 100 conference publications and journal articles. He is an Associate Editor of the Journal of Propulsion and Power. Dr. Lieuwen has held various leadership roles in the Air Breathing Propulsion technical committee of the American Institute of Aeronautics and Astronautics (AIAA) and the Combustion and Fuels committee of the American Society of Mechanical Engineers (ASME). Dr. Lieuwen has served on the organizing committees of several major international conferences sponsored by both AIAA and ASME. Dr. Lieuwen’s awards include the NSF CAREER Award, the AIAA Lawrence Sperry Award, and the ASME/ IGTI Turbo Expo Best Paper Award.
3.1
Key Combustion Issues Associated with Syngas and High-Hydrogen Fuels
Vincent G. McDonell Advanced Power and Energy Program University of California Irvine, CA 92697-3550 949-824-7302 ext 121 [email protected]
195
3.1-1 Key Combustion Issues Associated with Syngas and High-Hydrogen Fuels Combustion of syngas and high-hydrogen fuels requires attention to key combustion issues, especially if low emissions are to be achieved using these fuels. Current combustion systems operated on natural gas have evolved to the point where low single digit NOx emissions are possible with lean premixed strategies. However, the price of this evolution has been a significant increase in sensitivity to various perturbations such as changes in ambient conditions and variation in pipeline natural gas composition. In light of these observed sensitivities, strategies over an above lean premixed are continuing to be evaluated as discussed in “Combustion Strategies for Syngas and High Hydrogen Fuel”. To reduce risk and development time, it is desirable to apply the experience of developing low emissions combustion systems for natural gas to syngas and high-hydrogen fuels. However, the range of compositions found in syngas and high-hydrogen fuels varies more substantially than similar properties of pipeline natural gas. By way of example, consider the ranges of composition shown in Table 1. Table 2 summarizes the range and average values of the fuel constituents shown. As a result of the wider range of composition found in syngas and high hydrogen fuels, strategies well suited for low emissions performance on natural gas may not necessarily work best for syngas and hydrogen containing fuels. That said, it is important to note that the variation indicated in Table 1 and Table 2 is somewhat misleading. Specifically, if a given feedstock and gasifications process is considered, the variation found will be much less. By way of example, if the processes are limited to PSI Wabash, Tampa, El Dorado, and Motiva, a representation of variation found in coal/pet coke fed, oxygen blown gasification systems can be established as illustrated in Table 3. The other point to be made by way of introduction is that hydrogen poses the most significant challenge in terms of the combustion system. As a fuel, hydrogen behaves differently than a hydrocarbon in many ways including specific heat (hydrogen has a much higher specific heat than other gases), diffusivity (hydrogen has a much higher diffusivity than other gases), flammability limits (hydrogen has a wide range of volume concentrations over which it is flammable), and flame speed (hydrogen has a much higher laminar flame speed than do other gases). As a result, the presence of hydrogen creates issues for combustion that require a different perspective than would a hydrocarbon fuel. Further, mixtures of gases often exhibit non-linear behavior and little data are available on the types of mixtures found in syngas. With this in mind, the key issues that are associated with combustion of syngas and hydrogen containing fuels can be broadly classified into two major areas: reaction location and stability. These two areas are discussed in detail in “Static and Dynamic Stability”. Reaction location is an issue for all strategies and is related to the chemistry and time scales associated with the system. In strategies which involve premixing the fuel and oxidant (especially for lean strategies), the possibility of reaction evolving into the premixing region is a major concern. Given the high flame speeds of hydrogen, this concern must been examined carefully. Another issue related to the reaction location is ignition delay. With 1-5 msecs of premixing time available, at typical gas turbine inlet temperatures and pressures, ignition delay does not appear to be a major concern for the fuels of interest—however, the predictions of ignition delay have been developed largely in the absence of data at these conditions. As a result, better understanding of whether autoignition might be a factor is really needed to confirm this expectation. Stability can broadly be divided between static and dynamic regimes. It is often reasoned that the wide flammability limits of hydrogen can allow stable operation at leaner (and therefore cooler) reaction temperatures. This extension of the lean blow off or static stability limit is an inherent benefit that should be realized from the presence of hydrogen. However, the extent to which this limit can be achieved in practice and, furthermore, its sensitivity to variation in composition is a key issue. On the other hand, dynamic stability, which is less predictable than static stability, can arise for combustion systems due to various reasons. How fuel composition impacts the propensity of a system to exhibit dynamic stability issues is another concern that must be addressed.
Table 1 Typical Syngas Compositions.1
H2 CO CH4 CO2 N2+AR H2O
24.8 39.5 1.5 9.3 2.3 22.7
37.2 46.6 0.1 13.3 2.5 0.3
35.4 45.0 0.0 17.1 2.1 0.4
34.4 35.1 0.3 30.0 0.2 ---
Sierra Pacific 14.5 23.6 1.3 5.6 49.3 5.7
LHV BTU/ft^3 kJ/m^3
209 8224
253 9962
242 9528
210 8274
128 5024
183 7191
317 12492
Tfuel , F
570
700
250
200
1000
400
Tfuel , C
300
371
121
98
538
204
H2/CO
0.63
0.8
0.79
0.98
0.61
0.33
Constituents
Diluent
PSI
Tampa El Dorado Pernis
steam
Equivalent LHV BTU/ft^3 kJ/m^3
150 5910
N2
N2/steam
118 4649
113* 4452
steam 198 7801
steam 110 4334
ILVA 8.6 26.2 8.2 14.0 42.5 ---
-------
Schwarze Sarlux Pumpe 61.9 22.7 26.2 30.6 6.9 0.2 2.8 5.6 1.8 1.1 --39.8
Fife 34.4 55.4 5.1 1.6 3.1 ---
Exxon Motiva PIEMSA Tonghua Singapore Delaware 44.5 32 42.3 10.3 35.4 49.5 47.77 22.3 0.5 0.1 0.08 3.8 17.9 15.8 8.01 14.5 1.4 2.15 2.05 48.2 0.1 0.44 0.15 0.9
163 6403
319 12568
100
392
100
350
570
338
---
38
200
38
177
299
170
---
2.36
0.74
0.62
1.26
0.65
0.89
steam 200 7880
moisture
H2O
-----
* ---
241 9477
steam
248 9768
H2O/N
116 4600
150 5910
270.4 10655
N2
134.6 5304
0.46 n/a
129 5083
134.6 5304
* Always co-fired with 50% NG
Table 2 Summary of Compositions
Constituent Hydrogen Carbon Monoxide Methane Carbon Dioxide Nitrogen + Argon Water Hydrogen/Carbon Monxide Ratio
Min 8.6 22.3 0 1.6 0.2 0.1 0.33
Max 61.9 55.4 8.2 30 49.3 39.8 2.36
Volume %
Avg 31.0 37.2 2.2 12.0 12.2 7.8 0.86
Table 3 Composition Variation for Pet Coke/Coal Fired, Oxygen Gasified Fuel Streams.
Constituent Hydrogen Carbon Monoxide Methane Carbon Dioxide Nitrogen + Argon Hydrogen/Carbon Monxide Ratio
Min 32 45 0 13.3 0 0.65
Volume % Max 37.2 49.5 negligble 17.1 negligble 0.80
Avg 34.6 47.3 -15.2 -0.72
3.1-2 Notes _______________________ 1. D. M. Todd, “Gas Turbine Improvements Enhance IGCC Viability”, 2000 Gasification Technologies Conference, San Francisco, CA, October, 2000.
196
BIOGRAPHY
3.1 Key Combustion Issues Associated with Syngas and High-Hydrogen Fuels
Vincent G. McDonell Advanced Power and Energy Program University of California Irvine, CA 92697-3550 phone: (949) 824-7302 ext 121 email: [email protected]
Dr. Vincent G. McDonell is the Associate Director of the UCI Combustion Laboratory and is an adjunct Associate Professor at the University of California, Irvine. Dr. McDonell’s research focuses on gas turbine combustion systems and components including the design and characterization of devices for both liquid and gas fired applications. Research contributions range from micro-turbine generators, to central plant gas turbines, to propulsion gas turbines. Fuels experience ranges from natural gas to hydrogen to coal derived fuel gas and liquids. He has extensive experience in application of laser based and conventional diagnostics to a wide array of combustion devices and has particular expertise in the physics of two-phase transport and optical diagnostics for two-phase flows including phase Doppler interferometry. He also regularly applies CFD in the analysis and design of combustion systems and components. Dr. McDonell is a member of the Combustion Institute, ILASS-Americas, AIAA, and ASME. He currently serves on the executive committee of ILASS-Americas and the Western States Section of the Combustion Institute and also serves on the Academic Advisory Board the University Turbine Systems Research program. Research by Dr. McDonell has been documented in over 40 publications.
3.2.1.1
Conventional Type Combustion
3.2.1.1-1 Introduction Brayton Cycle The role of the combustor in a gas turbine engine is two-fold. First, the combustor transforms the chemical energy resident in the fuel into thermal energy for expansion in the turbine. Second, the combustor tailors the temperature profile of the hot gases at the exit plane in order to not compromise the material constraints of the turbine. To fulfill this two-fold role, the combustor is designed to mix fuel with air at elevated pressure and temperature, to both establish and sustain a stable continuous combustion reaction, and to mix the products of combustion to establish the desired exhaust temperature profile. The combustor processes are, as a result, a complex combination of fluid mixing, chemical kinetics, and heat transfer. To contain and control these processes, the design of the “conventional” combustor has evolved over seven decades for the production of propulsive thrust and electrical power. The thermodynamic path over which the gas turbine engine operates is the Brayton Cycle (Figure 1). The compressor [C] ingests and compresses ambient air to elevated pressures that vary in the range of a few to many tens of atmospheres depending on the engine design and application. The “Pressure Ratio” (ratio of outlet to inlet pressure of the compressor, P2/P1) is a major factor in establishing the overall thermodynamic efficiency of the engine. The higher the pressure ratio, the higher the overall thermodynamic efficiency.
Scott Samuelsen Professor of Mechanical, Aerospace, and Environmental Engineering Director Advanced Power and Energy Program University of California Irvine 92697-3550
Fig. 1. Gas Turbine Brayton Cycle for Electric Power Generation
phone: 949-824-5468 email: [email protected]
Fig. 2. Stationary Gas Turbine Electric Power Generator
209
Combustor Inlet Conditions The compression of the ambient air from State Point 1 to State Point 2 is accompanied by an increase in the temperature of the air. As a result, the air exits the compressor and enters the combustor at both an elevated pressure and an elevated temperature. In addition to air, fuel is also injected into the combustor at the inlet. The fuel (such as natural gas, coal syn gas, or petroleum liquids) is the source of energy required to “drive” the cycle.
Goal of the Combustor The goal of the combustor is to convert the chemical energy bound in the fuel into thermal energy. The thermal energy can then be expanded through a turbine [T] to produce (1) the power required to operate the compressor, and (2) the power required to turn a generator and produce electricity. To accomplish this goal, the combustor serves as the vehicle to: – – – –
Combine and mix the air and fuel entering the combustor, Ignite the mixture of fuel and air, Contain the mixture during the combustion reaction, and Tailor the temperature distribution of the hot gases at the exit plane.
Continuous Combustion The processes that occur within a gas turbine combustor (e.g., injection of the air and fuel, mixing of the air and fuel, combustion reaction) are “continuous” rather than intermittent, and occur at constant pressure. This is in contrast to the automobile spark ignited “Otto Cycle” engine where the combustion is intermittent and accompanied by a significant increase in pressure. The gases exit the gas turbine combustor as a steady flow and are then continuously expanded through turbine stages. After the final expansion stage, the spent gases are then exhausted into the atmosphere.
3.2.1.1-2 Combustor Features The design of gas turbine combustors has evolved over many decades with the final configuration based on the best of engineering judgment and intuitive reasoning. As demands have developed for efficiency and lower environmental impacts, engineering tools such as computational fluid dynamics1 and laser diagnostics2 have evolved to facilitate the design process. This notwithstanding, engineering judgment coupled with intuitively based empirical correlations, continues to serve as the anchor to modern design. Throughout the evolution of combustor technology, the basic requirements for combustor design have remained. In particular, the following five basic features are integral to the combustor design: a primary zone, a secondary zone, a dilution zone, various wall jets, and the management of heat transfer at the combustor boundary (Figure 3).
Fig. 3. Combustor Features
210
Scott Samuelsen 3.2.1.1-3 Primary Zone The air exiting the compressor enters the combustor through four major injection points, each of which has a particular role. Each injection point convects approximately one-quarter of the total air flow into the combustor. Two of these injection points (swirler, primary air wall jets) control both the structure of, and the mixing within, the primary zone. Swirler. The first entry point for the compressor air is through swirler vanes that are positioned at the front face of the combustor and typically surround the fuel injection port. The swirl vanes impact a circumferential velocity component to the air and thereby thrust the air radially outward as the air enters the combustor. This creates a pressure void at the center line and induces a backflow to fill the centerline pressure deficit. This effectively creates, as a result, a recirculation flow that extends approximately one duct diameter downstream and defines the “Primary Zone” of the combustor. The strength of the swirl is defined by the swirl number, SN:
(1) where: Gm = Axial Flux of Angular Momentum Gt = Axial Thrust Dsw = Diameter of Swirler The swirl number must exceed 0.6 in order to induce a recirculation zone. “Aerodynamic Spark Plug.” The fuel is injected at an angle to mix with the swirler air that is exiting the swirler. Mixing of the fuel and air is facilitated by the turbulence that is created by the passage of the air through the swirler. The resultant fuel/air mixture is then recirculated and mixed with energetic “hot products” of combustion that are pulled and entrained into the recirculation zone from downstream. These energetic species provide the ignition source for the fresh mixture of fuel and air. In effect, the recirculation zone combines as a combined aerodynamic “blender” and “spark plug.” Primary Air Jets. Wall jets affect the mixing, stoichiometry, and structure of the flows in gas turbine combustors. Due to this dominating role, a substantial literature has evolved to guide the design and estimate the behavior of jets injected into a crossflow.3 In a typical combustor design, two sets of air wall jets (primary and dilution) are prescribed (Figure 3). The primary air jets are located approximately one duct diameter downstream from the combustor inlet and serve two major functions. First, the jets bring closure to the recirculation zone by providing a strong force against which the primary zone cannot easily penetrate. Without the set of primary air jets, the dynamics of the recirculation zone would create aerodynamic fluctuations and result in pressure oscillations, undesirable noise, and elevated pollutant emission. Secondly, the primary jets bifurcate with a substantial percentage of the flow directed upstream to mix with the recirculating fuel/air mixture, and the remainder mixing downstream into the secondary zone (Figure 4). The primary jet flowing upstream augments the swirler air to establish the overall stochiometry of the primary zone.
Fig. 4. Bifurcation of Primary Jets
The stochiometry describes the actual fuel-to-air ratio compared to the chemically correct or “stoichiometric” ratio. A number of indices (e.g., theoretical air, excess air) can be used. For gas turbine combustion, the equivalence ratio (φ) is the index that is typically adopted:
φ= 211
(Fuel/Air)actual (Fuel/Air)stoichiometric
(2)
3.2.1.1 Conventional Type Combustion The primary zone is typically fuel rich (φ>1.0) in order to promote reaction stability (e.g., preclude blow-out). Mixing. Combustion is a complex coupling of fluid mechanics and chemical kinetics (Figure 1). A large scale, macro fluid mechanical structure (“recirculation zone”) mix the fuel and air within the primary zone and entrain hot, energetic species to ignite the fresh reactant mix. Chemical kinetics determine the paths and rate at which the reaction proceeds. The fluid mixing and chemical kinetics occur in parallel throughout the primary zone and over a range of scales. In particular, the zone of recirculation is at the macro scale and, within this zone, a range of turbulent eddy scales exists and persists. The size of the macroscale mixing associated with recirculation is on the order of the combustor diameter (Figure 3). Within the macroscale recirculation zone, mixing of the fuel, air, and recirculated energetic products occurs on the “microscale.” Whereas the macroscale recirculation zone is a “blender” on the scale of the duct diameter, the microscale mixing occurs within “mini-blender” packets that vary in (1) the concentrations of fuel and air (Figure 5), and (2) size. The microscale mini-blenders are turbulent eddies generated (1) at the physical boundaries of the inlet plane, and (2) within the shear that exists between the various flows in the primary zone. The most important shear layer (layer separating two streams of differing velocities) exists between the entering fuel and air streams, and within the steep velocity gradient associated with the macroscale recirculation zone. Each turbulent eddy will experience a finite lifetime (~tens of milliseconds) within the reaction zone before breaking up, mixing with adjacent eddies, and forming a new eddy. Some eddies containing unreacted fuel and air will ignite. Others will not, waiting to mix with other eddies to acquire sufficient energetic species of the necessary mixture ratio that is required for ignition. In traditional combustors, the fuel and air are injected separately (i.e., “non-premixed”). The reaction is often referred to as a “diffusion flame” and the combustor as a “diffusion combustor.” This is a misnomer. In a diffusion flame, the fuel is not premixed with the air prior to reaction, and the reaction occurs at the interface between the fuel and the air. Within the primary zone of a gas turbine combustor, the injection of reactants, the mixing of the reactants, the entrainment and mixing of energetic species, and reaction are occurring simultaneously throughout the volume of the recirculation zone. A variety of fuel/air packets are formed with a myriad of mixture ratios. As a result, mixing of the fuel and air indeed occurs before reaction of the individual packets. The extent to which, in the aggregate, the fuel and air mix prior to reaction depends upon the fuel properties, the fuel and air injection hardware, and the time for mixing prior to reaction. While not premixed (the fuel and air are injected separately), the reaction is not a diffusion flame. Instead, the reaction is a “partially-mixed” “distributed reaction.” To approached a premixed reaction, the fuel and air must be either (1) intensely mixed after injection in a zone that precedes reaction but precludes auto-ignition (“rapidly mixed, non-premixed”), (2) introduced over a spatially large area through a large number of discrete injection points (“spatially injected, non-premixed”), or (3) premixed prior to injection (“premixed”). Due to safety, non-premixed operation has been the preferred option. The need to reduce the emission of pollutant species, however, has sought a reaction in the primary zone that behaves closer to a premixed reaction. For stationary gas turbines, all three options listed above are being developed and deployed. For aero-propulsion applications, only the first two options are being developed and deployed. Gaseous fuels (e.g., natural gas, syn-gas) will mix more rapidly with the air than liquid fuels. Liquid fuels are injected as small droplets and must first evaporate into a vapor before mixing with the air can occur. (Some droplets may not completely evaporate and will react as a small diffusion flame.) Heat Release. The transformation of the chemical energy bound in the fuel to thermal energy is a two-step process. The first step is associated with the primary zone. Here, the hydrogen and carbon bonds in the fuel are converted relatively fast through a series of reactions to carbon monoxide (CO) and water (H2O) (Figure 6). Approximately two-thirds of the chemical energy bound in the fuel is released to thermal energy in this first phase. The radiative flux emanating from CO is light blue (Figure 7). In actual engines, this cannot be observed. In a laboratory model combustor with appropriate optical assess, the light blue emission is discernable at the edges surrounding the “white-light” associated with the long-duration exposure of the film (Figure 8).
Fig. 5. Microscale Eddies
212
Scott Samuelsen The CO produced retains one-third of the chemical energy. The release of the residual energy bound in the CO does not occur readily in the primary zone due to (1) the relatively slow kinetic rate for the oxidation of CO to carbon dioxide (CO2), (2) the relatively short residence time in the recirculation zone, and (3) the rich stochiometry of the primary zone. Herein is the role of the secondary zone.
Fig. 6. Heat Release Chemistry (Example for Methane, CH4, as the Fuel) Source: 4. Samuelsen, G. S., The Combustion Aspects of Air Pollution, Advances in Environmental Science and Technology, Vol. 5, pp. 219-322, John Wiley & Sons ,1975.
3.2.1.1-4 Secondary Zone The role of the secondary zone is to oxidize the CO to CO2. The principal elementary kinetic reaction that governs the oxidation is: CO + OH => CO2 + H
Fig. 7. Radiative Properties of the Primary Zone
213
(3)
3.2.1.1 Conventional Type Combustion
Fig. 8. Model Combustor Operating on JP-4 Source: Cameron, C.D., Brouwer, J., and Samuelsen, G.S., A Model Gas Turbine Combustor with Wall Jets and Optical Access for Turbulent Mixing, Fuel Effects, and Spray Studies, Twenty-Second Symposium (International) on Combustion, The Combustion Institute, pp. 465-474, 1988.
The strategy is to increase the sluggish forward reaction rate by (1) establishing an overall lean mixture ratio (e.g., φ~0.8) through the primary jet bifurcation, (2) retaining the temperature at an elevated level, and (3) providing the residence time needed to promote the oxidation. The emission from CO2 is purple (Figure 9). The effectiveness of the secondary zone is evident in Figure 8 where a purplish light emission, characteristic of the CO2 molecule, is observed between the primary and dilution jets.
Fig. 9. Radiative Properties of the Primary and Secondary Zones
3.2.1.1-5 Dilution Zone The role of the dilution zone is to reduce the temperature of the combustion products and mix the resultant gases in order to establish a temperature that will uphold the integrity of the turbine blades. This is accomplished by second major set of air jets. The dilution jet flow, approximately one-quarter of the total air flow exiting the compressor, is sufficient to reduce the overall equivalence ratio of the gases exiting the combustor to a very lean condition (e.g., φ~0.3) with a corresponding concentration of oxygen of 15% by volume. To protect the integrity of the turbine section, it is not sufficient to reduce the mean temperature. The radial and circumferential variation in local temperature from the mean can create hot spots and degrade, damage, and possibly destroy a turbine component (e.g., blade, stator, seal). As a result, the temperature profile at the exit plane must meet design criteria. The temperature profile is characterized by various indices including the “Pattern Factor,” the “Profile Factor,” and the “Turbine Profile Factor.”
214
Scott Samuelsen A combustor designer will work with the turbine design team to establish the exit plane temperature “design profile” (Figure 10). The temperature is reduced at the root (0% Blade Span) to protect the blade attachment to the shaft, and reduced at the 100 percent span point to manage the clearance at the wall. The peak temperature occurs closer to the 100 percent span point due to the larger circumferential area of the turbine that can manage the elevated heat flux. The actual temperature profile may deviate from the design profile. The Pattern Factor reflects the extent to which the maximum temperature deviates from the average temperature rise across the combustor {T3- T2}:
Pattern Factor =
{Tmax − T3 } {T3 − T2 }
(4)
Fig. 10. Exit Plane Temperature Profiles Source: Lefebvre, Arthur H., Gas Turbine Combustion, Second Edition, Taylor and Francis, p. 120, 1998.
The Profile Factor characterizes the extent to which the maximum circumferential mean temperature, Tmr, deviates from the average temperature rise across the combustor: Profile Factor
(5)
The Turbine Profile Factor addresses the maximum temperature difference by comparing the average temperature at any given radius around the circumference (T3r) and the design temperature for that same radius (T3des): Turbine Profile Factor
=
{T3r − T3des }max {T3 − T2 }
(6)
The goal is for the actual profile to match the design profile. The dilution jet penetration is the major force that directly determines the extent to which this match is achieved. In general, the combination of the number of dilution jets and the orifice size for each jet is selected such that the centerline of the dilution jets penetrates from the wall a distance that corresponds to 1/3 of the duct diameter.
215
3.2.1.1 Conventional Type Combustion 3.2.1.1-6 Heat Transfer A substantial design consideration in the management of the air flow exiting the compressor is to address the heat transfer demands of the combustor liner. The reaction within the combustor produces a substantial radiative flux of heat to the liner wall. Unless this heat is dissipated from the wall, the wall will be compromised and fail. To preclude this, approximately one-quarter of the compressor air flow is allocated to liner cooling. Designs for accommodating this cooling flow are small holes in the liner with louvers on the inside wall to direct the flow along the internal boundary (Figure 3). New designs incorporate liner materials with hundreds of closely spaced holes that promote a diffusive flux of air at all points along the liner.
3.2.1.1-7 Combustor Configurations Gas turbine combustors first evolved in a “can” configuration. Over time, a length-to-diameter ratio of ~3.0 has emerged as necessary to (1) physically accommodate the three zones (primary, secondary, and dilution), and (2) achieve the combustion efficiency, combustion stability, and pollutant emission required of viable, commercial systems. Due to the simplicity of this configuration, many modern stationary gas turbine engines today retain a can geometry. To accommodate the evolution of larger engines and the annular flow of air exiting compressors, a can-annular configuration evolved as the second-generation strategy for propulsion engines. The Pratt & Whitney JT8D was the epitome of this design, constituted for many years the major population of aero-propulsion engines, and powered the Boeing 727 and early series of Boeing 737s. The third-generation configuration is the “full-annular” geometry which provides an exact match in the open annular area to both the compressor exit and the turbine entrance. Modern aero-propulsion engines adopt this geometry universally in order to embrace many advantages including a reduced combustor length (and hence a higher thrust to weight ratio), and the ability to accommodate a wider range turndown (e.g, from idle to taxi, to cruse, to full power) with a low environmental signature. With the exception of some aero-derivative engines, stationary gas turbine designs have retained the can geometry for (1) ease of maintenance, and (2) ability to incorporate advanced low-emission combustor strategies (e.g., premixed injection, catalytic surfaces). Configuration options also include whether or not the combustor system will be “in-line” with the compressor and turbine. Due to the requirement of aero-propulsion engines to be efficiently packaged with a minimum length and overall weight, the inline configuration is standard. Stationary gas turbine engines are free from these constraints. As a result, engines designed from scratch for stationary applications are outfitted with can combustors that are often not in-line in order to support (1) ease of access, and (2) ease of maintenance. Aero-propulsion engines that are applied as well for stationary power generation will be often retrofitted with out-of-line can combustors as a substitute to the relatively elegant in-line annular configuration for the aero application.
3.2.1.1-8 Notes _______________________ 1. Mongia, H. C., Reynolds, R. S., and Srinivasan, R., “Multidimensional Gas Turbine Combustion Modeling: Applications and Limitations,” AIAA Journal, Vol. 24, No. 6, pp. 890-904, 1986. 2. McDonell, V.G. and Samuelsen, G.S., ”Measurement of Fuel Mixing and Transport Processes in Gas Turbine Combustion,” Measurement, Science, and Technology, Topical Issue on Measuring Techniques for Turbomachinery, Vol. 11, pp. 870-886, 2000. 3. Holdeman, J.D., “Mixing of Multiple Jets with a Confined Subsonic Crossflow,” Progress in Energy and Combustion Science, Vol. 19, No. 1, pp. 31-70, 1993.
216
BIOGRAPHY
3.2.1.1 Conventional Type Combustion 3.2.1.3 Rich Burn, Quick-Mix, Lean Burn (RQL) Combustor
Scott Samuelsen Professor of Mechanical, Aerospace, and Environmental Engineering Director Advanced Power and Energy Program University of California Irvine 92697-3550 phone: 949-824-5468 email: [email protected]
Professor Scott Samuelsen is Director of the Advanced Power and Energy Program (APEP) at the University of California Irvine and Professor of Mechanical, Aerospace, and Environmental Engineering. He directs as well the National Fuel Cell Research Center (NFCRC) and the UCI Combustion Laboratory (UCICL). His research is directed to advanced power systems including gas turbines, fuel cells, and fuels. He directs anchor research on advanced coal and natural gas power plants for the coproduction of electricity and hydrogen for the U.S. Department of Energy (DOE), distributed generation and information technology research for the U.S. Department of Defense (DoD) in support of energyefficient and environmentally-responsible power generation, advanced energy systems research for the California Energy Commission, and coal-gas and hydrogen-fueled gas turbine combustion studies. His energy expertise is based on forty years of combustion research working with strategic alliances involving industry with applications to gas turbine propulsion, gas turbine electronic power generation, and combustion distributed generation resources. He holds the Ph.D. degree from the University of California Berkeley.
3.2.1.2
Lean Pre-Mixed Combustion
3.2.1.2-1 Introduction Gas turbine designers are continually challenged to improve cycle efficiency while maintaining or reducing emissions. This challenge is made more difficult by the fact that these are often conflicting goals. The path to improved efficiency is higher working fluid temperatures, but higher temperatures promote NOx formation and at 2,800 F the threshold for thermal NOx formation is reached. Furthermore, reducing available oxygen to reduce NOx can result in higher carbon monoxide (CO) and unburned hydrocarbon emissions due to incomplete combustion. Moreover, increasing firing temperatures above 2,350 F represents a significant materials science challenge.1 To achieve lower pollutant emission rates, a variety of pre-formation and postformation control technologies have been utilized either individually or in combination, including: • Wet controls (water or steam injection) • Dry combustion controls (lean combustion, reduced residence time, lean premixed combustion, and two-stage rich/lean combustion) • Selective catalytic reduction • SCONOX catalytic absorption • Catalytic combustion (e.g. Xonontm ) • Rich Quench Lean Combustors • CO oxidation catalysts This section of the Handbook focuses on Lean Premixed (LPM) combustion, a pre-formation control strategy that has become the standard technique employed by gas turbine original equipment manufacturers (OEM), particularly for natural gas applications. OEMs have developed processes that use air as a diluent to reduce combustion flame temperatures and reduce NOx by premixing fuel and air before they enter the combustor. This lean premixed combustion process is referred to by a variety of trade names including General Electric’s and Siemens-Westinghouse’s Dry Low NOx (DLN) processes, Rolls-Royce’s Dry Low Emissions (DLE) process and Solar Turbines’ SoLo NOx process. When firing natural gas, most of the commercially available systems are guaranteed to reduce NOx emissions within the 15 to 25 parts per million by volume, dry (ppmvd) range, depending on the OEM, turbine model and application. A few OEM’s have guaranteed single digit NOx emissions.
3.2.1.2-2 Emissions Overview
William R. Bender Technology & Management Services, Inc. Gaithersburg, MD 20879 phone: (301) 670-6390 x144 email: [email protected]
217
The primary pollutants emitted by gas turbine engines are NOx, CO and to a lesser extent, unburned hydrocarbons (UHC). Sulfur dioxide, particulate matter (PM) and trace amounts of hazardous air pollutants may also be present when liquid fuels are fired. Both CO and UHC are the products of incomplete combustion. Given sufficient time and at high enough temperatures, these two pollutants will be further oxidized to carbon dioxide and water. In the proposed standards of performance for new stationary combustion turbines (40 CFR 60, subpart KKKK, dated February 18, 2005), EPA states, “Turbine manufacturers have significantly reduced CO emissions from combustion turbines by developing lean premix technology. Lean premix combustion design not only produces lower NOx than diffusion flame technology, but also lowers CO and volatile organic compounds (VOC), due to increased combustion efficiency.”2 The proposed rulemaking concludes that “Stationary combustion turbines do not contribute significantly to ambient CO levels.”3 Accordingly, the primary pollutant of concern from gas turbines continues to be NOx. There are two sources of NOx emissions in the exhaust of a gas turbine. Most of the NOx is generated by the fixation of atmospheric nitrogen in the flame, which is called thermal NOx. Thermal NOx production rates fall sharply as either the combustion temperature decreases, or as the fuel to air ratio decreases. Nitrogen oxides are also generated by the conversion of a fraction of any nitrogen chemically bound in the fuel. Emissions of NOx from fuel bound nitrogen are insignificant when firing natural gas, but must be considered when firing lower quality distillates and syngas.4
3.2.1.2-3 Regulatory Overview In the late 1970s, EPA established New Source Performance Standards (NSPS) for gas turbines (40CFR60, subpart GG) that limited NOx emissions from large utility gas turbines to 75 ppmvd (parts per million by volume, dry) when firing natural gas. In the 1980s, with the implementation of Best Available Control Technology (BACT), NOx emission limitations on gas were reduced to 42 ppmvd. With the implementation of EPA’s “top down” BACT strategy, NOx emissions further decreased to 25 ppmvd. A recent search of EPAs RACT/BACT/LAER Clearinghouse data (years 2000 to 2005) for natural gas-fired combustion turbines greater than 25 MW confirms the trend towards single digit NOx emissions.
3.2.1.2-4 Combustion Principles Fuel to Air Ratio A significant parameter used to characterize combustion is the fuel to air ratio (f/a), expressed either on a volume or mass basis.5 With precisely enough air to theoretically consume all of the fuel, combustion is referred to as having a “stoichiometric” f/a ratio. Adding more air produces combustion that is fuel-lean, and adding less air produces combustion that is fuel-rich. Because differing fuels have different stoichiometric f/a ratios, it is convenient to normalize the f/a ratio by the stoichiometric value, producing the term equivalence ratio Ø: Ø= (f/a)actual (1) (f/a)stoich (Source: note 5 - G.A. Richards, et al, p. 143)
By referring to the equivalence ratio, combustion using different types of fuel is readily described as lean if Ø < 1 or rich if Ø > 1.
Flame Temperature Another important combustion parameter is the flame temperature. Flame temperatures are determined by a balance of energy between reactants and products. In principal, the highest flame temperatures would be produced at Ø = 1, because all of the fuel and oxygen would be consumed. In practice, the effects of species dissociation and heat capacity shift the peak temperature to slightly above stoichiometric (Ø ~ 1.05). Fuel type is important in determining the flame temperature. To provide a sense of magnitude, the list below compares calculated adiabatic flame temperatures of two hydrocarbons, CO and H2. This list applies to stoichiometric combustion in ambient air: Table 1: Compared adiabatic flame temperature calculations.
Species Methane Propane Carbon Monoxide Hydrogen
Formula CH4 C3H8 CO H2
* Source: note 5 - G.A. Richards, et al, p. 144
Adiabatic Flame Temperature (K) 2223 2261 2381 2370
It should be noted that the methane flame temperature is approximately 150 K lower than hydrogen and CO. This distinction makes it somewhat easier to produce low-emissions from natural gas, which is mostly methane, compared to syngases containing undiluted H2 and CO.
3.2.1.2-5 Combustor Designs Diffusion Flames The combustion process in a gas turbine can be classified as diffusion flame combustion or lean-premix staged combustion.6 In diffusion flame combustion, both fuel and oxidizer are supplied to the reaction zone in an unmixed state. The fuel/air mixing and combustion take place simultaneously in the primary combustion zone. This generates regions of near-stoichiometric fuel/air mixtures where the temperatures are very high. At the beginning of gas turbine development, the primary design goal was to optimize performance while complying with applicable emission requirements. Initially, emphasis was placed on maximizing combustion efficiency while minimizing the emission of unburned hydrocarbons and CO. It was possible to fully realize these design goals by providing the diffusion flame with a relatively high combustion chamber volume in which all chemical reactions were allowed to take place completely without the addition of dilution air. To enable the mixing process to occur rapidly and to achieve a uniform temperature in the primary zone, several burners were arranged in a common flame tube. This combustion chamber design yielded optimum thermodynamic properties with low pressure losses and a combustion efficiency of practically 100 percent.
218
Bill Bender In the early 1970’s, when emission controls were introduced, the pollutant of primary concern to regulators shifted to NOx. For the relatively low levels of NOx reduction initially required, the injection of water or steam into the combustion zone produced the required reduction in NOx emissions with minimal performance impact. In addition, the emissions of other pollutants (CO, VOC) did not increase significantly. To comply with the greater NOx reduction requirements imposed during the 1980’s, further attempts were made to utilize increased quantities of water/steam injection to ensure compliance. These attempts proved detrimental to cycle performance and part lives, and the emission rates for other pollutants also began to rise significantly. Other control methodologies needed to be developed, which led to the introduction of the LPM combustor.
Lean Premixed (LPM) Combustion
219
The Gas Turbine Association defines Lean Premix Stationary Combustion Turbine as “Lean premixed stationary combustion turbine means any stationary combustion turbine designed to operate at base load with the air and fuel thoroughly mixed to form a lean mixture before delivery to the combustor.7 Mixing may occur before or in the combustion chamber. A lean premixed turbine may operate in diffusion flame mode during operating conditions such as startup and shutdown, low or transient loads and cold ambient.”8 Premixing prevents local “hot spots” within the combustor volume that can lead to significant NOx formation. In LPM systems, atmospheric nitrogen (from the combustion air) acts as a diluent, as fuel is mixed with air upstream of the combustor at deliberately fuel-lean conditions. The f/a ratio typically approaches one-half of the ideal stoichiometric level, meaning that approximately twice as much air is supplied as is actually needed to burn the fuel. This excess air is a key to limiting NOx formation, as very lean conditions cannot produce the high temperatures that create thermal NOx. Design of a successful LPM combustor requires the development of hardware features and operational methods that simultaneously allow the equivalence ratio and residence time in the flame zone to be low enough to achieve low NOx emissions, but with acceptable levels of combustion dynamics, stability at part-load conditions and sufficient residence time for CO burnout.9 In principle, the LPM strategy is quite simple: keep the combustion process lean at all operating conditions. In practice, this is not easily achieved. If the engine is already near the limit of lean operation at full power, it is not possible to reduce the combustor temperature rise on all of the fuel injectors, because the flame will be extinguished. To solve this problem, some of the fuel or air must be rerouted (or staged) to keep the flame within its operating boundaries. Typically, engine designers use either fuel or air staging to accomplish this goal. Fuel staging can be accomplished either radially or axially. Examples of radial staging include the use of pilot flames or reducing/eliminating fuel from some injectors completely. Axial staging injects fuel at two places along the combustion gas flowpath. Products from the first combustion zone are mixed with fuel and air in a subsequent combustion zone, providing an advantage for lean operation of the second zone. Finally, some engine designs use air staging (also known as variable geometry) to accomplish the goal of maintaining low flame temperatures. This approach can maintain the desired combustion zone temperature at all operating conditions, but adds the complexity of controlling the large volume flow of combustion air.
Fig. 1. Dry Low NOx Combustor Source: note 6, Davis & Black, p. 3, Figure 5.
Fig. 2. Fuel-Staged Dry Low NOx Operating Modes Source: note 6, Davis & Black, p.4, Figure 6.
3.2.1.2 Lean Pre-Mixed Combustion GEs Dry Low NOx system, as described in note 7 (L.B. Davis), has been selected to illustrate the evolution and operation of a LPM staged combustion system in response to regulatory changes and efficiency upgrades. The GE Dry Low NOx (DLN-1) combustor shown in figure 1 is a two-stage premixed combustor designed for use with natural gas and capable of operation on liquid fuel. As shown, the combustion system includes four major components: fuel injector system, liner, venture, and cap/centerbody assembly. These components form two stages in the combustor. In the premixed mode, the first stage thoroughly mixes the fuel and air and delivers a uniform, lean, unburned fuel-air mixture to the second stage. The DLN-1 system operates in four distinct modes, illustrated in figure 2: Modes of Operation for the DLN-1 system • Primary – Fuel to the primary nozzles only. Flame is in the primary stage only. This mode of operation is used to ignite, accelerate, and operate the machine over low- to mid-loads, up to a preselected combustion reference temperature. • Lean-Lean – Fuel to both the primary and secondary nozzles. Flame is in both the primary and secondary stages. This mode of operation is used for intermediate loads between two pre-selected combustion reference temperatures. • Secondary – Fuel to the secondary nozzle only. Flame is in the secondary zone only. This mode is a transition state between lean-lean and premix modes. This mode is necessary to extinguish the flame in the primary zone, before fuel is reintroduced into what becomes the primary premixing zone. • Premix – Fuel to both primary and secondary nozzles. Flame is in the secondary stage only. This mode of operation is achieved at and near the combustion reference temperature design point. Optimum emissions are generated in premix mode. At loads less than 20 percent of baseload, NOx and CO emissions from the DLN-1 were similar to those from standard (diffusion) combustion systems. Other OEM’s offer similar systems, with the notable exception being Alstom. Alstom’s sequential combustion DLN technology was developed originally by ABB for the GT24 and GT26 gas turbines. Combustion takes place in the primary DLN combustor (EVtm) followed by fuel addition in a second (SEVtm) combustion chamber located aft of the first row of turbine blades. This DLN technology was commercialized in 1997 and applies the thermodynamic reheat principal. The sequential combustion provides low NOx emissions due to the fact that the SEVtm combustor does not contribute to NOx production.10 The OEM’s continually strive to improve performance while complying with increasingly restrictive emissions requirements. As F-technology gas turbines became available in the late 1980s with their higher firing temperatures, the OEMs were forced to redesign their DLN systems to maintain emissions at acceptable levels (~25ppmvd). Studies conducted by GE concluded that air usage in the combustor other than for mixing with fuel would have to be strictly limited. A design that repackaged DLN-1 premixing technology but eliminated the venture and centerbody assemblies that required cooling air was implemented and called DLN-2. The DLN-2 combustion system is a single-stage dual-mode combustor that can operate on both gaseous and liquid fuels. On gas, the combustor operates in a diffusion mode at low loads (< 50 percent load) and in a premixed mode at higher loads. Oil operation on the DLN-2 combustor is in the diffusion mode across the entire load range, with diluent injection used for NOx control. The DLN-2 combustor system has a single burning zone formed by the liner and the cap face. In low emissions operation, 90 percent of the gas fuel is injected through radial gas injection spokes in the premixer, and combustion air is mixed with the fuel in tubes surrounding each of the five fuel nozzles. The premixer tubes are part of the cap assembly. The fuel and air are thoroughly mixed, flow out of the five tubes at high velocity and enter the burning zone, where lean, low NOx combustion occurs. The vortex breakdown from the swirling flow exiting the premixers, along with the sudden expansion in the liner, are mechanisms for flame stabilization. In the early 1990s, continued regulatory pressures led the OEMs to develop 9 ppm combustion systems. During this time, GE introduced the DLN-2.6 combustor for the Frame 7FA machine which allowed for approximately 6 percent additional air to pass through the premixers in the combustor. The change in air splits was accomplished through reductions in cap and liner cooling air flows, requiring increased cooling effectiveness. A key feature of the DLN-2.6 combustor was the addition of a sixth burner, located in the center of the five DLN-2 burners. By fueling the center nozzle separately from the outer nozzles, the f/a ratio could be modulated relative to the outer nozzles. Another key feature of the DLN-2.6 combustor was the elimination of the diffusion mode, which required additional loading and unloading strategies. GE’s H systemtm combustor called DLN-2.5 uses a simplified combustion mode staging scheme to achieve low emissions over the premixed load range. The most significant feature associated with this variant is that there are only three combustion modes: diffusion, piloted premix, and full premix.11 The modifications required to reduce the emissions from the Siemens Westinghouse 15 ppm DLN combustor to 9 ppm are predominately the use of a premixed pilot and support housing design changes.12
3.2.1.2-6 LPM Technological Challenges CO/NOx Tradeoff Since the optimum flame temperature of a LPM combustor is designed to be near the lean flammability limit, LPM combustor performance is characterized by a CO/ NOx tradeoff (figure 3).13 At the combustor design point, both CO and NOx are below target levels; however, deviations from the design point flame temperature cause emissions to increase. A reduction in temperature tends to increase CO emissions due to incomplete combustion. Conversely, an increase in temperature will increase thermal NOx formation.
220
Bill Bender The CO/NOx trade-off must be addressed during part-load operation when the combustor is required to run at an even leaner condition overall. The tradeoff also comes into play in development efforts to reduce LPM combustor NOx emissions by further reducing the primary zone design point temperature.
Cold Ambient Temperatures Below an OEM’s specified ambient temperature threshold (0 F or -20 F), emission rate warranties are generally not offered. LPM turbines have demonstrated reliable operation below the ambient temperature warranty level, but the impact on the emissions signature is not well documented. Emissions below the threshold can be impacted by specific engine control adjustments needed to maintain combustion stability and prevent oscillations. The most commonly deployed control system is to increase the pilot fuel level, which increases NOx and CO emissions. As a solution to extending the ambient temperature warranty range, OEMs are collecting cold ambient data in an effort to improve designs. In the interim, many OEMs offer the option to incorporate an inlet air heating system as part of the design package.
Fig. 3. NOx Production Rate Source: note 4, Pavri & Moore, p. 17, Figure 20.
Low and Transient Load Operation Low load or transient load events can affect the emissions performance of LPM gas turbines because of engine controls required to prevent combustor flameout. To prevent the formation of NOx, LPM combustors are designed to operate close to engine flameout temperatures when compared to conventional combustors. When load is reduced to a low level or increased/decreased rapidly, it is necessary to augment combustor flame stability to prevent flameout. Most OEMs augment combustor flame stability through a fuel distribution adjustment such as the addition of pilot fuel. The addition of pilot fuel creates a diffusion flame, which increases NOx, CO, and VOC emissions. When operating at sustained low load conditions, CO emissions may increase significantly as a result of incomplete combustion. Due to the lower temperatures in a LPM combustor at low loads and the introduction of pilot fuel, a rich stoichiometric fuel mixture results accompanied by incomplete combustion.
3.2.1.2-7 LPM Future Developments Manufacturers continue to develop LPM gas turbine combustion as the preferred approach to meet future emission requirements.14 From a life cycle cost perspective, preventing pollutant formation has been shown to be more cost effective than the use of postcombustion techniques. Work is in progress using the latest experimental and analytical tools to improve emissions and operating flexibility of LPM systems.
Combustor Liner LPM combustor liner cooling methods can have a significant effect on emissions. The current generation of LPM combustors employs a variety of liner cooling methods including film cooling (louver or effusion) and backside cooling. Many first generation LPM turbines use film cooling to maintain acceptably low combustor wall temperatures, but many manufacturers have or will make the transition to backside-cooled technology with their next generation of LPM turbines. Backside-cooled liners have been in use for some commercial products for several years. Compared to film cooling, backside cooled liners forego cooling air injection completely. Instead, combustor wall temperatures are controlled solely through convective cooling by a high velocity airstream on the cold side of the liner. In most instances, the high heat flux from the flame requires augmenting the backside convective process to keep the liner wall temperatures from becoming excessive. Turbulators in the form of trip strips, fins, and pins act to increase the cooling flow turbulence at the liner wall and augment the heat removal process. Those OEMs already utilizing backside cooling will optimize its design in order to warranty lower NOx levels.
221
3.2.1.2 Lean Pre-Mixed Combustion Fuel Injectors Incorporating LPM combustion into gas turbines also required significant change to the fuel injectors. LPM fuel injectors are significantly larger than conventional injectors due to the higher air flow through the injector swirlers and the required volume of the premixing chamber used to mix fuel and air. Both axial and radial swirlers have been used by the OEM’s to swirl the premix air. Fuel is injected either through the swirler vanes or fuel spokes. Most LPM fuel injector designs include a pilot fuel injection point. A pilot flame is used to stabilize engine operation during load transients and low load operation. Fuel injectors are being optimized to improve f/a mixing and reduce local hot spots while improving flame stability. Achieving an optimum f/a temperature profile exiting the injector is essential. Design modifications to the fuel injection points and the air swirler are being investigated both computationally and experimentally.
Combustor Air Management Several combustor air management systems have been employed in LPM combustion systems to avoid combustor flameout and expand the low emissions operating range. Each technique ultimately provides control of the primary zone airflow to maintain the primary zone f/a ratio near its optimum low emissions level during part-load operation.
Casing air bleed Some two-shaft gas turbines used for gas compression and mechanical drive bleed air from the combustor casing at part load. The case bleed method of variable geometry has proven effective in controlling CO emissions. A consequence of air bleed, however, is deterioration in engine part-load thermal efficiency, since compressed bleed air no longer enters the turbine section of the engine to produce power.
Inlet guide vanes Some gas turbines used for power generation maintain optimum primary zone f/a ratios by modulating the compressor inlet guide vanes (IGV). Closing the IGVs reduces the airflow through the engine compressor and combustor. Regulating IGVs for singleshaft engines to control combustor airflow has a very small reduction in part-load thermal efficiency.
Combustor/Injector staging To enhance stability, some LPM turbines use fuel injection in multiple axial stages, with airflow to the additional stages being variable. Other LPM designs use multiple injector heads fired as a function of load.
Control Systems The control system for LPM engines modulates the air and fuel management systems to keep the combustion primary zone temperature within a specified range while maintaining acceptable engine turn-down and low-load operating stability. Accurate control of the primary zone temperature is critical to controlling NOx and CO emissions, which is typically accomplished through power turbine inlet temperature as an indirect measurement of the combustor exit or turbine inlet temperature.
3.2.1.2-8 Dual Fuel Operation Many gas turbine installations require operation on both gaseous and liquid fuels without affecting operability or environmental performance.15 Liquid fuels are more difficult to mix and pose difficulties in achieving homogeneous f/a mixture distribution that is required for low- NOx combustion. Testing has shown that the pre-mixer enabled comparable environmental performance with both natural gas and No. 2 diesel fuel at representative temperatures and pressures.
3.2.1.2-9 Fuel Variability Concerns The Advanced Turbine Systems (ATS) program has produced significant advances in combustion technology.16 Careful development of premix systems allows state-of-the-art combustors to operate with NOx levels approaching single digit performance. Although this progress is notable, reliable attainment of ultra-low emissions is contingent upon tight control of manufactured components, engine operating parameters, and fuel specifications. Failure to operate a premix combustor within planned specifications can lead to problems that range failure to meet emissions targets to hardware failure caused by flashback or oscillating dynamics. This sensitive behavior presents a challenge for expanded deployment of low-emission combustors that have been optimized almost exclusively for operation on natural gas fuel. The desire to operate such combustors on various fuels or in unique engine cycles (e.g., highly humidified cycles, biomass gasification combined cycles) poses a new set of constraints for low-emission operation.
222
Bill Bender Hydrogen Pathways to “zero” emissions power plants include the utilization of hydrogen directly as a fuel. Because of its high combustion temperature, the development of hydrogen-fueled turbines with comparable performance to natural gas is problematic. Stable, efficient, low- NOx combustion requires rapid, homogeneous mixing of fuel and air, which is a challenge when firing natural gas and made far more difficult with highly reactive hydrogen.
Medium Heating Value Fuel A common source for medium heating value fuel (200 to 800 Btu/scf) is oxygen-blown gasification of coal or residual oil. Because these gases are “manufactured” from other fuels, they are commonly referred to as synthesis gas, or syngas fuels and typically contain significant quantities of H2 and CO. Compared to natural gas, the stoichiometry of these gas mixtures requires a smaller volume of air for complete combustion, producing higher flame temperatures. As a further complication, H2 has a very high flame speed and very short ignition delay. Thus it is very difficult to avoid flashback or autoignition in a premixed burner. The standard approach to premixing is unlikely to work for these fuels.
Low Heating Value Fuel For gas turbines, the usual source of low heating-value fuel (~200 Btu/scf) is air-blown gasification of coal or biomass gasification. A significant feature of low-heating value fuels is that they often contain ammonia which can greatly complicate NOx reduction. Because of the high dilution level, these fuels have lower flame temperatures and lower flame speeds than natural gas or medium heating-value fuels. From the standpoint of thermal NOx emissions, this is an advantage. Because the volume of fuel flow is so great, the combustor aerodynamics are significantly affected and must be re-designed. This may impact low-emission backup operation on conventional fuels. CO oxidation is also a concern.
3.2.1.2-10 Background Information Regulatory Overview
Clean Air Act Congress enacted the Clean Air Act (CAA) in 1970 to address growing concerns over the nation’s air quality.17 As a result of the Act, national ambient air quality standards (NAAQS) were established for criteria pollutants. Areas of the country that exceed the NAAQS are considered “non-attainment for that pollutant. The U.S. regulatory structure imposes more stringent air pollution control programs in non-attainment areas. The CAA is a Federal law covering the entire country; however, States and local governments are allowed to develop and implement more stringent air pollution rules than those mandated by the CAA. The CAA also established New Source Performance Standards (NSPS) for a variety of potential emissions sources, including gas turbines. NSPS
Emissions control requirements for oxides of nitrogen were first applied to gas turbines by the Los Angeles County Air Pollution Control District (LAAPCD) and the San Diego Air Pollution Control District (SDAPCD) in the early 1970’s. To comply with these regulations, water was injected into the combustor flame zone to reduce flame temperature. The consequent reduction in NOx amounted to about 40 percent when half as much water as fuel was injected into the reaction zone. The emission level achieved was approximately 75 ppmvd (parts per million by volume, dry) on oil. These results and other data were used by the U.S. EPA to develop New Source Performance Standards that went into effect in September 1979. The details of this NSPS are presented in 40CRF60 Subpart GG. Turbines with heat input over 10 million Btu/hr, generating less than 30 MW electrical output, and supplying less than one-third of their electrical output to an electric utility, are required to meet a NOx emission standard of 150 ppm, corrected for efficiency. Emergency turbines are exempt from this standard, as are certain other types of turbines. Electric utility turbines with a heat input above 100 million Btu/hr must comply with a NOx standard of 75 ppm, corrected for efficiency. Most turbines available today can achieve NOx emissions of 25 to 42 ppm or less without post-combustion controls. Thus, the existing NSPS is not typically a controlling regulation for gas turbines. In July 2004, EPA updated the NSPS for gas turbines in a direct final ruling. Most notably, the revised standards require new LPM turbines that commence construction after July 8, 2004 to use a NOx continuous emissions monitoring system (CEMS) or, owners can continuously monitor engine parameters that indicate when the turbine is out of LPM combustion mode. On February 18, 2005, EPA proposed standards of performance for new stationary gas turbines in 40CFR60, subpart KKKK. The new standards would reflect changes in NOx emissions control technologies and turbine design and are intended to bring the emission limits up to date with the performance of current combustion turbines.
223
New Source Review In addition to AAQS, the CAA established an air permitting program called New Source Review (NSR). NSR is divided into two primary programs: Prevention of Significant Deterioration (PSD) and Non-Attainment NSR. Each program applies to “major sources” and “major modifications,” but in non-attainment NSR, a source can be considered “major” at much lower thresholds.
3.2.1.2 Lean Pre-Mixed Combustion Best Available Control Technology (BACT) If a source triggers PSD review, then the owner must identify the appropriate level of emissions controls for pollutants that exceed specific thresholds. In attainment areas, the standard for evaluation is Best Available Control Technology (BACT). A BACT determination establishes an achievable emissions limitation taking into account the energy, environmental and economic impacts of applying the pollution control technology necessary to meet that limitation. EPA now requires that a “top-down” BACT analysis be followed for all PSD permit applications.18 As a result of its structure, BACT is a “living” standard, which becomes more stringent over time as new, more efficient, and lower cost control technologies become available. BACT for a specific application can only be defined in the context of current demonstrated technology and for a specific application. Because BACT determinations are site-specific, generic BACT requirements do not exist; however, there have been recent BACT NOx determinations as low as 2 to 5 ppm for gas turbines greater than 25MW. A recent search of EPAs RACT/BACT/LAER Clearinghouse data (years 2000 to 2005) for natural gas-fired combustion turbines greater than 25 MWs has resulted in the following NOx emission limits: Table 2: NOx emission limits.
Number of Permits 37 51 106 11 117 10 5
NOx Emission Limit (ppm) <9.0 9.0 <15.0 15.0 <25.0 25.0 >25.0
This table confirms the trend towards single digit NOx emissions. For certain gas turbines, NOx emission rates > 9 ppm can be met using LPM combustors. Permit limits less than 9 ppm require the application of post combustion controls. Lowest Achievable Emission Rate (LAER) Major sources/modifications in non-attainment areas are subject to a determination of LAER. LAER is defined as the most stringent emission limitation achieved in practice for the class or source category. The primary difference between BACT and LAER is that LAER does not allow economic impacts to be considered when evaluating pollution control technologies leading to an emissions limitation. Other Regulatory Programs Other regulatory programs that could potentially impact gas turbine emissions include the Clean Air Interstate Rule, the President’ s Clear Skies Initiative, and Maximum Available Control Technology (MACT) standards.19
3.2.1.2-11 Nitrogen Oxide Formation Tropospheric ozone has been and continues to be a significant air pollution problem in the U.S. and is the primary constituent of smog.20 Large portions of the country do not meet the ozone National Ambient Air Quality Standard (NAAQS) and thereby expose large segments of the population to unhealthy levels of ozone in the air. NO2 reacts in the presence of air and ultraviolet light (UV) in sunlight to form ozone and nitric oxide (NO). The NO then reacts with free radicals in the atmosphere, which are also created by the UV acting on volatile organic compounds (VOC). The free radicals then recycle NO to NO2. In this way, each molecule of NO can produce ozone multiple times. This may continue four or five times until the VOCs are reduced to short chains of carbon compounds that cease to be photo-reactive. In addition to the NO2 and ozone concerns, NOx and sulfur oxides (SOx) in the atmosphere are captured by moisture to form acid rain. Acid rain impacts certain ecosystems and some segments of our economy. All of these facts indicate the need to reduce NOx emissions, but to do so requires an understanding of the generation and control mechanisms. According to the Zeldovich equations, NO is generated to the limit of available oxygen (about 200,000 ppm) in air at temperatures above 1300C (2370F). At temperatures below 760C (1,400F), NO is either generated in much lower concentrations or not at all. There are two mechanisms by which NOx is formed in gas turbine combustors: 1. 2.
The oxidation of atmospheric nitrogen found in the combustion air (thermal NOx and prompt NOx), and The conversion of nitrogen chemically bound in the fuel (fuel NOx).
Thermal NOx Thermal NOx is formed by a series of chemical reactions in which oxygen and nitrogen present in the combustion air dissociate and subsequently react to form NOx. Prompt NOx, a form of thermal NOx, is formed in the proximity of the flame front as intermediate combustion products such as HCN, N and NH that are oxidized to form NOx. Prompt NOx is formed in both fuel-rich flames zones and
224
Bill Bender dry low NOx (DLN) combustion zones. The contribution of prompt NOx to overall NOx emissions is relatively small in conventional near-stoichiometric combustors, but this contribution is a significant percentage of overall thermal NOx emissions in DLN combustors. For this reason, prompt NOx becomes an important consideration for DLN combustor designs, establishing a minimum NOx level attainable in lean mixtures. The chemical mechanisms that produce NOx are listed below. These reactions represent the major pathways for NOx formation; see Nicol et al. for a more detailed description of the chemical pathways.21 Various authors have used different names for these pathways, or include different reactions. This is a result of advances in understanding the relative importance of these mechanisms. For example, until recently, the nitrous oxide path was simply included as an extension of the prompt mechanism22, but has emerged as an important chemical path in lean burning gas turbines and is now referred to as a distinct mechanism: Extended Zeldovich mechanism: (1) O + N2 ↔ NO + N (2) N + O2 ↔ NO + O (3) N + OH ↔ NO + H Nitrous oxide: (4) (5) (6) Prompt: (7)
N2 + O + M ↔ N2O + M N2O + O ↔ NO + NO N2O + H ↔ NO + NH N2 + CH ↔HCN + N
The prompt mechanism is followed by a sequence of reactions converting HCN to NO; reaction (7) is just the initiation. The detailed sequence was reported by Fenimore, and the prompt mechanism is sometimes referred to as “Fenimore-prompt” or just “Fenimore”.23 The CH reaction is also important for fuels containing nitrogen which can directly form the HCN species. The extended Zeldovich mechanism is also known as the thermal mechanism when O and H species are at equilibrium levels. The thermal route is a primary mechanism for NOx when flame temperatures are above approximately 1800K (2780F). Below this temperature, the thermal reactions are relatively slow. Thus, a common approach to NOx control is to reduce the combustion temperature so that very little thermal NOx can form. In the absence of thermal NOx, the other mechanisms become significant. Non-equilibrium concentration of O or H atoms in the flame region can produce NOx via reactions (1) to (3), and this is known as Zeldovich NOx. The nitrous oxide path depends on the intermediate species N2O which itself is generated by O-atom attack of nitrogen.
Fuel NOx Fuel NOx is formed when fuels containing nitrogen are burned. Molecular nitrogen, present as N2 in some kinds of natural gas, does not contribute significantly to fuel NOx formation. Some low-Btu synthetic fuels contain nitrogen in the form of ammonia (NH3). Other low-Btu fuels such as sewage and process waste-stream gases also contain nitrogen. When these fuels are burned, the nitrogen bonds break and some of the resulting free nitrogen oxidizes to form NOx. With excess air, the degree of fuel NOx formation is primarily a function of the nitrogen content in the fuel. The fraction of fuel-bound nitrogen (FBN) converted to fuel NOx decreases with increasing nitrogen content, although the absolute magnitude of fuel NOx increases. For example, a fuel with 0.01 percent nitrogen may have 100 percent of its FBN converted to fuel NOx, whereas a fuel with a 1.0 percent FBN may have only a 40 percent conversion rate. Natural gas typically contains little or no FBN. As a result, when compared to thermal NOx, fuel NOx is not a major contributor to overall NOx emissions from stationary gas turbines firing natural gas.
3.2.1.2-12 Conclusions OEMs continue to improve LPM technology; simultaneously, regulators continue to lower emissions requirements.24 R&D efforts continue to advance technology and provide valuable contributions to design and manufacturing techniques to further enhance performance while reducing emissions and overall plant costs. Leveraging advances made in natural gas-fueled turbines through the ATS Program is critical to achieving performance goals established for future coal-based systems, especially Integrated Gasification Combined Cycle (IGCC) plants and FutureGen. Gas turbines utilized in IGCC plants operate on syngas derived from gasification. Syngas typically contributes 15 to 20 percent to the volumetric flow through an advanced gas turbine to achieve the same heat input as natural gas. The additional mass flow theoretically increases gas turbine power output by 30 to 40 percent. However, aerodynamic issues currently limit power gains to values lower than those theoretically possible. DOEs Fossil Energy Turbine Technology R&D Program being implemented by NETL was recently expanded with the selection of ten new projects valued at $130 million. The new program will advance turbines and turbine subsystems for integrated gasification combined cycle (IGCC) power plants and address the use of hydrogen and syngas.
225
3.2.1.2 Lean Pre-Mixed Combustion 3.2.1.2-13 Notes _____________________________ 1. U.S. Department of Energy Office of Fossil Energy, NETL, Turbine Program, “Enabling Near-Zero Emission Coal-Based Power Generation,” October 2005. 2. Federal Register: February 18, 2005 (Volume 70, Number 33), pages 8314 – 8332. 3. Ibid. 4. R.Pavri and G.D. Moore, GE Power Systems, Gas Turbine Emissions and Control, GER-4211, 03/01 5. G.A. Richards, M.M. McMillan, R.S. Gemmen, W.A. Rogers, and S.R. Cully, “Issues for Low-emission, Fuel-flexible Power Systems” (U.S. Department of Energy); A. Biro, “Effect of flame-temperature on NOx emission during natural gas firing,” (Hungarian Academy of Sciences, 1998). 6. L.B. Davis and S.H. Black, “Dry Low NOx Combustion Systems for GE Heavy-Duty Gas Turbines,” GER-3568G: GE Power Systems, October 2000; R. Eldrid, L. Kaufman, and P. Marks, “The 7FB: The Next Evolution of the Gas Turbine,” GER-4194: GE Power Systems April 2001; R.D. Brdar and R.M. Jones, “GE IGCC Technology and Experience with Advanced Gas Turbines,” GER-4207: GE Power Systems, October 2000; F.J. Brooks, “GE Gas Turbine Performance Characteristics,” GER-3567H: GE Power Systems, October 2000; L.B. Davis, “Dry Low NOx Combustion for GE Heavy Duty Gas Turbines,” GER-3568A: GE Power Generation, 1983. 7. See note 4 (Pavri & Moore) and (Davis); also see note 5 (Richards and others). 8. B. Rising, letter to U.S. EPA, February 2005. 9. See note 6 (Davis & Black, 2000). 10. Alstom Power, Sequential Combustion, http://www.power.alstom.com/home/equipment_systems/turbine/gas_turbines/ GT24_andGT26__188MW_and_281MW.html. 11. R.K. Matta, G.D. Mercer, and R.S. Tuthill, “Power Systems for the 21st Century – “H” Gas Turbine Combined-Cycles,” (GE Power Systems, October 2000). 12. R. Bland, et al, “Siemens W501F Gas Turbine: Ultra Low NOx Combustion System Development,” (Siemens Westinghouse Power Corporation, 2004). 13. G. A. Richards, K.H. Casleton, D. J. Maloney, and R.S. Gemmen, “Federal Energy Technology Center, Addressing the Challenge of Low-Emission Combustion” (Department of Energy); also see notes 4 (Pavri & Moore) and (Davis) and note 5. 14. See notes 5 and 13. 15. See note 6 (Brdar & Jones) and (Davis); also see notes 5 and 13. 16. U.S. Department of Energy, Office of Fossil Energy, NETL, Advanced Turbine Systems, Advancing the Gas Turbine Power Industry; also see notes 2, 5 and 13. 17. U.S. Environmental Protection Agency, Office of Air Quality, Technical Bulletin EPA 456/F-pp-006R, “Nitrogen Oxides (NOx): Why and How They Are Controlled,” November 1999; U.S. Environmental Protection Agency, Office of Air Quality, AP 42, 5th ed. Compilation of Air Pollutant Emission Factors Vol.1: Stationary Point and Area Sources, Sec. 3.1; U.S. Environmental Protection Agency, Office of Air Quality, Technology Transfer Network, Clean Air Technology Center, RACT/BACT/LAER Clearinghouse; Energy and Environmental Analysis, Inc., Database on State Permitting Issues –Air Regulations (http://www.eea-inc.com/rrdb/DGRegProject/RegBack.html); N.H. Hydari, A.A. Yousuf, H.M. Ellis “Comparison of the Most Recent BACT/LAER Determinations for Combustion Turbines by State Air Pollution Control Agencies, June 2002; California Environmental Protection Agency, Air Resources Board, Report to the Legislature, Gas-Fired Power Plant NOx Emission Controls and Related Environmental Impacts, May 2004; ONSITE SYCOM Energy Corp., Cost Analysis of NOx Control Alternatives for Stationary Gas Turbines, 11/05/99. 18. J. Calcagni, U.S. Environmental Protection Agency, Memorandum dated June 13, 1989, Transmittal of Background Statement on “Top Down” BACT. 19. A. Jones, L. Witherspoon, and L. Cowell, Solar Turbines, Inc., Meeting Regulatory Challenges Through Advances in Gas Turbine Emission Control, 2004. 20. See note 7 (Davis & Black), (Pavri & Moore) and (Davis); also see note 17 (EPA Tech Bulletin, Nov. 1999), (CA EPA), and (ONSITE Sycom); note 19. 21. D.G. Nicol, R.C. Steele, N.M. Marinov, and P.C. Malte, “The importance of the nitrous oxide pathway to NOx in lean premixed combustion” ASME Journal of Engineering for Gas Turbines and Power :1995. 22. C.T. Bowman, “Control of combustion generated nitrogen oxide emissions: technology driven by regulations” (Twenty-Fourth Symposium on Combustion, The Combustion Institute, Pittsburgh, PA 1992). 23. C.P. Fenimore, “Formation of nitric oxide in premixed hydrocarbon flames,” (The Thirteenth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA 1971). 24. See notes 1 and 16.
226
BIOGRAPHY
3.2.1.2 Lean Pre-Mixed Combustion
William R. Bender Technology & Management Services, Inc. Gaithersburg, MD 20879 phone: (301) 670-6390 x144 email: [email protected]
William R. Bender is a Senior Associate with Technology & Management Services, Inc. (TMS). Mr. Bender has over 34 years of experience with gas turbine power systems as a Project Construction Manager, Project Manager, and Project Engineer. While with a large A/E, he was a Project Engineer on Florida Power & Light’s Martin combined cycle Project and Tampa Electric’s Polk Power Station. As a TMS employee, Mr. Bender is responsible for the planning and analyzing of fossil energy systems, policy initiatives and research, development and deployment programs in support of DOE Headquarters and NETL. His areas of expertise include coal, oil and gas power systems; technology and economic evaluation and assessments; energy and environmental policy analysis and multi-task project management.
3.2.1.3
Rich Burn, QuickMix, Lean Burn (RQL) Combustor
3.2.1.3-1 Introduction The Rich-Burn, Quick-Mix, Lean-Burn (RQL) combustor concept was introduced in 1980 as strategy to reduce oxides of nitrogen (NOx) emission from gas turbine engines.1 Later, in the 1990’s, the concept was targeted by the National Aeronautics and Space Administration (NASA) for the reduction of nitrogen oxides in next generation aero-propulsion engines. Today, the RQL is the anchor combustor technology in aeroengines deployed commercially by Pratt & Whitney under the name TALON (Technology for Advanced Low NOx). Due to safety considerations and overall performance (e.g., stability) throughout the duty cycle, the RQL is preferred over lean premixed options in aeroengine applications. In stationary applications, lean premixed combustor technology is the standard. Safety considerations are not as severe, the duty cycle is more constrained, and the reduction in NOx emission is more substantial in contrast to RQL technology. However, RQL combustor technology is of growing interest for stationary applications due to the attributes of (1) more effectively processing fuels of complex composition, and (2) processing fuels of varying composition. The latter is becoming of importance with the increasing international competition for fuels in general, the burgeoning interest in biomass fuels, the expanding use of “opportunity fuels” (land-fill gases, digester gases, well-head gases), and the growing use of liquefied natural gas to either complement domestic sources or serve as the sole source of natural gas to a large region of a country or the country as a whole. The California Energy Commission is engaged in RQL technology research, in cooperation with the U.S. Department of Energy, to explore the utility of RQL strategies as an alternative to combustors for niche applications in the stationary production of electrical power. The RQL concept is predicated on the premise that the primary zone of a gas turbine combustor operates most effectively with rich mixture ratios (Figure 1). First, a “rich-burn” condition in the primary zone (e.g., Ф = 1.8) enhances the stability of the combustion reaction by producing and sustaining a high concentration of energetic hydrogen and hydrocarbon radical species. Secondly, rich burn conditions minimize the production of nitrogen oxides due to the relative low temperatures and low population of oxygen containing intermediate species (Figure 2).
Scott Samuelsen Professor of Mechanical, Aerospace, and Environmental Engineering Director Advanced Power and Energy Program University of California Irvine 92697-3550 phone: 949-824-5468 email: [email protected]
227
Fig. 1. Rich-Burn, Quick-Mix, Lean-Burn Combustor (Ф, Equivalence Ratio)
The effluent emanating from the rich primary zone will be high in the concentration of partially oxidized and partially pyrolized hydrocarbon species, hydrogen, and carbon monoxide. As a result, the effluent cannot be exhausted without further processing. In particular, the addition of oxygen is needed to oxidize the high concentrations of carbon monoxide, hydrogen, hydrocarbon intermediates. This is accomplished by injecting a substantial amount of air through wall jets to mix with the primary zone effluent and create a “lean-burn” condition prior to the exit plane of the combustor. Ideally, this will result in the emission of an effluent comprised of the major products of combustion (CO2, H2O, N2, O2) and a non-zero concentration of criteria pollutants (e.g., NOx, CO, HC).
Fig. 2. Nitric Oxide Formation
Figure 3 RQL Strategy
A major challenge for the RQL is the selection of combustor liner material. In the primary zone, for example, the use of air for cooling the liner wall is precluded in order to avoid the generation of near-stoichiometric mixture ratios and the associated production of nitrogen oxides in the vicinity of the wall. As a result, the temperature and composition of gases in the primary zone create a demanding, reducing environment for the liner material. The concentrations of hydrogen alone and the concomitant demands of hydrogen embrittlement in particular have combined to require a major investment in materials research in support of RQL technology. A more demanding challenge is the design of the Quick-Mix section. A key to the success of the RQL is the efficacy of mixing the air with the effluent exiting the primary zone. The mixing of the injected air takes the reaction through the conditions most vulnerable for the high production of oxides of nitrogen (near stoichiometric conditions where both the temperature and oxygen atom concentrations are elevated). The challenge then is to rapidly mix air into the rich-burn effluent in order to rapidly create the lean-burn conditions (Figure 3). As a result, the label “Quick-Mix” is adopted to emphasize the requirement to rapidly mix the air and primary zone effluent. As a result, RQL research has historically focused on Quick-Mix section designs to establish the most rapid mixing.
3.2.1.3-2 Quick-Mix Zone Numerous jet in crossflow studies have been conducted under non-reacting conditions to yield insight on such flow field characteristics as jet structure and penetration, jet entrainment of crossflow fluid, and the flow field distributions resulting from jet mixing.2 Heated jets or a heated mainflow, or the doping of either the jets or mainflow with a tracer have allowed the measurement of scalars in the flow downstream of the jet orifices in order to quantify the convective and diffusive mixing efficacy. Early studies of jets in crossflow were motivated by the aerodynamics associated with (1) vertical/short takeoff and landing (V/STOL) aircraft and (2) primary and dilution jets on conventional gas turbine combustors with a focus on jet trajectory, centerline decay, and jet shape for unconfined single jets.3 The interest in the Quick-Mix zone of the RQL combustor has constituted the focus of jet in crossflow studies over the past two decades. A single round jet in a crossflow is presented in Figure 4. The jet enters the crossflow and is deflected downstream in response to the momentum of the cross flow. A recirculation zone can form in the near-wall downstream wake of the jet. The radial extent of jet penetration is governed by the angle of the jet relative to the crossflow, and the entry momentum of the jet in contrast to the momentum of the crossflow. A variety of nonreacting experiments have been used to establish empirical correlations for the maximum penetration Fig. 4. Single Jet in Crossflow of a single jet. For example:4
228
Scott Samuelsen (1) Where: Ymax = Maximum Radial Penetration of the Jet Centerline dj = Diameter of the Jet Entry Orifice ρ jets ⋅ Vjets 2 J = Jet-to-Crossflow Momentum Flux Ratio = ρ main ⋅ Vmain 2 ρ = Density V = Velocity θ = Entry Angle of the Jet
(2)
In the gas turbine combustor, the jets are confined and the interaction between multiple jets is a major factor in dictating mixing behavior. As a result, studies has been conducted to address the mixing behavior associated with the mixing of primary and dilution jets in conventional gas turbine combustors; and optimizing the mixing section in the RQL combustor.5 For multiple jets in a tubular duct, the correlation for the maximum penetration of a single jet must account for the effects of blockage:6
(3)
Where: MR = Jet-to-Crossflow Mass Flow Ratio
MR is much higher for an RQL combustor (~ 2.5) in contrast to the conventional combustor (~ 0.25). Since the density and momentum-flux ratios J are about the same in the two configurations, the biggest difference between the jets in conventional and RQL combustors is orifice size. Non-reacting studies have also been undertaken to evaluate geometrical features (e.g., orifice shape, number of orifices, axial staggering of orifices) and operating features (e.g., momentum flux ratio, density flux ratio, mass flow rate ratio) with the goal to optimize the mixing. Traditionally, such studies have defined “optimal mixing” as the shortest axial distance from the upstream edge of the jet orifice where a uniform radial profile is established of key mixing parameters (e.g., temperature, species concentration). The hypothesis is that the optimal mixing defined in this manner will minimize the production of nitrogen oxides. Due to the complex set of variables, many of the studies have benefited by a design of experiments statistical approach to explore the multiple factors that can affect jet mixing.7 In addition to non-reacting experiments and use of design of experiments methods, modeling has been effectively employed both independent of and in conjunction with the experiments.8 While a variety of jet orifice configurations has been studied (e.g., triangular, slanted, tear-drop), no option has been identified that penetrates significantly farther or faster than a single, round jet. For a cylindrical configuration, a NASA design method developed by Holdeman and co-workers defined a correlation that is used to design the jet mixing section of an RQL combustor utilizing round hole jets.9 The correlation, derived a study of jet-to-mainstream momentum-flux ratio, establishes the number of circular holes for optimum mixing:
n=
π 2J C
Where:
(4)
n = Number of Circular Jet Orifices to Optimize Mixing J = Momentum Flux Ratio C = Empirical Constant = 2.5 Reacting flow studies have also been conducted to complement the non-reacting studies and assess the impact of heat release on the mixing processes. Typically, a mixture of propane and air is used to generate a representative rich-burn effluent. A specially designed section is used to create a uniform presentation (e.g., temperature, velocity, composition, concentration) of the rich effluent to the mixing section. The injection of the quick-mix jet air results in the ignition of a reaction between the rich-burn effluent and the jet air (Figure 5). Measurements of temperature, species composition, and species concentration can then be made downstream of the jet orifices in order to establish the efficacy of mixing as a function of downstream distance.10 The results from the reacting experiments reveal that the non-reacting experiments provide a satisfactory description of the mixing of jets in a crossflow. Overall, the jets need to penetrate to the half radius in order to maximize the mixing and avoid either under-penetration or over-penetration.11
229
3.2.1.3 Rich Burn, Quick-Mix, Lean Burn (RQL) Combustor
Fig. 5. Laboratory Model Combustor
Jet mixing in a crossflow has been studied in two primary mainstream geometries. The cylindrical geometry has been the most extensively researched and is directly relevant to combustor can configurations. In contrast, the modern annular combustor configurations have spawned investigations of jets in the crossflow of rectangular geometries. For each, Holdeman has established the following procedures to design the most rapid mixing, Quick-Mix section:12 Cylindrical Geometry 1. 2. 3.
Typical Mass Ratio: Typical Momentum Flux Ratio: Optimal Number of Orifices:
2.5 60
n= 4.
π 2J C
(5) Orifice Size: Determined by the desired mass-flow ratio and the optimum number of orifices for the given momentumflux ratio.
Rectangular Geometry 1. 2. 3.
Typical Mass Ratio: Typical Momentum Flux Ratio: Optimal Orifice Spacing:
Where:
4. 5.
2.5 60
S/ H = C/ J
(6)
S = Orifice Spacing H = Channel Height
Orifice Size. For a given momentum-flux ratio, determined by the desired mass-flow ratio and the optimum orifice spacing. For a rectangular duct the number of orifices is infinite. For an annulus, the number of orifices will depend on the diameter and height of the mixing section. Orifice Configuration. Can be either in-line or staggered. The selection will depend on the application, and include such factors as momentum-flux ratio. In-line configurations are usually preferred as the orifices are smaller. The optimum spacing for staggered jets is four times the optimum spacing for in-line configurations. As a result, the orifice diameter for staggered jets must be doubled for the same total orifice area.
3.2.1.3-3 Formation of Nitrogen Oxides The hypothesis that the optimization of mixing in the Quick-Mix section will minimize the production of nitrogen oxides begs the question of proof. The first exploration of this relationship utilized the results of non-reacting experiments, and superimposed analytically the kinetics of NOx production. The results suggested that the best mixer may not necessarily minimize NOx emissions.13 In more recent research (Figure 6), oxides of nitrogen have been measured with the finding that the module designed to optimize mixing (12 holes) produces 15% more NOx than an 8-hole module (Figure 6b).14 These results also suggest that an aerodynamically “optimum” mixer may not minimize NOx. It is noteworthy that the data for all modules show that high concentrations of NOx occur in the wakes of the jets adjacent to the wall. This suggests that a significant production of NOx likely occurs in the wakes of the jets, downstream of the orifices.
230
Scott Samuelsen The research study reported in Figure 6 explored as well the effect of air preheat on NOx formation. The main air and jet air streams were independently heated in order to assess the relative influence of independently preheating each flow. Three preheat conditions are presented in Figure 7. The first set of conditions is for no air preheat. This serves as an anchor to which the results for the elevated inlet air temperature conditions can be compared. The second set of conditions is for jet air preheat only (no main air preheat). The third set of conditions is for both jet air and main air preheat, representing the case usually encountered in practical combustors. The results reveal the small impact of preheated jet air on NOx. The jet air comprises over 70 percent of the total air flow, but preheating only the jet air results in relatively small increases in NOx emissions compared to the case where both the main and jet air are preheated. The latter condition resulted in the largest NOx production for all the modules. The small effect of preheating the jet air is counter intuitive to the expectation that preheating jet air should promote NOx production via the thermal (Zeldovich) mechanism. The dominating influence of the main air preheat may be attributed to the total fixed nitrogen (TFN) production in the fuel-rich zone. In particular, the TFN generation in the fuel-rich zone, and its subsequent transformation to NOx in the mixing zone, may be influential in governing the total NOx emissions than expected.
Fig. 6. Composite NOx Emissions Data
Fig. 7. Effect of Air Preheat on NOx Concentrations
Source: See note 17.
Source: See note 17.
The distributions of equivalence ratios reveal that preheat has a negligible effect on jet penetration. The equivalence ratio distributions are quite uniform which is also expected as the 12 hole configuration as an optimum mixer. In addition, the O2 distributions serve as an indice for jet penetration and are virtually the same for all preheat conditions.
3.2.1.3-4 Conclusions The Rich-Burn, Quick-Mix, Lean-Burn (RQL) combustor has evolved over the past three decades as a major strategy for the reduction of oxides of nitrogen from gas turbine engines. The concept has the attribute of high combustor stability due to the rich primary zone. While the RQL is deployed commercially in aeroengine applications, lean premixed options have been selected for stationary applications in lieu of the RQL in order to achieve lower NOx emission. Niche applications in the stationary market, however, are driving a role for the RQL where fuels with complex compositions or fuels of varying composition are being encountered. This has prompted new research in the exploration of NOx formation in RQL configurations. The hypothesis that optimal mixing in the QuickMix section will lead to the minimization of NOx emission has been challenged by recent observations. In particular, the generation of nitrogen containing species in the Rich-Burn zone and subsequent processing in the Quick-Mix section may affect the emission of NOx. While the RQL concept is inherently a low-NOx generator, a further understanding of the primary zone chemistry and the coupling between the chemical kinetics and fluid mechanics in the Quick-Mix section may be required in order to optimize the RQL design. Fuels of varying composition and varying concentrations of fuel-bound nitrogen in stationary applications create a particular demand for this insight whereas the consistency of fuel composition in aeroengine applications allows insight derived from empirical evidence to be sufficient for the design of commercial RQL systems.
231
3.2.1.3 Rich Burn, Quick-Mix, Lean Burn (RQL) Combustor 3.2.1.3-5 Notes _________________________________ 1. Mosier, S.A., and Pierce, R.M., 1980. “Advanced Combustor Systems for Stationary Gas Turbine Engines, Phase I. Review and Preliminary Evaluation,” Volume I, Contract 68-02-2136, FR-11405, Final Report, U.S. Environmental Protection Agency. 2. Holdeman, J.D., 1993. “Mixing of Multiple Jets with a Confined Subsonic Crossflow,” Progress in Energy and Combustion Science, Vol. 19, pp. 31-70, August; Holdeman, J.D., Liscinsky, D.S., Oechsle, V.L., Samuelsen, G.S., and Smith, C.E., 1997. “Mixing of Multiple Jets With a Confined Subsonic Crossflow: Part I—Cylindrical Ducts,” Journal of Engineering for Gas Turbines and Power, Vol. 119, No. 4, October, pp. 852–862; Holdeman, J.D., Liscinsky, D.S., and Bain, D.B., 1999. “Mixing of Multiple Jets With a Confined Subsonic Crossflow: Part II—Opposed Rows of Orifices in a Rectangular Duct,” Journal of Engineering for Gas Turbines and Power, Vol. 121, No. 3, July, pp 551-562. 3. Margason, R.J., 1993. “Fifty Years of Jet in Cross Flow Research,” Presented at Computational and Experimental Assessment of Jets in Cross Flow, AGARD Conference Proceedings 534, April; Demuren. A.O: Modeling Jets in Cross Flow. NASA Contractor Report 194965. August 1994; 4. Lefebvre, A.H., 1999. Gas Turbine Combustion, Taylor and Francis. 5. See Note 2. 6. See Note 4. 7. Hatch, M.S., Sowa, W.A., Samuelsen, G.S., and Holdeman, J.D., 1995,“Jet Mixing into a Heated Cross Flow in a Cylindrical Duct: Influence of Geometry and Flow Variations,” Journal of Propulsion and Power, Vol. 11, No. 3, May-June, pp. 393–400; Kroll, J.T., Sowa, W.A., Samuelsen, G.S., and Holdeman, J.D., 2000. “Optimization of Orifice Geometry for Cross Flow Mixing in a Cylindrical Duct,” Journal of Propulsion and Power, Vol . 16, No. 6, November-December, pp 929-936. 8. See Note 2. 9. See Note 2. 10. Leong, M.Y., Samuelsen, G.S., and Holdeman, J.D., 1999. “Mixing of Jet Air With a Fuel-Rich, Reacting Crossflow,” Journal of Propulsion and Power, Vol. 15, No. 5, September-October, pp. 617-622; .Leong, M.Y., Samuelsen, G.S., and Holdeman, J.D., 2000. “Optimization of Jet Mixing into a Rich, Reacting Crossflow,” Journal of Propulsion and Power, Vol. 16, No. 5, September-October, pp. 729-735; Demayo, T.N, Leong, M.Y, Samuelsen, G.S., and Holdeman, J.D., 2003. “Assessing Jet-Induced Spatial Mixing in a Rich, Reacting Crossflow,” Journal of Propulsion and Power, Vol. 19, No. 1, January-February, pp. 14-21. 11. See Notes 7 and 10. 12. See Note 2. 13. See Note 7. 14. Samuelsen, G.S., Brouwer, J., Holdeman, J.D., Vardarkas, M.A, and Leong, M.Y., 2006, “The Effect of Air Preheat and Number of Orifices on Flow and Emissions in an RQL Mixing Section,” submitted for publication (see also NASA TM1999-209431, 1999).
232
BIOGRAPHY
3.2.1.1 Conventional Type Combustion 3.2.1.3 Rich Burn, Quick-Mix, Lean Burn (RQL) Combustor
Scott Samuelsen Professor of Mechanical, Aerospace, and Environmental Engineering Director Advanced Power and Energy Program University of California Irvine 92697-3550 phone: 949-824-5468 email: [email protected]
Professor Scott Samuelsen is Director of the Advanced Power and Energy Program (APEP) at the University of California Irvine and Professor of Mechanical, Aerospace, and Environmental Engineering. He directs as well the National Fuel Cell Research Center (NFCRC) and the UCI Combustion Laboratory (UCICL). His research is directed to advanced power systems including gas turbines, fuel cells, and fuels. He directs anchor research on advanced coal and natural gas power plants for the coproduction of electricity and hydrogen for the U.S. Department of Energy (DOE), distributed generation and information technology research for the U.S. Department of Defense (DoD) in support of energyefficient and environmentally-responsible power generation, advanced energy systems research for the California Energy Commission, and coal-gas and hydrogen-fueled gas turbine combustion studies. His energy expertise is based on forty years of combustion research working with strategic alliances involving industry with applications to gas turbine propulsion, gas turbine electronic power generation, and combustion distributed generation resources. He holds the Ph.D. degree from the University of California Berkeley.
3.2.1.4.1
Trapped Vortex Combustion
3.2.1.4.1-1 Trapped Vortex Combustion Benefits to IGCC Gas Turbines of Trapped Vortex Combustion • • • • • • •
The Trapped Vortex Combustion (TVC) technology has the potential to: Burn a wide variety of medium and low-BTU gases including hydrogen-rich gasified coal, biomass products, and landfill gas; Operate in a low NOx, lean premixed mode combustor environment on hydrogen-rich syngas to accommodate the high flame speed that is a characteristic of these fuels; Achieve extremely low NOx emissions without the added expense of exhaust gas after-treatment; Eliminate the costly requirement for high pressure diluent gas (nitrogen, steam or carbon dioxide) for NOx emissions control; Accommodate more types of gas turbines for IGCC applications by decreasing the mass flow through the turbine section; Improve the overall cycle efficiency of the gas turbine by decreasing the pressure drop through the combustor; and Extend the lean blowout limit offering greater turndown, (load following), with improved combustion and process stability.
3.2.1.4.1-2 The Challenges of IGCC Gas Turbine Combustion The Integrated Gasification Combined Cycle (IGCC) is emerging as a best available technology to utilize low quality energy resources, such as coal or oil, and meet emission limits not achievable by other conventional or advanced competing technologies. However, the success of the IGCC in the energy sector requires continuous enhancement in performance and reduction in capital costs. New, more efficient, gasification technologies are in demonstration; hot gas cleanup is improving; and gas turbines for IGCC applications are advancing in efficiency, capability and reliability. Commercially available gas turbines have been developed for the use of natural gas, (i.e. a methane-rich fuel with high calorific values of 800 to 1200 BTU/ scf). The gas turbines for these IGCC power plants have been adapted to burn syngas, a hydrogen-rich fuel1 with low calorific values of 100-300 BTU/scf, but their design features are not generally optimized for these fuel applications.
Robert Charles Steele Ramgen Power Systems 11808 Northup Way, Suite W-190 Bellevue, WA 98005 425-828-4919, ext. 288 [email protected]
233
The gas turbine encounters two major changes when transitioning from natural gas to syngas: • For the same fuel heat input, the fuel mass flow is four to five times greater than for natural gas, due to the lower heating value. • Premixed natural gas and air combustion systems have become common place for controlling NOx emissions. These systems are not used with syngas due to the high content of hydrogen and the potential for flashback of the flame into the fuel injection system. Diffusion flame or “non-premixed” combustors are used with syngas to control the NOx emissions by diluting the syngas with nitrogen, steam or carbon dioxide. The diluent reduces the flame temperature and consequently the formation of NOx. These two factors, greater fuel flow and the addition of diluent for NOx emissions control, substantially increase the overall mass flow through the turbine. This increase in flow creates backpressure to the compressor and can bring the engine close to surge conditions. Some gas turbines such as the GE 9001 EC are able to accommodate the increase in mass flow through the turbine expander. Unfortunately, the majority of gas turbines are not able to accept the overcapacity to the turbine expander.
3.2.1.4.1-3 Combustion of Syngas NOx emissions from coal based IGCC plants are as low or lower than those from the best conventional coal fired power plants. Nevertheless, NOx emissions from syngas gas fired turbines are inevitably compared to those of natural gas fired gas turbines without a good understanding of the differences in composition of the two fuels2. Dry (i.e. no addition of steam or water) Low NOx (DLN) combustors can achieve less than 10 ppmvd (parts per million by volume, dry, at 15% Oxygen) NOx emissions with natural gas fuel. DLN combustors rely on the premix principle that reduces the flame temperature, and subsequently the NOx emissions. DLN combustors are able to achieve much lower NOx emissions than diluted nonpremixed combustors because of an increase in premixing time prior to entering the combustion region. The high hydrogen content (up to 60% by volume) in syngas results in a flame speed that is up to six times faster than typical natural gas. The high flame speed of syngas makes the use of a DLN combustion system impossible because the flame will draw back into the premix zone and destroy the fuel injection hardware. The diluted non-premixed combustors have a chemical kinetic limit when too much diluent is added for further reduction of NOx emissions. The increase in diluent will cause flame instabilities in the combustion zone and eventual engine flame-out. The best case, practical NOx reduction limit for syngas combustors is between 10 and 20 ppmvd NOx. In order to further reduce NOx emissions in an IGCC power plant, the popular selective catalytic reduction (SCR) postcombustion control method will be required. Unfortunately, the SCR method which is very effective for sulfur-free natural gas, will not work if the sulfur cannot be removed from the syngas. Unlike natural gas, syngas does contain some sulfur which can be converted in the SCR to sulfur compounds and subsequently be deposited on the tube surfaces of the heat recovery steam generator. As an example, the Polk IGCC power plant located in Florida, owned by Tampa Electric Company, has experienced some sulfur deposits on the tube surfaces. The DOE reported that any additional deposits that would be generated by the addition of a SCR system would make the Polk plant inoperable on syngas in its current configuration.
3.2.1.4.1-4 DOE NETL IGCC Turbine Program In response to the challenges of burning syngas in IGCC gas turbines, the DOE has initiated a new multiyear program (Enabling Turbine Technologies for High-Hydrogen Fuels: DE-PS26-05NT42380) that will fund the development of technologies and products to serve the central station power generation market. The program will address key technologies needed to enable the development of advanced gas turbines and gas turbine-based systems that will operate cleanly and efficiently when fueled with coal-derived syngas. This is an investment in securing future U.S. electric power production through the use of the Nation’s largest fossil fuel energy resource. In order to address these technical challenges the participants in the program will pursue new combustor design approaches that include hydrogen premixing, catalytic combustion, and novel concepts such as TVC.
3.2.1.4.1-5 Trapped Vortex Combustion – Breakthrough Technology Trapped Vortex Combustion technology holds tremendous promise for IGCC gas turbine applications with improved efficiency, lower emissions, greater flame stability, added fuel flexibility, increased durability, and reduced capital costs. The TVC concept was originally conceived in the early 1990s for aero-propulsion applications with high through-put velocity requirements. It was not until the early 2000s that research and development organizations initiated the first design concepts for industrial applications. The TVC technology has the potential for product insertion into a wide variety of industrial applications including gas turbine power generation, manufacturing processing, chemical process heating, steam boiler systems, and incineration. The requirements of low fuel consumption and low pollutant emissions are paramount for all types of combustors, with the combustor primary zone airflow pattern of prime importance to flame stability, combustion efficiency, and low emissions. Many different types of airflow patterns are employed by non-TVC concepts, but one common feature to all is the creation of a toroidal flow reversal that recirculates (figure 1) and entrains a portion of the hot combustion products to mix with the incoming air and fuel to stabilize the flame. Although these designs have long been used in many practical combustion devices, there are limitations, especially for lean premixed applications.
Fig. 1. Non -TVC Swirl Stabilized Combustion (Courtesy of National Combustion Equipment, Inc.)
234
Robert Charles Steele Flame stability is achieved though the use of recirculation zones to provide a continuous ignition source which facilitates the mixing of hot combustion products with the incoming fuel and air mixture. Swirl vanes are commonly employed (figure 2) to establish the recirculation zones. This method creates a low velocity zone of sufficient residence time and turbulence levels such that the combustion process becomes self-sustaining. The challenge, however, is selection of a flame stabilizer which ensures that both performance (emissions, combustor acoustic and pattern factor) and cost goals are met. In contrast to conventional combustion systems which rely on swirl stabilization, the TVC employs cavities (figure 3) to stabilize the flame and grows from the wealth of literature on cavity flows. Much of the historical effort examines the flow field dynamics established by the cavities, as demonstrated in aircraft wheel wells, bomb bay doors and other external cavity structures. Cavities have also been studied as Fig. 2. Industrial Fuel/Air Swirler a means of cooling and reducing drag on projectiles and for scramjets (Courtesy of National Combustion Equipment, Inc.) and waste incineration. The actual stabilization mechanism facilitated by the TVC is relatively simple. A conventional bluff or fore body is located upstream of a smaller bluff body - commonly referred to as an aft body. The flow issuing from around the first bluff body separates as normal, but instead of developing shear layer instabilities which in most circumstances is the prime mechanism for initiating blowout, the alternating array of vortices are conveniently trapped or locked between the two bodies. In a TVC concept, the re-circulation of hot products into the main fuel-air mixture is accomplished by incorporating two critical features. First, a stable recirculation zone must be generated adjacent to the main fuel-air flow. If the vortex region, or cavity region, is designed properly, the vortex will be stable and no vortex shedding Fig. 3. Trapped Vortex Combustion will occur. This stable vortex is generally used as a source of heat, or hot products of combustion. The second critical design feature involves transporting and mixing the heat from the vortex, or cavity, region into the main flow. This is accomplished by using wake regions generated by bodies, or struts, immersed in the main flow. This approach ignites the incoming fuel-air mixture by lateral mixing, instead of a back-mixing process. By using geometric features to ignite the incoming fuel-air mixture, instead of pure aerodynamic features, the TVC concept has the potential to be less sensitive to instabilities and process upsets. This is particularly important near the lean flame extinction limit, where small perturbations in the flow can lead to flame extinction. The very stable yet more energetic primary/core flame zone is now very resistant to external flow field perturbations, yielding extended lean and rich blowout limits relative to its simple bluff body counterpart. Early research has demonstrated that the TVC configuration can withstand through-put velocities near Mach 1. This unique characteristic of the TVC technology provides a fluid dynamic mechanism that can overcome the high flame speed of hydrogen-rich syngas and potentially allow IGCC gas turbines to operate the combustor in premixed mode. This system configuration also has greater flame holding surface area and hence will facilitate the more compact primary/core flame zone essential to promoting high combustion efficiency and reduced CO emissions.
3.2.1.4.1-6 TVC Development Several TVC approaches from government research organizations and private companies are reviewed. Each design shares in the common fundamental features of the TVC technology but also exhibit the unique features that set them apart from each other. 1. Air Force Research Laboratory The Air Force program started in the early 1990s at the Air Force Research Laboratory (AFRL) in Dayton, Ohio.
235
3.2.1.4.1 Trapped Vortex Combustion
Fig. 4. AFRL First Generation TVC Source: See note 5 (Ren, Egolfopoulos, and Tsotsis 2002).
The phenomena of locked or trapped vortices has been known to reduce aerodynamic drag for years3, and the geometric features required to produce a locked or trapped vortex are the same features used to minimize drag. Hsu et al. in 1995 was first to report using this feature to stabilize reactions in gas turbine combustors for aero-propulsion applications4. Since then, several papers and patents have described the results from using this TVC concept to achieve stable and low combustor emissions5. The AFRL continues to investigate potential TVC applications for advanced military gas turbine engines6. The AFRL TVC development efforts have focused primarily on liquid fuel burning aero-propulsion applications and not on industrial natural gas or syngas burning applications. The developmental evolution of the TVC concept at the AFRL is extremely well summarized by Roquemore et al.7. The first generation TVC is shown in figure 4. The cavity is formed between the two disks in tandem. Katta and Roquemore used a timedependent, axisymmetric model to predict the results of reducing the drag of bluff-bodies in non-reacting flow and the experimental results of the first generation TVC8.
FLAMEHOLDERS
COMBUSTION CHAMBER MAIN FUEL & AIR
Fig. 5. AFRL Second Generation TVC Source: Same source as for fig. 4.
The second generation TVC design, shown in figure 5, was an axisymmetric can-type configuration with the cavity on the outside of the main burner. The depth of the cavity was approximately the same as that for the optimum first generation TVC. The third generation TVC shown in figures 6 and 7 was a two-dimensional sector designed for easy replacement and optical viewing of the cavities. The objective of the design effort was to develop a liquid fuel burning TVC concept for gas turbine engine applications.
236
Robert Charles Steele
Fig. 6. AFRL Third Generation TVC Source: Same source as fig. 4.
Fig. 7. AFRL Third Generation TVC Cavity Source: Same source as fig. 4.
The development program at AFRL concluded that the TVC offers significant improvements to aircraft gas turbine engines in lean blow out (LBO) and altitude relight when compared to conventional swirl stabilized combustors. Also, a wider operating range and the potential to achieve lower NOx emissions were demonstrated. The TVC concept can operate in a staged, main-pilot mode as well as in a rich burn–quick quench–lean burn (RQL) mode. Even though encouraging rig results have been obtained to date, no full engine test as been completed with an integrated TVC concept. 2. General Electric Company The General Electric (GE) Company has been developing TVC concepts for gas turbine engines since the mid 1990’s. At least ten GE TVC patents have been either filed and/or cleared since 1995. The majority of the patent work has been in the area of military gas turbine engines. More recently, GE has filed two TVC patents for low NOx emissions industrial gas turbine engine applications. The invention was made with support from the U.S. DOE. Aircraft Application GE Aircraft Engines and the AFRL have been jointly developing a novel TVC concept for military gas turbine engines since 19969. This effort represents an extension of earlier AFRL research with the third generation TVC concept. The work led to the fabrication of a rectangular sector test rig shown in figure 8 with a pressure capability of up 20.5 atmospheres and temperatures as high as 900 K. The performance evaluation covered all aspects of a gas turbine engine. The operating conditions with JP-8 fuel provided simulations of current commercial and military aircraft gas turbine engine cycles as well as some advanced cycles. Data was also obtained at selected conditions for the LM2500 marine Navy duty cycle using #1 Diesel. The TVC test rig demonstrated that ignition, blow out, and altitude relight were up to 50% improved over current swirl stabilized combustors. The NOx emissions were in the range from 40% to 60% of the U.N. International Civil Aviation Organization (ICAO) standard. The combustion efficiency was maintained at or above 99% over a 40% wider operating range than a conventional aircraft gas turbine engine combustor.
Fig. 8. GE TVC Sector Rig Source: See note 9 (Phillips).
Industrial Application GE Research is pursuing the application of TVC concepts to industrial gas turbine engines that can meet sub-9 ppmv NOx emissions. The objective of DOE Contract No. DE-FC26-01NT41020 is to explore advanced combustor concepts that show promise to meet future emissions requirements. The results of this DOE program are not published at this time. Any further information from GE regarding their low emissions TVC development effort was unavailable. 3. DOE National Energy Technology Laboratory The U.S. DOE is developing technologies for ultra-clean energy plants with efficiency and emission goals that are well beyond the capability of current gas turbine power plants. The DOE reports that ninety-percent of new power plants currently under construction will be fueled by a natural gas-based fuel. A key feature of these future power plants will be fuel diversity and flexibility. The gas turbine combustor designs will require the capability to operate on a wide range of fuels including hydrogen-rich synfuels. The DOE has also selected the TVC concept as a promising technology for future gas turbine combustor designs.
237
3.2.1.4.1 Trapped Vortex Combustion A collaborative effort began in 1999 between the AFRL and the NETL to evaluate the TVC concept for stationary power applications. The project was co-sponsored by the DOE Advanced Turbine Systems (ATS) program and the DOD Strategic Environmental Research and Development Program (SERDP). The primary intent was to assess the low-emissions capabilities of a novel RQL staged combustor shown in figure 910. The goal was to achieve NOx and CO emissions that are comparable to other commercial natural gas burning DLN systems. High BTU-fuels and fuels containing significant amounts of fuel-bound nitrogen were evaluated. NETL has continued to pursue the development of TVC concepts and combustor configurations with their own internal research program and through separate collaborative projects with GE and Ramgen Power Systems (RPS). The recently released multi-year turbine development program (DEPS26-05NT42380) is evidence that the DOE is committed to the advancement of novel combustor designs such as the TVC concept. Fig. 9. NETL RQL TVC NETL and RPS with support from the California Energy Commission Source: See note 10. (CEC Contract # 500-02-025) completed in early 2005 a series of rig tests that demonstrated less than 3 ppmv NOx at industrial gas turbine operating conditions without the need for a stabilizing catalyst or exhaust aftertreatment. The potential use of the RPS Advanced Vortex Combustion (AVC) technology with industrial gas turbines can best be described in the spring 2005 edition of the DOE Clean Coal Today: Researchers at NETL’s high pressure test combustion facility, in collaboration with Ramgen Power Systems, have completed testing of a fuel-flexible Advanced Vortex Combustion (AVC) concept that can achieve NOx emissions of less than 3 ppmv, and combustion efficiencies of over 99 percent. The Advanced Vortex Combustor is dynamically stable over a wide range of operating conditions, which makes this approach significantly more attractive than other lean premixed combustion approaches. In addition, the pressure drop associated with this combustion approach is significantly lower than a conventional gas turbine combustor, which translates into an improvement in the overall cycle efficiency. The relatively high velocities and low pressure drops achievable with this technology make the AVC approach an attractive alternative for hydrogen fuel applications. 4. Ramgen Power Systems RPS is pursuing the development of its unique AVC technology that has tremendous promise for improved efficiency, lower emissions, greater flame stability, fuel flexibility, increased durability, and reduced manufacturing costs. RPS is evaluating the potential for product insertion into a wide variety of industrial applications including gas turbine power generation and mechanical drive, manufacturing processing, chemical process heating, steam boiler systems, and incineration. The immediate AVC application is for low emissions stationary gas turbine engines. In 2002, RPS tested the first AVC concept (figure 10) with support from the DOE NETL (Contract No. DE-FC26-00NT40915) and achieved 9 ppmv NOX emissions on natural gas in lean premixed mode11. The testing was conducted at the GASL facility in New York. The development of AVC concepts for gas turbine applications has continued since the test program at GASL. RPS and NETL with support from the CEC have recently completed a joint project that has shown tremendous promise for the AVC technology Fig. 10. RPS AVC Configuration for extremely low emissions and acoustically stable Source: See note 11 (Little, Jr.). combustion. These test results indicate the potential to achieve unprecedented emissions levels of 1 ppmv NOx and 9 ppmv CO at industrial gas turbine operating conditions. At these levels, the AVC technology offers the potential to meet the stringent emission requirements in such regions as California and New York without further exhaust after-treatment. The tests were conducted in the NETL Low Emissions Combustion Test and Research (LECTR) facility at 10 atmospheres with combustor inlet temperature of 625°F. The combustor is air-cooled to closely simulate actual gas turbine conditions. The series of parametric tests allowed for variation in inlet pressure, fuel flow, air flow loading, lean blowout and pressure dynamics. A 4inch-by4inch quartz window allowed for flame visualization up to 5 atmospheres.
238
Robert Charles Steele The test results can be summarized as follows: • Data taken at 10 atmospheres and 625°F • Air-cooled combustor functioned as designed and without incident • Combustor ignited at ambient conditions • System was stable throughout start-up and shut-down • Flame structure was observed and recorded by video through the quartz window • Overall combustor pressure drop was less than 2.6% • NOx emissions are not pressure dependent under ultra-lean conditions • Lowest measured NOx with acceptable CO emissions: 3 ppmv NOx and 20 ppmv CO • Combustion efficiencies greater than 99.9% • Recorded combustion pressure oscillations with high frequency probe were insignificant RPS will continue to advance the development of its AVC technology and identify product insertion opportunities in industrial applications including gas turbine power generation and mechanical drive applications. The AVC technology is scalable to various sizes and heat load capabilities. 5. ALM Turbines Many turbine and combustor experts, including those at ALM, are increasingly optimistic about the promise of TVC12. It is ALM’s position that the TVC concept has many real potential advantages over both diffusion flame and DLN combustors, including lower emissions, multi-fuel capability, better flame stability, uniformity of flame, better dynamics, greater lean blowout limit offering greater turndown, higher efficiency due to lower combustor pressure drop losses, compactness, and lower manufacturing costs. Over the last few years, ALM has designed, manufactured and tested a number of proprietary prototype TVCs that have demonstrated many of the above mentioned advantages. ALM has been developing and testing its own proprietary version of TVC for both microturbines and large MW scale industrial turbines. In 2003, ALM and Alturdyne successfully designed, manufactured, and incorporated a TVC combustor into a Sunstrand T-62 APU. ALM has also designed, manufactured and rig tested two MW scale prototypes at ambient conditions that meet GE 7E 85MW operating conditions. The ALM TVC consists of two autonomous sections – a thermal nozzle and vortex (figure 11). The time of complete fuel burning, at the primary combustion temperature, is less than 2 milliseconds. ALM has achieved combustion at very low temperatures without use of any catalysts. The ALM design is notably different from all other TVC concepts. The concept utilizes its vortex and manages the recirculation flows in a different manner. The combustion mainly takes place in the thermal nozzle and not in the vortex. ALM uses a laminar boundary layer between the recirculation vortex flow and the fuel air mixture flow, to create certain desirable chemical reactions, and in turn a thermal nozzle effect which significantly improves the combustion process over other combustor designs. This laminar boundary layer, along with ALM’s unique use of the vortex distinguishes the ALM TVC from other TVC designs.
Fig. 11. ALM TVC Concept Source: See note 12 (Mair).
3.2.1.4.1-7 Notes __________________________________ 1. G. A. Richards, and J. Ciferno, “Carbon Dioxide Capture and Management in Energy Generation,” 4th Joint Meeting – U.S. Sections of the Combustion Institute, Philadelphia, PA, March 2005.
239
2. D.N. Anderson, “Effect of Hydrogen Injection on Stability and Emissions of an Experimental Premixed Prevaporized Propane Burner,” NASA Report # TM X-3301, October 1975; D.N. Anderson, “Emissions of Oxides of Nitrogen from an Experimental Premixed-Hydrogen Burner,” NASA Report # TM X-3393, May 1976; J.Y. Ren, W. Qin, F.N. Egolfopoulos, and T.T. Tsotsis, “Strain-Rate Effects on Hydrogen-Enhanced Lean Premixed Combustion,” 2001, Combustion and Flame 124 (2001): 717-720; J.Y. Ren, F.N. Egolfopoulos, and T.T. Tsotsis, “NOx Emission Control of Lean Methane-Air Combustion with Addition of Methane Reforming Products,” Combustion Science and Technology 174, No. 4 (2002): 181-205; G.S. Jackson, R. Sai, J.M. Plaia, C. Boggs, and K.T. Kiger, “Influence of H2 on the Response of Lean Premixed CH4 Flames to
3.2.1.4.1 Trapped Vortex Combustion Highly Strained Flows,” 2003, Combustion and Flame 132 (2003): 503-511; R.W. Schefer, D.M. Wicksall, and A.K. Agrawal, “Combustion of Hydrogen Enriched Methane in a Lean Premixed Swirl-Stabilized Burner,” Proceedings of the Combustion Institute 29 (2002): 843-851; T. Sidwell, K. Casleton, D. Straub, D. Maloney, G. Richards, S. Beer, and D. Ferguson, “Operation of a Pressurized Lean-Premixed Research Combustor and the Effects of Hydrogen Addition on NOx and Lean Extinction,” ASME Paper # GT2005-68752; J.N.Phillips, and R.J. Roby, “Enhanced Gas Turbine Combustor Performance Using H2-Enriched Natural Gas,” ASME Paper # 99-GT-115; Q. Zhang, D.R. Noble, and T. Lieuwen, “Characterization of Fuel Composition Effects in H2/CO/CH4 Mixtures Upon Lean Blowout,” ASME Paper # GT2005-68907. 3. B.H. Little Jr., and R.R. Whipkey, “Locked Vortex Afterbodies,” AIAA Journal of Aircraft 16, no. 5 (1979): 296-302; W.A. Mair, “The Effect of a Rear-Mounted Disc on the Drag of a Blent-Based Body of Revolution,” The Aeronautical Quarterly (November 1965): 350-360. 4. K.Y. Hsu, L.P. Goss, D.D. Trump, and W.M.Roquemore, “Performance of a Trapped-Vortex Combustor,” AIAA Paper # 950810. 5. K.Y. Hsu, L.P.Goss, and W.M. Roquemore, (1998), “Characteristics of a Trapped Vortex Combustor,” Journal of Propulsion and Power 14, no. 1 (1998): 57-65; V.R. Katta, and W.M. Roquemore, “Numerical Studies on Trapped Vortex Combustion Concepts for Stable Combustion,” Journal of engineering for Gas Turbines and Power 120 (January 1998): 60-68; Katta, V.R., Roquemore, W.M., “Study on Trapped-Vortex Combustor – Effect of Injection on Flow Dynamics,” Journal of Propulsion and Power, Vol. 14, No. 3, May-June 1998, pp. 273-281; G.J. Sturgess and K.Y. Hsu, “Entrainment of Mainstream Flow in a Trapped-Vortex Combustor,” AIAA Paper # 97-0261; C. Stone and S. Menon, “Simulation of Fuel-Air Mixing and Combustion in a Trapped-Vortex Combustor,” AIAA Paper # 2000-0478;K.Y. Hsu, C.D. Carter, V.R. Katta, and W.M. Roquemore, “Characteristics of Combustion Instability Associated with Tarpped-Vortex Burner,” AIAA Paper # 99-0488; W.M. Roquemore, et al. (2001), “Trapped Vortex Combustor Concept For Gas Turbine Engines,” AIAA Paper 2001-0483. 6. J. Zelina, D.T. Shouse, and R.D. Hancock, “Ultra-Compact Combustors for Advanced Gas Turbine Engines,” ASME Paper # GT2004-53155; R.C. Hendricks, R.C. Ryder, A. Brankovic, D.T. Shouse, W.M. Roquemore, and N.S. Liu, “Computational Parametric Study of Fuel Distribution in an Experimental Trapped Vortex Combustor Sector Rig,” ASME Paper # GT 200453225; V.R. Katta and W.M. Roquemore, “Simulation of PAHS in Trapped-Vortex Combustor,” ASME # GT2004-54165; G. Sturgess, J. Zelina, D. Shouse, and W.M. Roquemore, “Emissions Reduction Technologies for Military Gas Turbine Engines,” Journal of Propulsion and Power 21, no. 2 (March-April, 2005): 193-217. 7. See note 5 (Roquemore, Shouse, Burrus, Johnson, Cooper, Duncan, et al.). 8. See notes 3 and 5 (Katta and Roquemore, Jan 1998). 9. D.L. Burrus, A.W. Johnson, W.M. Roquemore, and D.T. Shouse, “Performance Assessment of a Prototype Trapped Vortex Combustor Concept for Gas Turbine Application,” ASME Paper # 2001-GT-0087. 10. D.L. Straub, T.G. Sidwell, D.J. Maloney, K.H. Casleton, G.A. Richards, W.A. Rogers, and G.M. Golden, “Simulations of a Rich Quench Lean (RQL) Trapped Vortex Combustor,” presented at the 2000 American Flame Research Committee (AFRC) International Symposium, Newport Beach, CA; D.L. Straub, K.H. Casleton, R.E. Lewis, T.G. Sidwell, D.J. Maloney, and G.A. Richards, “Assessment of a Rich Quench Lean (RQL) Trapped Vortex Combustor,” ASME Paper #GT2003-38569. 11. J. Bucher, R.G. Edmonds, R.C. Steele, D.W. Kendrick, B.C. Chenevert, and P.C. Malte, “The Development of a LeanPremixed Trapped Vortex Combustor,” ASME Paper #GT2003-38236. l2. M. Kalin, “Overview of the ALM TVC Combustor,” private communication with CEO of ALM Turbines, June, 2005.
240
BIOGRAPHY
3.2.1.4.1 Trapped Vortex Combustion
Robert Charles Steele Ramgen Power Systems 11808 Northup Way, Suite W-190 Bellevue, WA 98005 phone: 425-828-4919, ext. 288 email: [email protected]
Dr. Steele has almost 20 years experience in gas turbine combustion fundamentals and applications. He was the Combustion team leader for the Mars SoLoNOx engine at Solar Turbines. He joined Ramgen in 2000 and has been involved in the development of lean premixed “trapped” or “advanced” vortex combustion designs for gas turbine applications. He holds a M.S. in Aeronautics and Astronautics and a Ph.D. in Mechanical Engineering from the University of Washington. Having authored 30 technical publications, Dr. Steele is a prior member of the Combustion Institute, a member of the Combustion and Fuels committee of the ASME, and an Affiliate Adjunct Professor at the University of Washington.
3.2.1.4.2
Low Swirl Combustion
Robert K. Cheng Lawrence Berkeley National Laboratory MS70-108B, 1 Cyclorton Road Berkeley, CA 94720 phone: (510) 486-5438 email: [email protected]
241
3.2.1.4.2-1 Introduction Lean-premixed (LP) combustion technologies have been adopted by virtually every industrial gas turbine manufacturer as a Dry Low NOx (DLN) method to meet emissions regulations which are being implemented in the US and in many regions worldwide. But to meet more stringent ultra-low emissions standards being proposed, the DLN combustors have to operate at conditions near the lean limit of their stability envelopes where noise, instability, flame blowoff, and flashback can seriously affect engine performance. To mitigate these potential problems much effort has been devoted to explore passive control, e.g., fuel and/or air staging, and active control, e.g., feedback loop, strategies. Other alternatives invoke more costly exhaust gas clean up or catalytically assisted combustion. Undoubtedly, utilization of these new schemes would lead to more complex combustion devices consisting of tightly controlled sensors and actuators as well as many auxiliary components. For coal-based syngas engines the instability problems are further exacerbated due to the variability of the fuel contents. Therefore the injectors as well as the combustors have to be optimized or re-engineered to accommodate the changes in the combustion properties. Because most turbine combustors are designed for natural gas, they may not be readily adaptable or scalable to burn IGCC fuels. One promising solution to resolve fuel flexibility issues of IGCC turbines is a novel premixed combustion technology that operates on a unique low-swirl combustion (LSC) concept. Originally developed at the Lawrence Berkeley National Laboratory as a small laboratory research burner (about 15 kW) for fundamental studies, a good understanding of its operating principle has been obtained1. This patented combustion concept is based on exploiting the aerodynamic properties of the propagating premixed flames2. It is a simple, robust, and readily adaptable technology for industrial process burners and gas turbine combustors to meet stringent emissions targets without significantly altering their system configurations, efficiencies, turndown, and costs. LSC has been commercialized for industrial process heaters as low-swirl burners (LSB). Products of 150 kW to 7.5 MW (0.5 to 25 MMBtu/hr) with ultra-low emissions of 4 – 7 ppm NOx and CO (both @3% O2) have been available since late 2003. Central to the commercialization pathway was the scientific knowledge obtained from laboratory studies that has provided critical information for scaling as well as resolving system integration issues. LSC is also being adapted for natural gas turbines. Rig tests of prototype low-swirl injectors (LSI) for 10 MW size engines show it to be a very promising and cost-effective solution as “plug-in” injector replacements to enable current DLN turbines to meet the emission targets of < 5 ppm (@ 15% O2) for both NOx and CO. The LSC concept is readily adaptable for burning other hydrocarbons and hydrogen enriched fuels. Its operating principle is based on matching the flowfield to the turbulent premixed flame speeds of ultra-lean premixed flames. Laboratory measurements of flame speeds and flame temperatures for the alternate fuels will be necessary to obtain basic information for optimizing the LSC design. This strategy has already been applied to develop fuel-flexible industrial LSBs. Prototypes have been tested with propane, ethylene, natural gas diluted with flue gases (up to 40%), and refinery gases with large hydrogen constituencies (up to 50% H2). The main technical issue for adapting LSI to IGCC syngas turbines concerns the significantly different combustion properties of the two principal types of gasified coal fuels. Typical compositions of the syngas from oxygen blown coal gasification are 25% H2, 40% CO, 20% H2O, with a lower heating value of 200 BTU/ft3. With the addition of CO2 separation and sequestration, the syngas composition shifts to mostly hydrogen at 65-85% H2, and 15-35% H2O. These syngases have diverse combustion properties and the LSI needs to be optimized for the slower and faster burning flames (compared to natural gas) at operating conditions where the flame temperatures are sufficiently low to prevent NOx formation. Other concerns stem from the high H2 diffusivity and short auto-ignition delay time. Therefore, issues on system integration will need to address the impact on flashback, blow-off, light-off, shut-down, off-load, and load following. Currently, the research activities have been limited to proof-of-concept laboratory experiments using hydrogen and hydrogen/hydrocarbon blended fuels. More extensive laboratory studies will be necessary to develop basic LSI designs optimized for syngases and the accompanying scaling rules and engineering guidelines.
3.2.1.4.2-2 Principle of Low-swirl Combustion and Technology Transfer History Swirling flow burners have been essential to both premixed and non-premixed combustion systems because of their significant beneficial influences on flame stability, and combustion intensity, as well as the combustor performance. Until now, gas turbine combustors and industrial systems utilized a high-swirl type of burner in which the swirling motion generated by the injector (or burner) is sufficiently high to produce a fully developed internal recirculation zone at the entrance of the combustor. For conventional nonpremixed combustion, the role of the large recirculation zone, also know as the toroidal vortex core, is to promote turbulent mixing of the fuel and air. In premixed DLN systems, the recirculation zone provides a stable heat source for continuous ignition of the fresh reactants. Refer to the review of Syred and Beer for extensive background on the basic processes and practical implementation of highswirl combustors3. Low-swirl combustion is a relatively recent development. It is an excellent tool for laboratory research on flame/turbulent interactions4. Its operating principle exploits the “propagating wave” nature of premixed flames and is not valid for non-premixed combustion. Premixed flames consume the reactants in the form of self-sustained reacting waves that propagate at flame speeds controlled by the mixture compositions, the thermodynamic conditions, and turbulence intensities. In contrast, non-premixed diffusion flames do not propagate (i.e., move through the reactants) because burning occurs only at the mixing zones of the fuel and oxidizer streams. To capture a fast moving turbulent premixed flame as a “standing wave” that remains stationary, low-swirl combustion exploits a fluid mechanical phenomenon called a divergent flow. As the name implies, divergent flow is an expanding flow stream. It is formed when the swirl intensities are deliberately low such that vortex breakdown, a precursor to the formation of flow reversal and recirculation, does not occur. Therefore, the LSC principle is fundamentally different from the high-swirl concept of typical DLN gas turbines where strong toroidal vortexes are the essential flow elements to hold and continuously re-ignite the flames. The original LSB for laboratory studies known as the jet-LSB is shown in figure 15. This burner is essentially a cylindrical tube of 5.08 cm diameter fitted with a tangential air swirler section consisting of four small inclined jets of 0.63 cm in diameter. Reactants at a given fuel air equivalence ratio is supplied to the bottom of the tube. After passing through a turbulence generating plate, the reactants stream interacts with the tangential flow supplied through the jets. The size of the air-jets is kept small so that the swirling motions cling to the inner wall of the burner tube and do not penetrate into the center. When the flow exits the burner, centrifugal force due to the swirling motions causes the flow to expand and diverge. This divergent flow has a non-swirling core surrounded by a swirling shroud that weakens progressively downstream. Within the non-swirling center core, the adverse mean axial pressure gradient is accompanied by a linear decrease in the mean axial velocity. This velocity “down ramp” provides a very stable flow configuration for a premixed turbulent flame to freely propagate and settle at a position where the local flow velocity is equal and opposite to the flame speed. The flame does not flashback into the burner because it cannot propagate faster than the velocity at the burner exit. Blow off is also mitigated because the center non-swirling core provides a broad region where the flame naturally settles. More importantly, over mixture inhomogeneity or slight flow transients cause only a shift in the flame position so that the likelihood of catastrophic flameout is minimized. This is a robust self-adjusting mechanism for the flame to withstand transients and changes in mixture and flow conditions.
Fig. 1.
A jet-LSB demonstrates the principle of low-swirl flame stabilization.
242
Robert K. Cheng The flow feature crucial to flame stabilization is illustrated by the axial velocity profiles. Plotted in figure 2 are measurements obtained by laser Doppler velocimetry (LDV) published in Yegian and Cheng’s article where the effects of enclosure were evaluated by placing the jet-LSB inside quartz cylinders of 7.62 cm diameter, 20 and 30 cm in length with or without an exit constriction of 5.4 cm diameter to simulate typical combustor enclosures6. The experimental conditions were CH4/air flames with φ = 0.8 and 18.5 kW corresponding to a bulk flow velocity of U0 = 3.0 m/s. From the mean axial velocity (u) profiles (top), it can be seen that the velocity at the burner exit is slightly lower than U0 (about 70%). In the region just outside the exit (x < 20 mm), a linear decaying trend (with increasing x) shown on all the U/U0 profiles is the characteristic feature of flow divergence. The leading edge of the flame brushes are marked by the abrupt upturn in the profiles at 20 < x < 40 mm. These increases are due to combustion-generated flow acceleration. The minimum velocity on the U/U0 profiles corresponds to the velocity normal to the leading edge of the turbulent flame brushes and offers a convenient means to determine the turbulent flame speed, ST. Studies of ST using the jet-LSB show that a linear correlation exists between ST and turbulence intensity over a very broad range of turbulence intensities 7. This is an important property that enables the LSC concept to be scaled to the capacities of very large industrial combustion systems.
Fig. 2. Velocity profile showing that enclosures have little effect on the flame stabilization mechanism of a jet-LSB with CH4/air flame of φ = 0.8 at 18.5 kW (reproduced with permission from Combustion Science and Technology). Source: D.T. Yegian and R.K. Cheng, “Development of a Lean Premixed Low-Swirl Burner for Low NOx Practical Applications,” Combustion Science and Technology 139 no. 1-6 (1998): 207-227.
Technology transfer of LSC began with adaptation to residential pool heaters of 15 to 90 kW (50 to 300 KBtu/hr). These small domestic heaters are consumer products. To be price competitive they can only accept very simple and low-cost technologies that utilize rudimentary electronic controls. A LSB that has separate control for swirl jets and combustion air was deemed too elaborate to be economically and practically feasible. Therefore the key task was to develop a simpler burner that is easy to manufacture and requires few controls. The outcome of this work was a patented vane-swirler that has since been adapted for gas turbines. The main challenge in the swirler development process was a lack of relevant background knowledge on low-swirl flows. All prior research efforts on swirl combustion emphasized the generation of strong and stable flow recirculation. Extensive laboratory experimentation by LDV led to the design of figure 3. This LSB was sized for domestic heaters of 18 KW with a radius Rb of 2.54 cm. The unique feature of the swirler is a center by-pass channel (Rc = 2 cm) to allow a portion of the reactants to pass without being swirled by the outer annulus swirl vane section fitted with eight straight blades inclined at an angle α of 37.5°. The novelty of this design is the use of a perforated screen to control the flow-split between the unswirled core and the swirled annulus. The screen also produces turbulence in the center unswirled core. Because turbulence scales with flow velocity, it provides a crucial feedback control mechanism for the flame to accelerate and decelerate in response to changes in bulk flow, i.e. load change. The perforated screen fitted to the LSB in figure 3 has 3 mm holes arranged in a rectangular grid to give 81% blockage. By recessing this swirler assembly inside the burner tube at a distance, L, this burner produces the key flowfield features same as the jet-LSB.
243
3.2.1.4.2 Low Swirl Combustion
Fig. 3.
Schematics of a vane swirler developed for the low swirl burner.
As show in figure 4, the flame generated by the LSB is lifted with a bowl shape slightly different than the one produced by a jet-LSB. To demonstrate an exceptional feature of the LSC concept, all the components of the LSB in figure 3 were made of plastic to show that it does not receive significant heat from the flame. This has important practical significance because the burner suffers little or no material degradation due to a substantial reduction in thermal stresses. Subsequent to the development for pool heaters, several projects have been pursued to adapt LSB to industrial and commercial heaters. The efforts began by conducting parametric studies of LSB by varying Rb (2.54, 3.8, 5.1 and 6.35 cm) Rc (0.5 < R = Rc/Rb < 0.8), α (37o to 45ο), and L (1 < L/Rb < 4). Also investigated were other geometric variations such as the number of vanes, vane types (i.e. ,straight and aerodynamically shaped). types of center screens and their placement positions. The laboratory experiments and field tests were all performed with natural gas. The scientific foundation obtained for LSC has been most valuable to support data analyses and to devise solutions that meet specific operational and performance requirements. These studies proved that the LSB design is robust. To investigate turndown, the smallest LSB with Rb= 2.54 cm was fired in the open and generated stable flames from 10 to 600 kW that remained stationary despite the 60 to 1 change in input rate. At the lowest thermal input of 10 kW, the bulk flow velocity Uo corresponds to 1.7 m/s. This is the minimum allowable operating point for natural gas firing. Flashback becomes likely if Uo were reduced further because the velocity at the burner exit would be too close to the ST. The minimum Uo criterion to avoid flashback applies unequivocally to the larger burners because the LSB subscribes to constant velocity scaling. This simply means that the thermal input of LSB is directly proportional to Uo and Rb2. The effects of enclosure geometry on LSB performances was also investigated by testing several versions of the Rb = 6.35 cm LSB in boilers and furnaces at 150 kW to 2.3 MW. The results showed that vane shape and screen placement had little effect on flame noise, flame stability, and lean blow off. Most significantly, emissions of NOx depend primarily on φ. As shown in figure 5 by the NOx emissions from LSB of various sizes, the trends with φ are similar despite differences in thermal inputs and combustor geometries to show its capability to accept different fuels. Additional tests of the 6.35 cm LSB were performed with alternate fuels including natural gas diluted with up to 40% flue gases and refinery gases with hydrogen constituent up to 50%. Fig. 4. A vane-LSB firing at 18KW.
244
Robert K. Cheng
Fig 5. NOx emissions of LSB in furnaces and boilers of 300 kW to 1.8 MW.
In 2003, Maxon Corporation commercialized a line of industrial LSBs called M-PAKT burners. These products are developed for direct process heat applications of 0.3 to 1.8 MW (1- 6 MMBtu/hr) with a guarantee of 4 – 7 ppm NOx and CO (both at 3% O2) throughout its 10:1 turndown range. With these ultra-low emissions, the M-PAKT burners meet the most stringent air-quality rules in the U.S. As shown by the schematic in figure 6, the M-PAKT burner has a very simple compact design consisting of the swirler with air supplied by a blower through a plenum and a multi-port natural gas injector delivering fuel just upstream of the swirler. The control system is standard with conventional mechanical linkages and flow dampers. The exceptional performance of these commercial LSBs demonstrates that the implementation of LSC not only provides very effective emissions control but also improves system performance and reliability by eliminating the need for elaborate controls and intricate auxiliary components. The economic and operational benefit of this approach cannot be understated. In their continuing effort to commercialize LSC technology, Maxon engineers applied the scaling rules described in the next section to design a new LSB product of 7.5 MW (25 MMBtu/hr). The first installation was complete in February 2005. This large burner has a radius Rb of 25.4 cm and has an improved 20:1 turndown. It also incorporates a liquid fuel injector for dual-fuel firing.
3.2.1.4.2-3 Scaling rules and engineering guidelines The current size range of LSB from the smallest laboratory 7 kW unit of Rb = 1.27 cm to the largest 7.5 MW industrial burner represents a scaling factor of over 1000. This has been accomplished by using a set of scaling rules for the geometric variables and engineering guidelines for the LSBs to meet system integration and operational requirements. These design rules and guidelines are the significant outcomes of the technology development efforts. They represent the synthesis of the knowledge gained from extensive laboratory experiments and analyses based on theories on flame temperature, flame speeds, and reaction chemistry as well as combustion aerodynamics, turbulence, and turbulence/flame interactions. Though these rules and guidelines have been developed for natural gas, similar guidelines can be obtained for other fuels to account for the resulting changes in combustion properties.
245
Fig. 6. Emissions and schematic of a commercial LSB manufactured by Maxon Corporation, Muncie, Indiana.
3.2.1.4.2 Low Swirl Combustion Central to the scaling rule is the definition of a swirl number, S, for the LSB. It is derived from the formal definition of S based on the ratio of angular to axial flow momentum 8: (1)
where Gang is the angular momentum in the swirled section and G’x is the linear momentum flux through the unswirled center core and the swirled annulus. These terms can be calculated by integrating the mean axial, U, and the mean swirl, W, velocity components across the burner exit. However, this definition is not convenient because velocity data are not usually available. For engineering applications, a swirl number definition based on the geometry of the device is more amenable. With the assumption that the distribution of the axial flow remains flat, and U and W at the burner exit are kinematically related to the blade angle as tan α = U/W , the axial flux of angular momentum in the annular section is then written as follows: (2)
Here, Ua is a mean axial velocity supplied through the swirl annulus. By assuming flat axial velocity distribution, the linear momentum flux from the two regions of the burner is then calculated as follows: (3)
where Uc is a mean axial velocity through the center core. With Equation (1) as defined, the geometric swirl number for the vane swirl burner is then: (4)
Here, R is the ratio of centerbody to burner radii R= Rc / Rb. It is simplified further when Uc/Ua is expressed in terms of m the mass flux ratio (flow-split) m = m c / m a through the centerbody ( m c ) and annular ( m a ). The mass flux ratio is the same as the ratio of the effective areas of the center core and the swirl annulus and can be determined simply by the use of standard flow pressure drop procedure. Obviously, it is a more convenient form for engineering designs. The scaling rules were deduced from studying the influences of varying S, L and Rb on the burner operation. To start, the LSB prototype of figure 2 was used as a benchmark with its swirl number varied by the use of four different screens with 65 to 75% blockage. The swirl numbers were 0.4 < S < 0.44 corresponding to m of 0.8 to 1 meaning that 44 to 50% of the reactants bypassed the swirl annulus. These swirlers installed with L varied from 4 to 12 cm. The 16 LSBs with various combinations were tested with methane air flames at 5 < U0 < 25 m/s covering a thermal input range of 18 to 90 kW. All burners were found to be operable. Increasing S pulled the flame closer to the burner but the lean blowoff (LBO) remained relatively unaffected indicating that the performance of the LSB is not highly sensitive to a small variation in S. The differences were mainly with flame positions and the fuel/air equivalence ratio at lean blowoff, φLBO . Large swirler recesses generated a highly lifted flame but the overall flame stability remained relatively unchanged. A short recess distance produces higher φLBO indicating a compromise in the capability to support ultra-lean flames. Additional studies were performed to explore the impact of radius Rb as well as R (0.5 to 0.8) and α (30° to 45°). The swirl numbers of the burners with various combinations of Rb, R and α were varied by fitting them with screens of different blockages. The most significant finding was that the LSBs with larger diameters Rb operate at the same range of S (around 0.4 to 0.5) as the smaller burner. Their performances in terms of flame stability and φLBO were also identical. Moreover, decreasing R has no effects on emissions or performance but brings about a significant benefit in lowering the pressure drop of the LSB. This can be explained by the fact that reducing R enlarges the swirl annulus and lowers its drag. To maintain a swirl number of 0.4 to 0.5, a screen of lower blockage is required. For example, the screen used for the Rb = 6.35 cm with R = 0.5 has a 60% blockage compared to 65 to 81% needed for the R = 0.8 LSB. This combination effectively lowers the overall pressure drop of the burner. The drag coefficients determined for the different LSBs show them to depend only on R and independent of Rb. This knowledge is very important in the design of LSBs that will meet the various system requirements and efficiency targets.
246
Robert K. Cheng The scaling rules for the LSB were established from these results. They are the same regardless of burner radius (up to Rb = 25.4 cm). For stable and reliable operation, S of a LSB should be between 0.4 and 0.55. The swirler can have straight or curved vanes with angle α from 37 to 45o. The optimum center channel to burner radius ratios R can be from 0.5 to 0.8. Once α and R are defined, the blockage of the center channel screen can be varied to set S within the range of 0.4 to 0.5. In addition, L can be from 2 to 3 times the burner radius Rb. To determine the appropriate burner size, Rb, guidelines have been developed to optimize for the desired thermal input range, turndown, fuel pressure, fan power (pressure drop), combustion chamber size, and other physical constraints. The criterion for minimum thermal input is a bulk flow velocity of Uo = 3 m/s. This is simply the flashback point for natural gas (U0 ≈ 1.7 m/s) with a built-in safety factor. There is no restriction on the maximum thermal input owing to the high turndown (at least 20) available. To optimize for the fuel pressure and fan power, the drag coefficient for different R can be used. The optimum enclosure radius for the LSB is between 3 to 4 times Rb. Smaller enclosures restrict flow divergence and forces the flame to move inside the burner. Larger enclosures allow the flame to over-expand and generate internal flow patterns that negatively affect emissions. These rules and guidelines are easy to apply and are versatile enough to provide many design options to build simple and low cost LSBs for easy integration to existing or new systems.
3.2.1.4.2-4 Flowfield characteristics and their relevance to flame stability Low-swirl combustion exploits combustion aerodynamics. Understanding the flowfield characteristics of LSB and how these characteristics affect flame properties is central to the future development of the technology for IGCC turbines. Laboratory experiments remain an important step to provide an understanding of the underlying mechanisms that enable LSB to maintain its performance over a large range of conditions. The advent of Particle Image Velocimetry (PIV) has greatly facilitated the characterization of the LSB flowfields. In a recent investigation, a series of experiments were conducted to investigate the evolution of flow and flame features with increasing Uo. PIV captures in 2D the velocity distributions in a relatively large region. Our system was configured to have a 13 x 13 cm field-of-view covering the entire flame region of the laboratory burners. Shown in figure 7 is an example of the velocity vectors and turbulence stresses for a LSB with S = 0.53, Rb = 2.54 cm, R = 0.5, α = 37o and L = 6.3 cm burning a methane/air flame of φ = 0.8 and Uo = 5.0 m/s. The leading edge of the flame brush is outlined by the broken line. These velocity vectors show that the LSB flowfield is relatively uniform and free of steep velocity gradients. The background contours of the positive (red) and negative (blue) shear stresses also show that high turbulence stresses are relegated to the outer edges where the reactants mix with the room air. The center flow entering the flame brush is in fact relatively free of large shear stresses. The absence of high stresses means that the flames are much less vulnerable to stress induced non-uniform heat release and local quenching at ultra-lean conditions. This feature is very unlike those in other burners where steep velocity gradients are found at the flame stabilization region and the shear stresses can lead to premature flame blowoff. The evolution of the flowfield with Uo was investigated by applying PIV at 5 < U0 < 17.5 m/s. Shown in figure 8 are the normalized axial velocity profiles from the non-reacting cases. The fact that the profiles of the normalized axial velocity, U/U0, and the normalized two-component turbulent kinetic energy, q’/U0 (q’ = ((u’2 + v`2)/2)1/2) collapse to their respective trends shows that the LSB flowfield exhibits similarity feature. Similarity is also shown by radial profiles of U/Uo and V/Uo meaning that the key flowfield features are preserved at different bulk flow velocities.
247
Fig. 7. Velocity vectors and shear stresses (background contours) obtained in Rb = 2.54 cm LSB with methane air flame of φ = 0.8 and U0 = 5 m/s. White dash line marks the leading edge of the flame brush.
3.2.1.4.2 Low Swirl Combustion
a)
b) Fig. 8. Normalized centerline profiles of non-reacting flow produced by a laboratory LSB showing self similarity features.
Flowfield similarity explains why the flame maintains at a relatively fixed position regardless of U0. This can be illustrated by invoking an equality at the leading edge position of the flame brush, xf, (typically at 1.5 < x < 2.5 cm for this LSB). (5) Here, xo is the virtual origin of the linearly divergent portions of the axial profiles and has a negative value. As discussed earlier, ST of the LSB is linearly dependent on the rms velocity of the turbulence u’ such that ST = SL (1+ K u’) where SL is the laminar flame speed and K is an empirical constant that is in the order of 2.5 for methane. Substituting this into Eq. 5 and dividing both sides by Uo results in (6) The similar feature of the U/Uo profiles means that the normalized axial divergence rate (i.e., (dU/dx)/Uo) has a constant value ( ≈ 8 m-1 from data of Fig 8a). On the right hand side, (1 + K u’)/Uo tends to a constant value for large u’. This is because the turbulence at the flame stabilization point is isotropic such that u’ scales linearly with U0 as expected of turbulence produced by a perforated plate. Therefore, if SL is held constant, (i.e., at a fixed φ) the flame position xf attains a constant value that is independent of Uo and u’. As long as the flow similarity is preserved, the flame can be held at the same position. Changing φ will have an insignificant effect on xf because the range of SL for CH4 air flames (0.2 to 0.5 m/s) is small compared to the other values and constants in Eq (6). This analysis also shows that the turbulent flame speed ST is the important combustion parameter to consider when adapting the LSC for different fuels. However, measurements and predictions of ST are still active areas of fundamental research and data for the type of fuels IGCC turbines utilize are unavailable. But a lack of scientific ST data does not present a significant technical hurdle because the LSI developed for natural gas can be the benchmark to be adjusted for the slower and faster burning syngases. From Eq. (6) it can be seen that faster burning gases (e.g., high H2 constituents) need a lower divergence rate. Conversely, the slower burning gases (e.g., highly diluted syngas) need higher divergence rates. Of course, there are other combustion characteristics such as heat release ratios and preferential diffusion of the fuel components (e.g. between H2 and CH4) that need to be considered. From our studies of methane, ethylene, propane, and hydrogen flames, contributions from these other factors tend to be of secondary nature.
248
Robert K. Cheng 3.2.1.4.2-5 Development of low-swirl injectors for natural gas turbines To adapt low-swirl combustion for gas turbines the most pressing question is whether or not this flame stabilization method is operable at elevated temperatures and pressures. To seek an answer, a jet-LSB of Rb = 3.8 cm was constructed for proof-of-concept test at Solar Turbines, San Diego. Successful firing of this LSB at typical gas turbine combustor inlet temperatures (220 and 341°C), pressures (5 and 10 atm), and loads (0.6 to 1.2 kg/s air) showed the concept to be valid. Light-off of the LSB was easy even at elevated temperatures and pressures. The flame remained very stable and free of unacceptable pressure fluctuations towards lean blow-off. The trends of the jet-LSB emissions were also typical of other DLN injectors but the overall concentrations of NOx and CO were higher. This was caused by the poor mixing from a rudimentary fuel spoke injector and significant dilution from the nitrogen swirl jets. Following the proof-of-concept tests the next logical step was to design and evaluate a vane LSI prototype. The lack of relevant background knowledge on low-swirl flows and turbulent flame speeds at gas turbine conditions remained the main obstacle. To circumvent this difficulty and reduce engineering design efforts and prototype fabrication costs, a decision was made to pursue a fasttrack developmental path which exploited, as much as possible, the current injector hardware parts. Because the annulus swirler is the main component of Solar Turbines’ SoLoNOx Taurus 70 high-swirl injectors (HSI), the preferred option was to investigate if the HSI swirler could be converted to operate in the low-swirl mode.
Fig. 9. Schematics and photographs of the LSI (top a-c) and the HSI (bottom d-f)
The LSI (Fig. 9 a-c) converted from the HSI (Fig. 9 (e-f) has the same basic configuration as the LSB. The SoLoNOx swirler (Fig. 9c) has a modular design such that the solid centerbody can be easily removed to form the center-channel for the LSI (Fig 9b). Thus, the two key parameters of the LSI, α (40o) and R (0.63), are fixed by SoLoNOx swirler having an outer radius of 3.27 cm and a centerbody radius of 2.06 cm with 16 aerodynamically shaped curved vanes. To configure the other LSI parameters, S and L, the guidelines for atmospheric LSB were followed. The swirler recess of L = 9.5 cm satisfies the 2 < L/Rb, < 3 criterion. The swirl number, S, was set between 0.4 and 0.55 by the use of center channel screens with blockage of 50 to 73%. These screens were fitted to the LSI and tested at laboratory conditions with CH4/air mixtures at a range of stoichiometry and a fixed bulk flow velocity, U0 = 5 m/s (≈ 35 kW). The optimum was a 58% blockage screen that stabilized the flame at 1.5 to 2.5 cm downstream of the injector exit. From effective area measurements, m for this LSI was 0.3, thus 23% of the reactants passed through the center channel unswirled. This gave a swirl number S of 0.5 for the LSI. In comparison, the swirl number determined for the HSI was 0.73. Even though the difference in S between LSI and HSI is slight, figure 9f shows that the HSI flame is attached to the centerbody while the LSI flame of figure 9c is detached.
249
3.2.1.4.2 Low Swirl Combustion The flowfields of the LSI and the HSI flames were also investigated by PIV9. Figure 10 shows the results obtained for CH4/air flames with φ = 0.8 at 87 kW. Under these conditions, the bulk velocities for the LSI and HSI were 9.6 and 12 m/s respectively. To illustrate the differences between the flowfields of the two injectors, streamlines have been traced from the velocity vectors. For the LSI, its flowfield features are essentially the same as those of the LSB of figure 3. The spreading of the streamlines near the bottom corresponds to the divergent flow and their slight bends through the flame brush are associated with heat release effects. Downstream of the flame brush, the streamlines are relatively parallel showing a uniform flow of the hot products. The color flooded contours of the shear stresses in the background illustrate again that the flame does not experience high stresses. In contrast, the flowfield of the LSI is characterized by several high shear regions with the flame zones encountering the highest stress levels. These large velocity fluctuations with very steep instantaneous local gradients make the flame vulnerable to stress-induced non-uniform heat releases as well as local flame extinctions. Downstream of the flame, strong recirculation is shown by the streamlines forming into two separate loops. This is the crucial flow structure for HSI that provides a steady source of hot products for igniting the incoming reactants and holding the propagating flame. The PIV results clearly show that HSI flowfields are dominated by very large velocity gradients and vortex structures. With the LSI flowfield showing more uniformity, the flame characteristics and behavior produced by the two injectors are fundamentally different.
Fig. 10. Streamlines and shear stresses of HSI and LSI burner CH4/air flames at φ = 0.8 and 87 kW. The turbulent flame brushes are outlined by the white dash lines.
Subsequent to the laboratory tests, the LSI was evaluated to determine its operability at gas turbine conditions as well as its effectiveness in lowering emissions. The first set of tests was performed with preheated air at atmospheric pressure to observe flame positions, flame shift with Uo, and sensitivity to mixture homogeneity. Visual observation during these tests showed that the locations of the flames were not highly sensitive to U0, φ and initial temperature T0. Flashback did not occur throughout the test matrix. The flame size was similar to that of the HSI flame indicating that the LSI can use the same combustor liner. Tests performed with two premixers with +/- 3% and +/- 10% mixture uniformity showed no effects on NOx emissions, overall flame behavior and flame characteristics.
250
Robert K. Cheng Next the LSI was fitted to a standard film-cooled combustor liner and tested in a high pressure combustion rig. The combustor rig-test matrix covering To of 360 to 430°C, P0 of 5 to 15 atm and air flow of 0.4 to 1.33 kg/s represented the partial and full load conditions of Solar Turbines’ SoLoNOx engines of 5 to 7 MW. Successful firing of the LSI at these conditions showed that its operating range is fully compatible with that of the HSI. Throughout these tests, there were no indications of shifting in flame positions or flashback. Excessive acoustic amplitudes were not observed and peak rms acoustics pressures were generally below the allowable 3.4 kPa established for production hardware. Figure 11 shows NOx and CO emissions from the LSI (corrected to 15% O2) compared with the emissions of HSI. Due to the variation in T0 and P0, the combustion temperature was not proportional to φ. Instead, the emission data was plotted Fig. 11. NOx and CO emissions of natural gas LSI and HSI from against the theoretical adiabatic flame temperature Tad rig-tests at 5 < P0 < 15 atm and 360C < T0 < 430°C. (for the combustor-rig Tad was calculated for the primary zone). As can be seen, all the NOx emissions from LSI collapsed onto a consistent trend that crossed the 5 ppm threshold at Tad < 1920K. The strong correlation of NOx emissions with Tad showed that thermal NOx was the predominant reaction pathway. In comparison, NOx emissions of HSI were generally higher and only approached the 5 ppm NOx threshold before LBO. The over 2.5 times NOx emissions reduction offered by LSI represents a significant improvement of DLN technology. Moreover, these results imply that the LSI can operate farther away from LBO (higher Tad) where it is less prone to combustion oscillation. Although the CO emissions in figure 8b do not show a consistent trend, all but one point from the high T0 and P0 runs is below 5 ppm. Therefore, the LSI does not entail compromising CO for the sake of lowering NOx. The emissions from LSI rig-tests are very encouraging and show that the LSI has the potential to bring about a substantial reduction in NOx emissions from DLN gas turbines without incurring significant add-on cost or system complexity. The lowest NOx emission of < 2 ppm (at 15% O2) is comparable with those from the more costly and much less durable catalytic combustor. Despite the complexity of the NOx formation mechanisms, the flowfield characteristics of LSI may provide an explanation for why it is highly effective in reducing NOx. Recent studies of high-swirl combustion have shown a correlation between NOx emissions and the residence time within the recirculation zone. Trapped by the recirculation vortex, a portion of the combustion products resides for a prolonged period at high temperature, which allows the NOx formation mechanism to continue and increase in concentrations. The PIV results show clearly that the LSI does not have a recirculation zone to trap a large recirculating mass, therefore, the residence time of the hot products in a LSI is much shorter than in a HSI and truncates NOx formation in the post flame region. Continuing the effort to develop the LSI for SoLoNOx turbines, additional tests were conducted to show that the emissions and flame stability of LSI are not highly sensitive to mixture inhomogeneity. Its design can easily incorporate components needed for normal engine operations, e.g., plenum, liner, ignitor, pilot, and premixer. Therefore, adaptation of LSI to current engines does not require significant hardware modifications or change in control strategy or protocol. Currently, an engine-ready LSI prototype has been designed and has undergone a series of single-injector rig tests to verify its readiness for start up, shutdown, load change and offload. It uses the production swirlers made for Solar Turbine’s Taurus 70 engines integrated with a pilot and a multi-spokes fuel injector similar to the kind used in current HSIs. The prototype LSIs have been designed to be backward-compatible as “drop-in” replacements for the current HSIs.
3.2.1.4.2-6 Development of LSI for IGCC
251
The development of LSI for natural gas turbines shows that laboratory research is directly applicable to engineering of the turbine hardware. This process is efficient and cost effective and does not require extensive computational fluid dynamics efforts. The same approach is being taken to adapt the LSI concept for IGCC turbines. Currently, progress has been limited to proof-of-concept laboratory demonstration of LSI burning hydrogen and hydrogen/hydrocarbon blended fuels. Much more extensive laboratory studies will be necessary to obtain better understanding of the properties of the various syngas flames to configure the LSI to accommodate the changes in the reactivity, diffusivity and exothermicity. The laboratory studies will also be used for developing scaling rules and engineering guidelines that can be applied to IGCC gas turbines of different sizes. Due to the different constituencies and concentrations of the two principal types of gasified coal fuels, accommodating the variability in fuel properties is the main challenge in adapting the LSC technology for IGCC turbines. From oxygen-blown coal gasification, typical compositions of the syngas are 25% H2, 40% CO, 20% H2O. This fuel has a lower heating value of only 8.3 Mj/m3 compared to 42 Mj/m3 for natural gas. Due to the high level of dilution, the flame speeds of this syngas are much lower than those of natural gas and they also burn at lower flame temperatures. The key design issue is to stabilize these weak and not highly flammable flames without affecting engine operability. Mitigating NOx emissions will not be a significant challenge due to the lowered flame
3.2.1.4.2 Low Swirl Combustion temperatures. The second type of syngas is produced with CO2 separation and sequestration to change the compositions to mostly hydrogen at 65-85% H2, 15-35% H2O. This syngas is much more flammable with flame speeds several times higher than natural if it is not diluted by N2 from the air separation unit. The basic issues for LSI development are now shifted to stabilizing these potentially faster flames at very lean conditions where the flame temperatures are sufficiently low to prevent NOx formations. Additional issues arise due to the high diffusivity of H2 and its low flammability limit. Moreover, H2 rich mixtures have shorter auto-ignition delays than hydrocarbon mixtures. Therefore, integration of the LSI with fuel injection and premixer will be a significant part of the development. To gain insights on some of these issues, a study was performed with a LSB of Rb = 2.54 cm, R = 0.6 by burning blended mixtures of laboratory grade CH4 and H2 with concentrations progressively varied from pure CH4 to pure H2. The test conditions were the same (Uo = 10 m/s) as the conditions for the original laboratory experiments performed to develop the natural gas LSI. While this LSB was found to burn pure H2-air flames, a limiting flame phenomenon was found. Shown in figure 12 is the flame stabilization map obtained for the CH4-H2 mixtures consisting of the (LBO) limit as well as an additional limit where partially rim-attached flames were observed. The LBO limit is the boundary between the blowoff region and the stable lifted flame region. It has a decreasing trend towards lower φ values with increasing H2. This is consistent with increasing flammability due to the presence of H2. However, for a mixture with more than 20% H2, the flames transition from fully-lifted flames to partially rim-attached flames when φ is increased. The attached flame indicates that the mixing layer between the swirling reactants and the surrounding air has become flammable. This is caused by H2 diffusing preferentially into the surrounding ambient air when the overall fuel concentration increases with φ. Α partially attached flame is a consequence of the tail end of the lifted flame curled back upstream to ignite the mixing zones. While this phenomenon does not seem to affect the flame at the center region, it alters the overall flowfield dynamics and how the flame behaves within a combustor. The limit for partially attached flames also showed a decreasing trend with increasing H2 and thus the range of conditions for pure H2 flames was very small. This phenomenon showed that the highly diffusive nature of H2 needs to be carefully considered in the design of LSI for syngases. At test condition of U0 = 10 m/s, the 100% H2 flames are close to the flashback point. This is because of the high laminar flame speeds compared to those of methane. In accord with Eq. (6), the faster burning H2 flames also sit closer to the burner exit. The close proximity of the H2 flames can be one of the contributing factors for the onset of partial flame attachment. Therefore, to modify the current natural gas LSB for syngases with high H2 constituency, the flame needs to be maintained at a position further downstream. This can be accomplished by relaxing the normalized divergence rate (dU/dx)/Uo through lowering of the swirl number S. Conversely, for syngases with low heating values, the current LSB will generate flames that are further downstream and this is not optimum. To draw these flames closer to the burner exit, it would be necessary to increase the normalized divergence rate by increasing S. This illustrates that the flame speed is the important parameter for engineering the LSI for different fuels. Consequently, adjusting the LSI swirl number for each fuel is the first step toward development for IGCC turbines. Using the natural gas LSI as a benchmark, the strategy is to determine the S range for different fuels with various dilution levels. The results will be analyzed to obtain operating maps expressed in terms of S, Tad, and φ. Using the flame temperature and NOx emissions as the reference parameters, these maps will guide the development of fuel-flexible LSI prototypes that can accommodate several fuel types. Refinement of the designs to further optimize for differences in combustion characteristics such as exothermicity, preferential diffusion of the fuel components (e.g., between H2 and CH4) and other non-linear behavior can be accomplished through further rig tests with preheated air and at simulated gas turbine conditions. As LSI technology is still under development, issues concerning its full integration to gas turbine engines are still being investigated. The significant outstanding issue is, of course, the combustion oscillation characteristics of the LSI. The interesting question from both the scientific and technological perspectives is whether or not the absence of a large recirculation zone in the LSI flame will have an influence on the combustion oscillation characteristics. From field tests of atmospheric boiler and furnaces, the LSBs show relatively low tendencies to incite combustion oscillations. Rig-tests of the LSI at Fig. 12. Stability limits of a Rb = 2.54 cm, R = 0.6 LSB for CH4-H2-air flames at Uo = 10 m/s. simulated gas turbine conditions also indicate the absence of a strong characteristic acoustic signature from the flame. Though these observations are encouraging, the combustion oscillation characteristics of LSB and LSI need to be investigated more systematically to gain a fundamental understanding for addressing issues that may arise when the technology is adapted for more complex systems such as IGCC gas turbines.
252
Robert K. Cheng 3.2.1.4.2-7 Conclusions Low-swirl combustion (LSC) is a very promising simple, robust, and cost-effective solution to attain single-digit NOx operation. Already commercialized for industrial process heaters and currently being adapted for natural gas turbines, development of LSC for IGCC turbines can provide a simple and direct solution to resolve important complex coupling issues concerning fuel-variability, reliability, and emissions. LSC exploits a patented combustion aerodynamic process to burn ultra-lean premixed turbulent flames that emit very low levels of NOx. Originally developed for basic research, a good understanding of its operating principle has been obtained. The principle is fundamentally different than the high-swirl method used in all current DLN engines. Instead of relying on a very high level of swirl to generate a well formed vortex to “hold” the flame, LSC takes the opposite approach. By lowering the swirl intensity until a recirculation zone is not formed, it generates a divergent flow with a non-swirling center flow region. Linear velocity decay within the center region provides a stable flow configuration where the premixed flame freely propagates and maintains itself at the position where the local flow velocity is equal and opposite to the turbulent flame speed. Flash back is prevented because the flame cannot propagate faster than the velocity upstream. Blow-off is mitigated because the center flow region provides a broad range of velocities where the flame naturally re-settles during occasional swings in flow or mixture conditions. For practical implementations, the scientific background of LSC provides a sound foundation for the development of a patented vane swirler. It consists of an annulus swirl vane section found in many other swirler designs but with an open center channel that allows a portion of the reactants to remain unswirled. Laboratory experiments, prototype development and testing, and extensive analysis of the flow and emission data have produced a set of scaling rules and engineering guidelines that have been applied successfully to develop and commercialize low-swirl burners (LSB) of 17 kW to 7.5 MW for meeting ultra-low emission goals of < 7 ppm NOx and CO at 5% O2. Central to the rules and guidelines is a new definition of the swirl number based on the geometric variables that include the vane angle, ratio of the center channel radius to burner radius, swirler recess, and the flow split between the center core and swirled flow. By setting specific ranges for the swirl number and for the geometric variables, LSB can be configured to meet emissions goals as well as system integration, performance, and operational requirements. These rules and guidelines have also been applied to develop low-swirl injectors (LSI) for 5 to 7 MW gas turbine engines. To be compatible with existing engines, the LSI prototypes were made using the key components engineered for current production high-swirl injectors. Single injector rig-tests of the LSI prototypes showed them to emit < 5 ppm NOx and CO at 15% O2 at simulated part-load and full-load conditions. This represents a 2.5 times emissions reduction compared to current DLN high-swirl combustion technology. These results also showed that atmospheric laboratory experiments are directly applicable for designing and engineering of gas turbine hardware. Upon the completion of the single injector rig-tests, a set of pre-production LSI prototypes is being fabricated for engine tests in late 2006. These prototypes have been designed to utilize many key components from current production DLN high-swirl injectors and configured to be “drop-in” replacements for existing engines. Due to the robust low-swirl combustion mechanism, the operability of the engine such as lightoff, shutdown, off load, and load change are not expected to be affected when the new LSI are used. From system integration and cost standpoints, the most significant feature of both the LSBs and LSIs is that they can be fully compatible with current systems using some of the same components (i.e., plenum, swirler, air compressors, blowers, premixer, pilots and ignitor) but with significant performance improvements in terms of emissions, turndown, and stability. The laboratory studies also provide the scientific insights necessary for adapting the technology to accept different gaseous fuels. Measurements of the velocity distributions and turbulence statistics have shown that low-swirl produces self-similar flowfields that scale linearly with bulk velocities (i.e., input power). Because premixed turbulent flames propagate at turbulent flame speeds that also scale linearly with the turbulence intensity and bulk flow velocity, the flow and flame features are coupled optimally to give LSC its exceptional performances. Therefore, the adaptation of LSC to different IGCC syngases involves primarily the adjustment of the LSC flowfield via the swirl intensity and other geometric variables to accommodate for the changes in the flame speeds. Proof-of-concept laboratory experiments have already been performed and have shown that LSC can burn pure H2/air flames. The plan towards further development for IGCC includes experimental measurements of the flame speeds for syngases and developing guidelines for fuel-flexible LSIs.
253
3.2.1.4.2 Low Swirl Combustion 3.2.1.4.2-8 Notes ___________________________ 1. C.K. Chan et al., “Freely Propagating Open Premixed Turbulent Flames Stabilized by Swirl,”Proc. Comb. Inst., 24 (1992): 511518; B. Bedat and R.K. Cheng, “Experimental Study of Premixed Flames in Intense Isotropic Turbulence,” Combustion and Flame 100, no. 3 (1995): 485-494; R.K. Cheng, “Velocity and Scalar Characteristics of Premixed Turbulent Flames Stabilized By Weak Swirl,” Combustion and Flame 101, no.1-2 (1995): 1-14. 2. R.K.Cheng, 1998. Ultralean Low Swirl Burner. U.S. Patent 5735681, (The Regents of the University of California), filed March 19, 1993 and issued April 7, 1998; R.K. Cheng and D.T. Yegian, 1999. Mechanical Swirler for a Low-NOx WeakSwirl Burner. U.S. Patent 5,879,148, (The Regents of the University of California), filed April 16, 1997 and issued March 9, 1999 . 3. N. Syred and J.M. Beer, “Combustion in Swirling Flow: A Review,” Combustion and Flame 23 (1974): 143-201. 4. T. Plessing et al., “Measurement of the Turbulent Burning Velocity and the Structure of Premixed Flames on a Low Swirl Burner,” Proc. Comb. Inst. 28 (2000): 359-366; I. G. Shepherd et al., “Premixed Flame Front Structure in Intense Turbulence,” Proc. Comb. Inst. 29 (2002): 1833-1840; R.K. Cheng, et al., “Premixed Turbulent Flame Structures in Moderate and Intense Isotropic Turbulence,” Combustion Science and Technology 174, no.1(2002): 29-59; C. Kortschik, T. Plessing, and N. Peters, “Laser Optical Investigation of Turbulent Transport of Temperature Ahead of the Preheat Zone in a Premixed Flame,” Combustion and Flame 136, no.1-2 (2004): 43-50; L.P.H. de Goey et al., “Analysis of the Flame Thickness of Turbulent Flamelets in the Thin Reaction Zones Regime,” Proc. Comb. Inst. 30, no 1 (2005): . 859-866; also see note 2 above. 5. See note 2 above (Bedat & Cheng). 6. See note 1 above. 7. See note 2 above (Bedat & Cheng). 8. See note 4 above. 9. M.R. Johnson et al., “A Comparison of the Flowfields and Emissions of High-swirl Injectors and Low-swirl Injectors for Lean Premixed Gas Turbines,” Proc. Comb. Inst. 30 (2005): 2867 - 287
254
BIOGRAPHY
3.2.1.4.2 Low Swirl Combustion
Robert K. Cheng Lawrence Berkeley National Laboratory MS70-108B, 1 Cyclorton Road Berkeley, CA 94720 phone: (510) 486-5438 email: [email protected]
Dr. Robert K. Cheng is a Senior Scientist and the leader of the Combustion Technologies Group in the Environmental Energy Technolgies Division of Lawrence Berkeley National Laboratory. He received a B. S., a M. S., and a Ph. D. in Mechanical Engineering from the University of California at Berkeley. Since 1977, he has been leading experimental research on fundamental combustion fluid mechanics with an emphasis on lean premixed turbulent flames and has published over sixty papers on fundamental turbulent flame properties. His recent discoveries of novel flame stabilization concepts for ultra-low NOx combustion systems have generated three patents. These technologies are in various stages of development and commercialization for industrial heating equipment and gas turbines.
3.2.2.1
Fuel-Rich Catalytic Combustion
Dr. Lance Smith
Dr. Shahrokh Etemad
Dr. Hasan Karim
Dr. William C. Pfefferle Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, CT 06473 phone: (203) 287-3700 x217 email: [email protected]
265
3.2.2.1-1 Introduction Currently, gas turbines operating on natural gas offer the lowest achievable NOx emissions without exhaust-gas aftertreatment, as compared to other fuels. Commercially, this has been achieved through the use of leanpremixed combustion systems, allowing NOx emissions below 9 ppm (at 15% O2) to be guaranteed for natural gas operation, and emissions near 5 ppm to be demonstrated1. Lower emissions are needed, however, and are possible through the use of catalytic combustion2. Because natural gas has historically offered the greatest potential for low emissions, catalytic combustor development has until recently focused largely on methane and natural gas operation. As a result, the unique properties of methane have led to a number of development issues and design strategies, generally related to the behavior of Pd-based catalysts used for methane oxidation. In particular, for methane oxidation under fuel-lean conditions, only Pd-based catalysts among the PGM group are currently practical, because only they offer acceptable activity, lightoff temperature, and resistance to volatilization3. Unfortunately Pd-PdO catalyst morphology and its reactions with methane are complex, and lead to complex behaviors such as deactivation at high temperature (above about 750°C / 1380°F), hysteresis in reaction rate over heating and cooling cycles4, and oscillations in activity and temperature5. In addition, lightoff and extinction temperatures are well above 300°C (570°F) for fuel-lean reaction on Pd-based catalysts, thus requiring the use of a preburner in many engine applications6. Fuel-rich operation of the catalyst circumvents many of these issues and provides significant catalyst advantages, including a wider choice of catalyst type (non-Pd catalysts are active to methane under fuel-rich conditions), improved catalyst durability (non-oxidizing catalyst environment), and low catalyst lightoff and extinction temperatures. Catalyst extinction temperature is particularly low, and is generally less than 200°C (400°F) for the precious-metal catalysts used in the work reported here (that is, once the catalyst has been lit off, the catalyst remains lit at inlet air temperatures less than 200°C / 400°F), and a preburner is generally not required. A more complete discussion of fuel-rich versus fuel-lean catalyst behavior for methane oxidation is given by Lyubovsky et al.7. In addition to catalyst material challenges, commercial acceptance of catalytic combustion by gas turbine manufacturers and by power generators has been slowed by the need for durable substrate materials. Of particular concern is the need for catalyst substrates which are resistant to thermal gradients and thermal shock8. Metal substrates best fill this need, but their temperature must be limited to less than 950°C (1750°F) to assure sufficient material strength and long life. Downstream of the catalyst, combustion temperatures greater than about 1100°C (2000°F) are required for gas-phase reactions to complete the burnout of fuel and CO in a reasonable residence time (on the order of 10 ms). Thus, only a portion of the fuel can be reacted on the catalyst. A major challenge, then, is to limit the extent of reaction within the catalyst bed such that excessive heat does not damage the catalyst or substrate, yet release sufficient heat that downstream gas-phase combustion is stabilized under ultra-low emission conditions. For systems which lean-premix fuel and air upstream of the catalyst, the degree of reaction can be limited by chemical reaction rate upon the catalyst, or by channeling within the reactor such that only a limited fraction of the fuel contacts the catalyst. In all cases, however, it is imperative that uncontrolled gas-phase reactions do not occur within the catalystbed, since this implies a loss of reaction limitation and ultimate over-temperature and failure of the catalyst bed. Preventing such gas-phase reactions is especially challenging in applications to advanced, high-firing temperature turbines, where fuel/air ratios in the catalyst-bed can be well within the flammability limits.
3.2.2.1-2 Fuel-Rich Catalyst Systems
1320 1310 1300 1290 1280 1270 1260 1250 1240 1230 1220 1210 1200 300
400
500
600
700
800
2392 2372 2352 2332 2312 2292 2272 2252 2232 2212 2192 900 1000
PSR exit temperature o at imminent blowout, F
PSR exit temperature o at imminent blowout, C
Another approach to limiting the extent of reaction is to operate the catalyst fuel-rich. In this case, there is insufficient oxygen to fully oxidize all fuel in the catalyst bed, and the extent of reaction is therefore limited even if gas-phase reactions occur. To use a fuelrich catalyst bed in a catalytic combustion system, additional air is introduced downstream of the catalyst so that combustion completion can occur fuel-lean. Based on this concept, fuel-rich catalytic reactors were tested by NASA and contractors for liquid fuel applications, and showed good soot-free performance9. An examination of fuel-rich catalysis on a variety of liquid fuels was also conducted at Yale University under support from NASA10. Like the NASA results, this work showed soot-free catalyst performance on a range of fuel types, including a surrogate jet fuel. United Technologies Research Center11 also investigated fuel-rich catalytic reaction of liquid fuels, to reduce downstream thermal NOx generation by removing some heat of reaction prior to gas-phase combustion. For these liquid hydrocarbon fuel applications, ultra-low NOx emissions (< 3 ppm) have not been considered feasible because of these fuels’ propensity for autoignition during mixing with additional combustion air downstream of the catalyst. Even for natural gas fuel, previous systems have not permitted mixing of raw catalyst effluent with additional combustion air. For example, Acurex tested a two-stage natural gas combustion system having a fuel-rich catalyst stage followed by inter-stage heat extraction12. Additional combustion air was introduced only after heat extraction, and prior to a final fuel-lean catalytic combustion stage. Precision Combustion, Inc. (PCI), however, demonstrated that it is possible to directly mix catalyst effluent with additional combustion air without incurring autoignition13. This is possible because significant improvement in combustion stability is imparted to the downstream fuel-lean combustion process even at catalyst effluent temperatures well below the instantaneous autoignition temperature of the effluent. Thus, low-temperature low-emissions combustion requires only moderate catalyst temperatures, and autoignition can be avoided. This improvement in combustion stability with only moderate heat release on the catalyst can be quantified most simply by modeling the gas-phase combustion process as a zero-dimensional stirred reactor, as shown in figure 1. A simple Perfectly-StirredReactor (PSR) combustion model demonstrates that combustion stability is significantly improved when catalytic pre-reaction heats the combustor inlet gases from 400°C to 700°C, well below the instantaneous autoignition temperature. Details of the model are described by Smith et al. and model results for methane/air combustion are plotted here in figure 1 as flame temperature at incipient blowout (y-axis) versus combustor inlet temperature (x-axis)14. In this model, combustor inlet temperature is 400°C when there is no catalytic reaction. For the 400°C no-catalyst case lean blowout occurs at a flame temperature near 1305°C . With catalytic pre-reaction providing a 700°C inlet temperature to the combustor, however, the flame temperature at lean blowout drops by 55°C (100°F) to near 1250°C, a significant improvement. This improvement is achieved without risk of autoignition. For a catalyst effluent temperature (combustor inlet temperature) of 700°C, the estimated autoignition delay time is about 25 ms. This autoignition delay time was calculated using the correlation of Spadaccini and Colket for a representative natural gas composition (94.9% CH4, 3.1% C2H6, 0.65% C3H8, 0.3% C4H10, 0.1% C5H12, 0.1% C6H14, 0.05% C7 and higher-order hydrocarbons, and 0.8% diluent)15. For this calculation, pressure was 15 atm, equivalence ratio was 0.5 and the effect of vitiation from catalytic pre-reaction was neglected. Note that this 25 ms autoignition delay time at 700°C catalyst effluent temperature is far greater than the 2 to 5 ms residence time required to mix catalyst effluent with final combustion air, and is also greater than the typical 10-20 ms residence time of gas turbine combustors. Thus, complete mixing without autoignition is in fact possible downstream of the catalyst, yet combustion stability is markedly improved as a result of catalytic pre-reactions. The result is stable, low-temperature, ultra-low-NOx combustion.
o
PSR inlet temperature (post-catalyst), C
Fig. 1. Gas temperature (“flame” temperature) within PSR reactor at imminent blowout, as a function of PSR inlet temperature (catalyst exit temperature).
266
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle 3.2.2.1-3 Rich-Catalytic Lean-burn (RCL) Combustion Precision Combustion, Inc. (PCI) has developed a Rich-Catalytic Lean-burn combustion system (trademarked by PCI as RCL) based on this concept of stabilizing combustion of catalyst effluent having a temperature below the instantaneous autoignition temperature. The system catalytically reacts a portion of the fuel upstream of the combustor, thus preheating (and vitiating) the fuel/air mixture entering the combustor, and thereby improving combustion stability especially at low flame temperatures. NOx emissions are improved by operating at lower allowable (stable) flame temperatures, and turndown to low engine power can be improved by operating at still lower flame temperatures without excessive emissions of CO or unburned hydrocarbons. A schematic of PCI’s rich-catalytic combustion system is shown in figure 2. As shown, the combustion air stream is split into two parts upstream of the catalyst. One part is mixed with all of the fuel and contacts the catalyst, while the second part is used to backside cool the catalyst. At the exit of the catalytic reactor, the catalyzed fuel/air stream and the cooling air stream mix and then burn to completion to provide the final burner outlet temperature. As shown, combustion (fuel oxidation) occurs in two stages: a fuel-rich catalyst stage and an overall fuel-lean gas-phase combustion stage. Note that the catalyst is cooled only by primary combustion air, so that no heat is extracted from the system. With this approach, the fuel-rich mixture contacting the catalyst has insufficient oxygen to completely oxidize all of the fuel, thus limiting the extent of catalyst-stage reaction and enabling limitation of the catalyst-stage operating temperature to a safe value.
Catalyst Cooling
Combustion
Air
Burned Gas
Fuel
Premixer Catalytic Post-Catalyst Reactor Mixing Fig. 2. Schematic of Rich-Catalytic Lean-burn system, showing two-stage (rich-lean) combustion process.
3.2.2.1-4 Performance and Operating Characteristics of RCL Combustion PCI has tested the RCL catalytic reactor at pressures from 1 to more than 15 atm, in both sub-scale and full-scale tests and for multiple fuel types, providing design data over a wide range of operating conditions. Key performance parameters for RCL combustion are listed below, and measured values for these parameters are given in this section. From full-scale full-pressure rig tests using natural gas fuel: •
Combustor emissions and turndown
•
Combustor noise levels
•
Combustion system pressure drop
•
Catalyst operating temperatures
From sub-scale full-pressure rig tests to date: •
Catalyst lightoff and extinction temperatures
•
Alternate (non-methane) hydrocarbon fuel capability
•
Non-hydrocarbon fuel capability, e.g. syngas (discussed in later section)
In addition to these performance criteria, engine operational issues are also of interest, including method for engine start and catalyst lightoff, fuel staging needs, complexity of required controls, and transient capability such as load shifting and load rejection. To address these operational concerns, PCI and Solar Turbines have operated a modified Saturn engine using RCL combustion; these results are also presented here.
267
3.2.2.1 Fuel-Rich Catalytic Combustion 3.2.2.1-5 Full-Scale Full-Pressure Test at Solar Turbines Under DOE support and in cooperation with Solar Turbines, Inc., a full-scale fuel-rich catalytic reactor was fabricated and combustion tested to provide well-characterized experimental confirmation of the capabilities of Rich-Catalytic Lean-burn combustion for ultra-low-NOx power generation. The tests were performed in Solar Turbines’ full-pressure single-injector combustion rig using natural gas fuel, and are described in detail by Smith et al.,and summarized here16. A schematic of the catalytic combustor assembly, including premixer, catalytic reactor, and downstream combustor liner as tested in the rig is shown in figure 3. The catalytic reactor design is described generally in Smith et al. and includes backside cooling of the catalyst17. An annular reverse-flow premixer was fitted around the catalytic reactor, to provide a premixed fuel-rich mixture to the catalyst. Note that all fuel entered via this premixer, and all fuel contacted the catalyst. Catalyst cooling air bypassed the premixer, and entered from the left-hand side in figure 3. Downstream of the catalyst, but upstream of the combustor, the fuel-rich mixture and the catalyst cooling air were combined in a post-catalyst mixing duct (“post-mix” duct) to create a partially-reacted fuel-lean fuel/air mixture. The premixer, catalytic reactor, and post-mix duct together constitute what we call the “RCL-injector.” Conceptually, the RCL-injector replaces a conventional Dry Low-NOx (DLN) premixer/swirler arrangement, such as Solar’s SoLoNOx injector. For the tests reported here, a single Solar Turbines Taurus 70 SoLoNOx injector was replaced by an RCL-injector having a 3.0-inch diameter catalyst. An actual Taurus 70 engine would have 12 such injectors installed in its annular combustor. In Solar’s single-injector rig, the combustor liner was cylindrical, 8.0 inches in diameter, and backside-cooled. Four 0.5-inch diameter holes were located at the combustor’s downstream end, to allow dilution air to enter the combustor after cooling it. Some leakage air also entered the combustor at the injector insertion seal (grommet seal), where the RCL-injector was inserted into the combustor’s head end, or dome. A small amount of cooling air was also used to cool the combustor dome, and entered the combustor along the liner walls. A flameholding cone was installed at the exit of the post-mix duct, as shown in figure 3. Recirculation of hot combustion gases in the cone’s wake provided a flame anchor zone in the central part of the combustor. The expansion (dump) of the combustor liner’s dome also served to anchor combustion. In general, the RCL-injector catalyst is intended to improve combustion stability and turndown at the flame anchor point, but is not necessarily intended to provide gas-phase ignition. Solar’s torch igniter was used to ignite gas-phase combustion during rig testing.
Fig. 3. Assembly of 3-inch diameter fuel-rich catalytic reactor with 8-inch diameter combustor liner in Solar Turbines’ single-injector combustion test facility. Bulk flow is from left-to-right.
Combustor Emissions and Turndown Emissions performance and turndown were measured for the RCL combustion system depicted in figure 3, at nominal Taurus 70 operating conditions (16-17 atm pressure and 4 pps total airflow to the single-injector combustor). Measured NOx and CO emissions are plotted in figure 4 as a function of adiabatic flame temperature at the emissions rake. For the tests reported here, the emissions rake was located just upstream of the combustor dilution air holes, corresponding to about 30 ms combustor residence time. NOx and CO emissions are reported after correction to 15% O2 on a dry basis. UHC emissions are reported on a wet basis, corrected to 15% O2. It is also worth noting that, as discussed earlier, some leakage and cooling air entered the combustor between the RCL-injector and the emissions rake. For this reason, fuel/air ratio as measured at the emissions rake gave an adiabatic flame temperature that was about 130°C (230°F) lower than that based on fuel/air ratio measured at the RCL-injector exit. Imperfect mixing of this leakage air with the injector’s fuel/air mixture can increase NOx emissions to values slightly higher than expected for perfectly premixed combustion at the adiabatic flame temperatures measured at the emissions rake. In fact, the NOx emissions shown in figure 4 are about 1 ppm higher than expected at 1450°C (2650°F) based on perfectly premixed combustion. As shown in figure 4, the RCL combustion system achieved ultra-low emissions over a wide operating window of approximately 110°C (200°F) variation in flame temperature, with CO below 10 ppm and NOx below 3 ppm (and as low as 1 ppm). Unburned hydrocarbons (UHC) remained less than 2 ppm at all conditions shown in figure 4.
268
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle These high-pressure ultra-low-emissions results provide experimental confirmation of the ultra-low-NOx capability of the RCL combustion concept. In particular, they demonstrate that it is possible to mix fuel-rich catalyst effluent with final combustion air, without inducing autoignition, yet imparting significant combustion stability to the downstream combustion process to allow lowtemperature low-emissions combustion with wide turndown. Calculated Primary Zone Temperature (F) 2600
2650
NOx
4.5
NOx (ppm, 15% O2 dry)
2700
2750
2800
2850
20
CO
4.0
18 16
NOx
3.5 3.0
14 12
CO
2.5
10
2.0
8
1.5
6
1.0
4
0.5
2
0.0 2400
2450
2500
2550
2600
2650
CO (ppm, 15% O2 dry)
2550
5.0
0 2700
Adiabatic Flame Temperature at Emissions Rake (F) Fig. 4. NOx and CO emissions, as a function of adiabatic flame temperature at emissions rake. Data obtained at 16 atm pressure.
Combustor Noise Levels Combustion-driven pressure oscillations (noise) were also monitored during full-scale combustion tests at Solar, and remained less than 2.4 kPa (0.35 psi) peak-to-peak (less than 0.15% peak-to-peak of mean combustor pressure) at all conditions tested, indicating quiet operation. Low levels of combustion noise are expected, since gas-phase energy release in the combustor (the driving force for combustion noise) is reduced when a portion of the fuel is catalytically reacted prior to gas-phase combustion. Combustion System Pressure Drop Pressure drop through the fuel-rich catalytic reactor is a primary determiner of RCL-injector size for any given application. For the 3-inch diameter reactor tested at Solar, pressure drop through the entire combustion system at simulated Taurus 70 full load conditions varied from about 3.5% to 5% of combustor inlet pressure, depending on hardware modifications made to redirect airflow during rig testing. This pressure drop includes both the losses through the fuel-rich catalytic reactor and losses in the downstream combustor (pressure drop across flameholder, dump loss at combustor inlet, fundamental combustion loss, etc.). We estimate that losses in the downstream combustor account for about 0.5% pressure drop, with the remaining best-case 3% attributable to the catalytic reactor in Solar’s rig. Additional pressure loss data has been obtained for other full-scale RCL-injectors not tested in Solar’s rig, but designed for reduced pressure drop. Data from other rigs indicate that at Solar rig conditions pressure drop would be about 2%. Catalyst Operating Temperatures Figure 5 shows steady-state catalyst surface temperatures plotted against adiabatic flame temperature at the full-scale RCLinjector exit, as tested at Solar. As shown in figure 5, catalyst surface temperature increases only slightly as fuel flow is reduced, and all catalyst surface temperature measurements remain below 780°C (1430°F) over the complete range of operating conditions tested (1440-1700°C / 2620-3090°F range in adiabatic flame temperature). RCL catalyst temperatures do not vary significantly with fuel/air ratio because reaction rate (heat release) upon the catalyst surface is controlled by oxygen flow (air flow) under fuel-rich conditions, and because heat removal (heat transfer) from the catalyst is also determined primarily by air flow. Fuel flow has little effect on reaction rate and little effect on heat removal rate. This insensitivity of catalyst temperature to fuel/air ratio is advantageous in allowing combustor and turbine operation over a wide range of flame temperatures (including flame temperatures well above the low-NOx-emissions range), making the RCL system suitable even for advanced high-firing-temperature machines.
269
3.2.2.1 Fuel-Rich Catalytic Combustion 800 Catalyst Temperature (C)
750 700 650
P = 15 - 16 atm
600
Tinlet air = 440 C
550 500 450 400 1400
1450
1500
1550
1600
1650
1700
1750
Adiabatic Flame Temperature at RCL Injector Exit (C) Fig. 5. Catalyst surface temperature as a function of adiabatic flame temperature at RCL-injector exit.
3.2.2.1-6 Sub-Scale Test Data for Hydrocarbon Fuels Catalyst testing under controlled conditions is best conducted at sub-scale, where smaller-size equipment allows for accurate metering and control of flow and temperature. Thus, accurate values for catalyst lightoff and extinction temperature are obtainable. Sub-scale testing is also useful for evaluating new concepts, such as use of alternative fuels. Catalyst Lightoff and Extinction Temperatures Catalyst lightoff and extinction tests have been performed under well-controlled experimental conditions at sub-scale for pressures from 9 to 15 atm. For natural gas fuel having one or two percent ethane, PCI’s fuel-rich catalysts typically light off in the vicinity of 300°C. For natural gas fuel with greater than two percent ethane (or higher-order hydrocarbons) lightoff can occur at inlet temperatures below 280°C. This is shown in figure 6 below, which indicates a lightoff temperature between about 260 and 280°C for natural gas fuel, at 15 atm pressure. In figure 6, inlet gas temperature, catalyst surface temperature, and gas temperature exiting the module (following mixing of the catalytically reacted stream with the catalyst cooling air stream, but prior to gas-phase combustion) are plotted as a function of time in minutes. Lightoff occurs when the heat of reaction results in an increase in catalyst operating temperature and catalyst exit temperature as compared to the gas inlet temperature, as described in Section 3.2.2. of this Handbook
15 atm pressure 1470 F
1110 F
750 F
Fig. 6. Catalyst lightoff in a sub-scale high-pressure (15 atm) fuel-rich reactor operating on natural gas fuel. Inlet gas temperature (“T gas in”), catalyst surface temperature (“T catalyst”), and gas temperature exiting the module (“T gas out”) are plotted as a function of time in minutes.
270
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle 15 atm pressure
1470 F 1110 F
750 F 390 F
Fig. 7. Catalyst extinction does not occur until the fuel is shut off at an inlet air temperature less than 200 C. Data were obtained for the same sub-scale high-pressure (15 atm) fuel-rich reactor for which data were shown in Figure 6. Again, inlet gas temperature (“T gas in”), catalyst surface temperature (“T catalyst”), and gas temperature exiting the module (“T gas out”) are plotted as a function of time in minutes.
Following catalyst lightoff, the inlet air temperature can be reduced well below the initial lightoff temperature without extinguishing the catalyst. Thus, once lit (active), the catalyst remains lit (active) down to inlet temperatures approaching ambient. Following the catalyst lightoff event depicted in figure 6, the inlet air temperature was reduced to less than 200°C, but catalyst activity was not diminished. This is shown below in figure 7, which plots the same parameters as figure 6, now after several hours of testing following the initial lightoff. Here, still at 15 atm pressure and with the same flow of natural gas fuel, catalyst activity was maintained until the fuel was shut off at an inlet air temperature less than 200°C. Alternate Hydrocarbon Fuel Capability Sub-scale fuel-rich catalyst tests have also been performed using alternative (non-natural-gas) hydrocarbon fuels. In particular, liquid fuels have been tested (gasoline and Diesel No. 2 fuel) with performance similar to that obtained using methane or natural gas, and a simulated landfill gas fuel has also been tested.
COMBUSTOR BURNOUT SECTION
CATALYTIC REACTOR
BACK PRESSURE VALVE
WATER COOLED SAMPLE LINES
Fig. 8. Photograph of PCI’s 10 atm sub-scale combustion rig, for testing fuelrich catalytic reactors with downstream combustion. Bulk flow is from right to left in this photograph.
271
Sub-scale tests with alternative fuels were performed in a 10 atm combustion test rig at PCI equipped for obtaining combustion emissions measurements, as shown in figure 8. The gas-phase combustion section of this rig is nominally 2 inches in diameter and 18 inches in length, and is fabricated from high-temperature ceramic and housed in a steel pressure vessel as shown in the photograph of figure 8. Reactants enter the ceramic-lined combustor through an air-cooled fuel-rich catalytic reactor of nominally 3/4-inch
3.2.2.1 Fuel-Rich Catalytic Combustion diameter. Combustion emissions reported for the alternative-fuel tests were obtained from a water-cooled gas sampling probe location corresponding to 30 ms residence time in the ceramic-lined combustor downstream of the catalyst. A single sub-scale fuel-rich catalytic reactor was tested in PCI’s high-pressure combustion rig with three different fuels: methane, simulated landfill gas, and Diesel No. 2 (the first three listed in Table 3). No changes were made to the reactor or combustor for operation on these different fuels. For diesel fuel, however, a prevaporizer was added upstream of the reactor. Two different prevaporizers were used: initially we used a simple preheater to directly heat diesel fuel after adding less than 10% N2 (by weight) to assist in atomization; later we improved prevaporization by mixing steam with the diesel fuel. The latter was considered ideal for co-generation applications. Note that liquid fuel tests were generally performed at 6-8 atm pressure, based on limitations of the fuel prevaporizers. Prior to testing of the first three fuels listed in Table 3, a separate fuel-rich catalytic reactor of similar design was tested with gasoline and natural gas. For these initial liquid fuel tests gasoline was chosen for its high volatility, simplifying prevaporizer design. For all fuels, the most relevant measures of performance for PCI’s two-stage catalytic combustion system are catalyst temperatures in the first stage (catalyst-stage) and combustor emissions from the second stage (gas-phase combustion stage). We especially wish to compare these measures of performance for operation on various types of fuels.
Fig. 9 (left panel). Combined graph showing catalyst operation for both natural gas and liquid fuel (gasoline) at 7 atm pressure and 0.4 equivalence ratio (Φ), as a function of time. For both fuels, the inlet gas temperature (“Tgas in”) was initially held steady at about 350 C, and then ramped down to about 200 C without catalyst extinction.
Fig. 10 (right panel). Catalyst surface temperatures, and lightoff and extinction temperatures, for diesel fuel operation. Tests were performed with non-steam prevaporizer, at 6 atm pressure and 0.33 overall equivalence ratio (Φ). "Tinlet" represents air temperature entering the reactor, and "Tcatalyst" represents catalyst surface temperature.
In figure 9, catalyst temperatures for 7 atm operation on gasoline are compared to those for natural gas. As shown, catalyst operating temperature (“Tsurfaces”) and catalyst lightoff temperature are both very similar for the two dissimilar fuels. Note that for both fuels the overall equivalence ratio was 0.4 downstream of the catalyst, after mixing of catalyst cooling air with fuel-rich catalyst effluent. Catalyst lightoff and extinction temperatures for diesel fuel were tested using the first-generation prevaporizer at 6 atm pressure and 0.33 equivalence ratio, as shown in figure 10. For this test, temperature of the prevaporized fuel was between 350 and 380°C before mixing with air. For catalyst lightoff, inlet air temperature was ramped up from about 345°C until definitive lightoff occurred at about 360°C inlet air temperature, as indicated by a rapid increase in catalyst temperature. Prior to this event, some reaction occurred along the length of the reactor, as evidenced by catalyst temperatures nearly 75°C higher than the inlet temperature (e.g. 420°C versus 350°C). After lightoff, inlet air temperature was ramped down until sudden loss of activity (catalyst extinction) occurred at about 200°C inlet air temperature. Note that for all fuel types tested under fuel-rich conditions, catalyst extinction temperature was well below catalyst lightoff temperature. Also note the similar catalyst operating temperatures for diesel fuel as compared to gasoline and natural gas operating temperatures, as a result of fuel-rich operation of the catalyst. Although successful operation on diesel fuel was obtained using the non-steam prevaporizer, it could only provide enough prevaporized fuel to establish an overall equivalence ratio of 0.33 downstream of the catalyst. A steam prevaporizer was therefore designed and used, allowing catalyst temperatures and combustion emissions to be measured over a wider range of equivalence ratios (flame temperatures), as shown in figures 11 and 12. For diesel fuel operation using the second-generation prevaporizer, steady-state catalyst temperature data (“T surface”) are shown as a function of the reactor’s overall equivalence ratio in figure 11. Diesel fuel operating data were obtained at the prevaporizer’s maximum operating pressure of 6 atm, and at 430°C inlet temperature. Average gas temperature exiting the reactor (“T gas out”) is also shown. Note that “T gas out” and overall equivalence ratio are both defined after mixing of catalyst effluent with catalyst cooling air.
272
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle
Fig. 11. Catalyst performance with diesel fuel versus methane, for second-generation prevaporizer (~5:1 steam:fuel ratio by weight) operating at 6 atm pressure and 430°C inlet air temperature. Compare to methane tests at 9 atm and 440°C inlet temperature.
As shown in figure 11, catalyst operating temperatures are insensitive to operating condition (overall equivalence ratio) for both methane and diesel fuels, and in general very similar catalyst performance was obtained for both fuels. As stated earlier, catalyst temperatures do not vary significantly with fuel/air ratio because reaction rate (heat release) upon the catalyst surface is controlled by oxygen flow (air flow) under fuel-rich conditions. Thus, the oxygen available for reaction (the limiting reactant under fuel-rich conditions) is the same at all test conditions, with the result that heat release and temperatures in the catalyst bed are insensitive to equivalence ratio, and are very similar for both methane and diesel fuels despite a wide difference in reactivity between these two fuels. NOx emissions from the fuel-lean combustion zone downstream of the fuel-rich catalyst were also measured for both liquid fuels and gaseous fuels, as shown in figure 12. Note that NOx emissions from the diesel fuel are primarily due to fuel-bound nitrogen, except at the highest flame temperature tested (near 2900°F). The Diesel fuel tested was red-dyed Diesel No. 2 fuel, and its nitrogen content was measured at 188 ppm by weight by an independent laboratory. At this concentration, nearly complete conversion of fuelbound N to NOx is expected [4]. Thus, based on fuel-bound nitrogen content, the diesel fuel would emit at least 8.1 ppmv NOx (at 15% O2) when burned. Fuel-bound nitrogen for all other fuels was zero. NOx emissions are shown in figure 12 for three fuel types: methane, simulated bio-mass landfill gas (essentially diluted methane), and Diesel No. 2 fuel. Here, NOx emissions are measured on a dry basis and are corrected to 15% excess O2. NOx emissions are shown as a function of maximum measured flame temperature (via type S thermocouple) for each data point. For all data points obtained, CO and unburned hydrocarbon (UHC) emissions were less than 2 ppmv.
273
3.2.2.1 Fuel-Rich Catalytic Combustion
20
NOTE: For methane data measured and calculated flame temperatures generally agree within 50 degrees Fahrenheit.
18
NOx (15% O2) ppmv
16 14 12
Methane - measured NOx (32 ms, 9 atm) , calculated adiabatic flame temperature Biofuel - measured NOx (32 ms, 9 atm) , calculated adiabatic flame temperature Diesel (prevap, steam) - measured NOx (32 ms, 6 atm), measured flame temp Diesel (prevap, steam) - measured NOx (32 ms, 6 atm), measured flame temp Calculated NOx emissions from fuel-bound nitrogen in Diesel Prediction 10 atm, 30 ms
10 8 6 4 2 0 1900
2000
2100
2200
2300
2400
2500
2600
2700
2800
2900
Flame Temperature (°F) Fig. 12. NOx emissions for three fuel types (methane, biomass landfill gas, and Diesel No. 2 fuel). For all data points, CO/UHC emissions were less than 2 ppmv. For the Diesel No. 2 fuel, fuel analysis indicated that 8.1 ppmv NOx would be emitted as a result of fuel-bound nitrogen alone.
For methane and bio-mass landfill gas fuels, NOx emissions were below 3 ppm for measured flame temperatures below 2600°F. For diesel fuel, NOx emissions were about 10 ppm for measured flame temperatures below 2600°F, compared to the 8.1 ppm expected based on fuel-bound nitrogen alone. Thus, about 2 ppm NOx is likely formed by prompt (non-thermal) mechanisms at low flame temperatures (below 2600°F). At higher flame temperatures, NOx increases due to thermal formation mechanisms for all three fuels, as shown. The low NOx levels at low flame temperature indicates that well-mixed fuel-lean combustion was achieved downstream of the catalyst for all three fuels: methane, bio-mass landfill gas, and diesel.
3.2.2.1-7 Engine Test Results Based on the successful full-scale single-injector rig tests at Solar Turbines, a “cluster” of four RCL-injectors was installed in a modified (single can combustor) Solar Turbines Saturn engine, to assess controls compatibility and transient operation in an engine environment, including engine start, acceleration, and load variation. In addition, steady-state operating data were obtained, including NOx and CO emissions at the engine exhaust. The engine test also provided a basis for evaluating fuel-rich reactor robustness in an engine environment, over a range of operating conditions and demands (including start, acceleration, and load). Test Engine Specifications and Configuration Combustor Housing Compressor Discharge Air Pipe Combustor Primary Zone Air Pipe Dilution Air Pipe Burner Outlet Pipe (to turbine)
274
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle
Fig. 13. Side-mounted combustor configuration in modified Saturn engine, showing variable airflow control valves in primary zone air pipe and dilution air pipe.
Fig. 14. Photograph of four-RCL-injector assembly, prior to installation in Saturn engine.
The test engine was a modified version of a two-shaft recuperated Saturn T1200 engine, nominally rated at 750 kW (1000 hp) after modification. This engine was selected as a test bed because its external combustor configuration was amenable to modification. For catalytic combustor testing, the recuperator was removed, but the compressor discharge scroll and turbine inlet scroll were retained, allowing a single side-mounted combustor can to be installed. The overall combustor configuration is shown in figure 13. Note that variable airflow butterfly valves were fitted in the combustor primary zone air pipe and the dilution air pipe, to allow combustor air to be varied for best emissions at any given fuel flow (engine load). Also note that a preburner was located in the combustor primary zone air pipe below the butterfly valve, to temporarily increase catalyst inlet air temperature to about 350°C (660°F) to ensure catalyst lightoff. The preburner was turned off after catalyst lightoff, and before engine emissions were measured. All fuel and air entered the combustor through the four RCL-injectors (neglecting leakage air). The combustor liner was backside cooled with dilution air, before the dilution air entered the hot gas path 24 inches downstream of the combustor’s upstream end (the round plate through which the post-mix ducts are inserted, visible in figure 14, forms the combustor’s upstream end). The combustor liner itself was cylindrical and 15 inches in diameter. At full Saturn engine load, and assuming 1.3 pps airflow through each RCL-injector for ultra-low-emissions operation, combustor residence time was about 35 ms. Engine Operating Procedure Engine start-up data are shown in figure 15, with annotations, giving a graphical depiction of the start-up procedure. Note that there are three fuel circuits: a preburner fuel stage, which received about 25 kg/hr (55 pph) fuel during catalyst lightoff, and two RCLinjector fuel stages, which together received up to about 275 kg/hr (600 pph) fuel at load. RCL-injector fuel stage A supplied fuel to the top two injectors, while RCL-injector fuel stage B supplied fuel to the bottom two injectors. At cold crank conditions (29% gas producer shaft speed, Ngp) the preburner was ignited and adjusted to 260°C (500°F) outlet temperature, below the catalyst lightoff temperature. As seen in figure 15, the small preburner fuel flow provided little motive power to the engine and negligible increase in engine speed. Next, while still at 29% Ngp, fuel was introduced to the RCL-injectors and combustion was ignited by a torch igniter in the main combustor. With the starter motor still engaged, fuel flow was ramped up as the engine accelerated to 65% Ngp. At 65% Ngp the starter motor was disengaged and the engine controller added fuel to maintain a constant idle speed of 65% Ngp (no load). Preburner outlet temperature remained at 260°C (500°F), and the catalysts remained inactive.
275
3.2.2.1 Fuel-Rich Catalytic Combustion
225
Fuel mass flow (Wf, kg/hr)
200 175
90
70
Wf_A
fuel RCL injectors, ignite combustion, accel to 65% Ngp
125
80
Adjust combustor air valves for best emissions
idle
150
100
Ngp
Preburner Fuel (Wf_pbn) RCL Stage A fuel (Wf_A) RCL Stage B Fuel (Wf_B) Shaft Speed (Ngp)
60 50
Wf_B
100 75 50
40
cold crank
increase preburner fuel (RCL inlet temp) to light-off catalysts
preburner fuel on @ 29% Ngp
12:43:12
20
turn preburner fuel off @ 80% Ngp -catalysts remain active (lit off)
25 0 12:36:00
30
12:50:24
12:57:36
Wf_pbn 13:04:48
Gas producer shaft speed (Ngp, %)
250
10
0 13:12:00
Time
Fig. 15. Saturn engine start-up data, showing engine acceleration, catalyst activation by preburner (followed by preburner shutoff with continued catalyst activity), and operation at engine load by fuel supplied through four RCL-injectors.
Preburner temperature was then increased to about 350°C (660°F) to ensure catalyst lightoff. Engine speed was increased to 80% Ngp, the preburner was turned off, and the catalysts remained active. Engine speed was then increased to 90% and the variable airflow valves were adjusted to obtain optimum emissions. The valves served to vary the airflow to the RCL-injectors thus allowing control of NOx and CO emissions. Emissions data were taken as engine speed was reduced in increments of about 1% Ngp. The airflow valves were adjusted for best emissions at each speed. Engine controls were based on a Saturn T1202R design and used a state of the art Allen-Bradley microprocessor console to run the logic. For the RCL combustor engine tests, catalyst temperatures were not used in the fuel control algorithm. Instead, fuel control was performed according to standard DLN methods (primarily monitoring engine speed versus set point), with the addition of a preburner fuel control during initial start and catalyst lightoff. This was possible because catalyst temperature is insensitive to fuel/air ratio under fuel-rich conditions, as shown in figure 2.1.2 for the single-injector rig tests. In addition, the RCL catalyst is air-cooled by a large fraction of the total combustion air, and reactions on the catalyst are limited by available oxygen (fuel-rich); thus, the catalyst is resistant to flashback, autoignition, and overheating damage, and can operate safely without active temperature control. Engine Performance with RCL Combustor With RCL combustion, Saturn engine NOx emissions averaged 2.1 ppm with less than 10 ppm CO over an achievable engine operating range (82% to 89% Ngp), as shown in figure 16. Over this engine operating range, UHC emissions remained below 3 ppm, and combustion-driven pressure oscillations (CDPO) remained less than 0.7 kPa (0.1 psi) peak-to-peak (less than 0.15% peak-to-peak of mean combustor pressure). At 89% Ngp, combustor inlet air (compressor discharge air) was at 5.0 atm and 223°C (434°F). At 82% Ngp, combustor inlet air was at 3.9 atm and 191°C (376°F). For all data points shown in Figure 2.1.16 the preburner was turned off, the catalyst remained active at the available compressor discharge temperatures (as low as 191°C / 376°F), and NOx emissions remained below 3 ppm. Measured power output ranged from 237 kW (318 hp) to 453 kW (607 hp) over the 82% to 89% Ngp operating range, or about 32% to 61% load based on a 750 kW (1000 hp) nominal power rating for this modified engine. Engine load was delivered to a water dynamometer. Engine operation was limited to the 82% to 89% speed range. At less than 82% Ngp the compressor was at its surge condition, and the compressor bleed valve was opened to prevent surge. This reduced the airflow to the RCL-injectors thus increasing NOx emissions. At speeds greater than 89% Ngp operation was limited by locally hot temperatures within the scroll ducting downstream of the combustor. This limitation was not attributable to the RCL combustion technology but to inadequate mixing of combustor dilution air. Improving the test rig dilution mixing was deemed unnecessary to document the controllability of the RCL system. Table 1 summarizes the Saturn engine operating data at the low-end and high-end of the achievable operating range. In general, the results show good combustor performance (low emissions and low combustion-induced pressure oscillations, or CDPO) even at very low inlet temperatures. In addition, the Saturn engine operation shows the feasibility of engine start-up, acceleration, and operation at load using RCL combustion with simple engine controls. The engine was successfully started, accelerated, and powered at load by fuel injected through the four catalytic reactors, using conventional engine instrumentation and controls without instrumentation input from the catalyst.
276
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle
NOx / CO (ppm, @ 15% O2 dry)
10 9 8 7 6
NOx
5
CO
4 3 2 1 0 81
82
83
84
85
86
87
88
89
90
Ngp (gas producer shaft speed, % of max) Fig. 16. RCL combustor emissions during Saturn engine operation, showing ultra-low NOx and CO emissions over an achievable engine operating range of 82% to 89% speed. Table 1 Saturn engine operating data at low-end and high-end of achievable operating range. Note catalyst activity and ultra-low-emissions achieved at inlet temperatures as low as 191 C (376 F).
Engine Speed NOx Emissions CO Emissions CDPO (noise) Power Output Nominal Load Comb. Inlet Pressure Comb. Inlet Temp.
82% Ngp 2.2 ppm 9.5 ppm < 0.7 kPa pk-pk 237 kW / 318 hp 32% 3.9 atm 191 C / 376 F
89% Ngp 2.2 ppm 5.7 ppm < 0.7 kPa pk-pk 453 kW / 607 hp 61% 5.0 atm 223 C / 434 F
3.2.2.1-8 Ultra-Low NOx Combustion of Coal-Derived Syngas and High-Hydrogen Fuels While early development of PCI’s Rich-Catalytic Lean-burn (RCL) combustion system was focused on the use of natural gas fuel, fuel-rich operation of the catalyst is also advantageous for alternative fuels, including high-hydrogen fuels and coal-derived syngas. As with natural gas, combustion stability is improved (especially at low flame temperatures) by catalytically reacting a portion of the fuel upstream of the combustor, thus preheating (and vitiating) the fuel/air mixture entering the combustor. And, as with natural gas, this allows improved NOx emissions by operating the combustor at lower allowable (stable) flame temperatures. For low-Btu fuels such as blast furnace gas, flame stability improvement by catalytic pre-reaction of a portion of the fuel is especially important since combustion may be otherwise unsustainable even at the highest possible flame temperatures. With catalytic pre-reaction and concomitant pre-heating, however, fuels having heating values as low as 82 Btu/ft3 have been successfully combusted at PCI, with near-zero emissions of unburned fuel, CO, and NOx. For fuels with more moderate dilution levels, such as the syngas delivered from a coal gasifier, combustion of the raw fuel can occur without a catalyst, but low-emissions combustion is difficult as a result of the high levels of hydrogen. In particular, lean-premixed combustion for syngas fuels has not generally been considered feasible, because the high concentration of hydrogen leads to increased risk for flashback and flameholding in regions upstream of the combustor. Thus, syngas fuels are generally burned in a non-premixed mode, with NOx control accomplished by dilution of the fuel stream with water and/or nitrogen. This introduces combustion stability issues, however, such that low-single-digit NOx emissions have not yet been achieved in gas turbines burning coal-derived syngas. One solution to this combustion stability problem is to catalytically react some portion of the syngas fuel prior to gas-phase combustion, effectively providing a higher inlet temperature to the combustor. This is exactly the role of a catalyst in a lean-premixed combustion system, and it can also be applied to combustion of syngas fuel, allowing greater dilution of the fuel and reduced NOx emissions.
277
3.2.2.1 Fuel-Rich Catalytic Combustion For any fuel type, whether syngas, low-Btu gas, or hydrocarbon, fuel-rich catalytic reactor performance is insensitive to the fuel’s reactivity, because reaction rate (heat release) upon the catalyst surface is controlled primarily by oxygen flow (air flow) under fuel-rich conditions, and not by fuel flow or reactivity. Performance on different type fuels will therefore be similar when heat release per atom of oxygen reacted is similar, and when the fuel’s mass and thermal capacity is negligible in the fuel/air mixture. This is generally the case for hydrocarbon fuels, and the primary remaining issue for operation on heavy liquid fuels is prevaporization. For coal-derived syngas fuel, heat release per atom of oxygen reacted is similar to hydrocarbon fuels, and additionally prevaporization is not an issue. However, unlike fuels consisting only of hydrocarbons, the large volume (low Btu value) of syngas fuels is not negligible, particularly when the syngas is highly diluted with steam or nitrogen for NOx control. Thus, while fuel-rich catalytic reactor performance for syngas fuels can be made similar to that obtained on hydrocarbon fuels (on the basis of heat release per atom of oxygen reacted), the reactor design must consider the large volume flow of fuel that it must pass, as well as the oxygen flow required for the desired level of pre-reaction or pre-heat.
3.2.2.1-9 Sub-Scale Test Data for Syngas Fuel A sub-scale fuel-rich catalytic reactor was fabricated at PCI for testing with syngas fuel in PCI’s high-pressure sub-scale combustion rig (pictured in figure 8). For the syngas tests, two independently controllable air supplies are provided, both heated and at high pressure. The larger air supply (entering from the right in figure 8) provides catalyst cooling air, which becomes primary zone combustion air in the gas-phase combustor, and the smaller air supply (entering from the vertical pipe at the top-right of figure 8) provides air to the fuel-rich fuel/air mixture. For operation with syngas fuel, two heaters are also provided (but not shown in Figure 8); one heater heats N2 diluent just before it is mixed with fuel, and the second heater heats all other fuel components and CO2. Dry, oil-free, high-pressure air is supplied to the rig from compressors, at pressures up to about 145 psia entering the rig. At this flow rate, the rig inlet air can be heated to about 500°C. Fuel and diluent are supplied from bottles or Dewar flasks at high pressure, and are pressure regulated to the proper delivery pressure to the rig. All flows (air, fuel, and diluent) are metered with electronic mass flow controllers. Each fuel component is separately metered and then mixed with the other components to simulate the desired coal-derived syngas composition. Up to five fuel components can be introduced: H2, CO, CH4, CO2, and N2. For syngas operation, combustion rig startup was accomplished by bringing the reactor to fuel-rich conditions using unheated methane fuel, with some diluent addition to ensure proper mixing. When necessary, a small amount of H2 was temporarily added to the unheated methane to light off the reactor. Once the catalyst and combustor were lit and the rig was thermally stable, syngas fuel flow was ramped up while methane fuel flow was ramped down, holding catalyst equivalence ratio approximately constant. This startup procedure was economical and safe: it minimized the use of high-volume (costly) laboratory syngas fuel blend, and also avoided use of H2 during transient and ignition events, where there was a concern that unburned H2 might enter the exhaust stack and create an explosion hazard. This procedure is similar to syngas combustor startup in actual engine applications. Catalyst Lightoff Prior to high-pressure testing with syngas, atmospheric-pressure tests were performed to provide some initial experience in syngas fuel operation, and in catalyst and combustor behavior using syngas fuels. The results were used to help guide reactor design and test planning for the subsequent high-pressure tests. Catalyst lightoff tests were an important part of these early atmospheric pressure tests, and a lightoff temperature of 180°C was measured for fuel-rich reaction of syngas. This value was most carefully measured for a syngas mixture composed of 25% H2 and 35% CO (remainder diluent), but it was found that lightoff temperature is relatively insensitive to syngas composition unless CO levels drop to very low values, leaving essentially a high-hydrogen fuel having significantly lower lightoff temperature. In all cases, the 180°C or lower lightoff temperature is well below compressor discharge temperatures of industrial and large-frame turbines, and the need for a preburner is thereby avoided in expected syngas applications. High-Pressure Test Conditions For the high-pressure sub-scale syngas tests, “baseline” operating conditions were scaled from those published for the IGCC plant at Tampa Electric’s Polk Power Station. The Tampa Polk plant operates a GE 107FA combined cycle system on syngas generated from a Texaco oxygen-blown coal gasifier. Nitrogen injection reduces the effective heating value of the fuel, for NOx control. Data from references by the U.S. DOE, GE Power Systems, and Brdar and Jones were used to establish a baseline syngas fuel composition representative of the Tampa Polk IGCC power plant fuel after dilution with nitrogen for NOx control, as well as a baseline full-load firing temperature18. This baseline syngas composition was used for most sub-scale high-pressure testing at PCI, and is tabulated in the first row of Table 2 as noted. The baseline full-load flame temperature was determined to be about 2550°F (1400°C), and represents a maximum in NOx emissions for an expected syngas application of fuel-rich catalytic combustion. Additional tests were performed for a highly-diluted low-Btu syngas, to demonstrate the flame stability augmentation provided by the catalyst. The second row of Table 1 lists the composition of this low-Btu syngas mixture. Note that for all sub-scale tests at PCI, the nitrogen diluent was added to the syngas fuel prior to mixing with air, and well upstream of the catalyst.
278
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle Table 2 Syngas fuel compositions used for high-pressure sub-scale tests.
Baseline
H2 (%) 20
CO (%) 20
CO2 (%) 10
N2 (%) 50
LHV (Btu/ft3) 117
Low-Btu
15
15
10
60
88
Syngas Composition
Emissions measurements reported here were obtained from a water-cooled gas sample probe located 15 inches downstream of the catalyst, corresponding to about 50 ms residence time in the 2-inch inside-diameter ceramic-lined combustor. This represents the maximum residence time expected in a low-emissions gas turbine combustor, and therefore also represents the maximum expected NOx emissions for a given operating condition. Note that all emissions reported in ppm are parts per million by volume, and are corrected to 15% O2 dry. All measurements were made with a combustor inlet air temperature of 750°F (400°C) and a syngas fuel temperature of 570° F (300°C). Adiabatic flame temperatures were calculated based on fuel/air ratio as measured by the emissions analyzers (i.e. from gas samples extracted at the 15-inch gas sample probe location). Combustor Emissions and Turndown Figure 17 plots measured NOx and CO emissions as a function of adiabatic flame temperature at 10 atm pressure for the baseline syngas composition listed in row 1 of Table 2, having a Lower Heating Value (LHV) of 117 Btu/ft3. With this fuel composition, NOx emissions were 2.0 ppm (0.011 lbs/MMBtu) at the 2550°F (1400°C) flame temperature data point corresponding to baseline operation in the Tampa Polk IGCC plant at 100% load. As the fuel/air ratio was decreased, CO emissions remained near zero for flame temperatures greater than about 2250°F (1230° C), permitting a 300°F (150°C) turndown in flame temperature from the 2550°F (1400°C) baseline point, allowing ultra low emissions operation over a wide range of loads. These results − CO near zero, and NOx equal to or less than 2 ppm (0.011 lbs/MMBtu) for full load and below − easily met PCI’s emissions goals for these tests.
0.025
Syngas: 20% H2, 20% CO, 10% CO2, 50% N2 (LHV = 117Btu/ft3)
5.4 4.5
Combustor rig data, P = 10 atm
0.020 0.015
3.6 CO
Baseline Combustor Temp (2550 F)
NOx
2.7
0.010
1.8
0.005
0.9
0.000 0.0 2100 2200 2300 2400 2500 2600 2700 2800 Adiabatic Flame Temperature @ emissions probe (F) Fig. 17. Measured NOx and CO emissions for 10 atm baseline syngas tests, as a function of adiabatic flame temperature at the emissions probe.
279
NOx / CO (ppm @ 15% O2)
NOx ((lbs NOx)/MMBtu)
0.030
3.2.2.1 Fuel-Rich Catalytic Combustion
NOx ((lbs NOx)/MMBtu)
0.035 0.030
7.2
Except as noted: P = 10 atm; τres = 50 ms air @ 400 C (750 F), syngas @ 300 C (570 F) baseline syngas composition, LHV = 117Btu/ft (20% H2, 20% CO, 10% CO 2, 50% N2)
6.3 5.4
3
0.025
4.5
0.020
3.6
0.015
2.7
LHV = 117 Btu/ft3
3
LHV = 88 Btu/ft
Btu/ft
0.010 0.005
1.8 0.9
0.000 2100
2200
2300
2400
2500
2600
2700
NOx (ppm, 15% O2) - Baseline Syngas Only
0.040
0.0 2800
Adiabatic Flame Temperature @ emissions Probe (F) Fig. 18. Measured NOx emissions for two different syngas compositions having LHVs of 88 and 117 Btu/ft3.
Subsequent to the baseline tests, the heating value of the diluted syngas fuel was reduced to determine operability and emissions for highly-diluted low-Btu syngas fuels. NOx emissions are shown in figure 18 for the low-Btu syngas composition listed in row 2 of table 2, as well as for the baseline syngas composition. It is important to note that the right-hand vertical axis in figure 18 (NOx values in ppm) is only applicable to the baseline syngas composition, as marked. For the fuel composition with the lower heating value (88 Btu/ft3), NOx emissions in ppm are slightly lower than shown (for the low-Btu fuel, 0.011 lbs/MMBtu is equivalent to 1.6 ppm). It is worth noting that, as shown in figure 18, catalytic combustion allows stable operation with low emissions for the low Btu syngas case (88 Btu/ft3) even at flame temperatures as low as 2300°F (1260°C). CO emissions were less than 5 ppm in all cases, and were near zero for flame temperatures greater than 2200°F (1200°C).
3.2.2.1-10 Sub-Scale Test Data for High-Hydrogen and Low-Btu Fuels In addition to syngas, other fuels containing non-hydrocarbon heating values are also of interest for power generation with low emissions. Industrial process and waste gases are particularly of interest, and PCI has therefore tested its fuel-rich catalytic combustion system with two such fuels: a low-Btu Blast Furnace Gas (BFG) and a high-hydrogen refinery fuel gas. Both gases were simulated in the laboratory by blending gases obtained from bottles or Dewars, using the same equipment that was used for the syngas fuel tests. Again, tests were performed in PCI’s high-pressure sub-scale combustion test rig, pictured in figure 8. Blast Furnace Gas RCL combustion of an 82 Btu/ft3 blast furnace gas was tested using a catalytic reactor that had previously been used for syngas testing. Results for the blast furnace gas show that combustion of this gas is extremely stable following fuel-rich catalytic reaction, even at adiabatic flame temperatures as low as 2250°F (1230°C). For these tests the simulated blast furnace gas comprised 23% CO, 22% CO2, 1.4% H2, 0.6% CH4, and 53% N2, and entered the reactor after being heated to about 450°F (230°C). Air entered the reactor and combustor at about 660°F (350°C). Note that the high diluent fraction (low Btu value) of the blast furnace gas means that high fuel-lean equivalence ratios are needed in the combustor burnout zone to achieve the desired flame temperature. Therefore, tests were performed over a low range of adiabatic flame temperatures in the combustor burnout section, from about 2250°F(1230°C) to 2500°F (1370°C), representing maximum fuel flow capability of the rig for this blast furnace gas composition. For comparison, the stoichiometric flame temperature for this blast furnace gas is only about 2700°F (1480°C) for the inlet temperatures tested. Because the catalyst was able to stabilize complete combustion at such low flame temperatures, NOx and CO emissions were ultra-low for all conditions tested. NOx emissions for blast furnace gas operation are plotted in figure 19, as measured by sample extraction from the cooled probe located 15 inches downstream of the catalyst exit. For all conditions tested, NOx emissions were measured below 2.5 ppm on a raw basis (uncorrected) and below 1 ppm corrected to 15% O2 dry. CO emissions were near zero (< 1 ppm) for all conditions shown. Because measured oxygen concentrations after fuel-lean burnout were very low, varying between about 2.5% and 5.5% for the conditions shown, the standard emissions reporting correction to 15% O2 may be misleading. In fact, because of the high level of
280
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle diluent in the blast furnace gas, oxygen levels would never approach 15% in an actual engine application. Therefore, the raw NOx data are probably as relevant as the corrected NOx data, or perhaps more so, and both are therefore plotted in figure 19. In either case, NOx emissions were ultra-low as a result of the low blast furnace gas flame temperatures. 3
NOx @ 15% O2 dry
2.5
raw (uncorrected) NOx
2
1.5
corrected (15% O2) NOx
1
0.5
0 2200
2250
2300
2350
2400
2450
2500
Adiabatic Flame Temperature (F)
Fig. 19. NOx emissions, uncorrected and corrected to 15% O2, as a function of adiabatic flame temperature in the downstream combustion zone of the RCL combustor burning blast furnace gas (23% CO, 1.4% H2, 0.6% CH4, 22% O2, and 53% N2).
Refinery Fuel Gas Testing of RCL combustion of refinery fuel gas was also conducted using the same hardware configuration as previous tests for syngas and blast furnace gas. Results showed NOx emissions below 3 ppm for flame temperatures below 2800°F (1538°C). For the refinery fuel gas tests, the simulated refinery fuel gas comprised 30% H2 and 70% CH4, and entered the reactor without passing through a fuel heater. However, some fuel heat was obtained from hot combustor rig components so that the fuel plenum gas temperature measured about 347°F (175°C). Combustion air entered the reactor at about 734°F (390°C). Tests were performed over a range of adiabatic flame temperatures, from about 2400 to 3000°F (1315 to 1650°C) in the combustor burnout section, and at a pressure of about 10 atm. NOx and CO emissions were measured at each condition, as well as O2 and CO2. NOx emissions for the RCL combustion of refinery fuel gas are plotted in figure 20, as measured by sample extraction from the cooled probe located 15 inches downstream of the catalyst exit, corresponding to about 50 ms residence time. As shown, NOx emissions were measured below 3 ppm for flame temperatures less than about 2800°F (1540°C). CO emissions were less than about 1 ppm for all conditions shown. 9 8
NOx @ 15% O2 dry
7 6 5 4 3 2 1 0 2400
281
2500
2600
2800 2700 Adiabatic Flame Temp (F) TTeTemperature (F)
2900
3000
Fig. 20. NOx emissions, corrected to 15% O2, as a function of adiabatic flame temperature in the downstream combustion zone burning refinery fuel gas (30% H2 and 70% CH4).
3.2.2.1 Fuel-Rich Catalytic Combustion 3.2.2.1-11 Technology Status and Outlook For natural gas fuel, RCL combustion system has been demonstrated in engine testing of a modified Solar Turbines Saturn engine, and in full-scale full-pressure rig tests at Taurus 70 conditions and beyond, including F-engine type flame temperatures. The technology has been demonstrated viable in-engine, and the next step in technology maturity will be field demonstration for an extended period. Thus, the major challenges that lie ahead for natural gas applications will be durability demonstration and design integration for specific engine applications. RCL combustion system has also been tested at sub-scale with multiple alternative fuels, as listed in table 3. For all fuels tested, catalyst temperatures were well controlled and combustor emissions were held to single-digit values over a wide turndown range, confirming the benefit of catalytic reactions in achieving low emissions for multiple types of fuels. The results also confirm that fuel-rich operation of the catalyst allows similar catalyst and reactor performance with widely varying fuel types. For liquid fuel applications, the major outstanding development issue is prevaporization of the fuel prior to fuel-rich catalytic reaction. The primary concern here is autoignition and overheating of the premixer hardware. Likewise, the propensity for highhydrogen fuels to autoignite is also problematic, as in syngas fuels. In both cases, the air-cooled catalyst structure is able to tolerate autoignition and gas-phase reactions, but conventional premixing devices located upstream of the catalyst may not. Therefore, advanced premixing concepts are to be developed for both liquid fuel and high-hydrogen fuel (e.g. syngas) applications. Table 3 Composition of seven different fuel types tested.
Fuel Tested
Chemical Formula
Methane Landfill Gas Diesel No. 2 Gasoline Refinery Fuel Gas Syngas Blast Furnace Gas
CH4 0.65 CH4 + 0.35 CO2 Multi component not analyzed 0.70 CH4 + 0.30 H2 20% H2 + 20% CO + 10% CO2 + 50% N2 23%CO+22%CO2+1.4%H2+0.6%CH4+53%N2
C/H Ratio by Wt (%) 75 / 25 46.5 / 10.1 87.6 / 13.0 not analyzed 71.2 / 28.8 14.75 / 1.64 17.6 / 0.17
Fuel-Bound N by Wt (%) 0 0 0.0188 ~0 0 0 0
In addition, for IGCC applications with coal-derived syngas fuels, consideration must be given to the high volume flow of fuel that must pass through the catalytic reactor, to prevent excessive size or pressure drop penalties. Depending upon the application, this may require system level re-design and development of the catalyst and combustor. Finally, it is also recognized that syngas fuels carry trace levels of catalyst contaminants that may affect long-term catalyst durability. This needs to be examined and remediated if problematic. Long-term durability tests are required, preferably in an actual syngas slipstream at an operating IGCC plant, where real contaminants will be present.
3.2.2.1-12 Conclusions For natural gas operation, the Rich-Catalytic Lean-burn (RCL) combustion concept has been tested at gas turbine conditions, first in a full-scale full-pressure single-injector rig, and second in a modified industrial gas turbine. These constitute two significant experimental milestones: 1.
Experimental confirmation of the ultra-low-NOx capability of the RCL combustion concept. In particular, we confirm the ability to mix fuel-rich catalyst effluent with primary combustion air, without inducing autoignition, yet imparting significant stability to the downstream combustion process.
2.
Demonstration of RCL combustion feasibility for gas turbine engine operation. In particular, we demonstrate engine start-up, acceleration, and robust operation at load by fuel injected only through RCL-injectors (effectively a single fuel stage, with all fuel contacting the catalyst), and with simple engine controls that do not monitor catalyst temperature.
In summary, the data presented show that fuel-rich catalytic reactions can stabilize fuel-lean premixed combustion to provide stable, quiet combustor operation with ultra-low NOx and CO emissions. In addition, the air-cooled fuel-rich catalyst system can operate safely without active temperature control because its temperature is insensitive to fuel/air ratio. The RCL system also provides significant operational advantages as compared to earlier catalytic combustion systems. Most notably, the RCL reactor requires no preburner, is immune to issues of auto-ignition and flashback (and can therefore operate safely in high-firing-temperature machines such as F-class), and provides long catalyst life (as a result of the non-oxidizing fuel-rich catalyst environment).
282
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle Finally, fuel-rich operation of the catalyst allows operation on multiple types of fuels, as successfully demonstrated at sub-scale for gaseous and liquid hydrocarbon fuels, for low-Btu fuels and high-hydrogen fuels, and for simulated coal-derived syngas. For all of these fuels catalyst temperatures were well controlled, and ultra-low emissions were achieved.
3.2.2.1-13 Notes ________________________ 1. C.L. Vandervort, “9 ppm NOx / CO Combustion System for ‘F’ Class Industrial Gas Turbines,” ASME Paper No. 2000-GT-0086, ASME Turbo Expo May 8-11, 2000, Munich, Germany. 2. D.K. Yee, K. Lundberg, and C.K. Weakley, “Field Demonstration of a 1.5 MW Gas Turbine with a Low Emissions Catalytic Combustion System,” Journal of Engineering for Gas Turbines and Power 123 (2001): 550-556; L.L. Smith, H. Karim, M.J. Castaldi, S. Etemad, W.C. Pfefferle, V.K. Khanna, and K.O. Smith, “Rich-Catalytic Lean-Burn Combustion for Low-SingleDigit NOx Gas Turbines,” Journal of Engineering for Gas Turbines and Power 127 (2005): 27-35. 3. J.H. Lee and D.L. Trim, “Catalytic Combustion of Methane,” Fuel Processing Technology 42 (1995): 339-359; R.A. Dalla Betta, “Catalytic Combustion Gas Turbine Systems: the Preferred Technology for Low Emissions Electric Power Production and Co-generation,” Catalysis Today 35 (1997): 129-135; P. Forzatti and G. Groppi, “Catalytic Combustion for the Production of Energy,” Catalysis Today 54 (1999): 165-180. 4. R.J. Farrauto, M.C. Hobson, T. Kennelly, and E.M. Waterman, “Catalytic Chemistry of Supported Palladium for Combustion of Methane,” Applied Catalysis A: General 81 (1992): 227-237; R.J. Farrauto, J.K. Lampert, M.C. Hobson, and E.M. Waterman, “Thermal Decomposition and Reformation of PdO Catalysts; Support Effects,” Applied Catalysis B: Environmental 6 (1995): 263-270; J.G. McCarty “Kinetics of PdO Combustion Catalysis,” Catalysis Today 26 (1995): 283-293; N.M. Rodriguez, S.G. Oh, R.A. Dalla-Betta, and R.T.K. Baker, “In Situ Electron Microscopy Studies of Palladium Supported on Al2O3, SiO2, and ZrO2 in Oxygen,” J. Catalysis 157 (1995): pp. 676-686. 5. R.A. Dalla Betta, T. Shoji, K., Tsurumi, and N. Ezawa (1994). “Partial Combustion Process and a Catalyst Structure for Use in the Process,” U.S. Patent No. 5,326,253; T. Furuya, K. Sasaki, Y. Hanakata, T. Ohhashi, M. Yamada, T. Tsuchiya, and Y. Furuse (1995). “Development of a Hybrid Catalytic Combustor for a 1300°C Class Gas Turbine," Catalysis Today 26 (1995): 345-350; Y. Ozawa, Y. Tochihara, N. Mori, I. Yuri, T. Kanazawa, and K. Sagimori, “High Pressure Test Results of a Catalytically Assisted Ceramic Combustor for a Gas Turbine,” ASME Paper No. 98-GT-381, Stockholm, Sweden, 2-5 June 1998; R. Carroni, V. Schmidt, and T. Griffin, “Catalytic Combustion for Power Generation,” Catalysis Today 75 (2002): 287295. 6. S.T. Kolaczkowski, “Catalytic Stationary Gas Turbine Combustors: A Review of the Challenges Faced to Clear the Next Set of Hurdles,” Trans. I. Chem. E. 73 Part A (1995): 168-190; D.B. Fant, G.S. Jackson, H. Karim, D.M. Newburry, P. Dutta, K.O. Smith, and R.W. Dibble, “Status of Catalytic Combustion R&D for the Department of Energy Advanced Turbine Systems Program,” Journal of Engineering for Gas Turbines and Power 122 (2000): 293-300. 7. M. Lyubovsky, L.L. Smith, M. Castaldi, H. Karim, B. Nentwick, S. Etemad, R. LaPierre, and W.C. Pfefferle, “Catalytic Combustion over Platinum Group Catalysts: Fuel-Lean versus Fuel-Rich Operation,” Catalysis Today 83 (2003): 71-84. 8. R.E. Hayes and S.T. Kolaczkowski, Introduction to Catalytic Combustion (Amsterdam: Gordon and Breach Science Publishers, 1997); E.M. Johansson, D. Papadias, P.O. Thevenin, A.G. Ersson, R. Gabrielsson, P.G. Menon, P.H. Bjornbom and S.G. Jaras, “Catalytic Combustion for Gas Turbine Applications,” Catalysis 14 (1999): 183-235; also see note 6 (Kolaczkowski). 9. R.J. Rollbuhler, “Fuel-Rich, Catalytic Reaction Experimental Results,” 27th Joint Propulsion Conference, Sacramento, CA, 24-27 June 1991, NASA Technical Memorandum 104423, AIAA Paper No. 91-2463; T.A. Brabbs and S.A. Merritt, “Fuel-Rich Catalytic Combustion of a High Density Fuel,” NASA Technical Paper 3281 (1993). 10. G.O. Kraemer, (1996). “Fuel-Rich Catalytically Stabilized Combustion for Aircraft Engine Applications,” Ph.D. thesis, Yale University. 11. M.B. Colket, A.S. Kesten, J.J. Sangiovanni, M.F. Zabielski, D.R. Pandy, and D.J. Seery (1993). “Method and System for Combusting Hydrocarbon Fuels with Low Pollutant Emissions by Controllably Extracting Heat from the Catalytic Oxidation Stage,” U.S. Patent No. 5,235,804. 12. J.P. Kesselring, W.V. Krill, E.K. Chu, and R.M. Kendall. In proceedings of New fuels and advances in combustion technologies symposium, Mar. 26-30, 1979, New Orleans, LA.
283
13. W.C. Pfefferle, L.L. Smith, and M.J. Castaldi (2002). “Method and Apparatus for a Fuel-Rich Catalytic Reactor,” U.S. Patent No. 6,358,040. 14. See note 2 (Smith). 15. L.J. Spadaccini, and M.B. Colket, “Ignition Delay Characteristics of Methane Fuels,” Prog. Energy Combust. Sci. 20 (1994): 431-460.
3.2.2.1 Fuel-Rich Catalytic Combustion 16. See note 2 (Smith). 17. L.L. Smith, S. Etemad, M.J. Castaldi, H. Karim, and W.C. Pfefferle (2002). “Method and Apparatus for a Fuel-Rich Catalytic Reactor,” U.S. Patent No. 6,394,791. 18. United States Department of Energy (2000). Clean Coal Technology Topical Report Number 19, “Tampa Electric Integrated Gasification Combined-Cycle Project, An Update” July 2000; GE Power Systems (2002). “Gas Turbine and Combined Cycle Products,” available at www.gepower.com/corporate/en_us/assets/gasturbines_heavy/prod/pdf/gasturbine_2002.pdf; R.D. Brdar and R.M. Jones (2000). “GE IGCC Technology and Experience with Advanced Gas Turbines,” GE Power Systems Report No. GER-4207, October 2000.
284
BIOGRAPHY
3.2.2 Catalytic Combustion & 3.2.2.1 Fuel-Rich Catalytic Combustion
Dr. Lance L. Smith Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. Lance L. Smith is a senior research & development engineer in the Gas Turbine Products group at Precision Combustion, Inc. (PCI), and a visiting assistant professor in the Engineering Department at Trinity College. Dr. Smith has 14 years experience in combustion research and combustor development, including work in turbulent non-premixed combustion, premixed combustion and premixing, aerodynamic design of combustor components, pulsed combustion, and catalytic combustion. His academic research has been primarily experimental, with a focus on laser-based measurements in flames, including work conducted as a visiting researcher at Sandia National Laboratories and as a post-doctoral researcher at UCLA. Dr. Smith is a principal engineer of, and holds multiple patents for, the RCLTM catalytic reactor. At PCI, he works with OEM gas turbine manufacturers to develop integrated catalytic combustion systems for ultra-low emissions gas turbines. A graduate of Brown University (B.S., 1986) and of University of California, Berkeley (M.S., 1990 and Ph.D., 1994), Dr. Smith is an elected member of the Tau Beta Pi and Sigma Xi honor societies, and a member of the Combustion Institute.
Dr. Shahrokh Etemad Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473 phone: (203) 287-3700 x217 email: [email protected]
Dr. Shahrokh Etemad as Manager of Gas Turbine Products at Precision Combustion, Inc. (PCI) in North Haven, Connecticut, has full responsibility for technical and commercial development of two major products. He manages an advanced technology group to develop low-emissions combustion products in close collaboration with several OEM gas turbine engine manufacturers and the U.S. Department of Energy. Dr. Etemad is responsible for budgeting, funding opportunities, technology direction, proposal preparation and complete R&D operations including concept development, computational analysis, experimental testing, full-size performance demonstration and productionization. Prior to his present position at PCI, he worked for several years at Textron Lycoming and United Technologies, Carrier. Dr. Etemad has published 28 technical articles and holds 28 patents in the field of turbomachinery, combustion, and thermofluid systems. He earned bachelor’s and master’s degrees at Sussex University and University of London respectively, and received his Ph.D. from the University of Washington in 1984. He has been a member of ASME since 1995 and won the 2003 ASME Gas Turbine award.
Dr. Hasan Karim Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. Hasan Karim is a senior research & development engineer at Precision Combustion, Inc. (PCI), where his responsibilities include design, development, analysis, numerical and computational fluid dynamics, and testing of catalytic combustors for natural gas, syngas, and liquid fuel. He is the principal investigator for the catalytic combustor development project for the U.S. Navy and lead engineer for the catalytic pilot and catalytic combustor for downhole combustion programs. After receiving a bachelor’s degree from Indian Institute of Technology-Kharagpur in 1987, Dr. Karim earned his M.S. from New Jersey Institute of Technology in 1991, and his Ph.D. from Yale University in 1998. He is a co-inventor of air-cooled rich and lean reactor technology.
Dr. William C. Pfefferle Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. William C. Pfefferle invented the original catalytic combustor for gas turbine engines in the early 1970s and now holds over 90 U.S. patents. In 1986, Dr. Pfefferle co-founded Precision Combustion, Inc. (PCI), a Connecticut-based company dedicated to developing clean and efficient technology for clean air. His research has led to important industrial advances such as the RCL catalytic combustor for ground power gas turbine engines, which is now in late-stage development and evaluation by several major gas turbine manufacturers, and the Microlith catalytic reaction system, which forms the basis for paradigm-shift, high heat mass transfer catalytic reactors for fast-lightoff automotive catalytic converters and fuel processor reactors. With a B.S. in Chemical Engineering from Drexel University (1944) and a Ph.D. in Physical Chemistry from the University of Pennsylvania (1952), Dr. Pfefferle is a member of the American Chemical Society (ACS), and received the ACS 31st Northeast Regional Industrial Innovation Award. He was inducted into the New Jersey Inventor’s Hall of Fame in February 1990. He continues to work full-time to develop catalytic devices for clean and efficient energy.
3.2.2.2
Catalytic Combustion in Large Frame Industrial Gas Turbines
3.2.2.2-1 Introduction Large frame gas turbine engines employ three different types of combustion systems: diffusion flame, lean premixed combustion and catalytic combustion. In diffusion flame combustors the fuel and air are injected separately into the combustion zone where they mix and react. Because of the nature of the design, these combustion systems tend to have flame temperatures that are typical of stoichiometric combustion and therefore produce high NOx emissions. Obtaining reasonable emissions from a diffusion flame combustion system generally requires the injection of diluents into the combustion section to lower the flame temperature, typically either water or steam. At current F-class firing temperatures these systems can produce NOx emissions in the range of 25 ppm NOx. In the lean premixed combustion system, the fuel and air are allowed to premix upstream of the flame zone. This results in a significantly lower flame temperature than the standard diffusion flame combustor resulting in lower NOx emissions without the need to inject water or steam. The limitation on low emissions from the lean premixed combustion systems is the combustion instabilities which occur as the lean flammability limit of the mixture is approached. These instabilities can lead to large pressure fluctuation in the combustion chamber. At F class temperatures the lean premixed combustion system can obtain NOx emissions in the range of 7-9 ppm. The catalytic combustion system shows promise to achieve lower emissions because the combustion instabilities at the lean flammability limit are no longer a limiting factor. Although catalytic combustion systems have not yet been employed in large industrial gas turbines, results from current development are encouraging and emissions in the range of 2-3 ppm are achievable.
3.2.2.2-2 Catalytic Combustion Design
Walter Ray Laster Siemens Power Generation Corporation 4400 Alafaya Trail MC Q3 -042 Orlando, FL 32826 phone: (407) 736-5796 email: [email protected]
285
The major development effort for catalytic combustion in large frame gas turbine engines was initiated as part of the DOE ATS program1. The goal of the ATS program was the development of a high–efficiency, high-firing temperature engine (>1700 K) with NOx emissions less than 10 ppm for lean premixed systems and 5 ppm for the catalytic system. On this program the basic design of the catalytic combustor for a large industrial gas turbine was developed. Since this program, considerable progress has been made on the design. At the high firing temperatures of a typical gas turbine engine, it is not possible to design a pure catalytic approach where all of the fuel is reacted in the catalyst section. In the current design philosophy a hybrid catalytic two stage system is employed where the catalyst stage is followed with a homogeneous burnout region. Generally these systems will react 20-40% of the fuel in the catalytic stage. By reacting a portion of the fuel in the catalyst the stability of the flame in the homogeneous burnout zone is significantly improved. The hybrid catalytic combustion systems that have been investigated for large gas turbine engines are the lean catalytic lean burn (LCL) design and the rich catalytic lean burn (RCL) design. Figure 1 shows the basic concept of the LCL design. In this design all of the fuel and air are premixed upstream and enter the catalyst section under fuel lean conditions. At the end of the catalyst section any fuel not reacted is burned out in a homogeneous reaction zone. To insure proper catalyst activity, this concept requires an inlet temperature of fuel air mixture to the catalyst of approximately 500 C. Since this temperature is higher than the compressor exit temperature of a typical gas turbine engine, a preburner will be necessary to achieve the desired catalyst inlet conditions. Operation of the catalyst in the lean region requires very close control of the air fuel ratio in the vicinity of the catalyst to avoid high reaction rates and excessive catalyst temperatures. The lean combustion concept has been pursued by Catalytica in their patented Xonon technology. This technology has been commercially operated on a small scale in the Kawasaki 1.5 MW engine. On large frame engines this technology has been studied by General Electric and Siemens Westinghouse.
Figure 2 shows a basic concept of the RCL design. In this design the inlet air flow to the catalyst is separated into two streams. A portion of the air is mixed with the fuel and reacts on the surface of the catalyst under fuel rich conditions. The remaining air is used to backside cool the catalyst. The two streams mix at the catalyst exit and then react and burnout in the homogeneous reaction zone. By operating the catalyst in the fuel rich region, the reaction rate is limited by the rate of diffusion of oxygen to the catalyst surface. Therefore this design is able to tolerate wider variations in air fuel ratio within the catalyst region than the LCL design. In this design the preburner is no longer required as the fuel and air react at compressor exit temperature typical of gas turbine engines. The choice of catalyst material is critical for this design in order to insure proper catalyst lightoff. Precision Combustion, Inc and Siemens Westinghouse have pursued the RCL combustion design2.
Flam e
Catalys t B .S . Cooled Catalytic R eactor
Inlet A ir + Fuel @ 950F/500C
B urnout of CH4 and CO
M ixing M ix R eacted and N on-R eacted E ffluent
Fig. 1. LCL Combustion System Rich-Catalytic/Lean-Burn Module (ATCC3 - PCI)
Fuel Inlet, with A ir Fuel + A ir @ 750F/400C
Flam e B urnout of CH4 and CO
A ir Inlet A ir @ 750F/400C N o preburner
Catalys t A ir-Cooled Catalytic R eactor
M ixing M ix A ir w/ Partially R eacted Fuel/A ir
Fig. 2. RCL Combustion System
286
Walter Ray Laster 3.2.2.2-3 Rich Catalytic Combustion Applied to Large Gas Turbine Engines As part of the ATS program, Siemens Westinghouse performed an evaluation of both the RCL and LCL systems. Based on subscale module testing of both technologies, it was clear that either design could meet the emission targets of the catalytic program. The RCL design was chosen for two main reasons. By eliminating the pre-burner, the design was much more compact and could be easily fit into the existing envelope of the current gas turbine combustor without major modification to the casing. Also because of the operation in the rich region the design was much more tolerant of variations in both air and fuel flow. Over firing tests were performed on both designs and the RCL design was able to survive a severe over fuel transient without damage. This was not the case for the LCL design. Under the ATS program, the basic design of the RCL catalytic module was developed and the conceptual design of catalytic basket was developed. This design was continued for application to the lower firing temperature SGT6-3000E engine. Full scale basket testing was performed on this design at both E-class and F-class firing temperatures. During the initial development phase of the RCL design, testing was performed on the subscale module level. The full scale combustor basket was divided into 6 individual subscale modules each of which was designed to operate at 1/6 of the total combustor basket flow. The basic design of the RCL catalytic module is shown in figure 3. In this design the catalyst is composed of tubes with catalytic coating on the outside surface. These tubes are brazed to a plate on the upstream end and flared on the down stream end. A portion of the inlet air (~15%) enters the fuel mixing chamber where fuel is injected and allowed to mix before entering the catalyst region. This rich fuel air mixture flows along the outside of the tubes and is allowed to react on the catalyst surface. The remaining air enters the inside of the tubes and provides backside cooling for the catalyst surface. Both streams are allowed to mix down stream of the catalyst zone before they enter the homogeneous reaction zone. The ratio of air flow between the reacting fuel region and the cooling air is determined by pressure drops between the two flow paths. For a given design the fuel split is fixed. Figure 4 shows the catalytic module design used for testing. Catalytic module testing was performed at full engine conditions for both the STG-63000E and the STG-6-5000F engines. Module testing confirmed that emissions could be maintained at less than 2 ppm NOx and 10 ppm
287
Fuel Inlet Cooli ng Air Inlet ~85%
Reactive Mixture
Fuel-MixerFuel Manifold Gas Jets
Air Inlet (~15%)
45o
Mixing Zone
306 - Tubes with Catalytic Coating on the Outside Fig. 3. Rich Catalytic Module Design
Fig. 4. Rich Catalytic Module Design
Fig. 5. Catalytic Combustor for the Siemens SGT6-5000F Engine
3.2.2.2 Catalytic Combustion in Large Frame Industrial Gas Turbines CO for a wide range of conditions including both the E and F class engines. Throughout all conditions the catalyst and metal temperatures remained within limits. A full scale combustor basket was designed using 6 catalytic modules surrounded by a central pilot3. Figure 5 shows the combustor configuration envisioned for use in the Siemens gas turbine engine. The pilot is necessary to insure stable operation of the combustor basket at low loads. The goal of the design is to minimize or completely shut off the fuel flow to the pilot at baseload conditions. Although the pilot contributes to NOx emissions, this can be minimized by replacing the standard diffusion pilot design with a premixed pilot. Based on this concept, a full scale basket was fabricated and tested at the Siemens full pressure single basket test facility in Italy. This facility duplicates the geometry and flow conditions (pressure, temperature, air and fuel flow) of a single basket as installed in a Siemens gas turbine engine. Data was obtained for a range of firing temperatures encompassing the SGT-6-3000E and the SGT-6-5000F engines. The basket used for these tests is shown in figure 6. As expected, the emissions for the full basket tests were slightly higher than those of the module tests due to the pilot. For the SGT-6-3000E engine it was necessary to add dilution air to raise the combustor temperature in order to achieve proper CO burnout. At these conditions the catalytic combustion system was able to produce emissions of 3.3 ppm NOx and 7 ppm CO. When the temperature was increased to SGT-6-5000F conditions, the improved stability at higher temperatures enabled the combustor to run at a significantly lower pilot fraction. The resulting NOx and CO emissions were nearly the same as for the lower firing temperature conditions, 3.6 ppm NOx and 9 ppm CO. Basket and catalyst metal temperatures and combustor pressure oscillations were well below the design limits during the test. As part of this test program, an overfiring test was performed, with no damage to the catalytic module components.
Fig. 6. Full Scale Catalytic Combustor
3.2.2.2-4 Conclusions The rich catalytic combustion approach has been designed for application to the Siemens SGT-6-3000E and STG-6-5000F engine. Rig testing has shown that the design is capable of emissions in the range of 2 –3 ppm at SGT-6-5000F temperatures. This design has been shown to be robust with respect to variation in air and fuel flow. Additional work is underway to reduce the emissions. Current development on the RCL concept is focused on the fuel flexibility aspects of the design and the application to syngas and hydrogen fuels.
3.2.2.2-5 Notes _________________________ 1. D.B Fant,G.S. Jackson, H. Karim, D.M. Newburry, P. Dutta, K.O. Smith, and R.W. Dibble, “Status of Catalytic Combustion R&D for the Department of Energy Advanced Turbine Systems Program,” ASME Journal of Enginering for Gas Turbines and Power, (2000): 293-300. 2. L. Smith, et al, US Patent No. 6,174,159, “Method and Apparatus for a Catalytic Firebox Reactor,” 1999; W.C. Pfefferle, L. Smith, M.J. Castaldi, US Patent No. 6,358,040, “Method and Apparatus for a Fuel Rich Catalytic Reactor,” 2000. 3. D.M. Newburry, US Patent No. 6,415,608, “Piloted Rich-Catalytic Lean-Burn Hybrid Combustor,” 2000.
288
BIOGRAPHY
3.2.2.2 Catalytic Combustion in Large Frame Industrial Gas Turbines
Walter Ray Laster Siemens Power Generation 4400 Alafaya Trail MC Q3 -042 Orlando, FL 32826 phone: 407-736-5796 email: [email protected]
Dr Laster has a Ph.D. in Mechanical Engineering from Purdue University with a specialty in the area of combustion. He taught combustion and heat transfer at Texas A&M University. He has 12 years industrial experience in the gas turbine field at Siemens Westinghouse. He has been involved in the design of dry low NOx combustors and catalytic combustion systems. He has been involved in several projects related to alternatuve fuels for gas turbine systems.
3.2.2.3
Surface Stabilized Combustion
3.2.2.3-1 Introduction Surface-stabilized combustion is a simple approach that can maximize the emissions benefit of lean fuel/air premixing by increasing flame stability, and doing so in a compact and flexible manner. ALZETA Corporation is developing a surfacestabilized combustion system for industrial turbine applications capable of sub-3 ppm emissions of oxides of nitrogen (NOX) with simultaneous low emissions of carbon monoxide (CO) and unburned hydrocarbons (HC). The application of surface-stabilized combustion to gas turbines is being developed under the name nanoSTAR™. The development has been reported in a series of technical papers given at various ASME conferences1. Low emissions of oxides of nitrogen (NOX), as well as carbon monoxide (CO) and unburned hydrocarbons can be achieved with thorough fuel/air mixing and control of the adiabatic flame temperature of that mixture below about 1920 K (3000 °F). One of the great difficulties with such lean premixed systems has been maintaining flame stability in the narrow flame temperature range between high NOX production and lean flame extinction. Aerodynamically stabilized injectors have very narrow ranges of operation, necessitating multiple injector staging (up to four stages in some systems) or piloting2. When control of NOx emissions is achieved without the use of steam or water injection, it is referred to as a dry method, such Dry Low NOx, or DLN systems, have been successfully deployed to achieve sub-25 ppm NOx emissions in several gas turbine applications, and in some cases much lower. Surface-stabilized combustion is a simple approach that extends the operating range of lean premixed systems to achieve sub-3 ppm NOx emissions. The technology has advanced through proof-of-concept testing in pressurized rigs and demonstration in a one megawatt test engine. Prototype injectors for small industrial turbines have been designed, built, and rig tested. Multiple injectors have been tested in an annular combustor with varied combustion air inlet temperatures under atmospheric and elevated pressures while work is progressing toward an engine demonstration.
3.2.2.3-2 Technology The surface-stabilized combustion inherent in nanoSTAR injectors is best described as laminar blue-flame combustion stabilized by significant velocity gradients above a porous metal-fiber mat. The operation of this type of surface-stabilized combustion is characterized by the schematic to the left of figure 1, which shows premixed fuel and air passing through the metal fiber mat in two distinct zones.
B
Neil K. McDougald ALZETA Corporation 2343 Calle del Mundo Santa Clara, CA 95054 phone: (408) 727-8282 email: [email protected]
A
A Fuel/Air
Fig.1. Surface-Stabilized Combustion (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME])
289
Source: S. J. Greenberg, N. K. McDougald, and L. O. Arellano, “Full-Scale Demonstration of Surface-Stabilized Fuel Injectors for sub-Three ppm NOx Emission,” ASME Paper # GT2004-53629 (presented at the 2004 ASME Turbo Expo, Vienna, Austria, June 14-17, 2004).
In the porous-only zone true surface combustion (A) is realized. Under lean conditions this will manifest as very short laminar flamelets, but under rich conditions the surface combustion will become a diffusion-dominated reaction stabilized just over a millimeter above the metal matrix, which proceeds without visible flame and heats the outer surface of the mat to incandescence. This type of radiant surface combustion can be seen between the laminar flamelets to the right of figure 1. Portions of the metal fiber mat are perforated to allow higher mass flux (B). In these zones stretched laminar flames are established that are anchored by the adjacent surface combustion. This produces the distinctive flame pattern seen in the right-hand picture of figure 1. The specific perforation arrangement and pattern control the size and shape of the laminar flamelets. The perforated zones operate at flow velocities of up to 10 times the laminar flame speed producing a factor of ten stretch of the flame surface and resulting in a large laminar flamelets. The alternating arrangement of laminar blue flames and surface combustion, allows high firing rates to be achieved before flame liftoff occurs, with the surface combustion stabilizing the long laminar flames by providing a pool of hot combustion radicals at the flame edges. At atmospheric operation, nominal injector output would be 3.15 MW/m2 (1.0 million Btu/hr/ft2), so an injector with a fired area of .047 m2 (0.5 ft2) would have a capacity of 146.5 kW (500,000 Btu/hr). Assuming the firing rate of the injector increases linearly with pressure, the SFR remains constant as pressure increases. This results in a compact injector size for a given capacity in high pressure systems. Therefore the 146.5 kW (500,000 Btu/hr) injector at 0.1 MPa (1 atm) becomes nominally a 1,465 kW (5 million Btu/hr injector) at 1 Mpa (10 atm). Put another way, based on a gas turbine with a heat rate of 10,000 Btu/kilowatt-hour and a combustion pressure of 10 atmospheres, only about one square foot of injector surface area would be required for every megawatt of gas turbine output. NanoSTAR injectors are constructed of small metal fibers which are compressed and sintered, resulting in an all-metal structure. This porous pad is perforated to produce a proprietary arrangement of perforation zones. The perforated metal fiber pads have a very low pressure drop but excellent flow uniformity. They also display excellent durability in fired service. In an atmospheric cycling test, a nanoSTAR metal fiber pad withstood over 15,000 ignition/cooling cycles over a 30-day period without a significant loss in operability. Further material and oxidation studies are being conducted in order to estimate injector life which is expected to exceed 8000 hours. Figure 2 depicts an injector in a gas turbine combustor liner. COMBUSTOR LINER SINTERED METAL FIBER PAD
LINER DILUTION HOLES
PREMIXED FUEL/AIR
DISTIBUTOR
MOUNTING RING
SELECTIVE PERFORATIONS
Fig. 2. Surface-Stabilized nanoSTAR Injector (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]) Source: See fig. 1.
The laminar blue flame combustion zones created by the surface stabilization contribute to lower NOX emissions in three ways. The dominant mechanism is the expected benefit from using a fully premixed fuel and oxidizer, resulting in a uniform temperature across the reaction zone, and lean burning, resulting in reaction temperatures below the 1920 K (3000 °F) limit for thermal NOX formation. The second is the much lower residence time in the hot combustion zone. The peak temperatures are realized in the combustion front formed by each laminar flamelet which, like that of a Bunsen injector flame, is very thin. So the residence time in the peak flame temperature zone for a nanoSTAR injector is a fraction of that of a typical aerodynamically-stabilized injector. The third mechanism is a more rapid post-flame cooling of each blue-flame zone via the gas phase radiation mechanism. By spreading the flame over a larger surface, the gas layer thickness at any specific location on the injector is thin (relative to that of a conventional injector) and can more rapidly transfer energy as a result. These mechanisms combine in a nanoSTAR injector to produce lower NOX emissions than a typical lean premixed aerodynamically-stabilized injector. Figure 3 shows a comparison between nanoSTAR injector emission results from a high-pressure rig test and perfectly-premixed aerodynamically-stabilized emission results from a 1990 paper by Leonard and Correa3. In both cases the tests were conducted at 1.01 Mpa (10 atm) and 535-590 K (500-600 oF) inlet temperatures. A nanoSTAR injector firing in under atmospheric pressure in a quartz enclosure is shown in figure 4. In addition to lower emissions with a wide turndown window, nanoSTAR injectors can be designed to fit within existing combustor liners and fitted to existing fuel/air premixers without extensive modification to the combustion equipment or pressure case. Furthermore, they require no extraordinary control schemes or equipment beyond that which would be required for an aerodynamicallystabilized lean-premixed injector.
290
Neil K. McDougald
Fig. 3. Surface-Stabilized Compared to AerodynamicallyStabilized Emission Results at 1.01 MPa (10 atm) Pressure and 535-590 K (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]).
Fig. 4. Surface-Stabilized nanoSTAR Injector (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]). Source: See fig. 1.
Source: See fig. 1.
3.2.2.3-3 Experimental Results Calculated Adiabatic Flame Temperature (oF) 2900 2950 3000 3050 3150 3200 3250 The nanoSTAR injectors have been3100 extensively tested in sub-scale rigs over a broad range of inlet temperatures and operating pressures. Representative results are shown in figure 5 comparing single and dual injector results at 1.2 MPa (12 atm). These results Single Injector Dual Injectors - 3" Separation confirmed8 that injectors could be fired in close proximity without impacting emissions or operability clearing the way for full-scale annular combustor testing. 10
2750
2800
2850
)
2
6
4 10
2 x
)Emissions NO (ppm, corrected 15% O NO (ppm, corrected 15% O Emissions
x
291
Calculated Adiabatic Flame Temperature (oF) 2750
2800
2850
2900
2950
3000
3050
3100
3150
3200
3250
Single Injector Dual Injectors - 3" Separation
2 8 0 6
1800
1850
1900
1950
2000
2050
Calculated Adiabatic Flame Temperature (K) 4
2
0 1800
1850
1900
1950
2000
2050
Calculated Adiabatic Flame Temperature (K)
Fig. 5. Comparison of Sector Rig NOX Data with Single Injector Rig Data at 1.2 MPa (12 atm) Pressure and 640 K (690°F) Inlet Temperature (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]).
Fig. 6. Interior of Full Scale Annular Combustor during Atmospheric Testing (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]).
Source: See fig. 1. above.
Source: See fig. 1. above.
A full-scale annular combustor was fitted with twelve equally spaced nanoSTAR injectors. The assembly was tested under atmospheric conditions which allowed for visual observation of the fired injectors as in figure 6. Four thermocouple rakes recorded temperatures around the combustor outlet to create the outlet profile shown in figure 7. The outlet profile was uniform with an overall pattern factor of 0.16 that is well within acceptable limits.
10
3.2.2.3 Surface Stabilized Combustion
8 6
2
Emissions (ppm, corrected 15% O )
4 2600 16 2
Calculated Adiabatic Flame Temperature (oF) 2650
2700
2750
2800
2850
2900
14 0
2950
3000
NOx
12
CO
10 8 6 4
Source: See fig.1. above.
Emissions (ppm, corrected 15% O
Fig. 7. Plot of Temperature Contours at Exit Plane of Annular Combustor at Atmospheric Pressure and 650 K (700°F) Inlet Temperature (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]).
2 0
1700
1725
1750
1775
1800
1825
1850
1875
1900
Calculated Adiabatic Flame Temperature (K)
Fig. 8. Emissions Data Collected During Full Scale Pressurized Testing at 0.5-1.2 MPa (5-12 atm) Pressure and 475-700 K (400-800°F) Inlet Temperature (reproduced by permission of the publisher from American Society of Mechanical Engineers [ASME]). Source: See fig. 1. above.
The full-scale combustor was installed in a high pressure test cell and operated at pressures between 0.5-1.2 MPa (5-12 atm) with inlet air temperatures between 475-700°K (400-800°F). Emissions data collected during these tests are presented in figure 8. The results were consistent with previous sub-scale results and provided data necessary to design an engine-ready combustor.
3.2.2.3-4 Conclusions Surface stabilized combustion extends the lean premixed combustion stability allowing ultra-low emissions to be realized under small industrial gas turbine operating conditions. ALZETA’s nanoSTAR technology has progressed through a series of sub-scale and full-scale rig tests consistently demonstrating ultra-low NOx emissions of less than 3 ppm (corrected to 15% O2) over a broad range of operating conditions. The next stage in the development is an engine demonstration which should be completed by the end of 2005.
3.2.2.3-5 Notes _________________________ 1. C.K. Weakley, S. J. Greenberg, R. M. Kendall, N. K. McDougald, and L. O. Arellano, “Development Of SurfaceStabilized Fuel Injectors With Sub-Three ppm NOX Emissions,” ASME Paper # IJPGC2002-26088 (presented at the 2002 International Joint Power Generation Conference, Phoenix, AZ, June 24-26, 2002; Greenberg, S. J.); N. K. McDougald, C. K. Weakley, R. M. Kendall, and L. O. Arellano, “Surface-Stabilized Fuel Injectors With Sub-Three ppm NOX Emissions for a 5.5 MW Gas Turbine Engine,” ASME Paper # GT2003-38489 (presented at the 2003 ASME Turbo Expo, Atlanta, GA, June 16-19, 2003); S. J. Greenberg, N. K. McDougald, and L. O. Arellano, “Full-Scale Demonstration of Surface-Stabilized Fuel Injectors for sub-Three ppm NOx Emission,” ASME Paper # GT2004-53629 (presented at the 2004 ASME Turbo Expo, Vienna, Austria, June 14-17, 2004.). 2. C. L. Vandervort, “9 ppm NOX/CO Combustion System for “F” Class Industrial Gas Turbines,” ASME J. of Engineering for Gas Turbines and Power 123 (2001): 317-321. 3. G. L. Leonard and S. M. Correa, “NOX Formation in Premixed High-Pressure Lean Methane Flames,” Fossil Fuel Combustion Symposium 1990, S. N. Singh, ed. (New York: American Society of Mechanical Engineers, 1990): 69-74.
292
BIOGRAPHY
3.2.2.3 Surface Stabilized Combustion
Neil K. McDougald ALZETA Corporation 2343 Calle del Mundo Santa Clara, CA 95054 phone: (408) 727-8282 email: [email protected]
Since July 2000, Dr. McDougald has been leading ALZETA’s effort to develop nanoSTAR, an ultralow emissions combustion system for gas turbines. The nanoSTAR is a surface-stabilized lean premixed combustion technology that provides ultra-low NOx emissions while avoiding unstable combustion dynamics. As Director of Product Development, he is responsible for management and technical support for ALZETA’s existing and emerging combustion products.
3.2.2
Catalytic Combustion
Dr. Lance Smith
Dr. Shahrokh Etemad
3.2.2-1 Introduction The earliest work on what is now termed catalytic combustion was conducted by Pfefferle at Engelhard Corporation in the 1970s and introduced the use of both catalytic and non-catalytic combustion reactions in a temperature range amenable to both1. The original-type catalytic combustor is a ceramic honeycomb monolith containing catalytically-coated parallel channels and placed within a combustion chamber2. In this original-type catalytic combustor, surface reactions release heat and reactive intermediates into the boundary layer above the surface, eventually inducing gas-phase (non-catalytic) reactions. As a consequence, combustor operation can be at lean limits well beyond those feasible without the influence of a catalyst, and pollutant emissions can be extremely low. Early work on systems of this type were conducted at Engelhard, Acurex, Westinghouse, NASA, the Air Force, and elsewhere3. Active interest in catalytic combustion for power generation increased during the early 1990s as it became clear that continued pressure for reduced emissions could not be met simply by re-design of conventional combustors. A new approach of partial conversion in the catalyst bed and the use of metal catalyst substrates to circumvent thermal shock issues, revived catalytic combustion for power generation. Metal-substrate type catalyst beds were thus employed for catalytic combustion with increasing success during the 1990s, demonstrating the low NOx potential of catalytic combustion for gas turbine applications4. Ultimately, two very different systems emerged during this period: a fuel-lean catalyst system developed by Catalytica, Inc. and a fuel-rich catalyst system developed by Precision Combustion, Inc5. Engine tests of these two systems are described, respectively, in Yee et al. and Smith et al.6 . These systems are also described in greater detail in Sections 3.2.2.1.1 and 3.2.2.1.2 of this Handbook.
3.2.2-2 Role of Catalysis in Combustion
Dr. Hasan Karim
Dr. William C. Pfefferle Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, CT 06473 phone: (203) 287-3700 x217 email: [email protected]
255
In broad terms, a catalyst is used to promote a desired chemical reaction. Catalysts find a wide range of applications in the production of energy and power, but for combustion turbines there are three basic classes of reactions that one may desire to promote: fuel preparation such as reforming prior to combustion, fuel oxidation with heat release, and pollutant destruction. “Catalytic combustion” normally refers to fuel oxidation with heat release, particularly when the catalyst is placed inside an engine and within the combustor casing. We restrict our discussion here to catalytic combustion and exclude other catalytic processes such as fuel reforming or exhaustgas cleanup. In simple terms, the presence of a combustion catalyst enables complete combustion at lower temperatures than otherwise possible. This fact can be used for multiple benefits, but the primary motivation for low temperature combustion is reduced NOx emissions and/or increased combustor turndown. In particular, most non-catalytic combustors operate with peak flame temperatures higher than 1525°C (2780°F) to ensure adequate flame stability and margin from blowout. As is well known, NOx emissions even for perfectly premixed fuel-air flames at 1525°C (2780° F) can exceed the 3 ppm threshold (at 15% O2) targeted for many new power plants7. Catalytic combustors, however, can operate stably with flame temperatures far below 1525°C (2780°F), offering both reduced NOx emissions and improved combustor turndown.
3.2.2-3 Catalyst Materials for Combustion Applications By definition a catalyst promotes a chemical reaction, such as fuel with oxygen, but is itself neither consumed nor produced by the reaction. Precious metal catalysts are useful in promoting combustion reactions, and it is desirable to preserve such valuable catalysts by fixing them to a stationary, solid surface known as a substrate. The reactants, fuel and air, react on contact with the catalyst surface
forming combustion products. For solid fuels such as coal, this means that liquefaction or gasification of the fuel is required to enable contact with the solid catalyst surface in the presence of air. Regardless of fuel type, the need for the reactants to contact a solid catalyst surface also means that a high substrate surface area is desirable, as is a high rate of mass transfer to the solid surface. Thus, catalyst substrates typically take the form of small pellets, rods, wires, tubes, honeycombs, foams, or other high-surface-area shapes through which flow can pass. Channel diameters through such structures are typically between 1 mm and 1 cm in diameter, depending on size and pressure drop requirements. Substrates can be made of metal or ceramic materials, but must withstand the expected operating environment in a gas turbine engine, particularly with regard to thermal gradients and thermal shock8. Metal substrates best fill this need, but their temperature must be limited to less than 950°C (1750°F) to ensure sufficient material strength and long life. Ceramic materials can operate at higher temperatures, but issues of thermal shock failure during the transient operating conditions required for gas turbine operation have not been fully resolved. Thus, despite their inherent temperature limitations, metal substrates have been nearly universally adopted for gas turbine catalytic combustion, even for machines with firing temperatures hundreds of degrees higher than 950°C (1750°F). As a result, a prominent feature of successful catalytic combustor designs has been control or limitation of catalyst temperature without sacrifice in engine, combustor, or emissions performance. Once the reactants are in contact with the substrate surface, it is also desirable that catalytic reactions proceed quickly. This requires a large number of active catalyst surface sites per unit surface area of substrate. For this purpose a high-surface-area “support” material such as a porous ceramic washcoat may be applied to the substrate, creating a new “rough” or porous surface that can have thousands of times the surface area of the raw substrate. The catalyst is finely dispersed across this rough surface and throughout its pores, to give a high density of active sites for chemical reaction. Specific procedures for preparing such supported catalysts have been disclosed in the literature9. Figure 1 presents a cross-sectional view of an actual washcoat support, containing precious-metal catalyst, and bonded to an underlying metal substrate. A high density of active surface sites is especially important in achieving the lowest possible lightoff temperature for a given catalyst. For practical applications of catalytic combustion, lightoff temperature is a critical parameter that determines both system operation and system design. The lightoff process can be depicted graphically as suggested by Pfefferle and Pfefferle and as shown schematically here in figure 210. Prior to catalyst lightoff the temperature difference between the catalyst surface and the reactant gas stream is small (tens of degrees or less) and the reaction rate is slow enough that this temperature difference is sufficient to remove all heat released at the surface. Pre-lightoff operating conditions correspond to region I in figure 2. However, if gas and catalyst temperatures are increased, reaction rates will increase exponentially with temperature until Fig. 1. Microscope photograph of sectioned catalyst support and substrate. The reactants are consumed as quickly as they arrive at catalyst support is a ceramic washcoat, bonded to the underlying metal substrate. the surface. When this occurs, the surface reaction is The scale of this photograph is roughly 100 µm (0.004 in) in total height. said to be mass transfer limited, since mass transfer of fuel and oxygen to the surface now limit the overall reaction rate and temperature has little impact. At this condition the catalyst temperature will greatly III heat release exceed (by hundreds of degrees) the temperature of the reactant gas stream, and in fact, the catalyst will operate close to the adiabatic flame temperature Heat of the reactant gas stream if no external cooling is II Rate provided. This changeover to mass transfer limited (W) operation at greatly increased catalyst temperature is known as catalyst lightoff, and occurs rapidly (on heat removal the order of seconds) once the lightoff temperature is Ql.o. I reached. Figure 2 indicates how lightoff temperature Tgas Tl.o. Temperature (K) is defined and why lightoff is rapid. The solid curve represents heat release rate (in Watts, for example) Fig. 2. Representation of heat generation and heat removal as from exothermic reactions at the catalyst surface, a function of catalyst temperature, as during catalyst lightoff as a function of catalyst temperature. The dashed curve indicates the effect of catalyst temperature on the rate of heat removal from the catalyst surface,
reactants (gas)
100 µm (0.004 in)
catalyst / support substrate (solid)
256
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle predominantly by convection to the reactant stream. The dashed curve’s point of intersection on the x-axis is the gas stream temperature. Note that the solid curve is independent of gas stream temperature, but the dashed curve will slide left and right, depending on gas stream temperature, with a constant slope determined by the convective heat transfer coefficient. For steady-state operation heat removal must equal heat release, so operation must be at an intersection point of the solid and dashed curves. Furthermore, for this intersection point to be a stable operating condition, an increase in catalyst temperature must lead to greater heat removal than heat release. At lightoff, this stability condition fails, and lightoff temperature Tl.o. is defined at the point of tangency between the solid and dashed curves, as shown. At this point, any increase in catalyst temperature leads to greater heat release than heat removal, with the result that catalyst temperature rapidly increases until reaction rate is limited by mass transfer of reactants to the catalyst surface. Thus, catalyst temperature passes through the unstable region II, and stabilizes at its mass-transfer-limited value in region III where the dashed curve and the solid curves again intersect. Catalyst lightoff is therefore a rapid transient event, and can be measured for any catalyst material with only weak dependence on heat transfer coefficient. Note that some authors define lightoff temperature in terms of the gas temperature Tgas during lightoff, instead of catalyst temperature Tl.o. at lightoff as defined here. Catalyst materials for combustion applications have been carefully evaluated by a number of researchers11. Methane and natural gas fuels have been the recent focus of interest because natural gas is currently the low-emissions fuel of choice for power-generating gas turbines. Johansson and co-workers have reviewed the catalytic combustion literature and summarized lightoff temperatures for fuellean oxidation of methane on various catalysts, as listed in table 112. Dalla Betta (1997) reviewed and summarized fuel-lean methane reaction rates at 400°C (750°F) for a range of catalyst materials tested, and the more active of these are also listed in table 113. In both reviews, palladium (Pd) catalysts show the greatest activity for methane oxidation, and in general the consensus in the literature has been that, for fuel-lean combustion of methane, Pd-based catalysts are the only practical choice because only they offer acceptable activity, lightoff temperature, and resistance to volatilization. However, the adoption of Pd-based catalysts has come with difficulties resulting from the complex morphology and behavior of the Pd-PdO system and its reactions with methane, including deactivation at high temperature (above about 750°C (1380°F)), hysteresis in reaction rate over heating and cooling cycles, and oscillations in activity and temperature14. In addition, lightoff and extinction temperatures are well above 300°C (570°F) for fuel-lean reaction of methane on Pd-based catalysts, thus requiring the use of a preburner in many engine applications15. A wider choice of catalysts is possible for fuel-rich reaction of methane, and the complex behaviors of fuel-lean methane reaction on Pd are avoided. For example, Lyubovsky et al. compared fuel-lean and fuel-rich activity for three precious metal catalysts in an isothermal reactor, and demonstrated that all three could be practically used for fuel-rich catalytic reactors16. A comparison of measured catalyst activity for Pd, Pt, and Rh catalysts, under both fuel-rich and fuel-lean conditions, is shown here in figure 3 as a function of catalyst temperature17. Note that figure 3 does not depict thermal lightoff in the same sense as figure 2, since the data in figure 3 were obtained at isothermal conditions (heat transfer was essentially infinite, so the catalyst could not self heat). The first observation to make in figure 3 is that, for all three catalysts, methane conversion (a measure of catalyst activity) is generally much higher under fuel-rich conditions than under fuel-lean conditions. Furthermore, under fuel-lean conditions, only the Pd catalyst shows sustained measurable activity below 600°C (1110°F), whereas all three catalysts are always active in this range under fuel-rich conditions. Evidently the oxidizing environment of fuel-lean mixtures leads to oxidation of the catalyst and this has a pronounced effect on catalyst activity. Oxidation also leads to greater volatilization of Pt catalyst, limiting the usable life of Pt-based catalysts even at temperatures far below 950°C (1750°F). For fuel-rich combustion of methane, however, Pd, Pt, or Rh-based catalysts may be used without suffering from oxidation.
Table 1 Lightoff temperature and reaction rate for methane oxidation under fuel-lean conditions, for various catalyst materials
Catalyst Type Platinum Group Metals Perovskite Hexa-aluminate
Specific Examples of Catalyst Type Pd / Al2O3 Pt / Al2O3 La0.5Sr0.5CoO3 LaCoO3
Lightoff Temperature(C)
2
340 – 460 590 – 710 390 – 690
300 50
500 – 700
Sr0.8La0.2MnAl11O19 Zeolite Source: see note 4 (Dalla Betta) and note 8 (Johansson).
257
Reaction Rate (10-7 mol / g-s)
1
7 1 3
600 - 750
3.2.2 Catalytic Combustion
Fig. 3. Catalyst activity for methane oxidation under fuel-rich and fuel-lean conditions. Filled symbols represent increasing temperature, and closed symbols represent decreasing temperature (the selected temperature ramp was externally controlled by electric heaters).
For syngas and high-hydrogen fuels, catalyst lightoff of CO and H2 is of interest for both fuel-lean and fuel-rich conditions. Recent work at Precision Combustion, Inc., sponsored by the U.S. Department of Energy, has shown a precious-metal catalyst lightoff temperature of 180°C (350°F) for fuel-rich reaction of syngas. This value has been measured for a syngas mixture composed of 25% H2 and 35% CO (remainder diluent), but the lightoff temperature is relatively insensitive to syngas composition unless CO levels drop to very low values, thus leaving essentially a high-hydrogen fuel having significantly lower lightoff temperature. Light-off temperatures for fuel-lean syngas mixtures were also measured below 180°C (350°F). In all cases, the 180°C (350°F) or lower lightoff temperature is well below the compressor discharge temperatures of industrial and large-frame turbines, and the need for a preburner is thereby avoided.
Flame Temperature Considerations Actual flame temperatures for catalytic combustion will vary depending on the design approach used to integrate the catalytic reaction zone with the gas-phase combustion zone and also on the turbine inlet temperature requirement for the application engine. In any case, flame temperatures greater than about 1100°C (2000°F) are required for gas-phase reactions to complete the burnout of hydrocarbon fuels and CO in a reasonable residence time (on the order of 10 ms). Thus, systems which rely on catalyst-induced autoignition to stabilize combustion will normally have a minimum gas-phase combustion zone temperature of about 1100°C (2000° F), although the catalyst itself may be limited to temperatures well below this value. Systems which use conventional flameholding techniques without catalyst-induced autoignition (but with catalyst augmentation of flame stability) normally require higher flame temperatures to prevent blowout, with typical minimum flame temperatures exceeding 1300°C (2400°F) for hydrocarbon fuels. In most power generation applications, it is required that the turbine operates over a range of loads while meeting emissions regulations. The combustor must therefore be designed to operate with flame temperatures above the minimum achievable value at base load (full power) so that the machine can be turned down to a flame temperature near the minimum achievable value at the lowest required low-emissions load point. Thus, actual flame temperatures at base load are generally higher than the minimums listed in the previous paragraph. In addition, for modern high-efficiency engines having high turbine inlet temperatures, actual flame temperatures at base load may need to be higher than the minimums listed above simply to meet the required turbine rotor inlet temperature after the addition of necessary cooling air. For example, General Electric’s 7FA engine has a published turbine rotor inlet temperature of 1325°C (2420°F) (Eldrid et al., 2001) and a published temperature drop due to cooling air addition at the first stage nozzle (stator) of 140°C (280°F), giving a minimum flame temperature of 1480°C (2700°F) prior to the first-stage nozzle18. The above considerations establish the minimum required flame temperature at base load, but for ultra-low NOx operation a maximum flame temperature is also present, based on increasing formation of thermal NOx with increasing temperature19. It is generally accepted that at temperatures above about 1525°C (2780°F) NOx emissions increase markedly with temperature, and generally exceed 3 ppm (at 15% O2) for premixed flames of hydrocarbon fuels in air. Thus, for a given low-emissions application the design choice for base load flame temperature falls within a window that is generally less than 1525°C (2780°F) but higher than the minimum required. Based on such flame temperature considerations, catalytic combustion systems for gas turbines can be classified in three basic categories as shown in table 2. The categories are most directly stated in terms of turbine rotor inlet temperature requirements. For the first two categories in table 2 (low-temperature and high-temperature turbines), turbine rotor inlet temperatures are sufficiently below the 1525°C (2780°F) thermal NOx limit that catalytic combustion can be used to provide stable operation, thus providing ultra-low NOx emissions even at base load. For the third category (ultra-high temperature turbines); however, flame temperatures in excess of 1525°C (2780°F) may be required simply to meet the required turbine rotor inlet temperature after the addition of necessary cooling air. Therefore, NOx emissions for such applications may be higher than 3 ppm unless cooling air requirements are reduced (as in the steam-cooled H-engine for example), or unless alternative developments are advanced to allow low-NOx operation at higher flame temperatures (as discussed further below).
258
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle Note that in all cases, catalytic combustion provides additional benefits other than low emissions. In particular, catalytic combustion can be used to provide extended turndown by allowing stable combustion at low part-load flame temperatures and can also be used to reduce combustion dynamics (combustion-induced pressure oscillations) by shifting fuel energy release to the catalyst surface and away from the gas-phase combustion zone. Reduced combustion dynamics may be an especially important benefit in high-firingtemperature machines that are susceptible to combustion dynamics as a result of their inherent high rate of fuel energy release. Table 2 Categories for gas turbine applications of catalytic combustion, based on turbine rotor inlet temperature. The term 1-Stage means a single stage system combining both catalyst and gas-phase reactions. The term 2-Stage means that combustion occurs in two stages, sequential and separate, where the first stage is catalytic and the second is gas-phase. (See Figure 4 for schematic examples of 1-Stage and 2-Stage systems.)
Low-Temperature (Uncooled) Turbines
High-Temperature (Cooled) Turbines
Ultra-High Temperature (> F-Class) Turbines
Microturbines
Industrial D-, E-, F-Class
FB-Class G-Class
1-Stage or 2-Stage
2-Stage
2-Stage
1- or 2-Stage Ceramic
Low Oxygen 2-Stage (EGR)
1. Emissions
√
√
?
√
√
2. Turndown
√
√
√
√
√
3. Dynamics
√
√
√
√
√
Barrier Issues
-
-
-
Materials Develop.
Cycle & Machine Develop.
Engine Applications Approaches to Catalytic Combustion Benefits:
Specific considerations for each of the categories in table 2 are discussed in the following subsections, especially as related to catalytic combustion system design for each category.
Catalytic Combustion at Low Flame Temperatures (Uncooled Turbines) For uncooled turbines such as microturbines, turbine inlet temperatures are generally within catalyst material temperature limits (less than 1000°C (1830°F). This means that the catalyst and substrate can tolerate the maximum required flame temperature, and if complete combustion can be sustained, there is no need to separate the gas-phase combustion zone from the catalyst. Thus, a simple catalytic combustor for this application can comprise only a premixer and a catalyst bed, as shown schematically in the righthand panel of figure 4. This simple system is considered a single-stage system, although fuel is oxidized within this single stage by both catalytic reactions on the catalyst surface and gas-phase reactions in the fluid stream above the surface. Because these reactions can go to completion (all fuel converted to products) within the catalyst stage, this single-stage system is also known as a “total conversion” catalytic reactor.
Premixer
Catalytic Reactor
Air
Premixer Homogenous Reaction
Fuel
Partial Conversion
Catalytic Reactor
Air Fuel Total Conversion
Fig. 4. Schematic examples of 2-Stage (partial conversion catalyst) and 1-Stage (total conversion catalyst) catalytic combustors, in left and right panels, respectively
259
3.2.2 Catalytic Combustion Note that the single-stage “total conversion” catalytic reactor requires gas-phase reactions to complete the oxidation of all fuel in a reasonable length or residence time. Complete combustion (e.g. 99.9% combustion efficiency) without gas-phase reactions would require mass transfer of nearly all fuel (e.g. 99.9%) to the catalyst surface for reaction, which would require an impractically sized catalyst in terms of length, cross-sectional area, and/or channel size. Thus, the single-stage “total conversion” catalytic reactor must provide sufficient catalytic reaction and concomitant heating of the fuel/air mixture that gas-phase autoignition is induced throughout, providing combustion completion even if the fuel/air mixture is outside its flammability limits. Based on the adiabatic flame temperature of the fuel/air mixture passing over the catalyst, this autoignition requirement determines the minimum required catalytic conversion of fuel. Or, in the opposite sense, based on the practically achievable catalytic conversion of fuel, the autoignition requirement determines the minimum adiabatic flame temperature of the fuel/air mixture. A separate consideration is, as stated earlier, the adiabatic flame temperature of the fuel/air mixture must generally exceed at least 1100°C (2000°F) to provide complete burnout in a reasonable residence time. At 1100°C (2000°F) metal substrates cannot be used to support the catalyst, and a ceramic substrate must be used. Thus, the single-stage catalytic combustor is generally ceramics-based, and is prone to ceramics-related issues such as thermal stress and thermal shock. Early catalytic combustors of the type originally developed at Engelhard were single-stage ceramics-based systems20. Currently, however, metal substrates have gained favor in many catalyst applications, including catalytic combustion. Metals are especially suitable for gas turbine applications since they are able to withstand gas turbine demands such as thermal stress and thermal shock. But metal temperatures must normally be limited to less than about 950°C (1750°F) for long-term durability, and this precludes the use of the single-stage “total conversion” system described above. Instead, a two-stage “partial conversion” system must be used where only a portion of the fuel is reacted in the catalyst stage, and gas-phase combustion completion occurs downstream of the catalyst. This is shown schematically in figure 5, where there are two physically distinct stages having different temperature requirements (< 950° C (1750°F) in the catalyst stage to meet long-term durability requirements, and > 1100°C (2000°F) in the gas-phase combustor stage to provide combustion completion). The success of the two-stage system is dependent upon the catalyst stage’s ability to limit reactions (and thus temperature), so that the metal substrate within the reactor may operate below its maximum material temperature limit. In the catalyst stage, the degree of reaction can be limited by chemical reaction rate upon the catalyst, by mass transfer of reactants to the catalyst, or by channeling within the reactor such that only a limited fraction of the fuel can contact the catalyst. In all cases it is imperative that uncontrolled gas-phase reactions do not occur within the catalytic reactor, since this implies a loss of reaction limitation and ultimately the over-temperature and failure of the catalyst bed. While there are many design variations and Ultra low emissions Fuel / Air possibilities for the catalyst-stage, there are only two (NOx, CO, UHC) mechanisms for sustaining gas-phase combustion in the combustion stage: catalyst-induced autoignition (as implied Catalyst Gas phase burnout schematically in figure 5), or flameholding via backmixing of <950 C >1100 C hot combustion products (as shown schematically in figure Fig. 5. Schematic representation of metal-based two-stage 4, left panel). If catalyst-induced autoignition is used in the catalytic combustion system, showing catalyst stage requiring < two-stage system, the primary challenge is to prevent such 950 C (1750 F) temperature for long-term metal durability, and gasautoignition from occurring within the catalyst-stage, as it phase combustion stage requiring > 1100 C (2000 F) temperature does in the single-stage system described above. for combustion completion
Catalytic Combustion at Moderate Flame Temperatures (Low NOx Cooled Turbines) For modern, cooled turbines having turbine rotor inlet temperatures well above 1100°C (2000°F), the single-stage catalytic combustor of figure 4 (right-hand panel) is no longer feasible since flame temperatures prior to the addition of combustor and stator cooling air will generally exceed material temperature limits for available catalyst and substrate materials, even if ceramics are used. Thus, two-stage catalytic combustion is used in these machines, and metal catalyst substrates have generally been adopted for their robustness in the gas turbine environment. Catalyst-stage temperatures must therefore remain below 950°C (1750°F), while gasphase combustion temperatures may reach 1525°C (2780°F) before NOx emissions increase beyond acceptable levels. Ultra-low NOx emissions have been successfully demonstrated in recent engine tests of two different two-stage systems21. Note that if catalyst-induced autoignition (figure 5) is the intended means of sustaining gas-phase combustion, it becomes increasingly difficult to prevent autoignition or flashback from overheating the catalyst when gas-phase combustion temperatures increase to meet advancing turbine rotor inlet temperatures. At F-engine type conditions, where flame temperatures must reach 1480° C (2700°F) as discussed earlier22, no successful engine operation of a two-stage catalytic combustor has been reported using catalystinduced autoignition to sustain gas-phase combustion. However, successful results have been obtained over a wide range of flame temperatures, reaching 1480°C (2700°F) and beyond, by using conventional flameholding techniques (backmixing of hot combustion products, left-hand panel of figure 4) while maintaining catalyst-stage temperatures below the autoignition temperature of the fuel/air mixture23.
260
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle Catalytic Combustion at High Flame Temperatures (Tflame > 1525 C (2780 F)) In the highest firing-temperature machines used for power generation, such as the recent FB-class and G-class machines, gasphase combustion temperatures may need to exceed 1525°C (2780°F), before addition of necessary cooling air, in order to meet turbine rotor inlet temperature requirements. At the current level of machine, combustor, and catalyst materials development, this means that NOx emissions will likely exceed 3 ppm at 15% O2. In spite of this, there are possible developments to reduce NOx emissions at these high flame temperatures, and there are non-emissions-related benefits of catalytic combustion (improved turndown and combustion dynamics) that are available immediately. Combustion dynamics in particular have been problematic in low-emissions high-firing-temperature machines24, and there is great incentive to find an economical solution. One possible solution is catalytic combustion. Catalytic combustors have been reported to have low combustion dynamics (Smith et al., 2005; Schlatter et al., 1997), since gas-phase energy release in the combustor is the driving force for combustion-induced pressure oscillations (“combustion dynamics”) and these oscillations are reduced when a portion of the fuel is catalytically reacted prior to gas-phase combustion25. Thus, regardless of pollutant emissions levels, catalytic combustion may prove useful even when flame temperatures must be well in excess of 1525°C (2780°F) to meet high turbine inlet temperature requirements. In fact, it is at these high flame temperatures that combustion dynamics often become most problematic and a solution is most needed. In addition, non-catalytic premixed combustors now often employ piloting or fuel staging to tune out combustion dynamics, at the cost of increased NOx emissions. Thus, a catalytic combustor that operates with low combustion noise without piloting or fuel staging (that is, with more perfectly premixed combustion) may offer reduced NOx emissions as compared to an equivalent non-catalytic system, even at the same overall combustor outlet temperature. There is also the potential that catalytic combustion may allow low-NOx operation even at flame temperatures above 1525°C (2780°F). While it has generally been reported that NOx emissions exceed 3 ppm (at 15% O2) when flame temperatures exceed 1525° C (2780°F) for hydrocarbon fuels burning in air, these reports have typically focused on conventional flame stabilization techniques that use recirculation or backmixing of hot combustion products to sustain combustion. Schlegel and co-workers have shown through experiments and modeling that lower NOx emissions are possible for two reasons26. First, their work shows that NOx production is most significant in the stirred or backmixed flame stabilization zone, and that NOx emissions can be reduced by reducing the “size” (bulk residence time) of such backmixing zones. Catalytic combustion can facilitate such NOx reduction by improving flame stability and lean blowout, such that “smaller” (shorter residence time) backmixed zones are feasible. For example, model results by Schlegel and co-workers show that NOx emissions from a 1500°C (2730°F) combustion process can be reduced by more than 1 ppm by decreasing the stirred-zone residence time from 0.6 ms to 0.1 ms27. Because of this potential, a question mark was indicated in table 2 for the emissions benefit of 2-stage catalytic combustion in ultra-high firing-temperature machines. Second, Schlegel and co-workers have shown that by increasing the percentage of fuel that is reacted on the catalyst, NOx emissions are reduced because the chemical mechanisms for NOx formation are affected in the gas-phase combustion zone28. This effect is most pronounced at high levels of catalytic fuel conversion (> 50%, preferably > 80%) as shown in their 1500°C (2730°F) combustion experiments and modeling. At these fuel conversion levels, NOx reductions of more than 1 ppm are possible at 1500°C (2730°F) flame temperature. Note that materials development will be required to effectively implement this technique with long-term durability, since catalyst operating temperatures will greatly exceed the limits of currently available materials at such high levels of fuel conversion. Table 2 therefore indicates low emissions in a 1- or 2-stage ceramic-based catalytic combustor, with the caveat that materials development is required for implementation. The third approach to catalytic combustion listed in table 2 for ultra-high firing-temperature machines is a low-oxygen environment, where NOx emissions can remain below 3 ppm even at flame temperatures in excess of 1525°C (2780°F). Exhaust gas recirculation (EGR) is one possibility for obtaining such a low-oxygen environment and has been investigated for gas turbine emissions reduction in the past without the use of catalysts29. Because premixed combustion can be difficult to sustain under such low-oxygen conditions, a catalytic combustor will be beneficial in providing the needed flame stability, and could permit ultra-low NOx emissions even from ultra-high temperature machines. Development needs are great for such systems; however, since this concept requires a significant change in the working fluid of the engine. This is indicated in table 2 with regard to both cycle and machine development needs.
3.2.2-4 Systems for Catalytic Combustion
261
As discussed earlier, two very different catalytic combustion systems have been developed and engine tested in recent years30 , although both are based on the two-stage catalytic combustor concept and both use metal-based catalyst substrates. Thus, in both systems combustion completion occurs in a downstream gas-phase combustion stage that is physically separated from the upstream catalyst stage. The system developed by Catalytica, Inc. is much like that pictured schematically in figure 5, and is described in greater detail in Section 3.2.2.1.2 of this Handbook31. The Catalytica system premixes all combustion fuel and air upstream of the catalyst, and uses a partial conversion catalytic reactor to oxidize only a portion of the fuel in the catalyst stage. The temperature rise due to fuel oxidation
3.2.2 Catalytic Combustion in the catalytic reactor induces gas-phase autoignition in the downstream gas-phase combustion stage, where combustion is completed. The catalyst stage is operated fuel-lean and uses a Pd-based catalyst. A preburner is generally employed to ensure that the catalyst remains active (lit off) during low-emissions engine operation. The system developed by Precision Combustion, Inc. is shown schematically in figure 6, as originally developed for operation on natural gas32. This system is described in greater detail in Section 3.2.2.1.1 of this Handbook. Briefly, the Precision Combustion system is based on fuel-rich operation of the catalyst, and sustains fuel-lean gas-phase combustion downstream of the catalyst via recirculation-based flameholding. All the fuel and part of the air pass through the catalyst with the remainder of the air providing catalyst cooling. This cooling air then mixes with the catalyst effluent establishing a fuel lean flame. Fuel-rich operation of the catalyst provides greater catalyst activity than fuel-lean operation, such that a preburner is not normally required during low-emissions engine operation. Flashback and autoignition issues are also precluded because fuel oxidation and heat release in the catalyst stage are limited by available oxygen, and the system can therefore be operated safely even at the highest desired combustion-stage temperatures. Fuel-rich operation also allows similar catalyst and reactor performance with widely varying fuel types, since catalyst-stage reactions are starved of oxygen, thereby limiting the extent of fuel oxidation and heat Catalyst release regardless of the fuel’s intrinsic reactivity on Cooling Combustion the catalyst. The system has been successfully tested on multiple fuels, including Diesel No. 2, simulated Air refinery fuel gas and syngas, and other low-Btu fuels. Burned Gas Catalytic combustion systems can also be Fuel combined with non-catalytic lean-premixed systems to Premixer Catalytic Post-Catalyst offer unique capabilities in terms of performance, cost, Reactor Mixing pressure drop, and space requirements. For example, Precision Combustion, Inc. has developed a catalytic Fig. 6. Schematic of Precision Combustion’s two-stage catalytic combustion pilot system that combines its RCL system with non- system, as originally developed for operation on natural gas. Fuel-rich catalyst catalytic swirl-based fuel/air injection. This system effluent mixes with catalyst cooling air prior to fuel-lean gas-phase combustion, has been tested in cooperation with Solar Turbines and the system is therefore called Rich-Catalytic Lean-burn combustion (trademarked as RCL by Precision Combustion, Inc.) and is described by Karim and co-workers33.
3.2.2-5 Challenges for Catalytic Combustion While the art and science of chemical reactor design using heterogeneous catalysis is well known, the application of this technology to catalytic combustion has required significant development efforts due to both the exothermicity of combustion and the operational needs of gas turbine engines. A catalytic process for a gas turbine engine must contend with rapid heat release and high operating temperatures, variable feed composition (fuel/air ratio may change, and fuel quality and composition may vary), variable thermodynamic conditions (inlet pressure and temperature), and variable reactor residence time (space velocity). To provide this flexibility, the catalytic combustor is not generally required to maximize yield or selectivity to a particular product; instead, the catalyst is required to provide stability to a combustion process, usually by providing a temperature rise to a fuel/air mixture. For catalytic combustion, the overriding constraints are robustness and controlled catalyst temperatures over the wide operating range of the engine. For natural gas combustion, these challenges have been met and initial engine demonstrations have successfully shown that catalytic combustion is capable of driving a gas turbine engine and delivering ultra-low NOx emissions. For other fuels, however, further development and demonstration is required. For catalytic combustion of liquid fuels such as Diesel No. 2, development of an adequate and robust prevaporizer technology is required to prevent wetting of the catalyst surface and to allow reaction of mixed, prevaporized fuel and air on the catalyst surface. For solid fuels such as coal, gasification is required upstream of the catalyst, and there has been growing interest in catalytic combustion of gasified coal or syngas. Because high-hydrogen fuels such as coal-derived syngas are prone to autoignition and flashback, lean-premixed combustion has not been adopted for low-emissions combustion of these fuels34. Instead, non-premixed combustion is used, and diluents such as nitrogen, water, and/or carbon dioxide have been used to reduce flame temperatures and therefore NOx emissions35. Unfortunately, flame stability issues arise at high dilution levels, and NOx emissions below 3 ppm (at 15% O2) have not been reported for non-catalytic syngas combustion. Catalytic combustion, however, can stabilize combustion at lower flame temperatures, or with increased levels of dilution, and offers the potential for IGCC plant emissions below 3 ppm. Precision Combustion, Inc. under contract to DOE has recently demonstrated NOx emissions below 3 ppm in high-pressure, sub-scale catalytic combustion rig tests simulating Tampa Electric’s Polk Power plant operating conditions. However, further development is required to bring this technology to an engine, and eventually to commercialization. As with conventional combustion of syngas, and especially after diluent addition, the very low Lower Heating Values (LHV) of such fuels means that a very high volume flow of fuel is required. For catalytic combustion, this high volume flow must be accommodated within the catalytic reactor, and this raises issues of catalyst size and pressure drop, possibly requiring system level re-design and redevelopment of the catalytic combustor.
262
Dr. Lance Smith, Dr. Hasan Karim, Dr. Shahrokh Etemad, Dr. William C. Pfefferle The propensity for high-hydrogen fuels to autoignite, together with their wide flammability limits, require special attention to premixing design. In addition, syngas fuels carry trace levels of catalyst contaminants that may affect long-term catalyst durability, and this needs to be examined and remediated if problematic. Long-term durability tests are required, preferably in an actual syngas slipstream at an operating IGCC plant, where real contaminants will be present.
3.2.2-6 Conclusions Catalytic combustion has been established as providing low NOx emissions for modern gas turbines along with subsequent reduction in combustion dynamics and improved operability. Feasibility of catalytic combustion has been established for both high and low firing temperature gas turbines. For natural gas combustion, the challenges of catalytic combustion have been met and initial engine demonstrations have successfully shown that catalytic combustion is capable of driving a gas turbine engine and delivering ultra-low NOx emissions. Further field trials and engine operating experience will advance the catalytic combustion applications for gas turbines. For alternative fuels, particularly coal-derived syngas, further development is required; but, initial demonstrations have shown the potential for ultra-low emissions catalytic combustion in future IGCC power plants. In recent years, metal catalyst substrates have been adopted, together with a two-stage approach to catalytic combustion that maintains catalyst temperatures within their material limits. In addition, the two-stage approach allows gas-phase combustion temperatures to rise to the levels needed for good combustion efficiency and high efficiency turbine operation. Prior material limitations have been largely resolved through design concepts such as reactor backside cooling and reactor mode of operation as well as availability of high temperature metal alloys. Multiple approaches to catalytic combustion have been pursued over the years in the quest for ultra-low NOx emissions from power generating gas turbines. Low, single digit, NOx emissions with low combustion dynamics have been demonstrated in engine environments permitting a wider operating regime from part load to full load conditions. These benefits make catalytic combustion a viable, low cost approach as compared to selective catalytic reduction (SCR) to meet the low emissions requirements.
3.2.2-7 Notes ________________________
1. W.C. Pfefferle, R.V. Carruba, R.M. Heck, and G.W. Roberts, “Catathermal Combustion: A New Process for Low
Emissions Fuel Conversion,” ASME Paper No. 75-WA/Fu-1(1975). 2. Ibid. 3. D.N. Anderson, R.R. Tacina, and T.S. Mroz, “Performance of a Catalytic Reactor at Simulated Gas Turbine Operating Conditions,” NASA Technical Memorandum X-71747(1975); J.P. Kesselring, W.V. Krill, E.K. Chu, and R.M. Kendall. In proceedings of New fuels and advances in combustion technologies symposium, Mar. 26-30, 1979, New Orleans, LA; P.W. Pillsbury, “Update of FullScale Catalytic Burner Testing for Combustion Turbines,” ASME Paper No. 84-GT-54 (1984); T.J. Rosfjord, AIAA Paper No. 7646 (Washington DC, Jan. 1976). 4. R.A. Dalla Betta et al., Journal of Engineering for Gas Turbines and Power 119 (1997):, 844-851; P. Dutta, D.K. Yee, and R.A. Dalla Betta, ASME Paper No. 97-GT-497 (1997); S. Etemad, H. Karim, L.L. Smith, and W.C. Pfefferle, “Advanced Technology Catalytic Combustor for High Temperature Ground Power Gas Turbine Applications,” Catalysis Today 47 (1999): 305-313; D.A. Smith, S.F. Frey, D.M. Stansel, and M.K. Razdan, ASME Paper No. 97-GT-311(1997). 5. R.A. Dalla Betta, T. Shoji, D.K. Yee, and S.A. Magno. Catalyst Structure Employing Integral Heat Exchange. U.S. Patent 5,512,250; R.A. Dalla Betta, N. Ezawa, K. Tsurumi, J.C. Schlatter, and S.G. Nickolas, “Two Stage Process for Combusting Fuel Mixtures,” U.S. Patent No. 5,183,401 (1993); W.C. Pfefferle, L.L. Smith, and M.J. Castaldi, “Method and Apparatus for a FuelRich Catalytic Reactor,” U.S. Patent No. 6,358,040 (2002). 6. D.K. Yee, K. Lundberg, and C.K. Weakley, “Field Demonstration of a 1.5 MW Gas Turbine with a Low Emissions Catalytic Combustion System,” Journal of Engineering for Gas Turbines and Power 123 (2001): 550-556; L.L. Smith, H. Karim, M.J. Castaldi, S. Etemad, W.C. Pfefferle, V.K. Khanna, and K.O. Smith, “Rich-Catalytic Lean-Burn Combustion for Low-Single-Digit NOx Gas Turbines,” Journal of Engineering for Gas Turbines and Power 127 (2005): 27-35. 7. G. Leonard and J. Stegmaier, “Development of an Aeroderivative Gas Turbine Dry Low Emissions Combustion System,” Journal of Engineering for Gas Turbines and Power 116 (1994): 542-546.
263
8. E.M. Johansson, D. Papadias, P.O. Thevenin, A.G. Ersson, R. Gabrielsson, P.G. Menon, P.H. Bjornbom and S.G. Jaras, “Catalytic Combustion for Gas Turbine Applications,” Catalysis 14 (1999): 183-235; R.E. Hayes and S.T. Kolaczkowski, Introduction to Catalytic Combustion (Amsterdam:Gordon and Breach Science Publishers, 1997); D. Anson, M. DeCorso and W.P. Parks, “Catalytic Combustion for Industrial Gas Turbines,” International Journal of Energy Research 20 (1996): 693-711; S.T. Kolaczkowski, “Catalytic Stationary Gas Turbine Combustors: A Review of the Challenges Faced to Clear the Next Set of Hurdles,” Trans. I. Chem. E. 73 Part A (1995): 168-190.
3.2.2 Catalytic Combustion 9. J.G. McCarty, “Kinetics of PdO Combustion Catalysis,” Catalysis Today 26 (1995): 283-293; R. Burch and P.K. Loader, “Investigation of Pt/Al2O3 and Pd/Al2O3 Catalysts for the Combustion of Methane at Low Concentrations,” Applied Catalysis B: Environmental 5 (1994): 149-164; R.J. Farrauto, M.C. Hobson, T. Kennelly, and E.M. Waterman, “Catalytic Chemistry of Supported Palladium for Combustion of Methane,” Applied Catalysis A: General 81 (1992): 227-237; T. Kennelly and R.J. Farrauto, “Catalytic Combustion Process Using Supported Palladium Oxide Catalysts,” U.S. Patent No. 5,216,875 (1993); R.A. Dalla Betta, K. Tsurumi, and T. Shoji, “Graded Palladium-Containing Partial Combustion Catalyst and a Process for Using It,” U.S. Patent No. 5,248,251 (1993). 10. L.D. Pfefferle and W.C. Pfefferle, “Catalysis in Combustion,” Catal. Rev.-Sci. Eng. 29 (1987): 219-267. 11. W.S. Blazowski and D.E. Walsh, “Catalytic Combustion: An Important Consideration for Future Applications,” Combustion Science and Technology 10 (1975): 253-244; P. Forzatti and G. Groppi, “Catalytic Combustion for the Production of Energy,” Catalysis Today 54 (1999): 165-180; J.H. Lee and D.L. Trimm, “Catalytic Combustion of Methane,” Fuel Processing Technology 42 (1995): 339-359; F.H. Ribeiro, M. Chow, and R.A. Dalla Betta (1994). J. Catal., 146 (1994): 537. l2. See note 8 above (Johansson). 13. R.A. Dalla Betta, “Catalytic Combustion Gas Turbine Systems: The Preferred Technology for Low Emissions Electric Power Production and Co-generation,” Catalysis Today 35 (1997): 129-135. 14. R.J. Farrauto, J.K. Lampert, M.C. Hobson, and E.M. Waterman, “Thermal Decomposition and Reformation of PdO Catalysts; Support Effects,” Applied Catalysis B: Environmental 6 (1995): 263-270; T. Furuya, K. Sasaki, Y. Hanakata, T. Ohhashi, M. Yamada, T. Tsuchiya, and Y. Furuse, “Development of a Hybrid Catalytic Combustor for a 1300°C Class Gas Turbine,” Catalysis Today 26 (1995): 345-350; Y. Ozawa, Y. Tochihara, N. Mori, I. Yuri, T. Kanazawa, and K. Sagimori, “High Pressure Test Results of a Catalytically Assisted Ceramic Combustor for a Gas Turbine,” ASME Paper No. 98-GT-381 (Stockholm, Sweden, 2-5 June 1998); N.M. Rodriguez, S.G. Oh, R.A. Dalla-Betta, and R.T.K. Baker, “In Situ Electron Microscopy Studies of Palladium Supported on Al2O3, SiO2, and ZrO2 in Oxygen,” J. Catalysis 157 (1995): 676-686; R. Carroni, V. Schmidt, and T. Griffin, “Catalytic Combustion for Power Generation,” Catalysis Today 75 (2002): 287-295. 15. D.B. Fant, G.S. Jackson, H. Karim, D.M. Newburry, P. Dutta, K.O. Smith, and R.W. Dibble, “Status of Catalytic Combustion R&D for the Department of Energy Advanced Turbine Systems Program,” Journal of Engineering for Gas Turbines and Power 122 (2000): 293-300. 16. M. Lyubovsky, L.L. Smith, M. Castaldi, H. Karim, B. Nentwick, S. Etemad, R. LaPierre, and W.C. Pfefferle, “Catalytic Combustion over Platinum Group Catalysts: Fuel-Lean versus Fuel-Rich Operation,” Catalysis Today 83 (2003): 71-84. 17. M. Lyubovsky, private communication (2005). 18. T.C. Paul, R.W. Schonewald, and P.J. Marolda, “Power Systems for the 21st Century – H Gas Turbine Combined Cycles,” General Electric Power Systems Report No. GER-3935A (1996). 19. C.T. Bowman, “Control of Combustion-Generated Nitrogen Oxide Emissions: Technology Driven by Regulations,” in proceedings of Twenty-Fourth Symposium (International) on Combustion (1992): 859-878; S.M. Correa, “A Review of NOx Formation Under Gas-Turbine Combustion Conditions,” Combustion Science and Technology 87 (1992): 329-362. 20. See note 2 above. 21. See note 6 above. 22. R. Eldrid, L. Kaufman, and P. Marks, “The 7FB: The Next Evolution of the F Gas Turbine,” General Electric Power Systems Report No. GER-4194 (2001); also, see note 18 above. 23. See note 6 above (Smith). 24. D.E. Hobson, J.E. Fackrell, and G. Hewitt, “Combustion Instabilities in Industrial Gas Turbines – Measurements on Operating Plant and Thermoacoustic Modelling,” Journal of Engineering for Gas Turbines and Power 122 (2000): 420-428; T. Lieuwen and K. McManus, “Introduction: Combustion Dynamics in Lean-Premixed Prevaporized (LPP) Gas Turbines,” Journal of Propulsion and Power 19 (2003): 721; H.C. Mongia, T.J. Held, G.C. Hsiao, and R.P. Pandalai, “Challenges and Progess in Controlling Dynamics in Gas Turbine Combustors,” Journal of Propulsion and Power 19 (2003): 822-829. 25. J.C. Schlatter, R.A. Dalla Betta, S.G. Nickolas, M.B. Cutrone, K.W. Beebe, and T. Tsuchiya, “Single-Digit Emissions in a FullScale Catalytic Combustor,” ASME Paper No. 97-GT-57 (1997); also see note 6 above. 26. A. Schlegel, P. Benz, T. Griffin, W. Weisenstein, and H. Bockhorn, “Catalytic Stabilization of Lean Premixed Combustion: Method for Improving NOx Emissions,” Combustion and Flame 105 (1996): 332-340. 27. Ibid. 28. Ibid. 29. K.K. Botros, G.R. Price, and G. Kibrya, “Thermodynamic, Environmental and Economic Assessment of Exhaust Gas Recirculation for NOx Reduction in Gas Turbine Based Compressor Station,” ASME Paper No. 99-GT-173 (1999). 30. See note 6 above. 31. See note 5 above. 32. See note 5 above (Pfefferle). 33. H. Karim, K. Lyle, S. Etemad, L.L. Smith, W.C. Pfefferle, P. Dutta, and K. Smith. “Advanced Catalytic Pilot for Low NOx Industrial Gas Turbines,” Journal of Engineering for Gas Turbines and Power 125 (2003): 879-884. 34. R.M. Jones and N.Z. Shilling, “IGCC Gas Turbines for Refinery Applications,” General Electric Power Systems Report No. GER-4219 (2003). 35. Ibid.
264
BIOGRAPHY
3.2.2 Catalytic Combustion & 3.2.2.1 Fuel-Rich Catalytic Combustion
Dr. Lance L. Smith Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. Lance L. Smith is a senior research & development engineer in the Gas Turbine Products group at Precision Combustion, Inc. (PCI), and a visiting assistant professor in the Engineering Department at Trinity College. Dr. Smith has 14 years experience in combustion research and combustor development, including work in turbulent non-premixed combustion, premixed combustion and premixing, aerodynamic design of combustor components, pulsed combustion, and catalytic combustion. His academic research has been primarily experimental, with a focus on laser-based measurements in flames, including work conducted as a visiting researcher at Sandia National Laboratories and as a post-doctoral researcher at UCLA. Dr. Smith is a principal engineer of, and holds multiple patents for, the RCLTM catalytic reactor. At PCI, he works with OEM gas turbine manufacturers to develop integrated catalytic combustion systems for ultra-low emissions gas turbines. A graduate of Brown University (B.S., 1986) and of University of California, Berkeley (M.S., 1990 and Ph.D., 1994), Dr. Smith is an elected member of the Tau Beta Pi and Sigma Xi honor societies, and a member of the Combustion Institute.
Dr. Shahrokh Etemad Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473 phone: (203) 287-3700 x217 email: [email protected]
Dr. Shahrokh Etemad as Manager of Gas Turbine Products at Precision Combustion, Inc. (PCI) in North Haven, Connecticut, has full responsibility for technical and commercial development of two major products. He manages an advanced technology group to develop low-emissions combustion products in close collaboration with several OEM gas turbine engine manufacturers and the U.S. Department of Energy. Dr. Etemad is responsible for budgeting, funding opportunities, technology direction, proposal preparation and complete R&D operations including concept development, computational analysis, experimental testing, full-size performance demonstration and productionization. Prior to his present position at PCI, he worked for several years at Textron Lycoming and United Technologies, Carrier. Dr. Etemad has published 28 technical articles and holds 28 patents in the field of turbomachinery, combustion, and thermofluid systems. He earned bachelor’s and master’s degrees at Sussex University and University of London respectively, and received his Ph.D. from the University of Washington in 1984. He has been a member of ASME since 1995 and won the 2003 ASME Gas Turbine award.
Dr. Hasan Karim Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. Hasan Karim is a senior research & development engineer at Precision Combustion, Inc. (PCI), where his responsibilities include design, development, analysis, numerical and computational fluid dynamics, and testing of catalytic combustors for natural gas, syngas, and liquid fuel. He is the principal investigator for the catalytic combustor development project for the U.S. Navy and lead engineer for the catalytic pilot and catalytic combustor for downhole combustion programs. After receiving a bachelor’s degree from Indian Institute of Technology-Kharagpur in 1987, Dr. Karim earned his M.S. from New Jersey Institute of Technology in 1991, and his Ph.D. from Yale University in 1998. He is a co-inventor of air-cooled rich and lean reactor technology.
Dr. William C. Pfefferle Gas Turbine Group, Precision Combustion, Inc. 410 Sackett Point Road, North Haven, CT 06473
Dr. William C. Pfefferle invented the original catalytic combustor for gas turbine engines in the early 1970s and now holds over 90 U.S. patents. In 1986, Dr. Pfefferle co-founded Precision Combustion, Inc. (PCI), a Connecticut-based company dedicated to developing clean and efficient technology for clean air. His research has led to important industrial advances such as the RCL catalytic combustor for ground power gas turbine engines, which is now in late-stage development and evaluation by several major gas turbine manufacturers, and the Microlith catalytic reaction system, which forms the basis for paradigm-shift, high heat mass transfer catalytic reactors for fast-lightoff automotive catalytic converters and fuel processor reactors. With a B.S. in Chemical Engineering from Drexel University (1944) and a Ph.D. in Physical Chemistry from the University of Pennsylvania (1952), Dr. Pfefferle is a member of the American Chemical Society (ACS), and received the ACS 31st Northeast Regional Industrial Innovation Award. He was inducted into the New Jersey Inventor’s Hall of Fame in February 1990. He continues to work full-time to develop catalytic devices for clean and efficient energy.
3.2
Combustion Strategies for Syngas and High-Hydrogen Fuel
3.2-1 Introduction The technical challenges surrounding syngas and hydrogen fuel combustion have been outlined in section 3.1. Given the issues presented there, various options can be considered for combustor design and operation. First, it is critical to define the type of combustion system that will be used. There are two broad categories: diffusion flame combustors, and premixed combustors. These are described below, but before discussing the combustion strategies, it is useful to review how NOx pollutants are formed.
3.2-2 NOx Formation George Richards
National Energy Technology Laboratory 3610 Collins Ferry Rd. P.O. Box 880 email: [email protected] phone: (304) 285-4458
Nate Weiland
National Energy Technology Laboratory P. O. Box 10940 Pittsburgh, PA 15236 email: [email protected] phone: (412)386-4649
Pete Strakey
Energy Systems Dynamics Division National Energy Technology Laboratory 3610 Collins Ferry Rd. P.O. Box 880 Morgantown, WV 26507-0880 phone: (304) 285-4476 email: [email protected]
203
There are several routes to form NOx pollutants and these may be broadly catalogued as thermally-generated, flame-generated, or fuel-bound NOx. Different authors use different names to catalogue these mechanisms and there is still continuing research to understand the most prominent mechanisms at ultra-low NOx conditions. For example, in hydrogen fueled systems, the prominence of H radicals may contribute to NOx in a manner that is different than in systems fueled by natural gas.1 Thermal NOx is formed by oxidation of nitrogen in air and requires sufficient temperature and time to produce NOx. A rule of thumb is that below approximately 1700K, the residence time in typical gas turbine combustors is not long enough to produce significant thermal NOx. Where temperatures higher than 1700K cannot be avoided, it is necessary to limit residence time to control NOx formation, which favors very short combustor designs. Thermal NOx production also increases with the square root of operating pressure, making it more difficult to reduce in higher-pressure aeroderivative gas turbines. As the name implies, flame-generated NOx occurs in the flame front, created on the short time scale associated with primary combustion reactions. There are a variety of chemical mechanisms involved, all linked to intermediate combustion species that exist only in the reaction zone of the flame. It is important to understand that in practical combustors, the reaction zone is just a small portion of the total combustor volume –most of the combustor volume is sized to complete the relatively slow approach to equilibrium products (notably CO to CO2 oxidation). Thus, residence time in the whole combustor does not affect the flame-generated NOx produced – a significantly different behavior compared to thermal NOx. A convincing demonstration of this point was presented by Leonard and Stegmaier2 who studied multiple flame holders, operating conditions, and residence times from 2 to 100 milliseconds, demonstrating that the flame temperature alone (not residence time) determined the NOx production for emissions under 10 ppmv. Fig. 1, is useful to estimate the flame NOx produced at a given flame temperature, accounting for ideal, and “poor” premixing (not carefully defined in note 2). Note that the effect of poor premixing raises the NOx levels by as much as a factor of three. These data were recorded in turbulent flames, where combustion products are mixed with the fresh reactants right at the flame. It has been suggested that other combustion configurations, without significant stirring between the flame front and products, may reduce the flame generated NOx.3 This may be the basis for NOx reductions reported in the Low-Swirl Combustion section. Finally, fuel-bound NOx is produced by nitrogen species in the fuel reacting with air during combustion. For coal syngas, the most prominent fuel nitrogen species is ammonia, generated during gasification from nitrogen compounds in coal. The ammonia should ideally be removed from the fuel before entering the combustor, or it will be converted to NOx by most combustion strategies. Where this is not possible, rich-lean strategies have the most potential to reduce NOx pollutants. In this approach, combustion is first carried out under fuel-rich conditions, followed by completing combustion under fuel lean conditions. In fuel rich conditions, with sufficient residence times, the ammonia can be reduced to nitrogen and water, rather than
NOx (ppmvd) @15% O2
atmospheric oxygen. A number of studies have been conducted to evaluate rich-lean combustion as an approach to reducing fuel bound NOx. These studies have shown as much as 95% of the fuel ammonia can be reduced to nitrogen and water using rich-lean combustion, with the remaining 5% converting to NOx.4 Untreated syngas ammonia concentrations can exceed 1000ppm, where even 5% conversion would lead to 50ppm NOx, which is well above desired emissions levels. Thus, it is desirable to remove fuel ammonia during gas cleanup, rather than rely on combustion techniques to reduce it to water and nitrogen. 10
Poor Premixing
6 4 2 Ideal Premixing 1 1650
1700
1750
1800
1850
1900
1950
Flame Temperature (K) Fig. 1. NOx emissions, adopted from Leonard and Stegmaier. Source: See note 2.
3.2-3 Diffusion Flame Combustion In this style of combustion, fuel and air are introduced in separate passages, and the flame is stabilized where the fuel and air streams mix. Combustion reactions are typically so fast that fuel and oxidant consumption is limited by transport to the reaction zone (i.e., diffusion), and the reaction proceeds locally at nearly stoichiometric conditions. The Lewis number (Le) describes the ratio of thermal transport to species transport from this reaction zone. Where Le = 1, the temperature in the reaction zone will equal the adiabatic flame temperature because thermal energy diffuses away as fast as the reactants are supplied. The fuel species in hydrocarbon combustion typically have fuel Lewis numbers ( α mix / Dij fuel ) in the range of 0.9 to 1.2, meaning that diffusion flame combustors will have flame temperatures near the adiabatic flame temperature. These temperatures are high enough to oxidize nitrogen in air, producing appreciable NOx pollutants. Hydrogen itself has a fuel Lewis number as low as 0.4, making it even more difficult to reduce NOx because the peak laminar flame temperatures are higher than adiabatic due to differential diffusion effects. The effect of fuel Lewis number on flame temperature has been observed experimentally as well as with direct numerical simulations (DNS).5 Because of their high flame temperatures, diffusion flame combustors require some method to achieve low-NOx performance. An obvious technique is to dilute the fuel, lowering the adiabatic flame temperature. A common diluent is steam, which can both lower the flame temperature, and reduce the production of non-thermal NOx. The hydroxyl radical OH is increased by the presence of additional water, and these radicals favorably scavenge HCN fragments which might otherwise produce NOx. Steam dilution is already used on IGCC applications, but it is not completely desirable. The extra energy that is needed to make steam from water is not recovered in the turbine expansion, penalizing cycle efficiency, (but raising power output from the added mass flow). The additional steam in the exhaust produces a modest increase in the turbine nozzle heat transfer, raising metal temperatures. The protective thermal oxide layers in turbine material sets can be affected by increased moisture levels. Finally, steam consumption by stationary turbines should be minimized to conserve water resources. For these reasons, any further development of diffusion flame combustors for IGCC applications would ideally use nitrogen from the air separation plant, rather than steam. The amount of nitrogen available for flame dilution is established by the engine cycle and the ASU, and it can be shown that for example, hydrogen could be diluted up to about 50% with nitrogen in a typical IGCC configuration. Unfortunately, this level of dilution produces an adiabatic flame temperature around 2025 K, which is still too high for ultra-low NOx performance. Given the dilution limit on adiabatic flame temperature, it is important to consider other methods to reduce the diffusion flame temperature. As noted above, the diffusion flame temperature is set by the ratio of thermal diffusion away from the reaction zone to heat generated by reactants. If the reaction zone is “strained” by fluid shear, it is possible to change the balance between diffusion and reaction in the reaction zone, changing the flame temperature. Strongly sheared flows can locally extinguish the flame, providing opportunity for fuel air mixing before combustion is initiated elsewhere. This raises the possibility that strong shearing could be used to make a diffusion flame combustor behave more like a premixed combustor. The required levels of shearing (known as “stretch” or “strain”) have not been fully characterized. These concepts are discussed in the Highly-Strained Diffusion Flame Combustion section.
204
Pete Strakey, Nate Weiland, Geo Richards 3.2-4 Lean Direct Injection Lean Direct Injection (LDI) combustion was developed as a low NOx alternative to Lean Prevaporized Premixed (LPP) combustion for aircraft gas turbines, where the inherent flashback and dynamic instability concerns of LPP combustion are considered too great of a risk for flight application. In LDI combustors, liquid fuel is directly injected into the combustion chamber, where it is mixed with air in the shortest possible distance. The intent is to provide an essentially lean premixed fuel/air mixture that burns in a low-NOx flame, similar to LPP combustors, which are discussed in the Premixed Combustion section below.6 Low-NOx performance is compromised in an LDI combustor if the fuel and air are not perfectly mixed before combustion occurs, creating regions with higher fuel content that burn hotter and generate more NOx. Similarly, the mixture may burn upstream of the premixed zone in a diffusion flame, with combustion occurring at stoichiometric conditions that result in higher temperatures and NOx production. Nevertheless, flashback and auto-ignition concerns are nearly eliminated in LDI combustors, and they can operate over a wide turndown range with a high degree of static and dynamic stability using a wide range of fuels. The desire to burn high-hydrogen fuels in gas turbines used for power applications raises similar concerns of flashback and instability when operating in the Lean Premixed mode of combustion, so LDI combustors seem to be a natural fit for burning these fuels in a low NOx gas turbine system. To demonstrate the potential of LDI combustors, researchers at NASA Glenn have recently studied various low NOx LDI concepts for pure hydrogen combustion in aircraft gas turbine combustors.7 Five separate injector concepts from different manufacturers were tested at aircraft gas turbine conditions (4.8 – 13.6 atm, Tin = 600 – 1000 °F). At low combustor exit temperatures, it was possible to achieve very small NOx levels (~1 ppmv, wet, uncorrected). NOx emissions were primarily controlled by lowering equivalence ratios to limit combustion temperatures, and no hydrogen dilution cases were considered. One of the tested injectors at NASA Glenn was similar to those used in current IGCC gas turbines that burn syngas, where fuel is injected axially into a swirling airflow. Although this injector was very robust, it produced substantially higher NOx than the other tested injectors. Some of the other tested injectors were similar to those studied recently at GE Energy, where multiple fuel jets were injected at an angle into a central air jet.8 Their results show that more fuel injection ports per air jet reduce NOx emissions due to higher fuel jet momentum and mixing. Increasing the number and decreasing the size of the air jets is shown to reduce NOx by reducing the length of the combustion zone, although this comes at the expense of increased combustor pressure drop. Similar injector configurations studied at NASA Glenn had better NOx emissions, due in part to the shortened combustion zone. However, in some cases, this also led to overheating problems and injector failure, since the combustion zone was located much closer to the fuel and air injectors. Pressure drops in the NASA Glenn injectors were sometimes very large (4-25%). Redesign and optimization for power gas turbines could reduce these pressure drops. In addition, large pressure drops may have been required to reduce the flashback or flameholding potential in those injector designs that operated more in a premixed combustion mode than a diffusion combustion mode. As the injectors were tested on pure hydrogen, dilution with nitrogen will reduce flame speeds and may decrease the necessity for large injector pressure drops and high air velocities to avoid these issues.
3.2-5 Highly-Strained Diffusion Flame Combustion
205
Though not discussed explicitly in the above studies, successful LDI diffusion flame combustors use jets of fuel and air that introduce high strain rates in the combustion zone. In a pure diffusion flame, strain rate can be quantified by measuring or calculating the velocity gradients in the mixing flow field. In regions of high strain and fluid shear, mixing rates and bulk transport rates are faster than chemical reaction rates, thus local reactions are not allowed to go to completion before the flow carries the combustion radicals away from the reaction zone. The net result of this process is a reduction in peak flame temperature of a highly strained flame, which in turn reduces thermal NOx production. It should be pointed out, however, that thermal NOx is not only a function of temperature, but also of flame residence time and O-atom concentration in the reaction zone. Increasing the flame strain also tends to reduce the residence time in the flame, but it also can increase the O-atom concentration in the flame by an order of magnitude. This effect is shown in Figure 2, where intermediate strain rates tend to increase the production rate of NO due to the increased O-atom concentration, while at high strain rates, the reduction in flame temperature overcomes the influence of the O-atom concentration, and NO production rates are reduced. This points to increased strain rates as a possible path to reducing or effectively eliminating thermal NOx in a diluted diffusion flame, where dilution of the fuel alone does not reduce flame temperatures enough to satisfy ultra-low NOx emission goals. Increased strain rates are typically attained by increasing the fuel and/or air jet velocities to increase fluid shear, though at the expense of increased combustor pressure drop. In addition, the static stability of the flame is a strong function of these jet velocities, where too high of a jet velocity could cause the flame to blowout. Thus, flame stability concerns place limits on allowable levels of flame strain, particularly for diluted high-hydrogen content fuels, since flameholding ability is closely linked to the flame speed of the fuel/air mixture, which decreases as more diluent is added to the fuel stream. From this perspective, impinging fuel and air jet injector configurations9 hold an advantage over co-axial jet configurations, as forced mixing of the fuel and air should improve the flameholding abilities of these diffusion flames. Much more study could be done in this area to determine injector configurations that maximize flame strain while minimizing stability and combustor pressure drop concerns. In addition, the effect of strain rate on NOx emission from diffusion flames has only been partially quantified for simple diffusion flames, and there are no such studies in practical LDI-type diffusion flame combustors using hydrogen, syngas, and/or fuel diluents. Other areas requiring further study include the effects of increased flame strain on combustion efficiency and on in-flame NOx production mechanisms.
3.2 Combustion Strategies for Syngas and High-Hydrogen Fuel 0.12 0.1
2000
0.08
Tmax (K)
2500
1500
0.06
wNO
1000
0.04 10 XO
500 0 1
10 100 1000 Strain Rate (1/s)
0.02
3
Tmax
w NO (kg/m s) and X O
3000
0 10000
Fig. 2. Strain rate effects, adapted from Sanders et. al. wNO = NO formation rate, XO = O-atom mole fraction, Tmax = peak temperature Source: Sanders, J. P. H., Chen, J.-Y., and Gokalp, I., “Flamelet-Based Modeling of NO Formation in Turbulent Hydrogen Jet Diffusion Flames,” Combustion And Flame, Vol. 111, pp. 1-15, 1997.
3.2-6 Premixed Combustion As the name implies, premixed combustion is accomplished by mixing the fuel and air upstream of the flame. The fuel-air ratio normalized by the stoichiometric value is known as the equivalence ratio φ, and in many practical premixed turbine combustors, has a value of slightly more than 0.5. Thus, there is approximately ½ the fuel needed to burn all the air, or conversely twice as much air as needed to burn all the fuel. The excess air serves to dilute the combustion and keep the flame temperatures low enough to avoid thermal NOx formation. While the concept of premixed combustion is simple and effective at reducing NOx, it also has drawbacks. The combustor must operate in a very narrow range of equivalence ratio to avoid blowout at (typically) φ < 0.5 , and increasing NOx formation for φ somewhat greater than 0.6. The combustor controls must include some form of staging, since the range of desired exit temperatures usually cannot be achieved with such a small range of φ . For example, if four fuel injectors are used in a combustor, it is possible to reduce the heat input 50% keeping two injectors operating, but turning two off. The difficulty with this approach is that the air flow from inactive injectors can quench the boundary of the flame from operating injectors, raising CO emissions, but this can be addressed with good aerodynamic design. Staging in this manner is used on commercial engines.10 Beyond simply de-activating injectors, staging is also accomplished by operating some injectors at slightly richer equivalence ratios, to improve flame stability. This can also be accomplished using “pilots” on individual injectors. The pilot flame is typically supplied with some air for partial premixing, and the pilot fuel circuit is controlled to achieve stable combustion at the lowest possible NOx emissions, as described in the following section.
3.2-7 Tuning and Combustor Control Balancing the fuel delivery among various fuel circuits to meet operating requirements is known as “tuning” and has become a critical part of both commissioning and operating low-emission gas turbines. Various strategies have been used, or are being developed so that the emissions targets can be met with stable combustion. Because combustion stability is affected by inlet temperature and fuel composition, tuning may need to be adjusted to accommodate ambient environment temperatures and even fuel composition. In addition to controlling the fuel split, for some turbines, tuning may include adjustment of compressor inlet guide vanes or bleeding compressor flow11. This allows an adjustment of the combustor air flow at fixed compressor speed, providing another tuning option even on singlespeed (synchronous) gas turbines. It is important to understand that turbines must be able to contend with requirements for load rejection while low-emission combustors operate near the blowout condition. Without careful development, cutting the fuel during load rejection can lead to flame blowout, requiring (sometimes) unacceptable time to re-light and establish power, or making the engine unable to meet grid requirements. An interesting account of the development of combustor and control system required to meet stringent rejection requirements is given in the references.12 On some engines, fast acting valves are used to enhance lean-blowout performance13 and allow operation right near the limits of stable combustion. A more advanced concept is to modulate the fuel to counteract combustion oscillations, usually called active combustion control. Active control has been studied in many research projects14, but has only been deployed on one test engine15 and on one commercial engine installation16 to date. An important aspect of combustion tuning and control is diagnosing conditions in the combustor so that the control system can respond to maintain stable, low-emission operation. For example, it is possible to improve engine operation by monitoring combustion performance from available engine sensors.17 A number of recent papers have shown the potential of using flame optical signals, acoustic signals, or flame ionization to monitor and control the combustion process.18
206
Pete Strakey, Nate Weiland, Geo Richards 3.2-8 Oxy-Fuel Combustion As noted in section 1.3.1, advanced engine cycles using oxy-fuel combustion have been proposed as a means of capturing CO2 from engine operation. These oxy-fuel cycles require a different approach to combustor design because the combustion is ideally operated at stoichiometric conditions – having just enough oxygen to completely oxidize the fuel. Oxygen is produced from air separation, such that any excess oxygen is produced with an accompanying penalty to the overall cycle efficiency. In addition, after the water is condensed from the exhaust, any excess oxygen should be eliminated from the compressed CO2 to avoid corrosion in handling the CO2 gas. For these reasons, the combustor design must achieve very high combustion efficiency at conditions with little excess oxygen. This requirement places a premium on achieving high levels of mixing uniformity in the combustor, because even modestly unmixed fuel stream will be starved for oxygen. It should be noted that boiler designs also ideally operate near stoichiometric, but typically use 1-3% excess oxygen, and have relatively long residence times to complete fuel oxidation. For the oxy-fuel turbine, the excess oxygen would ideally be lower, with much shorter residence times (~30ms) to avoid excessively large pressurized combustion chambers. Oxyfuel combustion for power cycles has been studied in a number of papers.19 The easiest combustion strategy is to employ a diffusion flame combustor. The stability and simple operation of diffusion flame systems make them appealing for oxy-fuel systems. There is no need to control NOx, since the products are sequestered, and there is otherwise little nitrogen in the combustor. Even without sequestration, the peak flame temperature in diffusion flames can be controlled by the level of diluent added, thereby avoiding NOx formation. Nevertheless, a potential advantage of premixed combustion is that premixing the fuel and oxidant can reduce the unmixed streams of fuel and oxygen that are created in diffusion flame systems where relatively small fuel jets must penetrate and mix in the large combustion volume. There is relatively little fundamental data on premixed oxy-fuel flames diluted by water or CO220 such that proposed designs must include some margin with respect to fundamental issues like flame speed.
3.2-9 Notes __________________________
207
1. Konnov, A.A., Colson, G., De Ruyck, J. (2000). The new Route to Forming NO via NNH, Combustion and Flame, Vol. 121, pp. 548-550. 2. Leonard, G., Stegmaier, J. (1994). Development of an Aeroderivative Gas Turbine Dry Low Emissions Combustion System, ASME J. Eng. For Gas Turbines and Power, Vol. 116, pp. 542 – 546. 3. Sattelmayer, T., Polifke, W., Winkler, D., Dobbeling, K., (1998). NOx-Abatement Potential of Lean-Premixed Gas Turbine Combustors, ASME Journal of Engineering for Gas Turbines and Power, Vol. 120, pp. 48- 59. 4. Fietelberg, A. S., Lacey, M. A., (1997). The GE Rich-Quench-Lean Gas Turbine Combustor ASME 97-GT-127; Hasegawa,T., Sato, M., Ninomiya, T. (1997). “Effect of Pressure On Emission Characteristics In LBG-Fueled 1500CClass Gas Turbine, ASME 97-GT-277; Constant, D. R., Bevan, D. M, Cannon, M. F., Kelsall, G. J. (1997). Development of an LCV Fuel Gas Combustor for an Industrial Gas Turbine ASME 97-GT-38; Folsom, B.A., C.W. Courtney, Heap, M. P. (1980). “The Effects of LBG Composition and Combustor Characteristics on Fuel NOx Formation,” ASME J. Eng. Power, V102, pp459-467; Domeracki, W.F., Dowdy, T. E., Bachovchin, D. M. (1997). Topping Combustor Status for Second-Generation Pressurized Fluidized Bed Cycle Applications, ASME J. Eng. Gas Turbines and Power, Vol. 119, pp. 27 – 33. 5. Takagi, T., Xu, Z. and Komiyama, M., Preferential Dissusion Effects on the Temperature in Usual and Inverse Diffusion Flames, Comb. and Flame 106: 252-260 (1996); Gabriel, R. Navedo, J. E. and Chen R.,, Effects of Fuel Lewis Number on Nitric Oxide Emission of Diluted H2 Turbulent Jet Diffusion Flames, Comb. and Flame 121:525-534 (2000). 6. Tacina, R., Wey, C., Liang, P., and Mansour, A., “A Low NOx Lean-Direct Injection, Multipoint Integrated Module Combustor Concept for Advanced Aircraft Gas Turbines,” Clean Air Conference, Porto, Portugal, NASA/TM-20022111347; Tacina, R. R., Wey, C., Choi, K. J., “Flame Tube NOx Emissions Using a Lean-Direct-Wall-Injection Combustor Concept,” 37th Joint Propulsion Conference and Exhibit, Salt Lake City, Utah, July 8-11, 2001, AIAA-2001-3271. 7. Marek, C. J., Smith, T. D., and Kundu, K., “Low Emission Hydrogen Combustors for Gas Turbines Using Lean Direct Injection,” 41st Joint Propulsion Conference and Exhibit, Tuscon, Arizona, AIAA-2005-3776, July 10-13, 2005; GE Energy, “Premixer Design for High Hydrogen Fuels – Final Report,” DOE Cooperative Agreement No. DE-FC2603NT41893, November, 2005. 8. Ibid. 9. Ibid. 10. Joshi, N. D., Mongia, H. C., Leonard, G., Stegmaier, J. W., Vickers, E. C. (1998). Dry Low Emissions Combustor Development, ASME 98-GT-310; Lefebvre, A.H. (1998). Gas Turbine Combustion, 2nd ed, pp. 349, Taylor and Francis. 11. Sewell, J. B., Sobieski, P. A., (2005). Monitoring of Combustion Instabilities: Calpine’s Experience, in Combustion Instabilities in Gas Turbine Engines, Lieuwen, T. C. , Yang, V. [eds.], American Institute of Astronautics and Aeronautics, pp. 147 – 162.
3.2 Combustion Strategies for Syngas and High-Hydrogen Fuel 12. Myers, G., Tegel, D., Feigl, M., Setzer, F., Bechtel, W., Fitts, D., Couture, B., Tuthill, R. (2003). Dry, Low-Emissions For the ‘H’ Heavy Duty Industrial Gas Turbines: Full-Scale Combustion System Rig Test Results, ASME GT2003-38193; Feigl, M., Setzer, F., Feigl-Varela, R., Myers, G., Sweet, B. (2005). Field Test Validation of the DLN2.5H Combustion System on the 9H Gas Turbine at the Baglan Bay Power Station, ASME GT2005-68843. 13. Mongia, H.C., Held, T. J., Hsiao, G. C., Pandalai, R.P. (2003). Challenges and Progress in Controlling Dynamics in Gas Turbine Combustors. AIAA Journal of Propulsion and Power, Vol. 19, No. 5, pp. 822-829. 14. Cohen, J. H., Rey, N.M., Jacobson, C. A., Anderson, T.J. (1999). Active Control of Combustion Instabilities in a LiquidFueled Low-NOx Combustor. ASME Journal of Engineering for Gas Turbines and Power, Vol. 121, No. 2, pp. 281 - 284; Sattinger, S.S, Neumeier, Y., Nabi, A., Zinn, B. T., Amos, D. J., Darling, D. D. (1998). Subscale Demonstration of the Active Feedback Control of Gas Turbine Combustion Instabilities, ASME Paper 98-GT- 258; Jones, C. M., Lee, J. G., Santavicca, D. A. (1999). “Closed-loop Active Control of Combustion Instabilities Using Subharmonic Secondary Fuel Injection, Journal of Propulsion and Power, Vol. 15, No. 2, pp. 1-7. 15. Richards, G. A., Thornton, J. D., Robey, E. H., Arellano, L (2004). Open-Loop Active Control Of Combustion Dynamics On A Gas Turbine Engine, ASME IMECE2004-59702 16. Seume, J. R., Vortmeyer, N., Krause, W., Hermann, J., Hantschk, C.-C., Zangl, P., Gleis, S., Vortmeyer, D., and Orthmann, A., (1998). Application of Active Combustion Instability Control to a Heavy Duty Gas Turbine. ASME Journal of Engineering for Gas Turbines and Power, Vol. 120, No. 4, pp. 721 -726. 17. Angello, L. C., Castaldini, C. (2004).Combustion Instability Tuning Guidelines: Understanding and Mitigating Dynamic Instabilities in Modern Gas Turbine Combustors, ASME GT2004-54081. 18. Muruganandam, T., Seitzman, J.M. (2003). Optical Sensing of Lean Blowout Precursors in a Premixed Swirl Stabilized Dump Combustor. ASME GT 2003-38104; Lieuwen, T. (2004). Online Combustor Stability Assessement using Dynamic Pressure Data, ASME GT2004-53149; Benson, K., Thornton, J. D., Straub, D. L., Huckaby, E. D., Richards, G. A. (2005). Flame Ionization Sensor Integrated Into a Gas Turbine Fuel Nozzle, ASME Journal of Engineering For Gas Turbines and Power, Vol. 127 pp. 42 - 48 19. Chorpening, B. Richards, G. A., Casleton, K. H., Woike, M., Willis, B., Hoffman, L., (2005). Demonstration of a Reheat Combustor for Power Production with CO2 Sequestration. ASME Journal of Engineering For Gas Turbines and Power, Vol 127, pp. 740 – 747; Richards, G. A., Casleton, K. H., Chorpening, B. T., (2005). –CO2 and H2O Diluted Oxy-Fuel Combustion for Zero-Emission Power, Proc. IMecheE, Vol 219, Part A, J. Power and Energy, pp. 121 – 126. 20. Lewis, B., von Elbe, G., (1987). Combustion, Flames, and Explosions of Gases, 3rd ed. , ppAcademic Press; Koroll, G. W., Mulpuru, S. R., (1986). The Effect of Dilution with Steam and the Burning Velocity and Structure of Premixed Hydorgen Flames, The Twenty First Symposium (international) On Combustion, The Combustion Institute, pp. 1811-1819.
208
BIOGRAPHY
3.2 Combustion Strategies for Syngas and High-Hydrogen Fuel
George Richards National Energy Technology Laboratory 3610 Collins Ferry Rd. P.O. Box 880 email: [email protected] phone: (304) 285-4458
Geo Richards received his Ph.D. in mechanical engineering from Purdue University on the subject of gas turbine combustion. Since coming to the National Energy Technology Laboratory in 1988, he has conducted research on various topics in thermal science and energy production, with a particular emphasis on combustion dynamics. He currently leads the Energy System Dynamics Focus Area, providing technical direction for research groups investigating turbine combustion, carbon dioxide capture, high-temperature fuel cells, fuel processing, and stationary reciprocating engines. In addition to conducting his own research, Dr. Richards’ responsibilities include developing and executing cooperative research agreements with private industry and academia, and evaluating proposed concepts related to energy conversion. He also serves as a research advisor for both graduate and post-graduate investigators visiting NETL from academic institutions.
Nate Weiland National Energy Technology Laboratory P. O. Box 10940 Pittsburgh, PA 15236 email: [email protected] phone: (412) 386-4649
Nate Weiland graduated with a Bachelor’s Degree in Mechanical Engineering from Purdue University in 1997, received his Master’s Degree in Mechanical Engineering from Georgia Tech in 2000, and completed his PhD in thermoacoustics at Georgia Tech in 2004. He is currently an ORISE Post-doctoral Research Fellow at the National Energy Technology Laboratory, where he is investigating various gas turbine combustor concepts burning dilite diffusion hydrogen flames. His reserach interests include experimental, computational and theoretical studies of the interactions between acoustic, thermal, and chemical processes, and the development of novel devices utilizing these interactions.
Pete Strakey Energy Systems Dynamics Division National Energy Technology Laboratory 3610 Collins Ferry Rd. P.O. Box 880 Morgantown, WV 26507-0880 phone: (304) 285-4476 email: [email protected]
Pete Strakey received his Ph.D. in 1995 from the Pennsylvania State Univeristy in the field of Mechanical Engineering. The emphasis of his research is fluid dynamics, combustion and laser diagnostics. He spent 9 years at the Air Force Research Laboratory, Edwards AFB, CA working in the field of rocket propulsion, specifically high-pressure liquid rocket injector atomization and mixing. Since coming to NETL, he has been primarily involved in research on gas turbine combustion and the application of laser diagnostic techniques to combustion systems. He has also been involved in computational fluid dynamics (CFD) modeling of combustion systems as well as model validation. He has authored numerous technical papers on liquid atomization, combustion and laser diagnostics.
7XUELQH&RROLQJ'HVLJQ $QDO\VLV
4.1-1 Introduction Turbine cooling methods have provided the ability to increase turbine inlet temperatures above melting temperatures of turbine airfoil components with not only airfoil survivability, but also extending airfoil life. These cooling methods can be broadly classified into internal and external methods. Internal cooling methods include the use of geometric features placed in the flow path of internal channels within the turbine airfoils to promote turbulence, thereby enhancing convective heat transfer coefficients. These geometric features generally include ribs, pin fins, and impingement holes. External cooling methods include the use of film-cooling holes that are placed in the surface of the airfoils with the hole shapes and hole placement being the design issue. Because the flow fields across turbine vanes and blades vary relative to the position on the airfoil, one would expect that the cooling design would vary. Consider that the flow at the airfoil mid span is primarily two-dimensional while the flow at the airfoil edges is clearly influenced by the inner hub and outer casings of the turbine. The flows influencing the inner hub and outer cases often contain vortices that give rise to velocity components that are orthogonal to the primary flow direction. Not only do the cooling schemes vary in these regions, but the methods that are used to analyze these various sections also vary. Because these cooling schemes are relatively complex, the analysis methods employed are not straightforward. Section 4.1 is aimed at providing the reader with methods that are currently used to analyze complex turbine cooling schemes as well as a background for understanding relevant effects on the different cooling methods. Turbine airfoil geometries have also evolved over the years to reduce pressure losses across each stage resulting in three-dimensional airfoil designs. Section 4.2 provides the reader with an understanding of three dimensional airfoil geometries. - Karen Thole
%,2*5$3+< ,QWURGXFWLRQ $LUIRLO(QGZDOO+HDW7UDQVIHU
.DUHQ7KROH Professor and Department Head 0HFKDQLFDO(QJ'HSWPenn State Univ. University ParkPA16802-1412 SKRQH(814) 865-2519 HPDLO[email protected]
'U.DUHQ7KROHis the Department Head of0HFKDQLFDO(QJLQHHULQJDWPenn State University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¶VDUHDVRIH[SHUWLVHDUHKHDWWUDQVIHUDQGÀXLGPHFKDQLFVVSHFLDOL]LQJLQXQGHUVWDQGLQJKLJK IUHHVWUHDPWXUEXODQFHHIIHFWV2YHUWKHSDVWIHZ\HDUVVKHKDVGHYHORSHGDQXPEHURIXQLTXH WHVWLQJIDFLOLWLHVGLUHFWHGWRZDUGVJDVWXUELQHKHDWWUDQVIHULVVXHVLQFOXGLQJDFRPEXVWRUVLPXODWRU WKDWVLPXODWHVWKHÀRZ¿HOGHIIHFWVUHOHYDQWWRWKRVHHQWHULQJWKHWXUELQHVHFWLRQRIDQHQJLQH'U 7KROHKDVEHHQUHVSRQVLEOHIRUDWWUDFWLQJUHVHDUFKIXQGLQJDPRXQWLQJWRRYHUPLOOLRQIURPVXFK DJHQFLHVDWWKH'HSDUWPHQWRI(QHUJ\86$LU)RUFH3UDWW :KLWQH\0RGLQH0DQXIDFWXULQJ 6LHPHQV:HVWLQJKRXVHDQGWKH1DWLRQDO6FLHQFH)RXQGDWLRQ6KHKDVSXEOLVKHGRYHUSHHU UHYLHZHGSDSHUVZLWKDQXPEHURIWKHVHSUHVHQWDWLRQVJLYHQWRLQWHUQDWLRQDODXGLHQFHV
4.2.1
Cooling Design Analysis
4.2.1-1 Introduction The technology of cooling gas turbine components via internal convective flows of single-phase gases has developed over the years from simple smooth cooling passages to very complex geometries involving many differing surfaces, architectures, and fluid-surface interactions. The fundamental aim of this technology area is to obtain the highest overall cooling effectiveness with the lowest possible penalty on the thermodynamic cycle performance. As a thermodynamic Brayton cycle, the efficiency of the gas turbine engine can be raised substantially by increasing the firing temperature of the turbine. Modern gas turbine systems are fired at temperatures in excess of the material melting temperature limits. This is made possible by utilization of thermal barrier coating materials and by the aggressive cooling of the hot gas path (HGP) components using a portion of the compressor discharge air, as depicted in the aero-engine schematic of figure 1. The use of 20 to 30% of this compressed air to cool the high-pressure turbine (HPT) presents a severe penalty on the thermodynamic efficiency unless the firing temperature is sufficiently high for the gains to outweigh the losses. In all properly operating cooled turbine systems, the efficiency gain is significant enough to justify the added complexity and cost of the cooling technologies employed.
HP TURBINE VANE
COMPRESSOR DISCHARGE
COMBUSTION ZONE
HP TURBINE BLADE
Fig. 1. Aero-engine High Pressure Turbine and Combustor
Ron S. Bunker GE Global Research One Research Circle, K-1 ES-104 Niskayuna, NY 12309 phone: (518) 387-5086 email: [email protected]
295
Cooling technology, as applied to gas turbine components such as the highpressure turbine vanes and blades (also known as nozzles and buckets), is composed of five main elements: (1) internal convective cooling, (2) external surface film cooling, (3) materials selection, (4) thermal-mechanical design, and (5) selection and/or pre-treatment of the coolant fluid. Internal convective cooling is the art of directing coolant via the available pressure gradients into all regions of the component requiring cooling, while augmenting the heat transfer coefficients as necessary to obtain distributed and reasonably uniform thermal conditions. The enhancement of internal convective flow surfaces for the augmentation of heat transfer has occurred through a myriad of surface treatments and features as well as the forceful direction of flows via diverters, swirl devices, etc. The most common turbine airfoil interior surface features have been rib-rougheners or turbulators, and also pin-banks or pinfins, which continue to play a large role in today’s turbine cooling designs. Film cooling is the practice of bleeding internal cooling flows onto the exterior skin of the components to provide a heat flux reducing cooling layer. Film cooling is intimately tied to the internal cooling technique used in that the local internal flow details will influence the flow characteristics and temperature of the film jets injected
on the surface. Materials most commonly employed in cooled parts include high-temperature, high-strength nickel or cobalt-based superalloys coated with yttria-stabilized zirconia oxide ceramics (thermal barrier coating, TBC). The protective ceramic coatings are currently used to actively enhance the cooling capability of the internal convection mechanisms. The thermal-mechanical design of the components must integrate these first three elements into a package that has acceptable thermal stresses, coating strains, oxidation limits, creep-rupture properties, and aero-mechanical response. Under the majority of practical system constraints, this allows for the highest achievable internal convective heat transfer coefficients with the lowest achievable frictional coefficient or pressure loss. In some circumstances, pressure loss is not a concern and the highest available heat transfer enhancements are sought for cooling, while in other applications pressure loss may be so restricted as to dictate a very limited means of heat transfer enhancement. The last cooling design element concerns the correct selection of the cooling fluid to perform the required function with the least impact on the cycle efficiency. This usually is achieved through the use of compressor bleed air from the most advantageous stage of the compressor, but can also be done using off-board cooling sources such as closed-circuit steam or air, as well as intra-cycle and inter-cycle heat exchangers. In many respects, the evolution of gas turbine internal cooling technologies began in parallel with heat exchanger and fluid processing techniques, “simply” packaged into the constrained designs required of turbine airfoils (ie. aerodynamics, mechanical strength, vibrational response, etc.). Turbine airfoils are after all merely highly specialized and complex heat exchangers that release the cold side fluid in a controlled fashion to maximize work extraction. Actively or passively cooled regions of the hot gas path in both aircraft engine and power generating gas turbines include the stationary vanes or nozzles, the rotating blades or buckets of the HPT stages, the shrouds bounding the rotating blades, and the combustor liners and flame holding segments. Also included are the secondary flow circuits of the turbine wheelspaces and the outer casings that serve as both cooling and positive purge flows. The ever present constraints common to all components and systems include but are not limited to pressure losses, material temperatures, component stresses, geometry and volume, aerodynamics, fouling, and coolant conditions. An overview of the cooling design analysis system or method is presented in the generic summary diagram of figure 2. For the present purpose, the design analysis method is shown as a three level system, working from Level 1 outwards. Level 1 concerns the conceptual design of the components largely based on nominal target conditions and divorced from the surrounding systems constraints and competing requirements or trade-offs. Level 1 analysis can be performed based on 1D, 2D, or 3D complexities and details, and is primarily used to compare various options in design. Analysis at the conceptual level must still be detailed enough however to allow ranking and down-selection between options. Level 2 cooling analysis is the much more detailed inclusion of surrounding effects and constraints from aerodynamics, material properties, mechanical loads, lifing limitations, clearances etc. as depicted in the design cycle diagram of figure 3. The analyses performed in Level 2 often must be combined thermal-mechanical predictions using very detailed finite element models, sometimes even sub-models of certain component sections. Most Level 2 analyses are performed at one steady-state operating condition, e.g. 100% load. The result of Level 2 analysis, after various alterations and iterations, is the basic system design with balanced choices that satisfy the engine design goals. Level 3 analysis brings in the operational transient aspects to determine if requirements or constraints are violated under conditions such as normal start-up, fast start-up, trips, and hot restarts. Level 3 results can require that additional changes be made with new analyses at Levels 1 and 2. In all cooling system design analysis levels, engine experience design factors and known engine degradation factors must be included. As examples, such factors may include the use of –3σ material properties, knock-down factors on cooling augmentation, and loss of coatings or metal thickness. In addition, there is a Level 0 analysis not shown in figure 2. Level 0 is the preliminary design of the engine. The preliminary design deals mainly with the mission requirements, such as efficiency, cost-of-electricity, power sizing and number of starts. Level 0 sets the target goals on the cooling system, including the coolant consumption, turbine airfoil life, and inspection intervals.
Operational Transients Level 3
Level 2 Aero Work Combustor Profile Max T Oxidation LCF Creep HCF
Engine Experience Design Factors
Engine Degradation Max Flow
External Heat Transfer Coefficients
Internal Heat Transfer Coefficients
External Gas T Distribution
Coolant T
Adiabatic Film Effectiveness Emissivity
Cooling Design Analysis
Discharge Coefficients
Material
Wall Thicknesses
Leakages
Thermal Conductivities
Assembly
Level 1 Mass Balance Flow Rates Energy Balance Wall T’s
Fig. 2. Cooling Design Analysis System
Aero Design • Flowpath • Airfoils
Bearing Thrust Clearances
Repair Cost
Cycle • T41 • % Wc • SFC
Cost
Materials • Base metal • Coatings • Composites
Life • Mission mix • LCF
Repairability
Commonality
Servicing
Manufacturing
Durability
Thermal Design • Bulk Temp • Max. T • % Wc
AeroMechanical • HCF • Vibration
Mech. Design • Stresses • Creep
Fig. 3. Turbine Engine Design Cycle
296
Ron S. Bunker 4.2.1-2 Level 0 – Preliminary Cooling Design Analysis At this very early stage in the definition of an engine design, the cooling system is completely wrapped up in a single set of performance characteristic curves, usually presented in graphical format, known as Cooling Technology Maps. A generic cooling technology performance chart is shown in figure 4 for a turbine airfoil, either a vane or a blade. The technology curves shown on this chart present the gross airfoil cooling effectiveness versus a heat loading parameter, defined as non-dimensional quantities: Gross Cooling Effectiveness = (Tgas – Tbulk metal) / (Tgas – Tcoolant supply) Heat Loading Parameter = (mcoolant * Cp coolant) / 2 * Hgas * Agas The quantities in these terms are as follows: = average hot gas temperature (e.g. firing temperature for blade) = average metal temperature of entire airfoil with endwalls = temperature of coolant entering the airfoil = average external gas heat transfer coefficient (corrected for radiation) = external gas wetted surface area = coolant flow rate to airfoil = coolant specific heat.
100%
tion+ Convec +TB C oling Film c o
Convec
Convec ti
297
m c ooli
ng
n on c ooli
g
0 Heat Loading Parameter
2D 3D X
Suctio n side
s id e
Fig. 4. Cooling Technology Performance Chart
4.2.1-3 Level 1 – Conceptual Cooling Design Analysis Component design may take on one of several depths of analysis, from preliminary estimates, to detailed two-dimensional analyses, to complete three-dimensional computational predictions including the conjugate effects of the convective and radiative environments. Each mode of analysis has its use as the design progresses from concept to reality. Figure 5 shows a three-dimensional vane airfoil and endwalls reduced first to a two-dimensional, constant thickness cross-section of the aerodynamic shape, and then again to a one-dimensional basic
tion+Fil
Pre ssu re
The heat loading parameter ratios the overall hot gas heat flux (source) delivered to the component against the overall coolant capability to accept heat flux (sink). Since the gas and coolant temperatures are not in this term, the ratio is not unity, but does provide a relative scale for placement of past and current designs. The symbolic points on the chart represent various engine experience data points for different designs. Several curves will generally be present showing major levels of cooling technology. Such maps may also present extrapolated design points based on analysis only, or target design points for new engines. In this preliminary Level 0 design phase, cooling analysis is simply a matter of looking up the expected or projected coolant flow rates based on the cycle or mission design goals. Temperatures may be altered by various choices of cycles, surface areas by overall power requirements or aerodynamics, coolant specific heat by selection of cooling fluid, airfoil temperatures by cooling mass flow rate, and so forth. All of which lead to differing impacts on overall engine efficiency, emissions, life, and cost. A similar set of performance curves may be used to examine the effect of wheelspace and casing leakage flows from the secondary cooling circuits. Here, variations may be made in the complexity of seals to obtain lower overall leakages flows with potential consequences such as higher rotor rim material temperatures.
Gross Cooling Efficiency
Tgas Tbulk metal Tcoolant supply Hgas Agas mcoolant Cp coolant
X=0
X
Hot Air
X Coolant Air
1D
Flat plate external flow Internal cooling flow
Fig. 5. Simplified to Complex Cooling Design Analysis
4.2.1 Cooling Design Analysis flat plate representing flow from the leading edge stagnation point to the trailing edge. Preliminary design uses mostly bulk quantities and one-dimensional simplified equations to arrive at approximate yet meaningful estimates of temperatures and flow requirements. While the actual airfoil / endwall shape involves many complexities of accelerating and decelerating flows, secondary flows, and discrete film injection holes, a good estimate may still be obtained using fundamental flat plate relations. Two-dimensional design incorporates boundary layer analyses, network flow and energy balances, and some thermal gradient estimates to refine the results for local temperature and flow predictions suitable for use in finite element stress modeling. Three-dimensional design may use complete computational fluid dynamics and heat transfer modeling of the internal and external flow fields to obtain the most detailed predictions of local thermal effects and flow losses. Design analyses may of course also mix these methods, such as the use of CFD to predict the hot gas path pressures, velocities, and temperatures for the aerodynamic profile only, while the internal cooling and film cooling are predicted using semi-empirical correlations. One-Dimensional Analysis – Preliminary Design The simplest one-dimensional analysis may be best understood as an iterative sequence of several steps leading to an overall model that is approximately optimized for material thicknesses, cooling configuration, and cooling flow. Figure 6 shows the one-dimensional thermal model that applies to any discrete location on the airfoil. These steps include the following: 1.
Estimation of the external heat transfer coefficient distribution on the airfoil, which may include effects such as surface roughness and freestream turbulence. This estimate may include thermal radiative heat flux separately, or as part of an effective convective heat transfer coefficient;
2.
Calculation of the average adiabatic wall temperature due to film cooling;
3.
Calculation of the conductive material thermal resistances, e.g. TBC, bondcoat, and substrate.
4.
Estimation of the internal heat transfer coefficients due to cooling;
5.
Calculation of the required aggregate cooling flow rate; and
6.
Iteration of the solution to achieve target metal temperatures, thermal gradients, material thicknesses, etc., or to comply with target constraints.
Gas T(x,y,z,t) Film exit T
Metal T
“H” rad
TBC-Metal T
TBC T
The solution is iterative to account for fluid property changes with temperature, both internal and external to the airfoil, as well as temperature rise in the cooling fluid.
Adiabatic wall T
Coolant T(x,y,z,t) Hgas
haw
(k, t)TBC
(k, t)Metal
Hcoolant
Q
Fig. 6. General Thermal Resistance Model
Two-Dimensional Analysis – Conceptual Design The simple one-dimensional model is not of much use in conceptual design unless it is knit into a sectional or complete model representing the cooled airfoil. This means applying the simple analysis to many regions of the airfoil (wall elements) making up a 2D sectional view as depicted in figure 5. This is analogous to a finite element model construction, and in many cases can be achieved using a FEM approach. The elements can be disconnected from thermal conduction as a first estimate, or simply connected to include axial conduction effects within the airfoil section. Such conduction effects are more important in regions that are not well modeled by a single wall thickness, like the trailing edge. Taking this a step further, many radial sections of the airfoil may be stacked to form a pseudo-3D model of the nearly complete component (without endwalls, tip, or shank). Again, this can be accomplished with or without complete thermal conduction connections. These are each valid conceptual design modeling approaches with varying levels of accuracy. Note that such approaches do not typically integrate the airfoil and its endwalls, but treat these portions separately by similar analytical means.
298
Ron S. Bunker
•
• • • • •
Trailing Edge
Tip section
Fig. 7. Cooled HPT Blade (Bucket)
Typical internal cooling technologies including turbulators, pin-fins, turns, impingement jets, trailing edge holes, swirl cooling, vortex cooling, convoluted passages, tip purge holes, and basic number and sizing of passages must be readily (i.e. easily) manipulated to investigate design options and their effects on performance. Manipulation includes movement to new locations, change of size, change of number or spacing, addition to and subtraction from the component. Performance evaluation usually refers to cooling effectiveness and aerodynamic mixing losses at this stage of analysis. Film cooling holes and rows of holes need to be readily moved or altered in the design, including film hole angles and shaping. Rotational cooling circuit differences must be evaluated by altering the general passage layouts. Balancing of flow rates with coolant temperature rises and pressure losses must be performed readily. Changes in the external heat transfer coefficient distributions due to new estimates of freestream turbulence, surface roughness, film injection heat transfer coefficient augmentation, wakes / unsteadiness, hot-streaks / clocking, profile and pattern factors must be accommodated. Wall thickness and TBC coating thickness may also be changed in design at virtually any location.
These factors and more dictate that complex FEM and CFD analyses of cooled airfoils at the conceptual design phase are simply not practical. As an example, figure 8 shows three more blade cooling designs, none of which would be easily obtained from some initial generic design if FEM or CFD were used for every change and alteration being investigated1. In addition to these design manipulation requirements, the majority of current knowledge concerning internal cooling and film cooling is still contained in empirical and semi-empirical correlations. State-of-theart computational predictions are as yet not sufficiently advanced to provide prime reliant “data” for the design of cooled airfoils. As such, conceptual design methods must make use of a multitude of design correlations based on experimental data obtained by the original equipment manufacturers (OEM’s) and/or contained in open literature. Putting the foregoing discussion into practice, the two-dimensional or pseudo three-dimensional cooling analysis of the airfoil portion for a vane or blade is typically performed in the following manner.
299
Tip
Leading Edge
One may ask why a FEM approach is not always employed for the conceptual design of cooled airfoils, and also why the airfoil and endwalls are not always integrated into a single component. The answer is the same for both questions, and lies partially in historical design methods and partially in the state-of-the-art computational analysis. Looking at the turbine blade of figure 7, a candidate cooling circuit design can be very complex. In this example, the main portion of the blade is cooled using a turbulated fivepass serpentine circuit, the leading edge is cooled using a radial passage impinging through crossover holes into the concave stagnation region, and the trailing edge is cooled with a radial pin-bank array and aft ejection channels. Film cooling is employed heavily in the leading edge region and tip, with additional rows of film holes on both the pressure and suction sides of the blade. The blade has three distinct cooling circuits isolated in the shank cooling supply. This blade design, and for that matter any other, must be analyzed and modified with the following in mind:
US patent 4,753,575
US patent 4,753,575
US patent 5,931,638
Fig. 8. Diversity of Cooled Designs
4.2.1 Cooling Design Analysis •
Given a current prediction of the aerodynamics (static pressure distributions) and gas temperatures surrounding the airfoil sections, the external heat transfer coefficient distribution is calculated for each radial section using either boundary layer analysis or computational heat transfer. The distributions must account for some or all of the influencing factors including: Airfoil loading Subsonic boundary layer laminar and turbulent transitions Bypass transition Transonic shocks Surface roughness distribution Freestream turbulence Freestream approach swirl Rotational effects Boundary layer disturbances due to film coolant injections Boundary layer disturbances due to coating spallations Periodic unsteadiness and wake passing Secondary flow injections in hub and tip regions Radiative heat flux distributions
•
A detailed flow network model of the internal cooling circuits of the airfoil is built using the current known coolant supply pressure and temperature, and the external airfoil static pressure distribution as boundary conditions. The network flow model should allow compressible flow effects, though some models may be sufficient with incompressible flow only. The flow model is executed with an initial solution guess and iterated to convergence based upon the current boundary conditions and internal geometry. This cooling circuit model includes: Flow area distributions for each passage Detailed local geometry for each internal feature or repeating feature, such as turbulators, pin-fins, etc. Cooling passage aspect ratio distributions Impingement cooling geometry definitions locally Geometry details for all internal cooling holes Film cooling extraction locations Convective heat transfer coefficient correlations Coefficient of friction correlations Coefficient of pressure loss correlations for turns, holes, etc. Cooling fluid properties
•
The flow network model can also be arranged to contain a material shell representing the airfoil, such that the model may interact with the external conditions via thermal exchange. This merely requires definition of Wall thickness distributions for each section Internal dividing rib thickness distributions Protective coating thickness distributions Material property tables
•
In addition, the flow model is extended to include the flow and discharge of all film cooling holes. This is done by providing information on Film hole or film row exit locations Film hole sizes, shaping factors, spacing, and orientations Film hole discharge coefficient correlations Film hole or film row adiabatic effectiveness correlations Film injection mixing loss correlations Film hole internal heat transfer coefficient correlations
300
Ron S. Bunker This total airfoil model can be modified through relatively simple and quick adjustment of the several input distributions and boundary conditions. Execution of the model is straight forward as long as the boundary conditions and geometry parameters are realistic. It must be recognized that such a model contains multiple inlet and exit boundary conditions and parallel flow circuits, of which some flow circuits may be in communication. The complexity of the model must be sufficient to include/resolve all significant pressure losses. The output of the airfoil model can include predictions of all internal heat transfer coefficients, all flow distributions, individual film hole flow rates and mixing losses, total cooling flow rate, the external film temperatures, and of course the local material temperatures. This model can be further coupled to a prediction of the external heat transfer coefficients to update the heat loads for effects of film injection and wall temperature distributions. Once such a model is finalized upon a desired design and result, it may then be exercised to further study manufacturing effects on film hole discharge coefficients and turbulated cooling passages, tolerances for material properties, wall thicknesses, hole diameters, and core shifts, and special considerations for IGCC designs, including surface roughness, TBC spallation, and film hole blockage effects. Cooling Design Analysis Correlations A major consideration in the above cooling analysis is the provision of good correlations for both internal flows and film cooling under conditions representative of engines. These correlations are numerous as the variation of internal cooling geometries and film cooling parameters are vast. Because there are so many possible combinations and variations, design analysis is founded on several basic generic correlations from the open literature, and augmented by many geometry-specific correlations determined by OEM research. The following is a list of the primary correlation sources from open literature:
Impingement jet array heat transfer coefficients (Nusselt numbers) may be obtained from the correlation of Florschuetz et al. for average jet Reynolds numbers typical in engine design2. For square arrays of jets at somewhat lower Re numbers, the graphical data of Kercher and Tabakoff may be used3.
Impingement cooling that involves the use of individual jets, or slot type jets, or other non-standard configurations, may be determined by correlations in the summary paper of Martin4.
Simple fully-developed duct flow turbulent heat transfer may be estimated quite well by the Dittus-Boelter correlation, Nu = .023 Re0.8 Pr0.33 ,or other variants on this correlation that can be found in any modern textbook. Care should be taken to account for the wall-to-fluid temperature ratio.
Most fully-developed turbulated duct flow heat transfer correlations are of the format Nu = C * Ren Prm . The basic correlations for stationary turbulated ducts with transverse or angled rib rougheners can be found in Han et al.5. This research also includes the coefficients of friction.
Rotating passage heat transfer data with and without turbulators is contained in the NASA HOST program data sets6.
Pin bank internal heat transfer and pressure loss correlations are contained in the works by Metzger et al. and Van Fossen7.
Fundamental equations and correlations concerning various cases of idealized slot film cooling, such as might be encountered in various leakage flow paths, are summarized in the review of Goldstein8.
The best source of both adiabatic film effectiveness and heat transfer coefficient augmentation factors due to film injection for round and shaped holes is contained in the recent series of studies from the Institute for Thermal Turbomachinery at the University of Karlsruhe, Germany9. Such data is generally put into a simplified form to describe the centerline or laterally averaged adiabatic effectiveness as a function of distance and mass velocity ratio. Figure 9 shows several correlation formats that have been used.
ηaw = C1 / ( x/Ms )n ηaw = C1 / ( x/Ms + C2 ) ηaw = C1 Re0.2 / ( x/Ms )0.8 ηaw = C1 / { 1 + C2 ( x/Ms)0.8 } where
A broad set of data for discharge coefficients of film cooling holes is available from the research of Hay and Lampard and also from the ITS Karlsruhe group10.
ηaw = ( Trec – Taw ) / ( Trec – Trec coolant )
Aerodynamic film injection mixing losses may be estimated by the use of the method of Hartsel11.
Re = film jet Reynolds number
M = (ρV)coolant / (ρV)gas s = equivalent film row slot width
Fig. 9. Definitions of Adiabatic Film Effectiveness
301
4.2.1 Cooling Design Analysis
Internal Coolant Temperature
BC Uncertainty BC % Impact
Internal Heat Transfer Coefficient
Metal Thermal Properties
TBC Thermal Properties
Adiabatic Film Effectiveness
50 40 30 20 10 0
External Heat Transfer Coefficient
HPT Blade
External Gas Temperature
Other excellent sources of summarized data and correlations exist in the open literature, but it is up to the design team to determine what to use and how to use it in analysis. One such source is the Lecture Series accumulated by the von Karman Institute of Fluid Dynamics, Brussels. Specific lecture series that cover turbine cooling include Dailey et al., , Harasgama et al., and Glezer et al.12. While the above referenced correlations provide a good starting point for the most common methods of cooling, there are dozens of special regions, geometries, and circumstances in cooling design analysis that require case-by-case data. For these cases, the relevant literature is too large to mention here. Most of these cases deal with the so-called “edge” regions of the cooled components, including the endwalls, platforms, airfoil leading and trailing edges, blade tips, interfacial rails, fillets, and any isolated corners. All of these may be treated by the use of similar thermal-flow network models, or integrated into the airfoil model as special regions. Is this level of cooling analysis detail really required? Figure 10 shows the characteristic uncertainties in engine boundary conditions that affect the complete cooling design analysis of a HPT blade. Also shown is the percentage impact of each boundary condition on the final result (these add to 100%). It should be clear that no detail is unimportant here. Also clear is that the accuracy of certain data, such as the adiabatic film cooling effectiveness distribution, is of very high importance.
Fig. 10. Impact of Boundary Condition Uncertainties
Additional Factors Two additional considerations must be incorporated into the cooling design analyses as indicated in figure 2: 1.
Engine experience design factors such as film knockdown, coating of film hole interiors, hole spacings, etc; and,
2.
Engine degradation factors such as combustor gas profile changes, tip erosion, etc.
These factors account for past experience in both test engines and operational engines that cannot be obtained through research and design activities. These adjustments account for the unknown, or at least poorly understood, conversions from laboratory data and predictions to the reality of complex engine conditions. Another way to look at these factors is as “lessons learned”. For the cooling design analyses, experience factors will include film effectiveness realization or knockdown multipliers, film hole diameter reductions due to protective coating applications, minimum allowable hole spacings to avoid hole-to-hole cracking, reduction of internal heat transfer coefficients due to debris collection, typical TBC spallation sizes (if any), surface roughness distribution patterns, and any other generic or design-specific experience gained. Example engine degradation factors will include alterations to the hot gas temperature profiles or magnitudes due to combustor system operation, blade tip erosion, film hole blockages due to deposits, and even modified material properties with exposure at elevated temperatures. These additional factors are typically incorporated into the design process by one of two methods. First, the data from engine experience can be “data matched” to the design prediction to arrive at the required adjustment factors to be used in the design correlations. Second, modifications due to degradation can be carried through the design analysis in a statistical manner to determine magnitudes of change, as well as sensitivity coefficients.
4.2.1-4 Three-Dimensional Analysis The two-dimensional or pseudo three-dimensional analysis described above is very similar to the simple one-dimensional analysis in format, but includes all of the required detail to perform design manipulations and trade-off studies to arrive at a “final” cooled component design. Once this iterative process has produced a design that is sufficiently polished a more precise three-dimensional design analysis can be performed. The three-dimensional analysis primarily adds thermal-mechanical detail through the use of a full, accurate FEM of the component. The FEM is executed using mapped convective boundary conditions of local heat transfer coefficients and fluid temperatures from the 2D model results. The FEM solution presents the complete temperature distributions of the materials. The 3D analysis can also be performed completely through the use of computational modeling, with the prediction of external and internal flow fields and heat transfer coefficients, or coupled with the use of a conjugate model. This method of cooling design analysis must however be in agreement with the conventional design result, since the latter contains a great deal of empirical data and experience factors built in. Sufficient agreement is dictated by the sensitivity of the design to inaccuracies. For example, it is not generally sufficient for the predicted average heat transfer coefficient in a cooling passage to match the average correlation result.
302
Ron S. Bunker 4.2.1-5 Level 2 – Detailed Component and System Cooling Design Analysis Component Analyses As indicated in figure 2, Level 2 cooling design analysis is how and where the results of the Level 1 analysis interface with the other component and system goals and requirements. The Level 2 subjects noted in figure 2 do not comprise an exhaustive list, but do represent the diversity of requirements. These aspects of overall design, manufacturing, and operation apply to all of the cooled hot gas path components and their portions – vanes, blades, endwalls, platforms, shrouds, supports, and dovetails. There is no single cooling design analysis method that can be described here. Level 2 analysis must pass and receive results to/from the other engine design analysis packages in an iterative method until an acceptable total design solution is obtained. This may require many changes to the Level 1 design with subsequent re-analysis. As one example of the requirements and complexity of this process, the HPT blade tip region design interaction is considered, as presented in the review of Bunker13. (Reprinted by permission of the American Institute of Aeronautics and Astronautics, Inc.)
303
In designing blade tips, both cooled and uncooled, for proper operation within the larger turbine system one must consider the following major factors (in no particular order): • Stage and turbine aerodynamic efficiency are greatly affected by the blade tip design in terms of the resulting effective leakage clearance. The effective clearance, which may also be thought of as an effective overall tip discharge coefficient, is determined not only by the tip geometry, but also by the tip aerodynamic distribution, injected cooling flows, tip sealing arrangement, rotational speed, shroud surface treatments, and much more. As a first estimate, each stage can be thought of as having an isolated tip region aerodynamically, but the reality of multistage turbines is that all stages must be designed together to obtain maximum benefit. Another important aspect of the aerodynamic efficiency directly tied to blade tips is the mixing loss associated with the tip leakage flows as they combine with the high momentum suction side passage flow. • Stage thermal efficiency, and then also overall turbine efficiency, is strongly affected by the amount of chargeable cooling air used to maintain blade tip integrity and life. In highly cooled HPT blades, the tip region alone may account for as much as 20% of the total blade cooling flow. • Bulk material temperature limits must be considered for the entire blade structure. While the tip region is generally not subject to the same limitations as the rest of the blade in this respect, the tip design does influence the resulting bulk temperatures of the lower blade sections through the overall cooling design. The tip may also present enough weight to require lower bulk temperatures in the main blade sections to avoid creep rupture issues. • Maximum local material temperatures are typically a major concern for blade tips as these regions are the most difficult to cool. Temperature limits will be placed on the metal substrate, the bond coat, and the thermal barrier coating (TBC) to avoid, for example, excessive oxidation, high coating strains, and melt infiltration of surface deposits, respectively. • Tip sealing methods vary widely, but all methods attempt to reduce the effective tip clearance. The type of sealing arrangement is intimately tied to the other system design aspects. In many ways, the sealing design is the result of which system design parameters are given the most emphasis. • Casing out-of-roundness (ie. non-cylindrical) will be transmitted through the structure response to the hot gas flow path roundness bounding the blade tips. This leads to non-uniform tip gaps around the circumference, and potential tip rubs. • Shroud segment variation, such as bowing, can result from the thermal gradients present in the design, again leading to nonuniform tip gaps either radially and/or axially. • Approaching and leaving disturbances in the flow around blade tips can affect both the aerodynamics and the cooling. Approaching disturbances are most notably associated with the wakes and shocks being shed from the upstream vane row, which to some degree must influence the tip flow and heat transfer by the introduction of unsteady effects. Approaching and leaving disturbances may be encountered in tip designs that involve shroud recesses and axial flow gaps between the stationary shrouds and attached tip shrouds. • Gas temperature profiles are the result of the particular combustion system design, the operational point, and mixing through the subsequent stages. The radial gas temperature profile may have severe impact on the blade tip, both in respect to the temperature field itself and the pressure distribution. Stronger radial flows may bring hotter gases to the blade tip than desired, while gas temperatures may drive strong material thermal gradients and cause lower cooling effectiveness. • Aeromechanics must be considered in the overall blade structural design, and the tip region must be included in this response. • Stresses, both mechanical and thermal, are key in turbine blade survival. Blade tips must typically deal with very high thermal stresses locally. Higher cooling effectiveness in the tip can alleviate thermal stresses, but must be weighed against the cost to the cycle efficiency. As noted earlier, the blade tip design will influence the weight distribution in the entire blade, which must then be dealt with in the allowable stresses, as well as the low cycle fatigue (LCF) and high cycle fatigue (HCF) responses. This effect will also be transmitted into stress requirements for the blade shank, dovetail, rotor disk posts, and the rotor disk. • Operating conditions must be considered at various limiting points in the engine cycle, because these change the gas and coolant flow rates, temperatures, and pressures. A blade tip design focused solely on steady state takeoff conditions may not be well suited for cruise conditions. A balanced or optimized cycle design must be sought. • Transients play a major role in the durability and life of any effective blade tip design. The relative displacements, radial and axial, of the rotor and stator systems during various transients will determine the ultimate steady state operating clearances, as well
4.2.1 Cooling Design Analysis as the potential for detrimental interference. • Durability is desired for both the blade tip and the opposing shroud as a system. In the long term, durability may be associated with oxidation, while in the short term durability is a matter of survival in the face of tip rubs (intentional or unintentional), plugged cooling holes, and thermal stresses. • Materials and material loss must be planned in blade tip design. Blades and blade tips are not automatically designed with the highest temperature capability material, nor the highest strength material. The tip material may be different from the rest of the blade. The compatibility of the tip and shroud materials must be considered, for example, if a highly abrasive shroud should damage a relatively weak tip material. • Cumulative damage of blade tips is typically experienced in certain characteristic locations in each design type. Uniform damage or material loss is not the general rule. The change in tip geometry with characteristic damage and loss will alter the aerodynamics and heat transfer, usually leading to accelerated loss. • Exhaust gas temperature (EGT) is directly and strongly affected by blade tip clearance. Any improvement in effective tip sealing will preserve valuable EGT margin. • Cost of new parts and cost of repair depend on the complexity of tip design. • Blade weight impacts the blade root stresses, LCF life, and blade creep. This is not limited to a simple matter of centrifugal stresses, but can also have severe effects on the overall aerodynamic design, changing the reaction and work of the stage. • Thrust bearing location and bearing housing distortion affect axial motion and disk sag, which in turn are transmitted through the rotor to the blade tip potentially creating larger clearances on one side of the turbine and rubs on the other side. • Rotor and stator systems should be thermally matched to minimize variations in blade tip clearances during transients. Active clearance control systems can aid in this goal by providing fast thermal response of the shroud radial location. • Blade tips are commonly at least partially damaged or worn in the course of operation. The ability or inability to repair blade tips becomes an important factor in lifetime cost. The complexity of a blade tip design impacts the decision to provide more or less cooling to balance the cost of repairs. Unrepairable blade tips result in scrapped blades. While this summary of system design aspects may appear quite detailed and daunting for such a relatively small region of the turbine, there is one requirement that exceeds all others – the blade tip system design must never cause such severe damage as to liberate blades or pieces of blades in operation. As in the other interacting system relationships within the turbine, prior design and operational experience must guide and temper improved designs. Combustor-Turbine System Analysis The turbine has a special relationship with the combustion system. Turbine cooling design analysis is directly influenced by the type of combustor system, the combustor exit conditions, and the change in combustor conditions at various cycle points. The combustion system operation and its design relative to the turbine has potential impact in at least six main respects: • • • • • •
Hot gas temperature profiles Hot streak clocking relative to the turbine Turbulence characteristics Airfoil backflow margins NOx emissions Fuel type
Each different combustor system design has it’s own set of characteristic radial and circumferential gas temperature profiles. The “set” of profiles refers to the fact that the full power radial profile differs from any part-power profile. For example, some systems have annular combustors, some have can-annular combustors, and others have dump combustors. Full annular combustors may be single, dual, or even triple annular systems with respect to the number of fuel nozzle rings present. In such cases, combustor nozzle staging may be used for differing power requirements. Another major difference arises between the low NOx systems of power turbine engines and aero-engines, the former employing very little dilution or film flow injection within the combustors, and the latter utilizing a great deal of dilution and film injection. Most power generation turbines tend toward very flat radial profiles, while aero-engines tend to have more peaked radial profiles that may change peaking location with power condition. Power turbines may also have radial temperature profile changes as operation is changed from diffusion mode to premixed mode. The key for turbine cooling design is to know as much as possible about the combustor system exit conditions for all operating conditions, and to carry this information through to the design for each cycle point. Combustion systems have circumferential gas temperature and pressure profiles as well, due to the discrete nature of virtually all designs with respect to air/fuel injection and flame holding. While radial profiles are caused by the combined effects of fuel nozzles and combustor dilution / cooling flows, circumferential profiles or pattern factors are caused primarily by the number and spacing of the fuel nozzles. Since the turbine inlet vanes are also of a finite number, this leads to the interesting aspect of hot streak clocking. The combustor hot streak may be aligned directly on a vane leading edge, or midway between two vanes. In fact, the hot streaks may be variable around the entire vane ring depending on the relative count of fuel nozzles and turbine inlet vanes. Different unsteady gas conditions may be incident upon the rotating blade row. The center hot streak may pass through the passage with little vane interaction,
304
Ron S. Bunker while the leading edge hot streak may be greatly modified by interaction with the vane and its cooling flows. There are of course immediate consequences for the vane, but this also translates through to the blade. As with the hot streak effects, combustion system turbulence and swirl flow are additional complicating factors. The turbulence intensity levels, distributions, and length scales will not be the same as those generated by the grids used in simplified studies. The combustor exit flow, in addition to temperature profiles, may also contain significant swirl content. These factors may not be entirely washed out by the inlet vane row. Some studies have indicated that combustor exit average turbulence intensity over the entire region is as high as 30%. Of great concern in all gas turbine designs is the attainment of single digit NOx emission levels. The cooling of low-emissions combustor liners is achieved primarily through the use of convective backside heat transfer, with little or no injection of coolant into the hot gas path. Given the high levels of flow required to perform this cooling, the pressure drop allocated to the combustion system is an important factor. A typical combustion system may use up to 7% of the available pressure from the compressor. This cooling system pressure loss is roughly equivalent to 1% in cycle efficiency, a very significant amount. It is therefore of great concern to designers to achieve the greatest possible cooling effectiveness with the lowest possible pressure loss. It is equally important to the design to achieve a greater cooling effectiveness while matching the pressure loss required by the compressor and turbine design. In this respect, lower pressure loss combustion systems can impose higher loading on the turbine inlet nozzle, and can also present problems in meeting backflow margin requirements. Additionally, since lower NOx emissions can be obtained by “stealing” cooling air from the turbine, this puts pressure on turbine cooling design to use less air. Fuel flexibility is another clear objective in power turbines, with the desire to use gas, liquid distillates, various syngases, and even heavy oils. The operation of a turbine on multiple fuels presents multiple scenarios for the cooling design analysis. In most cases, this means analysis for the most severe cases. Future turbine systems may conceive of controllable turbine cooling to accommodate such changes in operation. These several issues concerning the combustor-turbine system all point to the requirement that the cooling design analysis must not only be performed for changing conditions due to the combustor, but in some cases will even lead to vane-to-vane differences in the cooling analysis.
4.2.1-6 Turbine Secondary Cooling Circuit Analyses
• • • • • • • •
305
• • • • • • • • • • •
Shrouds & hangers
Lab seals Nozzle diaphragm
Inboard air supplies
While much attention is given to the cooled airfoils of the turbine, the secondary flow cooling circuits deserve equal scrutiny and diligence to arrive at a total engine design solution. Figure 1 shows the secondary flow circuits typical of an aero-engine HPT, and figure 11 shows an example of the secondary flow regions for a heavy frame turbine. Secondary circuits of the turbine include the following:
Outboard air supplies
Upper wheelspaces
Lower wheelspaces or disk cavities inboard of the hot gas path Lower wheelspaces Supply circuits from the compressor discharge region to the inboard turbine flows Fig. 11. Heavy Frame Turbine Secondary Flow Regions Upper wheelspaces including buffer and trench cavities around the angel wings Supply circuits from the compressor discharge to the outer turbine casing flows Outboard nozzle and shroud cooling air plenums and connections All rotating seals in these areas, e.g. labyrinth and brush seals All stationary seals in these areas, e.g. labyrinth and cloth seals Component interface leakage pathways and their seals, such as nozzle-to-combustor gaps, nozzle-to-nozzle gaps, shroud-toshroud gaps, nozzle-to-shroud gaps, and blade-to-blade gaps (spline seals, C-seals, W-seals, leaf seals, etc.) Rotating orifices Stationary orifices Pre-swirl supply nozzles Inducers and cover plate systems Blade dovetail / shank leakages Bolt leakages Nozzle support leakages Outboard-to-inboard cooling circuits routed through turbine airfoils Nozzle diaphragm chambers Supply flows bled from earlier compressor stages Shroud hanger system flows and leakages
4.2.1 Cooling Design Analysis
The cooling design analysis for the secondary flow systems is performed in much the same way as that of the turbine airfoils, the main difference being that most of these flow circuits do not directly interact with the hot gas path. Because there is no “external” hot gas flow involved, the thermal-fluid design analysis of these regions becomes an elaborate flow network model with thermal boundary conditions at the hardware surfaces. Just as in the turbine airfoil analysis, the secondary flow models may be simple or complex depending on the design stage. Bucket Nozzle Ultimately, some regions will require complex CFD analysis to resolve full details. The primary location where this level of design analysis is required concerns the points at which the secondary flows do interact with the hot gas path. One such region is the forward wheelspace sealing cavities between the turbine inlet nozzle and the first stage blade, as Hot Hotfluid fluidmay mayenter enter depicted in figure 12. A source of cooling air is supplied the theregion regionbetween between from the inboard location and routed through the stationarythe thetwo twoanglewings anglewings and rotating seal cavities, in this case a buffer cavity and then the andmix mixwith withcolder colder wheelspace wheelspacepurge purge trench cavity at the turbine flow path. Aside from this flow fluid. fluid. circuit, there are several other leakage pathways influencing the region, as depicted in figure 13. In addition, the exit of the flow circuit sees a very three dimensional flow that varies in the circumferential direction due to nozzle wakes and blade leading edge effects. Such interaction regions can involve substantial mixing of the cold and hot flows. A more detailed Fig. 12. Purge Flow Circuit for Turbine Wheelspace knowledge or prediction level of the heat transfer coefficients and gas temperatures in these regions is required. Secondary flow design analysis begins with overall, large Bucket Nozzle Hot gas network models representing the compressor discharge and bleeds to the eventual exit flows into the turbine flow path, Slashface leakages Platform accounting for all key flow areas, lengths, restrictions, and leakage discharge coefficients, using approximate thermal boundary conditions for heat transfer. More detailed models are made to examine separate portions of the flow circuits and add greater Shank leakage Nozzle support fidelity to the boundary conditions. Open literature sources may leakages be used for most of the required information concerning flow restrictions, friction coefficients, and discharge coefficients. Some commercial flow network solvers contain correlations for much of this information. Heat transfer boundary conditions can be estimated by simple forced duct flow and Dovetail leakage natural convection correlations in most locations other than the radial disk flow, radial cavity flow, and labyrinth seal regions. Purge cooling flow A good summary of the flow and heat transfer in radial rotating disk and disk cavity systems for various situations is that of Fig. 13. Cooling Flows and Leakages in Buffer Cavity Region Owen and Rogers and the subsequent literature publications of Owen and co-workers14. Labyrinth seal flow and heat transfer data for planar and stepped geometries may be found in the research of the ITS Karlsruhe group, such as that of Waschka et al.15. The thermal condition of the hardware surrounding the secondary flow circuits must be included in the final design analysis. These boundaries cannot in most cases be treated as adiabatic. For example, the bucket dovetails are connected to the wheel in the disk-posts. While the forward and aft surfaces of the dovetail and disk-post are exposed to the secondary flows of the upper wheelspace, the primary cooling flow of the bucket is routed between the bottom of the dovetail and the wheel, and the coolant flows inside the dovetail to the airfoil. This forms an additional network that connects the secondary flow circuit and the coolant circuit of the buckets. This internal cooling of the bucket dovetail and shank will thermally affect the response of the disk-post and wheel. Even the cooling of the bucket airfoils and platforms has an influence on the top portion of the wheel, serving to conduct energy from the hot gas path down into the wheel. This latter effect is usually analyzed by applying lumped or equivalent thermal masses to the top of the wheels or bucket shanks to act as heat sources. Detailed thermal models of the airfoils, supports, and wheels are rarely if ever done in the same model. In a similar manner, the outer shrouds and their hangers must be modeled together to provide the complete prediction of flows and thermal response. Individual wheels may be modeled, or the entire turbine rotor system. In fact, at some detailed design level, the entire turbine rotor must be thermally analyzed as one in order to correctly predict all clearances. Going one step further, the so-called “unit rotor”, which is the combined compressor-turbine-generator rotor must also be analyzed with thermal boundary conditions, albeit with a less detailed application of conditions.
306
Ron S. Bunker 4.2.1-7 Level 3 – Transient Operational Cooling Design Analysis All of the foregoing cooling design analyses are commonly applied to steady-state operating conditions at some well defined point in the cycle deck or operational map of the turbine engine. The reality of turbine operation however is that both slow and fast transients must considered in the design process. Slow transients include normal startup and shutdown, or load following operations, while fast transients include quick starts for peaking power, engine trips, and hot restarts. Within even the normally slow startup procedure for a heavy frame turbine, there are several intermediate operating points and transients, such as turning gear operation, low RPM holding, and ~80% power point warm up. Other transients may include specific operating domains dictated by the combustion system, water washing, and power augmentation (e.g. water injection to post-combustion). Figure 14 shows an approximate transient growth behavior for the turbine rotor and stator during a fast start (< 30 minutes). The transient growth of the rotor is a combination of all portions making up the rotor, with contributions from centrifugal and thermal effects. The transient growth of the stator and casing outboard of the rotor is thermally dominated, occurring at a different rate than the rotor. The cooling design analysis of all transients is performed using a sufficient number of steady-state analyses and their associated boundary conditions. Each steady-state analysis is performed using the Level 2 methods discussed in the pervious section. The boundary conditions of these several operating points, flow rates, pressures, gas temperature profiles, heat transfer coefficients, and film effectiveness, are used to form the anchor points of the transient analysis. Since the number of steady-state analysis points is typically limited, the boundary conditions at several intermediate steps must be obtained by interpolation. As the basic fluid dynamic and thermal domains of the hot gas and cooling flows also change with operating conditions, these interpolations are performed using explicit or ad-hoc rules. The exact nature and definition of these rules are very dependent on the turbine design and operation, and as such are specific to the OEM’s. Transient analyses of individual components, such as the turbine airfoils, follow the same general guidelines. Usually, the concerns associated with these components are not the same as those of the overall turbine stator and rotor systems. Instead, issues with clearances, leakage gaps, binding, and hot gas backflow or ingestion are scrutinized. In addition, transient effects on peak material stress and strain are important, as evidenced by the potential for TBC spallation under severe thermal transients. The transient cooling design analysis for hot gas path components may therefore focus on certain transients, or portions of transients, known to be of greatest concern.
Steady State
Ro to r
RPM
RPM
S ta tor
∆R
Trip
Potential Tip Rub
Disk Thermal Growth
Blade Thermal Growth Blade & Disk Centrifugal Growth 0 Cold Start
5
10
15 Time (min)
20
Fig. 14. Transient Rotor and Stator Growth for Fast Startup. (From Bunker13, reprinted by permission of the American Institute of Aeronautics and Astronautics, Inc.)
307
4.2.1 Cooling Design Analysis 4.2.1-8 Notes ________________________ 1. J. L.Levengood and T. A. Auxier, “Airfoil with Nested Cooling Channels,” U.S. Patent 4,753,575, 1988; D. A. Krause, D. J. Mongillo, F. O. Soechting, and M. F. Zelesky, “Turbomachinery Airfoil with Optimized Heat Transfer”, U.S. Patent 5,931,638, 1999. 2. L. Florschuetz, C. Truman, and D. Metzger, “Streamwise Flow and Heat Transfer Distributions for Jet Array Impingement with Crossflow,” Journal of Heat Transfer 103 (1981): 337-342. 3. D. Kercher, and W. Tabakoff, “Heat Transfer by a Square Array of Round Air Jets Impinging Perpendicular to a Flat Surface Including the Effect of Spent Air,” Journal of Engineering for Power, 92 (1970): 73-82. 4. H. Martin, “Heat and Mass Transfer Between Impinging Gas Jets and Solid Surfaces,” Advances in Heat Transfer 13 (1977): 1-60. 5. J. C. Han, J. S. Park, and C. K. Lei, “Heat Transfer Enhancement in Channels with Turbulence Promoters”, Journal of Engr. for Gas Turbines and Power 107 (1985): 628-635. 6. T. J. Hajek, J. H. Wagner, B. V. Johnson, A. W. Higgins, and G. D. Steuber, “Effects of Rotation on Coolant Passage Heat Transfer,” NASA Contractor Report 4396 (1991). 7. D. E. Metzger, R. A. Berry, and J. P. Bronson, “Developing Heat Transfer in Rectangular Ducts with Staggered Arrays of Short Pin Fins,” Journal of Heat Transfer 104 (1982): 700-706; G. J. VanFossen, “Heat Transfer Coefficients for Staggered Arrays of Short Pin Fins,” Journal of Engineering for Power 104 (1982): 268-274. 8. R. J. Goldstein, “Film Cooling,” Advances in Heat Transfer 7 (1971): 321-379. 9. M. Gritsch, A. Schulz, and S. Wittig, “Adiabatic Wall Effectiveness Measurements of Film-Cooling Holes with Expanded Exits,” Paper 97-GT-164 (IGTI Conference, Orlando, Florida [1997]); M. Gritsch, A. Schulz, and S. Wittig, “Heat Transfer Coefficients Measurements of Film-Cooling Holes with Expanded Exits,” Paper 98-GT-28 (IGTI Conference, Stockholm, Sweden [1998]); C. Saumweber, A. Schulz, and S. Wittig, “Free-Stream Turbulence Effects on Film Cooling with Shaped Holes,” Paper GT-2002-30170 (IGTI Turbo Expo, Amsterdam, Netherlands [2002]); J. Dittmar, A. Schulz, and S. Wittig, “Assessment of Various Film Cooling Configurations Including Shaped and Compound Angle Holes Based on Large Scale Experiments,” Paper GT-2002-30176 (IGTI Turbo Expo, Amsterdam, Netherlands [2002]). 10. N. Hay and D. Lampard, “Discharge Coefficient of Turbine Cooling Holes: A Review,” Journal of Turbomachinery 120 (1998): 314-319; M. Gritsch, A Schulz, and S. Wittig, “Discharge Coefficient Measurements of Film-Cooling Holes with Expanded Exits,” Paper No. 97-GT-165 (IGTI Turbo Expo, Orlando [1997]); M. Gritsch, C. Saumweber, A. Schulz, S. Wittig, and E. Sharp, “Effect of Internal Coolant Crossflow Orientation on the Discharge Coefficient of Shaped Film-Cooling Holes,” Journal of Turbomachinery 122 (2000): 146-152. 11. J. E. Hartsel, “Prediction of Effects of Mass Transfer Cooling on the Blade Row Efficiency of Turbine Airfoils,” AIAA Paper 72-11 (AIAA Aerospace Sciences Meeting, San Diego, California [Jan. 17-19, 1972]). 12. G. M. Dailey, M. Taslim, D. L. Rigby, P. Sagaut, M. Cakan, B. Han, R.J. Goldstein, and J.M. Buchlin, “Aero-Thermal Performance of Internal Cooling Systems in Turbomachines,” Von Karman Institute for Fluid Dynamics Lecture Series VKI-LS 2002-01 (2002); S.P. Harasgama, J.C. Han, S. Dutta, H. Iacovides, G. Rau, J.M. Owen, and M. Wilson, “Heat Transfer and Cooling in Gas Turbines,” Von Karman Institute for Fluid Dynamics Lecture Series VKI-LS 1996-01 (1996); B. Glezer, N. Harvey, C. Camci, R. Bunker, and A. Ameri, “Turbine Blade Tip Design and Tip Clearance Treatment,” Von Karman Institute for Fluid Dynamics Lecture Series VKI-LS 2004-02 (2004). 13. Bunker, R.S., “Axial Turbine Blade Tips: Function, Design, and Durability,” to be published in the AIAA Journal of Propulsion and Power, March-April, 2006 special issue on Turbine Science & Technology. 14. J. M. Owen and R.H. Rogers, Flow and Heat Transfer in Rotating Disc Systems, vol. 1 (Somerset, England: Research Studies Press, 1989); J. M. Owen and R. H. Rogers, Flow and Heat Transfer in Rotating Disc Systems,vol. 2 ( Somerset, England: Research Studies Press, 1989). 15. W. Waschka, S. Wittig, and S. Kim, “Influence of High Rotational Speeds on the Heat Transfer and Discharge Coefficients in Labyrinth Seals,” Paper No. 90-GT-330 (IGTI Turbo Expo Conference, Brussels, Belgium, 1990); W. Waschka, S. Wittig, S. Kim, and T. Scherer, “Heat Transfer and Leakage in High-Speed Rotating Stepped Labyrinth Seals,” AGARD Conference Proceedings 527, Heat Transfer and Cooling in Gas Turbines (1993).
308
BIOGRAPHY
4.2.1 Cooling Design Analysis
Ron S. Bunker GE Global Research One Research Circle, K-1 ES-104 Niskayuna, NY 12309 phone: (518) 387-5086 email: [email protected]
Dr. Bunker is an internationally recognized research engineer in the field of Gas Turbine Heat Transfer. Dr. Bunker received his PhD in Mechanical Engineering from Arizona State University in 1988. After a one-year post-doctoral research fellowship from the Alexander von Humboldt Foundation of Germany, Dr. Bunker joined GE Aircraft Engines in Cincinnati. In 1993, Dr. Bunker joined the GE Global Research Center. He has worked on R&D activities focused on turbine vane and blade internal and external heat transfer. The main thrust of efforts during the most recent years has been new technology development for the Advanced Turbine System “H” power plant. Dr. Bunker is a Fellow of the American Society of Mechanical Engineers and Associate Technical Editor for the Journal of Turbomachinery. Dr. Bunker has been awarded 35 US patents and is the author of 75 technical publications.
4.2.2.1
Airfoil Film Cooling
4.2.2.1-1 Introduction
Fig. 1. Schematic of film cooling configurations on a vane Source: (from http://lttwww.epfl.ch/research/htprojects/filmcool.htm)
Film cooling is a major component of the overall cooling of turbine airfoils. An example of a film cooled turbine vane is shown in figure 11. From the schematic of the airfoil in figure 1, it is evident that there are holes placed in the body of the airfoil to allow coolant to pass from the internal cavity to the external surface. The ejection of coolant gas through holes in the airfoil body results in a layer or “film” of coolant gas flowing along the external surface of the airfoil. Hence the term “film cooling” is used to describe the cooling technique. Since this coolant gas is at a lower temperature than the mainstream, the heat transfer into the airfoil is reduced. The adiabatic film effectiveness has a predominant effect in the design of the overall airfoil cooling. Consequentially, in this section details of film cooling performance are reviewed.
4.2.2.1-2 Fundamentals of Film Cooling Performance
David G. Bogard Mechanical Engineering Department University of Texas at Austin Austin, TX 78712 email: [email protected]
309
The primary process by which film cooling reduces the heat transfer to the wall is by reducing the gas temperature near the wall, i.e. reducing the driving temperature potential for heat transfer to the wall. As the coolant flows from the coolant holes, it mixes with the mainstream gas resulting in an increase in coolant temperature. A typical example of this is presented in figure 2 which shows measurements of the temperature profile along the centerline of a coolant jet as it flows downstream of the coolant hole. In this figure the temperature contours are presented as normalized θ contours where θ is defined as:
θ=
T∞ − T T∞ − Tc
(1)
where T is the local temperature, T∞ is the mainstream temperature, and Tc is the coolant temperature at the exit of the hole. Note that θ = 1 is the normalized initial coolant temperature and θ = 0 is the normalized mainstream temperature. The θ contours in figure 2 show that coolant quickly increases in temperature as it flows downstream. The coolant temperature at the wall will be at the adiabatic wall temperature, Taw, and this temperature is generally assumed to be the driving temperature potential for heat transfer into the wall. Generally a normalized form of Taw, referred to as the adiabatic effectiveness or film effectiveness, is used to characterize the film cooling performance. The film effectiveness, η, is defined as follows:
θ 1.5 1 y/D
0.5 0
-4
-2
0
2
x/D
4
6
8
10
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Fig. 2. Thermal profiles showing the coolant distribution flowing from a film cooling hole.
(2)
Where Tc,exit is the coolant temperature at the coolant hole exit. For perfect film cooling performance, the film effectiveness would have a value of η = 1.0, i.e. Taw would be equal to the coolant temperature at the exit of the hole; while a value of η = 0 would indicate that the film cooling has not reduced the gas temperature at the wall. In practice, η values decrease rapidly downstream of the coolant holes due to the strong turbulent dispersion of the coolant jet. As mentioned above, typically Taw is presumed to be the driving temperature potential for heat transfer into the wall. Consequently, the heat flux into the wall with film cooling, q ′′f , is determined using the heat transfer coefficient with film cooling, hf, defined as follows: (3) To evaluate the performance of the film cooling in reducing the heat flux to the wall, q ′′f should be compared to the local heat flux to the wall that would occur without film cooling, i.e. q0′′ that is determined based on the heat transfer coefficient without film cooling, h0, using the following:
q0′′ = h0 (T∞ − Tw )
(4)
Examining equations (3) and (4), it is apparent that a reduced temperature for Taw relative to T∞ will result in a reduced heat flux to the wall. However, these equations also highlight that there is potentially a difference in heat transfer coefficients for the film cooling case and the no-film cooling case. In fact, the disturbance caused by the injection of coolant often causes an increase in the heat transfer coefficient. This increase in heat transfer coefficient causes an increase in heat transfer to the wall, and hence is detrimental. Consequently the overall performance of the film cooling configuration needs to be evaluated in terms of the a net heat flux reduction which takes into account decreased gas temperature provided by the coolant film and the increased heat transfer coefficient due to the coolant injection process. This net heat flux reduction, ∆qr, is obtained by combining equations (3) and (4) resulting in the following:
(5) which can be rewritten as:
∆qr = 1 −
hf η 1 − h0 φ
(6)
where φ is the non-dimensional metal temperature for the operational turbine airfoil, and is defined as follows:
φ=
T∞ − Tw T∞ − Tc, internal
(7)
where Tc,internal is the coolant temperature inside the internal cooling passages of the turbine airfoil. Note that φ is an unknown that is not generally determined in the laboratory experiment, and a value for φ must be assumed in order to estimate a net heat flux reduction using equation (6). A typical value for operational film cooled turbine airfoils is φ = 0.6, and this value is generally assumed when analyzing laboratory data.
310
David G. Bogard 4.2.2.1-3 Correlations of Film Cooling Performance
z/d
The primary measure of film cooling performance is the film effectiveness, η, since this has a dominating effect on the net heat flux reduction. Furthermore, industrial designers typically will focus on the laterally averaged film effectiveness, η , which is the average η over a line normal to the flow and extending a distance equal to the pitch between holes. Besides the simplification in processing film 56 effectiveness results by using only laterally averaged data, there is a physical rationale for using only the laterally averaged film effectiveness. Recall that η represents the normalized adiabatic wall temperature 54 η which corresponds to the gas temperature adjacent to the surface. As the coolant jet flows downstream of the coolant hole there is a large spatial 1 52 variation of gas temperature near the wall as is evident by the contour 0.9 plots η shown in figure 3. However the large conductivity of the metal 0.8 turbine airfoil causes a much more uniform distribution of the “metal 50 temperature”. Consequently the laterally averaged film effectiveness 0.7 is a reasonable representation of the effect of the coolant jet2, and most 0.6 of the correlations for film effectiveness presented in this section are in 48 0.5 terms of laterally averaged cooling effectiveness. However, it should be 0.4 recognized that for purposes of understanding the physical processes of 4 6 0.3 coolant dispersion, and for validation and improvement of computational predictions, the spatial distribution of η is important information. 0.2 Ideally a film of coolant would be introduced to the surface of an airfoil using a slot angled almost tangential to the surface in order to provide a uniform layer of coolant that remain attached to the surface. However, long slots in the airfoil would seriously reduce the structural strength of the airfoil, and hence are not feasible. Consequently coolant is typically introduced to the airfoil surface using rows of holes. The film cooling performance is dependent on the hole geometry and configuration of the layout of the holes. Furthermore, various factors associated with the coolant and mainstream flows, and the airfoil geometry, also significantly affect the cooling performance. A listing of the various factors influencing film cooling performance is presented in table 13. Considering the many factors listed in table 1, the difficulty in predicting film cooling performance is evident. The effects of these factors are discussed in the following subsections.
0.1
44
0
42
40
38 -36
-34
-32
-30
-28
-26
-24
-22
x/d
Fig. 3. Typical film effectiveness contours.
Film Effectiveness at Varying Blowing Ratios In the following description of film cooling performance, a baseline geometry of cylindrical holes spaced 3d apart and inclined 30º to the surface and aligned in the flow direction is used. A comprehensive study of the film effectiveness for this configuration was done by Baldauf et al. using a flat, smooth surface test facility4. Results for a range of blowing ratios are presented in figure 4. The blowing ratio, M, is the ratio of the coolant mass flux to the mainstream mass flux and is defined as follows:
M =
ρ cU c ρ ∞U ∞
(8)
where ρc and ρ∞ are the coolant and mainstream density, respectively, and Uc and U∞ are the coolant and mainstream velocity, respectively. Figure 4 shows that the level of η increases systematically with an increase in M until M = 0.6, but for M ≥ 0.85, the peak level of η begins to decrease, and the position of the peak moves downstream. The initial increase in η with increasing M is expected due to the greater mass flow of coolant. The decrease in η for M ≥ 0.85 is due to the coolant jet separating from the surface. This is graphically illustrated in the sequence of thermal profile measurements presented in figure 5 (generated from data from Thole, Sinha, Bogard & Crawford5) showing the non-dimensional temperature along the centerline of a coolant jet exiting a cylindrical coolant hole inclined 35º to the surface. Three blowing rates are presented, but they are identified in terms of the momentum flux ratio I which is defined as follows:
311
4.2.2.1 Airfoil Film Cooling Table 1 Factors Affecting Film Cooling Performance
Coolant/Mainstream Conditions
Hole Geometry and Configuration
Mass flux ratio*
Shape of the hole*
Airfoil Geometry Hole location - leading edge - main body
Injection angle and compound angle of the coolant hole *
Momentum flux ratio*
- blade tip - endwall
Mainstream turbulence*
Spacing between holes, P/d
Surface curvature*
Coolant density ratio
Length of the hole, l/d
Surface roughness*
Approach boundary layer
Spacing between rows of holes and number of rows
Mainstream Mach number Unsteady mainstream flow Rotation * Factors that have a significant effect on predictability of film cooling performance.
a)
b)
c)
Fig. 4. Distributions of η for varying blowing ratios presented as a function of the streamwise distance x/d (reproduced with permission from Journal of Turbomachinery). Source: reproduced from Figure 2(b) in Baldauf et al. (see note 4).
Fig. 5. Thermal profiles showing three states of coolant jets: attached, detached then reattached, and fully detached (reproduced with permission from Hemisphere Publishing Corporation). Source: See note 5.
312
David G. Bogard The three profiles presented in figure 5 represent samples of three states for the coolant jets6: (a) fully attached coolant jets shown in fig. 5a, (b) coolant jets that detached then reattached shown in fig. 5b, and (c) coolant jets that were fully detached shown in fig. 5c. Clearly as the coolant jets begin to detach the coolant temperature at the wall decreases (θ increases) as the core of the coolant jet travels above the surface. The range of momentum flux ratios for each of these flow states was found to be I < 0.4 for fully attached jets, 0.4 < I < 0.8 for detached/reattached jets, and I > 0.8 for fully detached jets for flat surface flows7. Clearly, whether or not the coolant jets are attached strongly affects the cooling performance. To first order, the film effectiveness performance for varying blowing ratios can be scaled using the parameter x/MSe where Se is the “equivalent slot length” with Se = Ahole/P where Ahole is the cross-sectional area of the coolant hole and P is the pitch between holes8. The η distributions for the Bauldauf et al. data shown in figure 4 presented in terms of the x/MSe parameter are shown in figure 69. At first this does not appear to collapse the data; but, if results are considered only for 0.2 < M < 0.85, then there is a good collapse of the η profiles. These measurements were made using coolant with a density ratio of DR = 1.8, and consequently the blowing ratio of M = 0.85 corresponds to a momentum flux ratio of I = 0.4. As will be shown below, coolant jets with I > 0.4 are in blowing regimes where there is detachment of the coolant jets. Consequently, the η performance scales well with x/MSe when the coolant jets are attached, i.e. I ≤ 0.4. For prediction of film effectiveness for higher blowing ratios, Baldauf et al. developed more sophisticated correlation techniques that will not be detailed here10.
Fig. 6. Distributions of η for varying blowing ratios presented as a function of the x/Mse parameter (reproduced with permission from Journal of Turbomachinery). Source: reproduced from Figure 7 (a) in Baldauf et al. (See note 4.)
Film Effectiveness at Density Ratios Typically the coolant to mainstream density ratio for engine conditions is DR ≈ 2, but often experimental measurements of film cooling performance are conducted with density ratios that are much smaller, even with DR ≈ 1. Because of this range of density ratios used in testing, it is valuable to understand how the coolant density ratio affects film cooling performance. When testing with lower density ratios, coolant flows at a given mass flux ratio will have higher velocity and momentum flux ratios. Recall that coolant jet separation is primarily a function of momentum flux ratio, so lower density coolant jets will tend to separate before higher density ratio jets. Consequently the maximum film effectiveness for lower density ratio coolant jets is less than for the higher density ratio jets, but the difference in film effectiveness levels is generally small. For example, Sinha et al., Pederson et al., and Baldauf et al. found that the maximum laterally averaged film effectiveness was nominally 20% higher for coolant DR ≈ 2 compared to DR ≈ 1.2 near the hole (x/d < 20) but was essentially the same farther downstream11. These tests were for smooth, flat surfaces. Tests for a vane leading edge, pressure side and highly curved suction side showed similar film effectiveness for low and high density coolant, but the low density ratio coolant has 10% lower film effectiveness in some cases12. For low momentum flux ratios where coolant jets are fully attached, film effectiveness performance for low density coolant is essentially the same as for high density coolant when compared at the same mass flux (blowing) ratio. However, at higher momentum ratios where the coolant jets begin to detach, I > 0.4, the film effectiveness for low and high density ratio coolant jets are most similar for similar I. However, for showerhead blowing, film effectiveness for low and high density ratio coolant is best matched using M for all blowing ratios13.
313
4.2.2.1 Airfoil Film Cooling Heat Transfer Coefficients The disturbance to the flow caused by coolant injection might be expected to increase heat transfer coefficients downstream of the coolant holes. Generally this is true, but the increase in heat transfer coefficient relative to the no-blowing case is relatively small, less than 5% beyond x/d = 5, for momentum flux ratios of I < 0.314. For higher momentum flux ratios the heat transfer coefficient can be increased by 10% to 20%, but these higher momentum flux ratios are not likely to be used because of poor film effectiveness. Most studies of heat transfer coefficients were done with low density ratio coolant, but results showed that the effects on the heat transfer coefficient were not very sensitive to the density ratio, with the lower density ratio coolant causing a larger increase due to the higher momentum for lower density ratio coolant15.
4.2.2.1-4 Effects of Hole Geometry and Configuration on Film Cooling Performance As described in table 1, there are many hole geometry and configuration variables that affect film cooling performance. Compound angle injection and shaped holes have major effects on film cooling performance and will be discussed in this section. This is a summary of a more comprehensive review of the effects of the varying hole configurations presented in ”Gas Turbine Film Cooling”16. Film Cooling with Compound Angle Holes
0.30
(a) x/D = 3 0.20
Φ = 0° Φ = 90° Smooth surface, Tu� = 0.3%
0.10
0.00
Laterally averaged effectiveness
For the baseline case described above, the coolant holes were angled such that the exiting coolant jets are parallel with the mainstream direction. When the coolant hole is angled to the mainstream direction, this is referred to as “compound angle” injection. Compound angles can be as much as 90º, i.e. normal to the mainstream direction. Coolant injected at a compound angle is quickly turned to the mainstream direction, but will generally have a broader distribution of coolant. Furthermore, the coolant presents a broader profile to the mainstream so that the mainstream has a larger impact on the jet more effectively turning the jet towards the wall. This inhibits jet separation, and results in better film effectiveness for the compound angle holes at higher blowing ratios. Film effectiveness performance for 90º compound angle holes compared to of 0º (streamwise oriented holes), shown in figure 7, illustrates this point. These data are for cylindrical holes spaced 6.5d apart on a smooth flat test surface with low mainstream turbulence levels. Maximum film effectiveness for the 90º compound angle holes was similar to that for the 0º holes and occurred at a similar momentum flux ratio. However the 90º compound angle holes sustained high film effectiveness for very high blowing ratios. For momentum flux ratios greater than I = 1.0, the film effectiveness for the 90º compound angle holes was a factor of 2 to 3 higher than that for the streamwise-oriented holes. Although the film effectiveness for compound angle holes is significantly better than for streamwise-oriented holes at high momentum flux ratios, the net heat flux reduction for compound angle holes is similar to the streamwise- oriented holes17. This is illustrated in figure 8 for 90º compound angle holes. At the higher momentum flux ratio of I = 1.1 the average ∆q r over the 90d distance downstream of the coolant holes was about the same for 90º and 0º compound angle holes. The similarity of the net heat flux reduction even though the film effectiveness is much greater for 90º compound angle holes is due to a greater increase in heat transfer coefficient for these holes compared to streamwise-oriented holes. Even though the average increase in heat transfer coefficient by the compound angle holes was only 10%, this was sufficient to offset the improved film effectiveness. 0.30
0.30
(b) x/D = 25
(c) x/D = 90
0.20
0.20
0.10
0.10
0.00
0.00
0.0
0.5
1.0
1.5
2.0
2.5
Momentum flux ratio, I
Fig. 7. Comparison of streamwise and laterally directed holes in terms of laterally averaged effectiveness as a function of momentum flux ratio for a smooth surface and low free-stream turbulance Source: See note 14 (Schmidt & Bogard).
Film Cooling with Shaped Holes Improved film effectiveness can be achieved if the exit of the hole is expanded so that coolant is slowed through a diffuser. Examples of shapes investigated in the open literature are shown in figure 9. There are two advantages for such a “shaped hole”: the coolant exit velocity is reduced and a broader jet cross-section is presented to the mainstream flow. Both these characteristics will reduce the tendency for the coolant jet to separate. This results in good film effectiveness levels for shaped holes at very high blowing ratios as shown in figure 10. These data were obtained with a row of coolant holes angled 30º with the surface and spaced 4d apart. The spatially averaged film effectiveness, η , was based on a average from x/d = 2 to 22. The blowing ratio for this figure is based on the average
314
David G. Bogard velocity of the coolant at the inlet to the coolant hole, so the mass flow of coolant for the cylindrical and shaped holes are the same for the same M. Film effectiveness for cylindrical holes begins to decrease for M > 0.7 which corresponds to a momentum flux ratio of I > 0.3 given that the density ratio for these tests was DR = 1.7. This decrease is due to separation of the coolant jets. In contrast the film effectiveness for the shaped holes continues to increase for blowing ratios up to M = 2.5 (I = 3.7) showing that the diffusing hole shape is very effective in keeping the coolant jets attached. 0.4
I=0.3, I=0.3, I=1.1, I=1.1,
— ∆ qr
0.3
Φ Φ Φ Φ
= 0° = 90° = 0° = 90°
0.2
0.1
0 0
20
40
x/D
60
80
100
Fig. 8. Comparison of streamwise and laterally directed holes in terms of net heat flux reduction for a smooth surface and high free-stream turbulence. Source: See note 14 (Schmidt & Bogard).
Fig. 9. Schematics of different cooling hole shapes (reproduced with permission from Journal of Turbomachinery). Source: C. Saumweber, A. Schulz, and S. Wittig, “Free-Stream Turbulence Effects on Film Cooling with Shaped Holes,” Journal of Turbomachinery 125 (2003): 65-73.
315
4.2.2.1 Airfoil Film Cooling
0.5
0.4
0.3
�
Cylindrical Hole, Tu=3.6%, L=2.7D Cylindrical Hole, Tu=7.5%, L=2.7D Fan-Shaped Hole, Tu=3.6%, L2.7D Fan-Shaped Hole, Tu=7.5%, L2.7D
0.2
0.1
0 0
0.5
1
1.5
2
2.5
M
Fig. 10. Comparison of spatially averaged cooling effectiveness for cylindrical holes and shaped holes (reproduced with permission from Journal of Turbomachinery). Source: same as for fig. 9.
4.2.2.1-5 Airfoil Surface Effects on Film Cooling Performance Surface curvature and surface roughness are significant factors affecting film cooling performance. Clearly for turbine airfoils strong convex curvature exists around the leading edge and along the suction side of the airfoil. Sometimes strong concave curvature is encountered on the pressure side of the airfoils. Surface roughness varies with the length of operation of the engine; new airfoils are relatively smooth, but after some period of operation the surfaces can become quite rough due to erosion, spalation of thermal barrier coatings, and deposition of contaminants. The following is a brief review of these surface effects. Surface curvature Several studies have shown that surface curvature can significantly change film effectiveness; convex curvature increasing η and concave curvature decreasing η at typical operational blowing ratios18. The effects of varying strengths of curvature are demonstrated in figure 11 in which the laterally averaged film effectiveness, η , at x/d = 40 are presented for a range of curvatures, 46 < 2r/d < 126, with zero pressure gradient (r is the radius of curvature for the surface). These studies indicated that an increased convex curvature (decreasing 2r/d) greatly enhances film effectiveness, while concave curvature decreases film effectiveness except at high momentum flux ratios. These effects of surface curvature can be explained by the wall normal pressure gradients that necessarily exist with wall curvature. When the momentum of the jet tangential to the wall is less than the mainstream momentum the normal pressure gradients drive the coolant jets towards or away from the wall for convex and concave curvature, respectively. For convex curvature, the inward pressure broadens the coolant distribution by pressing the jet to the wall, and keeps the jet attached for higher momentum flux ratios. For concave curvature the opposite occurs, i.e. the coolant jets are pushed away from the wall. Surface Roughness Significant increases in surface roughness during typical operating conditions have been reported by several studies19, with maximum roughness levels as high as as Rek = 500 where Rek is the equivalent sandgrain roughness Reynolds number20. Given that “fully rough” conditions exist when Rek > 70, this roughness level is extremely large. Also, maximum roughness heights were observed to greater than 250 µm, which is 0.5d for typical coolant hole diameters. Surface roughness degrades film cooling performance by increasing the heat transfer coefficient and potentially reducing film effectiveness. Heat transfer coefficients can be increased by as much as 50% to 100%21. Studies of the effects of surface roughness on film effectiveness using flat surface facilities22 showed small reductions (<10%) of average film effectiveness for lower blowing ratios, and small increases for high blowing ratios. However, a study of roughness effects on film effectiveness on the suction side of a vane23 showed surface roughness decreased film effectiveness by as much as 25% at the optimum blowing ratio, but increased film effectiveness as much as 50% at higher blowing ratios. The decrease in film effectiveness at the optimum blowing ratio was primarily due to the roughness upstream of the coolant holes. The upstream roughness doubled the boundary layer thickness and significantly increased turbulence levels which resulted in more separation of the coolant jets and increased dispersion of the coolant.
316
David G. Bogard
η
Fig. 11. Effect of convex and concave curvature on film effectiveness (reproduced with permission from Journal of Turbomachinery). Source: see note 18 (Schwarz, Goldstein, and Eckert).
4.2.2.1-6 Mainstream Effects on Film Cooling Performance There are a number of mainstream factors that can affect film cooling performance including approach boundary layers, turbulence levels, Mach number, unsteadiness, and rotation24. Because of the very high levels of mainstream turbulence exiting the combustor and entering the turbine section, turbulence levels have the largest effect on film cooling performance. Mainstream turbulence levels exiting the combustor can be higher than Tu = 20% and have been found to be nominally isotropic in simulated combustor studies25. Furthermore the integral length scale of the turbulence is large relative to the coolant hole diameters, i.e. Λf/d > 10 (based on Λf values given in Radomsky and Thole26). Primarily due to the acceleration of the mainstream as it passes around the first vane, the local turbulence levels reduce to less than 5% on the suction side of the vane, and to about 10% for much of the pressure side. These are still relatively high turbulence levels, and it is important to recognize the effects on film cooling performance. High mainstream turbulence levels degrade film cooling performance by increasing heat transfer coefficients and generally decreasing film effectiveness. Simulations of the large scale turbulence with levels of Tu = 10% to 17% showed an increase in heat transfer coefficient of 15% to 30%, respectively27. The effects of high mainstream turbulence levels on film effectiveness are shown by the laterally averaged film effectiveness levels for Tu = 0.3%, 10%, and 20% shown in figure 12. Results in figure 12 were obtained using a flat surface test facility with a row of cylindrical holes spaced 6.5d apart, with an injection angle of 30º and aligned with the mainstream direction. Smooth and rough surfaces were tested. The coolant density ratio was DR = 2.0. For a smooth surface with low turbulence levels the optimum momentum flux ratio was I = 0.3. At this momentum flux ratio, a turbulence level of Tu = 17% caused a factor of two decrease in film effectiveness near the hole, and almost a complete loss of cooling for x/d > 25. The optimum momentum flux ratio for high mainstream turbulence conditions was about I = 1.1, substantially higher than would have been expected from low mainstream turbulence tests. At this higher momentum flux ratio the film effectiveness for the high mainstream turbulence case was higher than for the low mainstream turbulence case. This difference was attributed to the higher mainstream turbulence mitigating the effect of coolant jet separation by returning some of the coolant towards the surface with the increased coolant dispersion caused by the higher turbulence levels. These results show the importance of accounting for realistic mainstream turbulence levels when predicting film cooling performance.
317
4.2.2.1 Airfoil Film Cooling 4.2.2.1-7 Airfoil Leading Edge Film Cooling Film cooling of the leading edge of vanes and blades is distinctly different than film cooling of the aft-body of the airfoils because coolant is injected into a stagnation region rather than into a cross-flow. Furthermore, the heat loads are typically much larger along the leading edge, so generally a dense array of coolant holes is used around the leading edge. This array of holes around the leading edge is referred to as the “showerhead” and generally consists of six to eight rows of holes for vanes and three to five rows of holes for blades. Holes are typically aligned radially, i.e. normal to the mainstream direction, with injection angles relative to the surface ranging from 20º to 45º.
a)
b)
c)
Fig. 12. Effect of freestream turbulence level on laterally averaged effectiveness as a function of momentum flux ratio for a smooth surface and low free-stream turbulance Source: D.L. Schmidt and D.G. Bogard, “Effects of Free-Stream Turbulence and Surface Roughness on Film Cooling,” ASME Paper 96-GT-462, 1996.
Fig. 13. Film cooling performance for a simulated blade leading edge with three rows of holes. Mainstream turbulence was Tu = 10%. Stagnation line coolant holes at x/d = 0. Performance in terms of (a) laterally averaged film effectiveness, (b) laterally averaged heat transfer coefficient augmentation, and (c) laterally averaged net heat transfer reduction. Source: See note 2.
318
David G. Bogard Film cooling performance for a simulated blade leading edge is presented in figure 13 in terms of the laterally averaged film effectiveness, η , heat transfer coefficient increase, hf/h0, and net heat flux reduction, ∆qr 28. These data were measured using a simulated blade leading edge with a three-row coolant hole configuration with “laid back” shaped holes oriented radially, an injection angle of 20º, and a spacing between holes of 7.6d. Blowing ratios were based on the approach velocity to the leading edge and ranged from M = 1.0 to 2.5. As shown in figure 13, film effectiveness continues to increase with increasing blowing ratio. Coolant injection caused a 10% to 35% increase in heat transfer coefficients. Finally the net heat flux reduction mirrored the film effectiveness performance. High levels of net heat flux reduction can be attributed to the high levels of film effectiveness.
4.2.2.1-8 Notes _____________________________
319
1. Figure from web site: http://lttwww.epfl.ch/research/htprojects/filmcool.htm 2. B.D. Mouzon, E.J. Terrell, J.E. Albert, and D.G. Bogard, “Net Heat Flux Reduction and Overall Effectiveness for a Turbine Blade Leading Edge,” ASME paper GT2005-69002, 2005. 3. D. G. Bogard and K.A. Thole, “Gas Turbine Film Cooling,” accepted AIAA Journal of Propulsion and Power, 2006. 4. S. Baldauf, M. Scheurlen, A. Schulz, and S. Wittig, “Correlation of Film-Cooling Effectiveness from Thermographic Measurements at Enginelike Conditions,” Journal of Turbomachinery 124 (2002): 686-698. 5. K.A. Thole, A. Sinha, D. G. Bogard, and M. E. Crawford, “Mean Temperature Measurements of Jets with a Crossflow for Gas Turbine Film Cooling Application,” Rotating Machinery Transport Phenomena, J. H. Kim and W. J. Yang, ed. Hemisphere Publishing Corporation, New York, New York, 1992. 6. Ibid. 7. Ibid. 8. R. J. Goldstein, “Film Cooling,” Advances in Heat Transfer 7 (1971): 321-380. 9. See note 4 above. 10. Ibid. 11. A.K. Sinha, D.G. Bogard, and M.E. Crawford, “Film Cooling Effectiveness Downstream of a Single Row of Holes with Variable Density Ratio,” ASME Journal of Turbomachinery 113, no. 3 (1991): 442-449; D.R. Pedersen, E. Eckert, and R. Goldstein, “Film Cooling with Large Density Differences Between the Mainstream and the Secondary Fluid Measured by the Heat-Mass Transfer Analogy,” ASME Journal of Heat Transfer 99 (1977): 620-627; also see note 4 above. 12. Cutbirth, J. and Bogard, D., “Effects of Coolant Density Ratio on Film Cooling,” ASME Gas Turbine Expo, GT2003-38582, Atlanta, Georgia, June, 2003, pp 1-10; M. I. Ethridge, J.M. Cutbirth, and D.G. Bogard, “Scaling of Performance for Varying Density Ratio Coolants on an Airfoil with Strong Curvature and Pressure Gradients,” ASME Journal of Turbomachinery 123, (2001): 231-237. 13. See note 12 above (Cutbirth). 14. V.L. Eriksen and R. Goldstein, “Heat Transfer and Film Cooling Following Injection Through Inclined Circular Tubes,” ASME Journal of Heat Transfer 96, no.1 (1974):. 239-245; Schmidt, D.L. and Bogard, D.G., “Effects of Free-Stream Turbulence and Surface Roughness on Laterally Injected Film Cooling,” Proceedings of the 32nd National Heat Transfer Conference, HTD-Vol. 350, vol. 12, pp. 233-244, 1997. 15. S. Baldauf, M. Scheurlen, A. Schulz, and S. Wittig, “Heat Flux Reduction From Film Cooling and Correlation of Heat Transfer Coefficients from Thermographic Measurements at Enginelike Conditions,” Journal of Turbomachinery 124 (2002): 699-709 16. See note 3 above. 17. B. Sen, D.L. Schmidt, and D.G. Bogard, “Film Cooling with Compound Angle Holes: Heat Transfer,” ASME Journal of Turbomachinery 118, no. 4 (1996): 800-806; also see note 14 (Schmidt). 18. S. Ito, R. Goldstein, and E. Eckert, “Film Cooling of a Gas Turbine Blade,” Journal of Engineering for Power 100 (1978): 476-481; S. Schwarz, R. Goldstein, and E. Eckert, “The Influence of Curvature on Film Cooling Performance,” Journal of Turbomachinery 112 (1990): 472-478. 19. J.P. Bons, R. Taylor, S. McClain, and R.B. Rivir, “The Many Faces of Turbine Surface Roughness,” Journal of Turbomachinery 123 (2001): 739-748; D.G. Bogard, D.L. Schmidt, and M. Tabbita, “Characterization and Laboratory Simulation of Turbine Airfoil Surface Roughness and Associated Heat Transfer,” Journal of Turbomachinery 120 (1998): 337-342. 20. D.G. Bogard, D. Snook, and A. Kohli, “Rough Surface Effects on Film Cooling of the Suction Side Surface of a Turbine Vane,” ASME Paper No. EMECE2003-42061, 2003.
4.2.2.1 Airfoil Film Cooling 21. J.L. Rutledge, D. Robertson, and D.G. Bogard, “Degradation of Film Cooling Performance on a Turbine Vane Suction Side Due to Surface Roughness,” ASME Gas Turbine Expo, GT2005-69045, 2005; also see note 19 (Bogard). 22. R.J. Goldstein, E.R.G. Eckert, H.D.Chiang, and E. Elovic, “Effect of Surface Roughness on Film Cooling Performance,” Journal of Engineering for Gas Turbines and Power 107 (1985): 111-116; D.L. Schmidt, B. Sen, and D.G. Bogard, “Effects of Surface Roughness on Film Cooling,” ASME Paper No. 96-GT-299, 1996. 23. See note 20 and 21. 24. See note 3. 25. R.W. Radomsky and K.A. Thole, “Flowfield Measurements for a Highly Turbulent Flow in a Stator Vane Passage,” Journal of Turbomachinery 122 (2000): 255-262. 26. Ibid. 27. See note 19 (Bogard). 28. J. E. Albert, F. Cunha, and D. G. Bogard, “Adiabatic and Overall Effectiveness for a Film Cooled Blade,” ASME Paper GT2004-53998, 2004.
320
BIOGRAPHY
4.2.2.1 Airfoil Film Cooling
David G. Bogard Mechanical Engineering Department University of Texas at Austin Austin, TX 78712 email: [email protected]
Dr. David Bogard is a Professor of Mechanical Engineering at the University of Texas at Austin, and holds the John E. Kasch Fellow in Engineering. He received his B.S. and M.S. degrees in Mechanical Engineering from Oklahoma State University, and his Ph.D. from Purdue University. He has served on the faculty at the University of Texas since 1982. Dr. Bogard has been active in gas turbine cooling research since 1986, and has published over 100 peer-reviewed papers. He was awarded the ASME Heat Transfer Committee Best Paper Award in 1990 and 2003, and is a fellow of the ASME.
4.2.2.2
Enhanced Internal Cooling of Turbine Blades and Vanes
Je-Chin Han
4.2.2.2-1 Introduction Gas turbines play a vital role in the today’s industrialized society, and as the demands for power increase, the power output and thermal efficiency of gas turbines must also increase. One method of increasing both the power output and thermal efficiency of the engine is to increase the temperature of the gas entering the turbine. In the advanced gas turbines of today, the turbine inlet temperature can be as high as 1500°C; however, this temperature exceeds the melting temperature of the metal airfoils. Therefore, it is imperative that the blades and vanes are cooled, so they can withstand these extreme temperatures. Cooling air around 650°C is extracted from the compressor and passes through the airfoils. With the hot gases and cooling air, the temperature of the blades can be lowered to approximately 1000°C, which is permissible for reliable operation of the engine. It is widely accepted that the life of a turbine blade can be reduced by half if the temperature prediction of the metal blade is off by only 30°C. In order to avoid premature failure, designers must accurately predict the local heat transfer coefficients and local airfoil metal temperatures. By preventing local hot spots, the life of the turbine blades and vanes will increase. However, due to the complex flow around the airfoils it is difficult for designers to accurately predict the metal temperature. Figure 1 shows the heat flux distribution around an inlet guide vane and a rotor blade. At the leading edge of the vane, the heat transfer coefficients are very high, and as the flow splits and travels along the vane, the heat flux decreases. Along the suction side of the vane, the flow transitions from laminar to turbulent, and the heat transfer coefficients increase. As the flow accelerates along the pressure surface, the heat transfer coefficients also increase. The trends are similar for the turbine blade: the heat flux at the leading edge is very high and continues decrease as the flow travels along the blade; on the suction surface, the flow transitions from laminar to turbulent, and the heat flux sharply increases; the heat transfer on the pressure surface increases as the flow accelerates around the blade.
Lesley M. Wright
Turbine Heat Transfer Laboratory Department of Mechanical Engineering Texas A&M University College Station, Texas 77843-3123, USA email: [email protected] Fig. 1. Cross-Sectional View and Heat Flux Distribution of a Cooled Vane and Blade
321
Due to the complex flow, designers need data that will aid them in the development of efficient cooling designs. They need detailed hot gas path heat transfer distributions. Heat transfer and film cooling data are also needed for the airfoils. The surface heat transfer on a stator vane is affected by the combustor-generated high turbulence, the laminar-to-turbulent transition, acceleration, film cooling flow, platform secondary flow, and surface roughness. These factors as well as the rotational, centrifugal forces and blade tip clearance and leakage must be considered for the rotating blades. After the potential hot spots on the airfoil surface are identified, the internal cooling schemes can be developed. Designers need new internal heat transfer data to improve current rotor blade cooling performance. They also need detailed flow and heat transfer data to understand the flow physics and to improve the current internal cooling designs. Many techniques have been developed to enhance the heat transfer in these passages. The cooling passages located in the middle of the airfoils are often lined with rib turbulators. Near the leading edge of the blade, jet impingement (coupled with film cooling) is commonly used. Jet impingement is also used throughout the cross-section of the stator vanes. Pin-fins and dimples can be used in the trailing edge portion of the vanes and blades. These techniques have also been combined to further increase the heat transfer from the airfoil walls. A number of traditional cooling concepts are used in various combinations to adequately cool the turbine vanes and blades; these techniques are identified and described throughout this chapter. In addition, newly developed, advanced cooling concepts are also introduced as possible cooling alternatives. The interested reader is referred to Gas Turbine Heat Transfer and Cooling Technology by Han et al. for a more in depth description of turbine blade heat transfer and cooling1. In addition, Lakshminarayana reviewed recent publications involving turbine cooling and heat transfer, and Dunn put together a detailed review of convective heat transfer and aerodynamics in axial flow turbines2. A symposium volume discussing heat transfer in gas turbine systems is also available by Goldstein3.
4.2.2.2-2 Enhanced Internal Cooling of Turbine Vanes A typical cooled turbine vane is shown in figure 2. As shown in the figure, the vane is hollow, so cooling air can pass through the vane internally. The coolant is extracted from the internal channel for impingement and pinfin cooling. Jet impingement is a very aggressive cooling technique which very effectively removes heat from the vane wall. However, this technique is not readily applied to the narrow trailing edge. The vane trailing edge is cooled using pin-fins (an array of short cylinders). The pin-fins increase the heat transfer area while effectively mixing the coolant air to lower the wall temperature of the vanes. After impinging on the walls of the airfoil, the coolant exits the vane and provides a protective film on the vane’s external surface. Similarly, the coolant traveling through the pin-fin array is ejected from the trailing edge of the airfoil.
Impingement Cooling Impingement cooling is commonly used near the leading edge of the airfoils, where the heat loads are the greatest. With the cooling jets striking (impinging) the blade wall, the leading edge is well suited for impingement cooling because of the relatively thick blade wall in this area. Impingement can also be used near the mid-chord of the vane. Figure 2 shows jet impingement located throughout the cross-section of an inlet guide vane. Several aspects must be considered when developing efficient cooling designs. The effect of jet-hole size and distribution, cooling channel cross-section, and target surface shape all have significant effects on the heat transfer coefficient distribution. Jet impingement near the mid-chord of the blade is very similar to impingement on a flat plate; however, the sharp curvature at the leading edge of the vane must be considered when utilizing impingement in this region.
Fig. 2. Schematic of a Turbine Vane Cross-Section with Impingement and Trailing Edge Pin-Fin Cooling
Jet Impingement from Multiple Jets As shown in figure 2, many jets are used to increase the heat transfer from the vane wall. It has been shown by Metzger et al. that multiple jets perform very differently from a single jet striking a target surface4. They concluded that for multiple jets, the Nusselt number is strongly dependent on the Reynolds number, while there is no significant dependence on the jet-to-target plate spacing. The difference is due to the jet cross-flow from the spent jets. Studies by Florschuetz et al. and Koopman and Sparrow showed that the mass from one jet moves in the cross-jet flow direction, and this flow can alter the performance of neighboring jets5. The crossflow attempts to deflect a jet away from its impinging location on the target plate. In situations with very strong cross-flow and sufficiently
322
Je-Chin Han and Lesley M. Wright large jet-to-target plate spacing, the cross-flow can completely deflect the jet away from the impingement surface. Florschuetz and Su reported that cross-flow decreases the overall heat transfer from the impingement surface6. They determined that cross-flow enhances the convective heat transfer, but the enhancement from the jets decreases, as the jets are deflected. Because the enhancement from the impingement jets is much greater than the convective enhancement, the overall Nusselt numbers decrease in the presence of cross-flow. A typical test model used by Florschuetz et al. is shown in figure 37. As shown in this figure, the coolant jets impinge on the target surface from the jet plate in an inline array. As the coolant travels along the test surface, the spent air from the upstream jets effects the heat transfer coefficient distributions of the downstream jets, and this effect increases as more spent air accumulates on the target surface.
Fig. 3. A Typical Test Model for Impingement Cooling Studies
Correlations based on experimental data were developed by Kercher and Tabakoff and Florschuetz et al. to estimate the heat transfer enhancement from an array of impinging jets8. Although the correlations are in different forms, they both demonstrate the dependence of the heat transfer enhancement on the amount of cross-flow. Florschuetz et al. also showed the cross-flow effect is much stronger in staggered arrays of jets than an inline array9. Bailey and Bunker extended the correlation developed by Florschuetz et al. to include the effect of jet spacing10. The correlation has been extended to include dense impingement arrays. Huang et al. controlled the direction of the cross-flow and obtained detailed distributions of the heat transfer coefficients for three target plates11. Their results clearly indicate when the cross-flow travels in two opposite directions, the heat transfer enhancement on the target plate is much greater than when the cross-flow is restricted to one direction. This study was extended by Ekkad et al. to include the effect of coolant extraction for film cooling12. The heat transfer enhancement on the target plate decreases near the edges due to the decreased coolant flow (for film cooling). Wang et al. investigated cross-flow through a confined space; they also considered cross-flow traveling in one direction and two directions13. This study also concluded that increasing cross-flow results in degraded heat transfer; however, the heat transfer coefficient distribution is much more uniform. The heat transfer coefficient distributions on target plates with stretched arrays of impinging jets were studied by Gao et al.14. This array varies from the traditional square array in which the jets are evenly spaced. They concluded the existing correlations for square arrays over-predict the effect of cross-flow in the target surface. The presence of initial cross-flow also effects the heat transfer enhancement from the target plate. The cross-flow described above is created by the spent flow from the jets. Therefore, the first row of jets is not affected by the cross-flow. However, in many situations, cross-flow may develop upstream of the first row. The flow from upstream of the impingement jets can significantly alter the flow near the jets, and thus alter the heat transfer coefficients on the target surface. Florschuetz et al. investigated the effect of initial cross-flow on the heat transfer enhancement15. The results of this study were similar to those mentioned above describing cross-flow: the heat transfer enhancement on the target plate decreases when initial cross-flow is present.
Jet Impingement on a Curved Surface The above studies investigated the heat transfer on flat target plates. The results obtained for flat plates can be applied to impingement near the mid-chord of the blade. However, the effect of target surface curvature must be considered when implementing jet impingement near the leading edge of the airfoil. The curvature of the airfoil creates different cross-flow behavior, and therefore, the heat transfer coefficients on the curved surface are different than those on the flat surface. Chupp et al. studied impingement on a curved surface, and this group concluded that the average Nusselt number ratio increases as the curvature of the target plate increases16. The effect of target surface shape was also pursued by Bunker and Metzger17. They concluded that a sharper nose radius yields a more
323
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes uniform Nusselt number distribution compared to a smooth-nosed chamber. This study was also extended to include the effect of coolant extraction for film cooling18. When the bleed from the pressure and suction surfaces is equal, the greatest reduction in the Nusselt numbers occurs.
Pin-Fin Cooling Due to manufacturing constraints in the very narrow trailing edge of the blade, pin-fin cooling is typically used to enhance the heat transfer from the blade wall in this region. The pins typically have a height-to-diameter ratio between ½ and 4. In a pin-fin array heat is transferred from both the smooth channel endwall and the numerous pins. Flow around the pins in the array is comparable to flow around a single cylinder. As the coolant flows past the pin, the flow separates and wakes are shed downstream of the pin. In addition to this wake formation, a horseshoe vortex forms just upstream of the base of the pin, and the vortex wraps around the pins. This horseshoe vortex creates additional mixing, and thus enhanced heat transfer. Many factors must be considered when investigating pin-fin cooling. The type of pin-fin array and the spacing of the pins in the array effect the heat transfer distribution in the channel. The pin size and shape also have a profound impact on the heat transfer in the cooling passage. Because pin-fins are commonly coupled with trailing edge ejection (as shown in figure 2), the effect of this coolant extraction must also be considered.
Pin Array and Partial Length Pin Arrangement There are two array structures commonly used. One is the inline array and the other is the staggered array. Figure 4 shows a typical experimental test model with a staggered array of pin-fins. Metzger et al. used staggered arrays of circular pins with 1.5 to 5 pin diameter spacing in a rectangular channel19. A closer spaced array (smaller x/D) shows a higher heat transfer coefficient. Their observations clearly indicate that addition of pin-fins significantly enhances the heat transfer coefficient. However, the addition of pins also increases the pressure drop in the flow channel. Chyu et al. showed that the heat transfer coefficient on the pin surface for both arrays is consistently higher than that of the channel endwall20. The pin surface heat transfer is observed to be 10 to 20 percent higher for the presented case. Experimental results have been correlated by Metzger et al. and VanFossen to predict the Nusselt number in channels with pin-fin arrays21. The average Nusselt number in a channel with short pin-fins is primarily dependent on the Reynolds number of the flow, and a weaker dependence is shown for the pin spacing.
Fig. 4. A Typical Test Model and Secondary Flow for Pin-Fin Cooling Studies
Arora and Abdel-Messeh studied the effects of partial length pins in a rectangular channel22. The surface containing pins is not affected by the pin tip clearance. Whereas the opposite surface, that does not have pins, shows a decrease in heat transfer coefficient with an increase in the pin tip clearance. The friction factor is lower for partial pins compared to full-length pins. In general, the heat transfer coefficient decreases with partial length pins.
324
Je-Chin Han and Lesley M. Wright Effect of Pin Shape and Array Orientation Metzger et al. studied the effects of pin shape and array orientations23. They reported the effect of flow incident angle on oblong pins. All incident angles except 90 yield higher Nusselt numbers than circular pins. The γ= 90° array yields significantly lower Nusselt numbers, especially toward the lower end of the Reynolds number range. The γ= ±30° array has the highest Nusselt numbers, about 20 percent higher than the circular pin array on the average. Except for γ= 90°, the pressure drop for oblong pins are significantly higher than circular pins. This increase in the friction factor is associated with the flow turning caused by oblong pins. The pin shapes mostly studied are straight cylinders. However, the casting or other manufacturing processes cannot make perfect cylinders and these manufacturing imperfections may affect the heat transfer performance. Chyu studied the effect of a fillet at the base of the cylindrical pin24. Straight cylinders in staggered array formation have the highest heat transfer followed by filleted cylinders in the staggered formation. It is interesting to note that the fillet cylinder inline formation has better heat transfer than the straight cylinders in the inline formation. Though a staggered array gives higher heat transfer coefficients, performance of the inline straight cylinders is best among the group and the fillet cylinders in staggered formation are the worst. In a different experimental work, Goldstein et al. studied the effect of stepped diameters on mass transfer coefficients25. The diameter of the pin is axially varied. The base diameter is greater than the center diameter and no fillet radius is provided. The array configuration is staggered. Results show that the mass transfer increases or remains the same compared to a straight cylinder pin array when the radius is varied, but the pressure drop reduces significantly for the stepped diameter cylindrical pins. Chyu et al. used cube and diamond shaped pins to enhance the heat transfer coefficient from a surface26. The cube-shaped pins have the highest mass transfer coefficients among the shapes considered and round pins have the lowest mass transfer coefficients. Corresponding pressure loss coefficients are higher for the cube and diamond shaped pins relative to the circular pins.
Effect of Flow Convergence and Turning The flow channel in the trailing edge of an airfoil has a reducing cross-section, and therefore, the flow in the channel accelerates. The results are row averaged and the accelerating flow shows an increase in the heat transfer coefficient27. Chyu et al. used mass transfer technique to study the effect of perpendicular flow entry in two pin-fin configurations28. They show that the turning inlet configuration always results in lower average Sherwood numbers. The reduction is about 40-50% for the inline array and 20-30% for the staggered array.
Pin-Fin Cooling With Ejection The trailing edge pin-fin channel normally has ejection holes through which the spent coolant exhausts to the main stream flow. Kumran et al. investigated the effects of the length of coolant ejection holes on the heat transfer coefficient in pin-fins29. The length of the ejection hole can significantly alter the discharge rate of coolant. More coolant ejection reduces the Nusselt number significantly from no ejection. This decrease in the heat transfer coefficient can be explained by the fact that coolant mass is extracted from the coolant channel before its cooling capacity is fully utilized. Results indicate that the correlation based on the local Reynolds number can predict the heat transfer coefficient distribution for lower coolant ejection but does not adequately predict the heat transfer coefficients at higher ejection rates. Hwang and Lu investigated a converging channel with ejection30. They also found that increasing the ejection degrades the endwall heat transfer near the tall wall opposite of the ejection, and the heat transfer on the channel endwall surface near the ejection holes is increased. They also concluded that square, diamond, and circular pin-fin arrays enhance the heat transfer equally in channels with large ejection flows.
Dimple Cooling In recent years, dimples have been considered as an alternative to pin-fin cooling. Dimpled cooling is a very desirable alternative due to the relatively low pressure loss penalty (compared with pins) and moderate heat transfer enhancement. A typical test section for dimple cooling studies is shown in figure 5; this figure also shows the dimple induced secondary flow. These concave dimples induce flow separation and reattachment with pairs of vortices. The areas of high heat transfer include the areas of flow reattachment on the flat surface immediately downstream of the dimple. The heat transfer in the dimpled channel is typically 2 to 2.5 times greater than the heat transfer in a smooth channel with a pressure loss penalty of 2 to 4 times that of a smooth channel. These values show little dependence on Reynolds number and channel aspect ratio. However, the dimple size, dimple depth (depth-to-print diameter ratio = 0.1 to 0.3), distribution, and shape (cylindrical, hemispheric, teardrop) each effect the heat transfer distribution in the channel. Recent studies have investigated the influence of these factors on the heat transfer in rectangular channels31. Dimples have also been investigated in a circular channel and similar levels of heat transfer enhancement and frictional losses were measured32. Syred et al. compared the heat transfer enhancement due a single dimple on both flat and curved surfaces33. From this study it was shown that the surface curvature significantly influences the heat transfer enhancement. The heat transfer is further enhanced on a surface that is concavely shaped (compared to a flat surface); however, a convexly curved surface with a dimple decreases the level of heat transfer enhancement.
325
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Fig. 5. A Typical Test Model for Dimple Cooling Studies with a Conceptual View of Dimple Induced Secondary Flow
Compound and New Cooling Techniques Several internal heat transfer enhancement techniques are discussed in previous sections. Most common methods of heat transfer augmentation in gas turbine airfoils are ribs, pins, and jet impingement. It is shown that these enhancement techniques increase heat transfer coefficients, but can combining these techniques increase the heat transfer coefficient more? Several researchers have combined these heat transfer enhancement techniques to improve the heat transfer coefficient. However, it is not always recommended to combine more than one heat transfer augmentation technique. In addition to compounding more than one heat transfer enhancement technique, there are attempts to incorporate new concepts (e.g., jet swirlers and heat pipes) in the turbomachinery cooling. An introduction of several new cooling techniques with their applications is presented here.
Impingement on Ribbed, Pinned, and Dimpled Walls Azad et al. studied the impingement effect on dimpled and pinned surfaces34. Because the dimples and pins are circular depressions and protrusions, respectively, these two target surfaces offer an interesting comparison of the heat transfer enhancement. At lower Reynolds number the pinned surface performs better than the dimpled surface. At higher Reynolds numbers, the dimpled surface performs better than the pinned surface for a certain flow orientation. Taslim et al. reported a significant increase in the heat transfer enhancement on a curved target surface roughened with conical bumps35. This study was extended to include film cooling holes, similar to the showerhead-type film cooling on the leading edge36. They concluded that the presence of leading edge extraction also significantly increases the heat transfer on the target surface. The heat transfer is further increased if the “racetrack” jet holes are used rather than traditional round jet holes37
Combined Effect of Swirl and Impingement A new jet impingement and swirl technique was investigated by Glezer et al.38. A preliminary test showed significant improvement in the heat transfer performance. Based on that study, a new airfoil has been designed with swirling impingement in the leading edge. This new airfoil is tested in a hot cascade test section. Results indicate that screw shaped swirl cooling can significantly improve the heat transfer coefficient over a smooth channel and this improvement is not significantly dependent on the temperature ratio and rotational forces. Moreover, it was concluded that optimization of the internal passage geometry in relation to location and size of the tangential slots is very important in achieving the best performance of the screw-shaped swirl in the leading edge cooling. Pamula et al. studied the heat transfer enhancement by a combination of impingement and cross flow-induced swirl in a two-pass channel39. Results show that the new impingement system, from the first pass to the second pass, using cross flow injection holes produce significantly higher heat transfer on the second pass walls.
Combined Effect of Swirl Flow and Ribs Kieda et al. experimentally investigated the single-phase water flow and heat transfer in a rectangular cross-sectioned twisted channel40. Several aspect ratios and twist pitches were used. Results indicate that in a cooling application, this twisted channel performs similar to a ribbed pipe. Zhang et al. used different types of inserts to study the combined rib and twisted tape inserts in square ducts41. Four test configurations were used: twisted tape, twisted tape with interrupted ribs, hemi-circular wavy tape, and hemi-triangular wavy tape. The twisted tape with interrupted ribs provides higher overall heat transfer performance over the twisted tape without ribs and
326
Je-Chin Han and Lesley M. Wright hemi-circular wavy tape. The performance of the hemi-triangular wavy tape is comparable with the twisted tape plus interrupted ribs. Hemi-circular wavy tapes show the lowest heat transfer performance in this group.
Combined Effect of a Ribbed Wall with Grooves or Pins Zhang et al. studied the heat transfer and friction in rectangular channels with a rib-groove combination42. The Stanton numbers for the ribbed-grooved walls are higher than that for the only ribbed walls at similar rib spacing values. Metzger et al. indicated that the addition of ribs does not change the heat transfer coefficient from the pin-mounted surface43. However, the heat transfer coefficient on the rib surface is significantly higher than the pin-mounted surface.
New Cooling Concepts Heat pipes have very high effective thermal performance44. Therefore, they can transfer heat from high temperature to the low temperature regions. This concept may be used in the airfoil cooling. Heat is removed from the initial stage stator airfoils and the heat is delivered at a later stage to heat up the main flow. This way the heat extracted can be recycled to the main flow. In a concept developed by Yamawaki et al., the heat is conducted away from the hot airfoil to the fin assembly45. This passive heat extraction reduces the required cooling air. Most heat pipe applications are designed for the stator airfoils, where it is easier to mount the connecting pipes or fins. Recently, Kerrebrock and Stickler proposed a design to incorporate heat pipe in the rotor46. The concept of cooled cooling air systems, through a heat exchanger, for turbine thermal management was reported by Bruening and Chang47. Results show that the use of a cooled cooling air system can make a positive impact on overall engine performance for land-based turbines. Commonly a closed loop steam cooled nozzle with thermal barrier coatings (TBC) is used in order to reduce the hot gas temperature drop through the first stage nozzle (Corman and Paul48. A closed looped with mist/steam cooling was reported by Guo et al.49. Results show that an average heat transfer enhancement of 100% can be achieved with 5% mist (fine water droplets) compared to the steam cooling; similar results were reported by Li et al.50 .
4.2.2.2-3 Enhanced Internal Cooling of Turbine Blades Figure 6 shows several techniques to cool a modern gas turbine blade. The blade consists of serpentine cooling passages lined with rib turbulators. Jet impingement is used to cool the leading edge of the blade, and pin-fin cooling with ejection is used near the trailing edge. Although the techniques used to cool the blades are similar to those used to cool the vanes, the heat transfer trends in the vanes and blades are very different. Because the blades are rotating, the flow of the coolant in the passages is altered. Therefore, the effect of rotation on the internal heat transfer enhancement must be considered.
327
Fig. 6. Schematic of a Modern Gas Turbine Blade with Common Cooling Techniques
Fig. 7. A Typical Test Model for Turbulated Cooling Studies with Rib Induced Secondary Flow
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes Rib Turbulated Cooling Rib turbulators are the most frequently used method to enhance the heat transfer in the internal serpentine cooling passages. The rib turbulence promoters are typically cast on two opposite walls of the cooling passage. Heat that conducts from the pressure and suction surfaces through the blade walls is transferred to the coolant passing internally through the blade. The heat transfer performance of the ribbed channel depends on the channel aspect ratio, the rib configurations, and the Reynolds number of the coolant flow. Many fundamental studies have been conducted to understand the coolant flow through a stationary ribbed channel51. The studies show as the coolant passes over a rib oriented 90° to the mainstream flow, the flow near the channel wall separates. Reattachment follows the separation, and the boundary layer reattaches to the channel wall; this thinner, reattached boundary layer results in increased heat transfer coefficients in the ribbed channel. This rib induced secondary flow is shown in figure 7. If the rib turbulators are skewed to the mainstream flow direction, counter-rotating vortices are created. Figure 7 shows in a channel with angled ribs, two counterrotating vortices are formed in the cross-section of the cooling passage. However, if V-shaped rib turbulators are used, four vortices are generated. The additional set of counter-rotating vortices associated with the V-shaped ribs results in more heat transfer enhancement in a channel with V-shaped ribs than angled ribs. The ribs also create turbulent mixing in the areas of flow separation. With this additional mixing, the heat is more effectively dissipated from the wall, and thus additional heat transfer enhancement. Because only the flow near the wall of the cooling channel is disturbed by the ribs, the pressure drop penalty by ribs affordable. Han and Han and Park developed correlations for both the pressure penalty and heat transfer enhancement in ribbed channels52. Given the Reynolds number of the coolant flow and the rib geometry (e/D, P/e, W/H, and α), the average friction factor in a channel with two opposite ribbed walls, f , and the centerline average Stanton number on the ribbed walls, Str, can be determined from the correlations. Figure 8 demonstrates the correlations developed for cooling passages with 90° ribs. The friction roughness function, R, is only a function of rib spacing for the range of the roughness Reynolds number, e+, shown. Based on the rib spacing (P/e), R can be calculated, and substituted into the following equation to determine f, the four ribbed wall friction factor.
Fig. 8. Friction Factor and Heat Transfer Coefficient Correlations for 90° Ribs
2 R = f
1
2
2e 2W + 2.5 ln ⋅ + 2.5 D H +W
(1)
From f (the friction factor in a channel with ribs on all four walls) and the channel geometry (H/W), the average friction factor in a channel with ribs on two walls, f , can be calculated using Eqn. 2.
328
Je-Chin Han and Lesley M. Wright
(
H f = f + f − fs W
)
(2)
The friction factor in a channel with smooth walls, fs, is known from the existing Blasius correlation for smooth channel flow. Using the four ribbed wall friction factor, f, the rib height, e/D, and the Reynolds number of the coolant flow, Re, the roughness Reynolds number, e+, can be calculated using the definition shown in figure 8. From e+ the top figure can be used to obtain G, the heat transfer roughness function, and Eqn. 3 can be used to calculate the Stanton number on the ribbed walls, Str.
(3)
The correlation shown in figure 8 is for a Prandtl number of 0.703. Because G is inversely proportional to the Stanton number, a low heat transfer roughness function implies high heat transfer from the cooling passage wall. With the understanding that skewed ribs yield higher heat transfer enhancement than orthogonal ribs, these correlations were extended to include the effect of the rib angle. Figure 9 shows the correlations taking into account the rib angle, α. From the rib angle (α), rib spacing (P/e), and channel aspect ratio (W/H), the roughness function, R, can be determined. Eqn. 1 can used to calculate f, and Eqn. 2 is used to determine the friction factor in a channel with two ribbed walls. Similar to channels with 90° ribs, R and e+ are then used to determine the Stanton number on the ribbed walls. These correlations can be used over a wide range of channel aspect ratios and rib configurations; however one should refer to the original papers for specific restrictions of the correlations.
Fig. 9. Friction Factor and Heat Transfer Correlations in Rectangular Ribbed Channels
Because ribs are the most common heat transfer enhancement technique for the serpentine cooling passages, many studies have been conducted to study the effects of channel cross-section, rib configuration, and coolant flow Reynolds number. As shown in figure 6, the aspect ratio of the channels changes from the leading to the trailing edge of the blade. Near the leading edge of the blade, the channel may have an aspect ratio around ¼, but near the trailing edge, much broader channels are present with aspect ratios around 4. Multiple studies have shown that by skewing the ribs, so they are angled into the mainstream flow, the heat transfer coefficients can be further enhanced. Placing the ribs with an attack angle between 30° and 60° results in increased heat transfer and reduces the pressure penalty. Most studies focus on Reynolds numbers ranging from 10,000 to 80,000, but for today’s advanced gas turbines the coolant in the channel can have a Reynolds number up to 500,000. The height of the ribs is typically 5-10% of the channel hydraulic diameter, and the rib spacing-to-height ratio varies from 5 to 15. In addition, a limited number of studies have focused on the more closely spaced ribs with much larger blockage ratios.
329
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes With angled ribs performing superior to orthogonal ribs, many researchers have extended their studies to include a wide variety of rib configurations. Han et al. showed that V-shaped ribs (figure 10) outperform the angled ribs; for a given pressure drop, the Vshaped ribs give more heat transfer enhancement53. Numerous other studies have shown the same conclusion that V-shaped ribs perform better than the traditional angled ribs in a variety of channels and flow conditions54. In an effort to further increase the heat transfer performance of the rib turbulators, discrete rib configurations were introduced. Figure 10 shows discrete (or broken) ribs are similar to the traditional ribs, but they are broken in one or more locations, so the rib is not continuous. In the majority of cooling channels, discrete ribs were shown to outperform the continuous angled or V-shaped ribs and Cho et al.55. Han and Zhang compared the performance of many high performance rib configurations, and the comparison is shown in figure 1156. As shown in this figure, the broken 45° angled ribs create more heat transfer enhancement than the continuous 45° angled ribs at a given friction factor ratio, and this conclusion can be extended to other broken versus continuous rib configurations.
Fig. 10. High Performance Rib Turbulators for Turbine Blade Internal Cooling
Fig. 11. Comparison of Heat Transfer Performance for Broken and Non-Broken Rib Configurations
The majority of ribs used in experimental studies have a square cross-section; however, studies have investigated the heat transfer enhancement of various profiled ribs. Delta-shaped ribs were studied by Han et al. and these ribs were also shown to result in higher heat transfer enhancement than the traditional angled ribs57. Bunker and Osgood investigated the performance of ribs leaning into or away from the flow58. They concluded the traditional square ribs give greater heat transfer enhancement and less frictional losses than the ribs leaning into or away from the flow. When the blades are cast, the ribs are unlikely to have sharp edges as the previous studies have considered. The ribs are likely to have rounded edges, and this was taken into account by Taslim and Spring59. From their experimental work, they concluded the effect of rounding decreases the level of heat transfer enhancement in the cooling channel. Ribs with a higher aspect ratio (taller ribs) are more sensitive to the rounding effect; whereas, square ribs are only slightly affected by the rounded edges. However, the pressure drop is significantly less in the channels with rounded ribs. Studies have also focused on ribbed channels with more blockage. Bailey and Bunker and Taslim and Spring showed that increasing in the effective blockage, the heat transfer coefficients can increase60. However, this increase comes at the cost of a significant increase in the pressure penalty. Therefore, if the heat loads are extremely high and the high frictional losses can be tolerated, the additional heat transfer enhancement would be beneficial. An additional factor that should be considered when determining the heat transfer distribution in cooling channels is the decreasing coolant flow rate due to extraction for film cooling. Most modern turbine airfoils have ribs in the internal coolant channel and film cooling for the outside surface. Therefore, some of the cooling air is bled through the film cooling holes. The presence of periodic ribs and bleed holes creates strong axial and spanwise variations in the heat transfer distributions on the passage surface. Shen et al., Ekkad et al., and Thurman and Poinsette each studied the heat transfer enhancement by ribs in the presence of coolant extraction61. These studies showed that the heat transfer coefficients in the near-hole regions increase, but no broader impact by these holes is noticeable in these results. Ekkad et al. showed that the regional-averaged Nusselt number ratios for different rib orientations are almost identical with and without bleed hole extraction, as shown in figure 1262. This indicates that 20 to 25% reduction of the main flow can be used for film cooling without significantly affecting the ribbed channel cooling performance. It was also shown that the heat transfer near the holes can be further enhanced if the ribs are placed near the bleed holes63.
330
Je-Chin Han and Lesley M. Wright
Fig. 12. Heat Transfer Enhancement in Cooling Passages with and without Film Coolant Extraction Holes
Rotational Effect on Rib Turbulated Cooling
331
Heat transfer in rotating coolant passages is very different from that in stationary coolant passages. Both Coriolis and rotating buoyancy forces alter the flow and temperature profiles in the rotor coolant passages and affect their surface heat transfer coefficient distributions64. It is very important to determine the local heat transfer distributions in the rotor coolant passages under typical engine cooling flow, coolant-to-blade temperature difference (buoyancy effect), and rotating conditions. Effects of coolant passage cross-section and orientation on rotating heat transfer are also important. As sketched in figure 13, the secondary flows in a two-pass channel are different for radial outflow and radial inflow passes65. Since the direction of the Coriolis force is dependent on the direction of rotation and flow, the Coriolis force acts in different directions in the two-passes. For radial outward flow, the Coriolis force shifts the core flow towards the trailing wall. If both the trailing and leading walls are symmetrically heated, then the faster moving coolant near the trailing wall would be cooler (therefore heat transfer would be enhanced) than the slower moving coolant near the leading wall (i.e., heat transfer would be decreased). Rotational buoyancy is caused by a strong centrifugal force that pushes cooler heavier fluid away from the center of rotation. In the first channel rotational buoyancy affects the flow in a similar fashion as the Coriolis force and causes a further increase in flow and heat transfer near the trailing wall of the first channel; whereas, the Coriolis force favors the leading side of the second channel. The rotational buoyancy in the second channel tries to make the flow distribution more uniform in the duct.
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Fig. 13. Conceptual View of Coolant Flow through a Two-Pass Rotating Channel
Wall Heating Condition Effect on Rotating Coolant Passage Heat Transfer From the above analyses, the rotation effect on channel heat transfer comes from the Coriolis and centrifugal forces. The centrifugal force is known as rotation buoyancy when there is a temperature difference between the coolant and the channel walls under rotating conditions. Since the temperature difference between the coolant and the channel walls varies along the coolant passages, so does the rotation buoyancy. Therefore, it is expected that the channel wall heating conditions would affect rotor coolant passage heat transfer. The channel heating conditions imply that the channel walls may be at the same temperature (or heat flux) in both streamwise and circumferential directions, or the trailing wall temperature may be higher than the leading wall temperature in real turbine blade cooling applications. Han et al. studied the uneven wall temperature effect on rotating two-pass square channels with smooth walls66. They concluded that in the first pass, the local uneven wall temperature interacts with the Coriolis force-driven secondary flow and enhances the heat transfer coefficients in both leading and trailing surfaces, with a noticeable increase in the leading side, as compared with the uniform wall temperature case. However, the uneven wall temperature significantly enhances heat transfer coefficients on both leading and trailing surfaces. Parsons et al. and Zhang et al. studied the influence of wall heating condition on the local heat transfer coefficient in rotating two-pass square channels with 90° ribs and 60° ribs on the leading and trailing walls, respectively67. They concluded that the uneven wall temperature significantly enhances heat transfer coefficients on the first-pass leading and second-pass trailing surfaces as compared with the uniform wall temperature condition.
Combined Effect of Rotation and Rib Shape on the Heat Transfer in Rotating Channels The above studies investigating the effect of rotation on the heat transfer in cooling channels only consider the heat transfer in channels with square ribs. However, the shape of the rib can significantly alter the heat transfer trends, as demonstrated in stationary channels. Acharya et al. investigated the heat/mass transfer in a square, two-pass rotating channel with various profiled ribs placed on the leading and trailing surfaces68. It was shown that certain profiled ribs provide better heat transfer enhancement than the conventional square ribs. The smooth sidewalls of the channel also see significantly more enhancement than the smooth walls of channels with square ribs.
Effect of Channel Cross-Section and Channel Orientation on Rotating Channel Heat Transfer The first studies of heat transfer in rotating channels were performed on square channels oriented normal to the direction of rotation. Wagner et al. reported that the heat transfer coefficients on the trailing surface of the first pass can be enhanced 2-3 times that of a non-rotating channel, while the leading surface experiences a declination of up to 50%69. Opposite trends were present in the second pass of this smooth channel. The cooling channel was lined with angled turbulators, and it was found there is less of an effect of rotation in a ribbed channel than a smooth channel70. Because the heat transfer enhancement of the ribbed channel is already 3.5 times greater than that of a smooth channel, rotation does not provide the same percentage of enhancement in the cooling channel with ribs. However, the additional enhancement is significant and should be considered. Similar to the non-rotating channel, 45° angled ribs provide more enhancement than 90° ribs in a rotating channel. Park et al. conducted naphthalene sublimation experiments to examine the effects of rotation on the local heat and mass transfer distribution in a two-pass ribbed square channel71. They also found that the overall heat and mass transfer in a rotating channel with ribbed surfaces was not affected by the Coriolis force as much as that in a rotating channel with smooth surfaces.
332
Je-Chin Han and Lesley M. Wright When considering rotating channels, the orientation of the channels must be considered. One can see in figure 14 that the orientation of the channel changes as on moves away from the middle of the blade. Johnson et al. studied the effects of rotation on the heat transfer for smooth and 45° ribbed serpentine channels with channel orientations of 0° and 45° to the axis of rotation72. They found that the effects of Coriolis and buoyancy forces on heat transfer in the rotating channel are decreased with the channel at 45° compared to the results at 0°. This implies that the difference in heat transfer coefficient between the leading and trailing surfaces due to rotation will be reduced when the channel is angled to the axis of rotation. Parsons et al. investigated the effect of channel orientation on the heat transfer in a rotating, square two-pass channel with 60° angled ribs73. The also concluded that difference between the heat transfer coefficients on the leading and trailing surfaces decreases when the orientation of the channel changes from 90° to 45°. Dutta and Han conducted an experimental study of regionally averaged heat transfer coefficients in rotating smooth and ribbed two-pass channels; this study also investigated the effect of channel orientation on the heat transfer distributions in this square channel74. They found the effect of rotation is reduced for non-orthogonal alignment of the heat transfer surfaces with respect to the plane of rotation. They also concluded that the discrete V-shaped ribs have better heat transfer performance than the 90° ribs and the 60° angled ribs in the rotating channels. Al-Hadhrami and Han used parallel and crossed 45° angled ribs in rotating two-pass square channels to study the effect of channel orientation on heat transfer75. They concluded that the parallel 45° angled ribs are better than the crossed 45° angled ribs. They also confirmed the difference between the leading and trailing wall heat transfer coefficients is reduced for the channel with a 45° angle to the axis of rotation. As shown in figure 14, the cross-section of the cooling channels can vary depending on where the cooling channel is located within the blade. Studies showed the heat transfer distributions in square channels can vary significantly from those in rectangular channels. Therefore, it is necessary to study the effect of rotation on heat transfer in rectangular channels. As with the square channels, the orientation of the channel varies depending on the location in the blade.
Fig. 14. Typical Cooling Passage Size and Orientation with Conceptual Views of the Rotation Induced Secondary Flow
333
Fig. 15. Advanced Rib Configurations Studied in Rotating Cooling Channels
Guidez reported the effect of rotation on heat transfer in a straight rotating rectangular channel (AR = 2:1) with smooth walls76. Soong et al. conducted heat transfer experiments in rotating smooth straight duct with different aspect ratios ranging from 0.2 to 5.077. They concluded that the aspect ratio of the duct was a critical parameter in the secondary flow patterns. Taslim et al. investigated the varying effect of rotation in square (AR = 1:1) and rectangular (AR = 2:1) with 45° crossed ribs78. Azad et al. and Al-Hadhrami et al. studied heat transfer in a two-pass rectangular rotating channel (AR = 2:1) with 45° angled ribs and 45° V-shaped ribs, respectively, including the effect of channel orientation with respect to the axis of rotation79. The heat transfer trends in these rectangular channels are similar to those of square channels: the heat transfer on trailing surface of the first pass is enhanced with rotation while the heat transfer on the leading surface decreases (and the opposite is true in the second pass). However, the difference between the leading and trailing surfaces decreases in the 2:1 rectangular channels. As the orientation of the channels varies from 90° to 45°, the effect of rotation continues to decrease. Similar to non-rotating channels, these studies concluded that the 45° V-shaped ribs perform better than the 45° crossed V-shaped ribs, and subsequently better than 45° angled ribs (figure 15).
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes Moving closer to the trailing edge, the aspect ratio of the cooling channels continues to increase, as the channel orientation is also changing. Griffith et al. studied heat transfer in a single pass rectangular channel (AR = 4:1) with smooth and 45° angled ribbed walls, including the effect of channel orientation with respect to the axis of rotation80. Results show that the narrow rectangular passage exhibits much higher heat transfer enhancement for the ribbed surface than the square and 2:1 channels previously investigated. Also, channel orientation significantly affects the leading and side surfaces, yet does not have much affect on the trailing surfaces for both smooth and ribbed surfaces. Therefore, this investigation has determined that spanwise variations in the heat transfer distribution of rectangular cooling passages exist, and that the enhancement is a function of channel orientation with respect to the axis of rotation, surface configuration (such as smooth or 45° angled ribbed walls), and channel aspect ratio. Lee et al. also studied the heat transfer in a single pass rectangular channel (AR = 4:1)81; they compared the heat transfer performance of V-shaped and angled rib turbulators with and without gaps (figure 15), in a channel oriented at 135° to the axis of rotation. The results show that V-shaped rib configuration produces more heat transfer enhancement than the angled rib configurations for both the stationary and rotating cases. There is only negligible difference in heat transfer enhancement between the parallel and staggered rib configurations for both the stationary and rotating cases. The results also show that the V-shaped ribs with gaps produce overall less heat transfer enhancement than the V-shaped ribs without gaps; whereas the angled ribs with gaps produce overall greater heat transfer enhancement than the angled ribs without gaps for the stationary case, and clearly the same enhancement for the rotating case. Most importantly, for narrow rectangular rib-turbulated channels oriented at 135° with respect to the plane of rotation, heat transfer enhancement on both the leading and trailing surfaces increases with rotation. This is quite different from the square channel where rotation enhances the trailing surface heat transfer but reduces the leading surface heat transfer for the radial outward flow case. This provides positive information for the cooling designers. Wright et al. expanded the study of the 4:1 channel with ribs to examine not only the heat transfer performance, but the overall thermal performance82. They investigated channels with six rib configurations: angled, V-shaped, W-shaped, discrete angled, discrete V-shaped, and discrete W-shaped (figure 15). Figure 16 shows that similar to the stationary channels, the rotating channel with discrete V-shaped ribs produced more heat transfer than the V-shaped and angled ribbed channels. However, the W-shaped and discrete W-shaped ribs yielded more heat transfer enhancement than the discrete V-shaped ribs. However, the increased heat transfer enhancement in the W-shaped and discrete W-shaped ribs came at the cost of an increased pressure penalty; these two configurations resulted in greatest pressure drop of the six configurations considered. Therefore, the overall performance of the discrete V-shaped and W-shaped ribs is comparable in the rotating channels. The traditional angled ribs exhibited the worst overall performance.
Fig. 16. Heat Transfer Enhancement, Frictional Losses, and Thermal Performance of Various Rib Configurations in a 4:1 Rotating Cooling Passage
The cooling channels near the leading edge of the blade typically a smaller aspect ratio compared to those near the center of the blade. Cho et al.studied the effect of rotation on heat transfer with a naphthalene sublimation technique in a two-pass rectangular channel (AR = 1:2) with smooth and 70° ribbed walls83. They found that the effect of rotation diminishes in the second pass with inward flow due to the strong influence of the 180° turn. This is quite different from the heat transfer trends found in the two-pass rectangular channels with aspect ratios greater than unity. Agarwal et al. reported the effect of rotation on heat transfer with naphthalene sublimation technique in two-pass rectangular channels (aspect ratio = 1:1 and 1:4) with smooth and 90° ribbed walls84. They concluded that the 1:4 rectangular channel provides lower levels of heat transfer enhancement along the trailing wall and higher levels of heat transfer degradation along the leading wall compared to the 1:1 square channel. Fu et al. also experimentally investigated the heat transfer trends
334
Je-Chin Han and Lesley M. Wright in cooling channels located near the leading edge of the blade85. The regionally averaged heat transfer coefficients in rotating cooling channels with aspect ratios of 1:4 and 1:2 lined with 45° angled ribs were obtained. This study concluded the rotation effect increased heat transfer on the trailing wall but decreased the heat transfer on the leading wall in the first pass of both the 1:4 and 1:2 channels. In the second pass, the difference of the heat transfer between the leading and trailing walls was reduced under the rotating condition when compared to the first pass. This study included the effect of channel orientation, and it was shown that the 45° channel orientation creates less heat transfer difference between the leading and trailing walls than the 90° channel orientation for both aspect ratio ducts. Rotation has a relatively small effect in the second pass of the 1:2 and 1:4 channels. It was suggested that the 180° turn induced vortices dominate the rotation induced vortices for the low aspect ratio ducts. When the results from this study were compared to others with various aspect ratios, it could be seen that the leading wall heat transfer has a strong dependence on the buoyancy parameter for 1:2 and 1:4 ducts in the first pass with radial outward flow, but it has a weak dependence on the buoyancy parameter for 2:1 and 4:1 ducts. Increasing the buoyancy parameter reduced the heat transfer on the leading walls for the 1:2 and 1:4 ducts in the first pass. The 1:4 duct has the largest heat transfer difference between the leading and trailing walls in the first pass. Although the majority of rotating studies have focused on the heat transfer in square or rectangular cooling channels, a limited number of studies have focused on the heat transfer in channels with other cross-sections. Channels with a triangular cross-section might be used on some portion of the blade in order to provide compact channel structure and good cooling efficiency. Clifford et al. studied the mean heat transfer in a straight triangular-sectioned rotating duct with smooth walls86. Harasgama and Morris compared the effect of rotation on heat transfer in straight circular, triangular, and square duct with smooth walls87. Dutta et al. studied the effect of rotation on the heat transfer coefficients in two-pass triangular channels with smooth and ribbed walls88. For the locations in the first pass, the triangular-duct heat transfer coefficients are mostly contained within the upper and lower limits imposed by the square duct. This is because in a triangular duct, there is less space for the coolant to form secondary flow by rotation. However, in the second pass, the leading surface of the triangular duct shows much higher heat transfer coefficients than the square duct. This is due to more mixing and favorable secondary flow in the 180° turn for the triangular duct geometry. Rathjen et al. investigated the heat/mass transfer in a twopass rotating channel with a near-engine cross-section; the first pass of the channel (radial outward) was trapezoidal, and the second pass (radial inward) had a larger trapezoidal cross-section89. They showed strong gradients after the 180° turn as the flow is forced to follow the shape of the blade.
Fig. 17. Typical Experimental, Two-Pass Test Section with 45° Angled Ribs with a Conceptual View of the Rib and
335
Turn Induced Secondary Flow
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes With the internal cooling passages of the blade can take on a variety of shapes, sizes, and orientations, it is interesting to investigate the effect of rotation in more detail. Results drawn from Fu et al. are presented to show the isolated affect of rotation on the heat transfer enhancement in channels with aspect ratios ranging from 1:4 (near the leading edge) to 4:1 (in the trailing edge region)90. Specifically, they studied two-pass channels with aspect ratios of 1:4, 1:2, 1:1, and 2:1, and they included the work of Griffith et al. to incorporate a 4:1 single pass cooling channel91. As shown by the sample test section in figure 17, each of the two-pass test sections consists of twelve regions in the stream-wise direction; with region one at the inlet of the test section, and region twelve at the outlet. Because the hydraulic diameter of each channel is different, comparisons are made at specific regions (1 through 12), rather than specific x/D locations, and the regions are labeled in figure 17. Figures 18-21 show the Nusselt number ratios at various regions in rotating channels with smooth walls, oriented at 90° to the direction of rotation, and figures 22-25 compare the heat transfer enhancement in cooling channels with smooth walls with various channel orientations (incorporating the probable channel location and orientation in an actual blade). This comparison is extended to rotating channels with 45° angled ribs in figures 26-33. The Nusselt number ratio shown is the measured Nusselt number in the rotating channel (NuR) – to – the Nusselt number measured in the identical stationary channel (NuS). With this altered definition of the Nusselt number ratio, the effect of rotation is isolated from other factors influencing the heat transfer enhancement. In addition, the Nusselt number ratio is presented as a function of the local buoyancy parameter.
Fig. 18. Nusselt Number Ratio Comparison at Region 4 (1st Pass, Fully Developed) in Smooth Rotating Channels (β = 90°)
Fig. 19. Nusselt Number Ratio Comparison at Region 11 (2nd Pass, Fully Developed) in Smooth Rotating Channels (β = 90°)
Figure 18 shows the Nusselt number ratios in region 4, on the leading and trailing surfaces (figure 18a), and the inner and outer surfaces (figure 18b); region 4 should represent fully developed flow in the first pass of each channel. The effect of rotation is most clearly seen on the trailing surface of the various channels. As the buoyancy parameter increases, the Nusselt number ratio increases in each channel, and the buoyancy parameter approaches zero, the Nusselt number approaches unity, as the effect of rotation diminishes. As explained earlier (and depicted in figure 13), the Coriolis and buoyancy forces combine to enhance the heat transfer coefficients on the trailing (destabilized) surface. Interestingly, the Nusselt number ratios on the trailing surfaces of the various channels all collapse to the same trend. The increase of the heat transfer coefficients on the trailing surface comes at the expense of the degradation of the heat transfer coefficients on the leading surface. In general, as the buoyancy parameter increases, the Nusselt number ratios decrease; indicating the adverse effect of rotation. However, the Nusselt number ratios cover a larger range than the ratios on the trailing surface. Positive information for turbine designers is shown in figure 18(b). As the buoyancy parameter increases, the Nusselt number ratios on both the inner and outer surfaces at region 4 increase. Figure 19 moves to region 11 in the second pass; this region is far away from the 180° turn, so it should represent fully developed flow in the second pass. As shown in figure 19, the overall effect of rotation is reduced (when compared to the first pass) in second pass as the Coriolis and buoyancy forces act is opposite directions. Although the effect of rotation is expected to decrease, previous studies have shown, it is not eliminated, and this is further confirmed in figure 19. The leading surfaces of each channel experience heat transfer enhancement as the buoyancy parameter increases. However, the heat transfer enhancement due to rotation is much less on the leading surface of the second pass, than the trailing surface of the first pass. As shown in figure 18a, the enhancement due to rotation on the trailing surfaces increases the Nusselt number ratio up to (and beyond in the 4:1 channel) two times that of the stationary channel. However, on the leading surfaces in the second pass, the maximum heat transfer enhancement due to rotation does not exceed 1.5. As expected the trailing surfaces are adversely affected by rotation, with the 1:1 and 2:1 channels experiencing the greatest declination. The heat transfer coefficients on all of the inner and outer surfaces of region 11 are slightly enhanced with rotation, as shown in fiugre 19(b). However, the level of enhancement in this region is less than the enhancement in the first pass at region 4.
336
Je-Chin Han and Lesley M. Wright
Fig. 20. Nusselt Number Ratio Comparison at Region 6 (1st Pass Turn) in Smooth Rotating Channels (β = 90°)
Fig. 21. A Nusselt Number Ratio Comparison at Region 7 (2nd Pass Turn) in Smooth Rotating Channels (β = 90°)
In addition to considering the fully developed regions in the two-pass channel, the level of heat transfer enhancement in the sharp turn should also be considered. Figure 20 shows the Nusselt number ratios at region 6 for the leading, trailing, inner, and outer surfaces. Similar to region 4, the trailing surfaces at region 6 all experience heat transfer enhancement due to rotation. While the Nusselt number ratios for the leading surfaces are generally below unity. The variation between the leading and trailing surfaces is less than the difference in region 4, due to the onset of the sharp turn. As figure 20(b) shows, the heat transfer coefficients on both the outer and tip surfaces of region 6 are enhanced with rotation.
337
Figure 21 shows the Nusselt number ratios in region 7 (the second half of the sharp turn). The Nusselt number ratios on the leading surfaces in all the channels increase with the increasing buoyancy parameter, similar to region 11. With the exception of the 1:4 channel, the heat transfer coefficients on the all of the trailing surfaces decreases with rotation. The effect of rotation on the outer and tip surfaces of region 7 is less than the effect of rotation in region 6. With the exception of the tip surface in the 2:1 channel, both the outer and tip surfaces are slightly elevated above the Nusselt number ratio of unity, and increasing the buoyancy parameter does not result in a significant increase in the Nusselt number ratios. With the flow redirection created in the stationary channel, the heat transfer coefficients are naturally elevated. Therefore, the additional secondary flow induced by rotation only slightly increases the already elevated Nusselt numbers.
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Fig. 22. Nusselt Number Ratio Comparison at Region 4 (1st Pass, Fully Developed) in Smooth Rotating Channels (β = 45° or 135°)
Fig. 23. Nusselt Number Ratio Comparison at Region 11 (2nd Pass, Fully Developed) in Smooth Rotating Channels (β = 45° or 135°)
The effect of channel orientation is combined with the channel aspect ratio and local buoyancy parameters in figures 22 – 25. The 1:4 and 1:2 channels have orientations of 45°, the 2:1 and 4:1 channel are oriented at 135°, and the 1:1 channel maintains its previous orientation of 90°. As shown in figure 22, in the fully developed region of the first pass (region 4), the level of heat transfer enhancement on the trailing surfaces, due to rotation, is the same as in the channels with the 90° orientation. However, the degradation due to rotation on the leading surfaces is less in these skewed channels. Unlike the inner and outer surfaces of the channels with the 90° orientation, not all surfaces of every channel experience heat transfer enhancement with the increasing buoyancy parameter. The heat transfer enhancement (or declination) is strongly dependent on the channel aspect ratio. As shown in figure 14, the channel orientation affects the rotation induced vortices. In the 1:4 and 1:2 channels with the 45° channel orientation, the vortices are impinging on the inner surfaces of the first pass; thus increasing the heat transfer coefficients. However, the 135° of the 2:1 channel results in the vortices impinging on the outer surface, further increasing the heat transfer coefficients on the outer surface, rather than the inner surface. The Nusselt number ratios at region 11 in the smooth channels with skewed orientations are shown in figure 23. The heat transfer enhancement on the leading surfaces of these skewed channels is similar to those with the normal orientation. However, the Nusselt numbers on the trailing surfaces of the 1:4 and 1:2 channels are increased with rotation; this differs from the orthogonal rotating
338
Je-Chin Han and Lesley M. Wright channels. The Nusselt numbers on the trailing surface 2:1 channel with the 135° orientation decrease with the increasing buoyancy parameter (similar the 2:1 channel with the 90° orientation). The level of heat transfer enhancement on the inner and outer walls in the second pass of these channels with various orientation is similar to the orthogonally rotating channels, with the exception of the inner wall of the 2:1 channel. The coolant is forced away from the corner between the inner and trailing walls to the corner between the leading and outer wall of the second pass. As the heat transfer coefficients on the trailing surface of the 2:1 channel sharply decrease, so do the heat transfer coefficients on the inner surface. The additional complexity of flow through the sharp 180° turn is shown in figures 24 and 25. The level of heat transfer enhancement is the same on the leading and trailing surfaces, as well as the outer and tip surfaces of region 6. The dominant flow behavior through the turn is not strongly effected by the channel orientation; therefore, the Nusselt number ratios are not strongly effected as the channel orientation shifts from 90°, and these conclusions can be extended to region 7 of the channels, as well (figure 25). The internal cooling passages of turbine blades generally do not have smooth surfaces. Therefore, it is important to consider not only channels with smooth walls, but cooling channels with rib turbulators. As described in previous sections, angled rib turbulators are commonly used in the blades. Fu et al. experimentally studied the effect of buoyancy on the heat transfer enhancement in channels with rib turbulators as they did for smooth, rotating passages. Figures 26 – 29 show rotating – to – stationary Nusselt number ratios at various regions in orthogonal cooling passages with angled ribs92.
Fig. 24. Nusselt Number Ratio Comparison at Region 6 (1st Pass Turn) in Smooth Rotating Channels (β = 45° or 135°)
Fig. 25. Nusselt Number Ratio Comparison at Region 7 (2nd Pass Turn) in Smooth Rotating Channels (β = 45° or 135°)
Fig. 26. Nusselt Number Ratio Comparison at Region 4 (1st Pass, Fully Developed) in Rotating Channels (β = 90°) with 45° Angled
Fig. 27. Nusselt Number Ratio Comparison at Region 11 (2nd Pass, Fully Developed) in Rotating Channels (β = 90°) with 45°Angled Ribs
Ribs
339
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes Figure 26 shows the effect of rotation on the heat transfer coefficients at region 4 in the cooling channels with 45° angled ribs. The general trends for the trailing (enhancement) and leading (declination) surfaces is same as for the smooth channel, as figure 26(a) shows. However, the effect of rotation on the trailing surface is less in the ribbed channels than the smooth channels. The Nusselt number ratios in the stationary channel are elevated to the presence of the ribs, and additional enhancement due to rotation is less significant in these ribbed channels. The declination of the heat transfer coefficients on the leading surface is more severe in these ribbed channels than the smooth channels; this is markedly clear in the 1:4 and 1:2 channels. From the numerical predictions of Su et al., the interaction of the rotation and rib induced secondary flow creates flow reversal, and a small cell of relatively hot air is trapped near the leading surface93. Within this cell, less mixing occurs with the core of the coolant flow, and therefore, these leading surfaces are adversely affected by rotation. Figure 26(b) shows that the Nusselt number ratios on both the inner and outer surfaces in region 4 are enhanced with rotation. Figure 27 shows the effect of rotation is reduced in the second pass at region 11 when compared to the first pass at region 4. In addition the effect of rotation on the leading and trailing surfaces is less in these ribbed channels than the smooth channels. The heat transfer enhancement on the inner and outer surfaces at region 11 is strongly dependent on the channel cross-section. The 1:4 channel is most strongly affected by rotation; both the inner and outer surfaces experience heat transfer enhancement. The Nusselt number ratios of the remaining channels are not significantly affected by rotation, as the Nusselt number ratios are approximately unity over the range of buoyancy parameters. The inner surface of the 2:1 channels is the exception as a sharp decrease in the Nusselt number ratio is observed.
Fig. 28. Nusselt Number Ratio Comparison at Region 6 (1st Pass Turn) in Rotating Channels (β = 90°) with 45° Angled Ribs
Fig. 29. Nusselt Number Ratio Comparison at Region 7 (2nd Pass Turn) in Rotating Channels (β = 90°) with 45° Angled Ribs
Figures 28 and 29 show the effect of rotation in the turn of the ribbed channels. Similar to region 4, figure 28(a) shows Nusselt number ratios on the trailing surface are increasing with buoyancy parameter, while they decrease on the leading surface; the greatest effect of rotation is seen in the 1:4 channel. Meanwhile, the outer and tip surfaces of each channel are increasing with the increasing effect of rotation. Figure 29 shows in the second half of the turn (region 7), all of the surfaces experience heat transfer enhancement with the increasing buoyancy parameter. The level of enhancement on the outer and tip surfaces is strongly dependent on the channel aspect ratio. With an understanding of the heat transfer enhancement in orthogonally rotating channels with rib turbulators, the additional complexities in skewed channels with angled ribs can be discussed. Figures 30 – 33 show the Nusselt number ratios in cooling channels with angled ribs with non-orthogonal rotating angles. As with the smooth channels, the 1:4 and 1:2 channels have orientation angles of 45°, the 2:1 and 4:1 channels have orientation angles of 135°, and the 1:1 channels maintains an orientation of 90°. As shown in figure 30, the enhancement due to rotation is less in the skewed channels than the normal channels of region 4. The declination in the Nusselt numbers is significantly less in the channels with skewed orientations than the declination in the normal channels. The orientation of the channel does not significantly effect the heat transfer enhancement on the trailing, inner, or outer surfaces. The effect of channel orientation on the heat transfer coefficients decreases in the second pass. Figure 31 shows that the trends and level of enhancement on the leading and trailing surfaces of the skewed channels are the same as those of the normal channels. This can be anticipated as the effect of rotation decreases in the second pass, and altering the orientation angle further lessens the effect of rotation on the cooling channel heat transfer coefficients. The diminished effect of rotation is clearly seen on the inner walls of the 1:4 and 2:1 channels. Significant enhancement is seen on the inner wall of the 1:4 channel while extreme declination is seen on the inner wall of the 2:1 channel (figure 27); however, neither of these trends are seen the skewed channels (figure 31).
340
Je-Chin Han and Lesley M. Wright
Fig. 30. Nusselt Number Ratio Comparison at Region 4 (1st Pass, Fully Developed) in Rotating Channels (β = 45° or 135°) with 45° Angled Ribs
Fig. 31. Nusselt Number Ratio Comparison at Region 11 (2nd Pass, Fully Developed) in Rotating Channels (β = 45° or 135°) with 45° Angled Ribs
Fig. 32. Nusselt Number Ratio Comparison at Region 6 (1st Pass Turn) in Rotating Channels (β = 45° or 135°) with 45° Angled Ribs
Fig.33. Nusselt Number Ratio Comparison at Region 7 (2nd Pass Turn) in Rotating Channels (β = 45° or 135°) with 45° Angled Ribs
Similar to the smooth channels, figures 32 and 33 show a minimal effect of channel orientation in the turn of the two pass channels. The leading and trailing surfaces are affected less in the skewed channels than the normal channels in region 6; however in region 7, there is only negligible difference between the Nusselt number ratios on the leading and trailing surfaces. The heat transfer coefficients on the outer and tip surfaces of the 2:1 channel with the 135° orientation are positively influenced by the skewed channel orientation. The effect of channel orientation is further discussed by considering the average heat transfer enhancement of the two pass channels. Fu et al. presented the overall Nusselt number ratio (measured Nusselt number to the Nusselt number for turbulent flow through a smooth tube, Dittus-Boelter correlation) in two pass smooth and ribbed channels94. The overall average Nusselt number ratio was taken as the 12 region average of both the leading and trailing surfaces. Figure 34 shows the effect of rotation is greater in the 1:4 and 1:2 smooth channels than the 1:1 and 2:1 smooth channels. It should be noted that Fu et al. conducted their experiments at constant rotational speed (550 rpm)95. Therefore, the rotation number varies inversely with the Reynolds number of the coolant flow. Therefore, the rotation number is the highest at the lowest Reynolds number. Similar to the smooth channels, the effect of rotation is most apparent in the 1:4 channel.
341
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Fig. 34. Channel Averaged Nusselt Number Ratios for NonRotating and Rotating Channels
Fig. 35. Overall Friction Factor Ratios for Non-Rotating and Rotating Channels
To further evaluate the viability of the various cooling channels, the frictional losses should be considered along with the heat transfer enhancement. Figure 35 shows the friction factor ratios measured by Fu et al. for the various aspect ratio channels96. The friction factor ratios in the smooth channels are greater than unity due to the large pressure loss incurred in the 180° turns. The friction factor ratios in the ribbed channels vary significantly depending on the channel aspect ratio. The maximum friction factor is 5.8 in the 1:4 channel, while the friction factor in the 2:1 channel can be 11.5 times greater than the friction factor in a smooth tube as given by the Blasius correlation. Physically the ribs in each of the four channels are the same size. Therefore, the blockage due to the ribs in the 2:1 channel is 5 times greater than the blockage in the 1:4 channel, and at a Reynolds number of 25,000, the increased blockage results in a friction factor ratio increase of 2.3 times. To heat transfer enhancement and the pressure penalty are combined in figure 36 to evaluate the overall thermal performance in each channel. As expected the 1:4 channels have the greatest thermal performance, as they have the lowest friction factor ratios.
Numerical Simulations of Cooling Channel Heat Transfer Enhancement
Fig. 36. Overall Thermal Performance for Non-Rotating and Rotating Channels
Thus far a wide variety of experimental results have been discussed for the enhancement of cooling channel heat transfer. However, numerical predictions have helped researchers better understand the complex flow phenomena in the cooling passages. The kε, low Reynolds k-ε, the two-layer k-ε, and the low Reynolds number k-ω turbulence models have been used to predict the flow behavior and heat transfer enhancement in rotating cooling channels. However, problems with each of these models can lead to inaccurate predictions of the complex flow in rotating channels. Prakash and Zerkle showed the k-ε model is unable to predict the flow and heat transfer behavior in two pass rotating channels with angled ribs due to the isotropic turbulence assumption and the near surface wall function required by this model97. However, Lin et al. showed the low Reynolds k-ω model is an adequate model for rotating two pass cooling channels with angled ribs98. This model replaces the wall function in the near wall region with turbulence dissipation specified in terms of the turbulent kinetic energy.
342
Je-Chin Han and Lesley M. Wright The advanced Reynolds stress model and the second moment closure model have also been applied to rotating cooling channels. The second moment closure model requires more computing power than the previously mentioned models as six additional transport equations are solved in the three-dimensional turbulent flow field. Also, the eddy diffusivity in the momentum transport equation is replaced with the source terms developed from the turbulent Reynolds stress tensor. At the expense of additional computational time, this model provides accurate predictions for flow in two pass rotating channels with angled rib turbulators.
343
The second moment closure model used by Chen et al. has been applied to internal cooling channels with a wide range of turbulator configurations and channel aspect ratios99. The model was applied to a single pass rotating channel with a square cross-section by Jang et al.100. The predicted heat transfer coefficients in channels with either 45° or 60° angled ribs compared favorably with experimental data. Al-Qahtani et al.extended these predictions to include 2:1 two pass channels and 4:1 single pass channels with 45° angled ribs101. Su et al. used this same second moment closure model to predict the heat transfer coefficients in rotating 4:1 single pass channels with V-shaped ribs102. Figure 37 shows the velocity vectors and temperature contours in non-rotating, single pass 4:1 channels (with angled and V-shaped ribs) and a non-rotating, 2:1 two pass channel with angled ribs. The figures clearly show near the wall, the coolant follows the rib turbulators. The dimensionless temperature contours show the lowest coolant temperatures occur at the most upstream point on the ribs. In addition, the wall temperature is relatively high just downstream of the ribs, due to the flow separation that occurs as the coolant flows over the turbulators. The secondary flow and temperature contours in a two pass, rotating channel with an aspect ratio of 2:1 is shown in figure 38103. In this channel, with smooth walls, Fig. 37. Velocity Vectors and Temperature Contours: (a) AR = 4:1 the shift of the coolant toward the trailing surface of the One-pass Channel with 45° Inclined Ribs, (b) AR = 4:1 One Pass first pass is clearly seen. With the core of the coolant Channel with V-shaped Ribs, (c) AR = 2:1 Two-Pass Channel with forced to the trailing surface, a steep temperature gradient 45° Inclined Ribs (Re = 10,000) near the trailing wall forms. The thinner boundary layer on this destabilized trailing surface results in enhanced heat transfer coefficients due to rotation. The opposite behavior occurs in the second pass, as the core of the mainstream coolant is forced to the leading surface. However, the temperature contours downstream in the second pass are influenced by the flow through the 180° turn, so the flow is not symmetrical in the channel crosssection. Similar to the experimental studies which cover a wide range of channel aspect ratios, Su et al. performed a complementary study to numerically predict the heat transfer coefficients in the cooling channels with aspect ratios varying from 1:4 to 4:1104. Figure 39 shows the secondary flow and dimensionless temperature contours in the first pass of rotating cooling channels with 45° angled ribs. This study includes parameter variations which are unachievable in most experimental facilities. They found that for the low Reynolds number and low rotation number cases, the rotation effect on the Nusselt number and friction factor ratios is more significant in the 1:2 channel than those observed in either the square or 1:4 channel due to the presence of strong turn-induced vortices. However, in the under engine-like conditions with high rotation and high Reynolds numbers, the effect of rotation decreases as the channel aspect ratio changes from 1:1 to 1:2 and finally to 1:4. Although the previously mentioned studies are only a sampling of those utilizing the second moment turbulence model, they have demonstrated the power of this advanced second-order Reynolds stress turbulence model. The models have been validated with existing experimental data, and upon validation, they have been extended to high Reynolds number and high rotation number flows which cannot be obtained in most experimental facilities. Over a wide range of flow conditions, these predictions provide detailed velocity, pressure, and temperature profiles which are essential to understanding the complex flow behavior in these rotating channels.
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Fig. 38. Secondary Flow and Temperature Contours in Rotating AR = 2:1 Smooth Duct (Re = 10,000, Ro = 0.22)
Fig. 39. A Secondary Flow in the First Passage (before the 180° turn) of Rotating Ribbed Channels with Different Aspect Ratios (Re = 10,000, Ro = 0.14)
344
Je-Chin Han and Lesley M. Wright Flow Field Measurements in Rotating Cooling Channels Heat transfer is a side effect of the flow field. Flow in a rotating channel is significantly different from flow in a non-rotating channel. The secondary flow in rotation redistributes velocity and also alters the random velocity fluctuation patterns in turbulent flows. Lezius and Johnston examined flow instability in a straight rectangular channel with smooth walls caused by rotation105. They reported that rotation increases flow velocity and turbulence near the unstable trailing wall and reduces the turbulent fluctuations significantly near the stable leading wall. Elfert measured velocity and turbulence distributions in a rotating circular pipe106. Rotation shifts the bulk flow toward the trailing side, and the turbulence profile shows a different distribution due to rotation. Tse and Mcgrath used laserDoppler velocimetry to measure rotating flow in a two-pass channel with smooth and ribbed walls107. Tse and Steuber investigated the mean flow characteristics in the first and second passages of a rotating four-pass coolant passage with 45° ribs of semi-circular cross section using LDA108. Cheah et al. used the LDA to measure the velocity and turbulence quantities in a rotating two-pass channel109. Bons and Kerrebrock measured the internal flow in a straight smooth-wall channel with particle image velocimetry (PIV) for both heated and non-heated cases110. Schabacker et al. and Chanteloup et al. also used PIV to measure the flows in stationary two-pass square ducts with 45° turbulators111. Son et al. measured the flows in two-pass channels with smooth and 90° ribbed walls using PIV112. Liou and Chen measured the developing flow in a two-pass channel with LDA113. Liou et al. studied fluid flow in a rotating two-pass duct with in-line 90° ribs using LDA114. Liou et al. also performed LDV and pressure measurements for in-line detached 90° ribs115. Liou and Dai measured pressure and flow characteristics in a rotating two-pass square duct with 45° angled ribs of square cross section by using the LDA116. Liou et al. measured flow and pressure fields in a rotating two-pass duct with staggered 45° ribs of rounded cross section117. Both Liou and Dai and Liou et al. reported that the absence of periodic fully developed flow condition in their tests118. The above-mentioned flow measurements help to understand the flow physics and serve to explain the heat transfer results obtained in two-pass rotating channels with smooth and ribbed walls.
Rotational Effect on Jet Impingement Cooling Several studies were mentioned above that investigate jet impingement cooling. Although a number of impingement studies have been completed, only a few studies consider the effect of rotation on impingement cooling. Epstein et al. studied the effect of rotation on impingement cooling in the leading edge of a blade119. They reported that the rotation decreases the impingement heat transfer, but the effective heat transfer is better than a smooth rotating channel. The zero staggered cooling jets (i.e., jet direction is perpendicular to rotation direction) show lower heat transfer coefficients compared to that with a staggered angle. Mattern and Hennecke reported the effect of rotation on the leading edge impingement cooling by using the naphthalene sublimation technique120. Their experiment did not include the rotating buoyancy effect. The jet direction has an offset angle with respect to the rotation direction. They found that the rotation decreases the impingement heat transfer for all staggered angles. The effect of rotation is least when jet direction has an angle of 45° to rotation direction. However, a maximum of 40% reduction in heat transfer is noted when jet direction is perpendicular to rotation direction. This may be because the Coriolis force creates a swirl action on the spent flow and also deflects the jet when jet direction is parallel to rotation direction. Glezer et al. studied the effect of rotation on swirling impingement cooling in the leading edge of a blade121. They found that screw-shaped swirl cooling can significantly improve the heat-transfer coefficient over a smooth channel and the improvement is not significantly dependent on the temperature ratio and rotational forces. Parsons et al. studied the effect of rotation on impingement cooling in the mid-chord region of the blade122. A central chamber serves as the pressure chamber, and jets are released in either direction to impinge on two heated surfaces. The jet impinging directions have different orientations with respect to the direction of rotation. They reported that the rotation decreases the impingement heat transfer on both leading and trailing surfaces with more effect on the trailing side (up to 20% heat transfer reduction). Akella and Han studied the effect of rotation on impingement cooling for a two-pass impingement channel configuration with smooth walls123. The difference from the earlier experiment by Parsons et al. is that spent jets from the trailing channel are used as cooling jets for the leading channel124. Therefore, the cross-flow in the trailing side is radial outward; for the leading side, it is radial inward. They reported that irrespective of the direction of rotation, the heat transfer coefficient on the first-pass and second-pass impinging wall decreases up to 20% in the presence of rotation. Akella and Han included 45° angled ribs in the target surfaces of their two-pass impingement channel with rotation125. They reported that jet impingement on the ribbed wall can provide 10-50% more heat transfer compared to that on the smooth wall for jet Reynolds number increasing from 4,000 to 10,000. This is because the angled rib- induced secondary flow gets stronger with higher cross-flow at higher jet Reynolds number. They also found that the rotation decreases impingement heat transfer on the first-pass and second-pass ribbed wall. Parsons et al. extended their earlier work to include heat transfer on a smooth target wall with film coolant extraction from a channel with four heated walls126.
Rotational Effect on Pin-Fin Cooling Pin-fin cooling has been investigated for many years, but only recently has the effect of rotation been considered in channels with pin-fins. Recently, Willett and Bergles studied the effect of rotation on heat transfer in narrow rectangular channels (AR = 10:1) with smooth and with typical pin-fin array, respectively, including channel orientation effect with respect to the plane of rotation127. They found that the heat transfer enhancement in the pin-fin channel due to rotation and buoyancy was less than the enhancement in the smooth channel. They showed that heat transfer enhancement mainly is due to pin-fin flow disturbance; pin-fins significantly reduce the effect of
345
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes rotation, but they do not eliminate the effect. Wright et al. studied the effect of rotation on heat transfer in narrow rectangular channels (AR = 4:1 and 8:1) with typical pin-fin array used in turbine blade trailing edge design and oriented at 150° with respect to the plane of rotation128. Results show that turbulent heat transfer in a stationary pin-fin channel can be enhanced up to 3.8 times that of a smooth channel; rotation enhances the heat transferred from the pin-fin channels up to 1.5 times that of the stationary pin-fin channels. Most importantly, for narrow rectangular pin-fin channels oriented at 150° with respect to the plane of rotation, heat transfer enhancement on both the leading and trailing surfaces increases with rotation. This provides positive information for the cooling designers.
Rotational Effect on Dimple Cooling Because dimple cooling has only recently been considered as an internal cooling technique, the number of studies available in open literature is limited. The majority of dimple studies were described above, and they are only applicable for stator cooling designs. Only a few studies focus on rotor blade dimple cooling. Zhou et al. studied heat/mass transfer in a rotating square channel with typical dimple array129. They found that heat transfer enhancement for the stationary dimple channel is around two times that of the smooth wall value; however, rotation enhances heat transfer on the trailing dimple surface and reduces heat transfer on the leading dimple surface in a similar manner as the rotational effect on the trailing and leading surfaces of the square channel with ribs. Griffith et al. studied heat transfer in rotating rectangular channels (AR = 4:1) with typical dimple array on both leading and trailing walls, including the effect of channel orientation with respect to the plane of rotation130. The results show that rotation enhances heat transfer on both trailing and leading surfaces of the narrow dimpled channel in a similar trend as the rotational effect on the trailing and leading surfaces of the narrow rectangular channel with ribs or pins; however, the heat transfer enhancement of the ribbed or pinned channel exceeds that of the dimpled channel. Also, the dimpled channel oriented at 135° with respect to the plane of rotation provides greater overall heat transfer enhancement than the orthogonal dimpled channel.
4.2.2.2-4 Concluding Remarks With modern gas turbines operating at extremely high temperatures, it is necessary to implement various cooling methods, so the turbine blades and vanes survive in the path of the hot gases. Simply passing coolant air through the airfoils does not provide adequate cooling; therefore, it is necessary to implement techniques that will further enhance the heat transfer from the airfoil walls. The internal heat transfer can be enhanced with jet impingement, pin-fin cooling (used in the trailing edge), and internal passages lined with turbulence promoters. The most commonly used turbulators are ribs. The heat transfer distribution in cooling channels with ribs has been studied for many years because a number of factors combine to affect the heat transfer. Because the flow through non-rotating and rotating channels is very different, many studies have focused on the effect of rotation in ribbed channels. Although the majority of studies concentrate on channels with a square cross-section, a limited number of investigations are now available for rectangular cooling channels with ribs. A relatively new alternative to rib turbulators are dimples. A limited number of studies are available to address the heat transfer in stationary channels with dimples, and even fewer studies are available that address the effect of rotation in cooling channels with dimples. It is also important to investigate the effect of rotation on jet impingement and pin-fin cooling, and in the recent years, such studies have become available. More studies are needed for the blade-shaped coolant passages with high performance turbulators and with or without film cooling holes under realistic coolant flow, thermal, and rotating conditions. Also, more studies are needed for rotating impingement cooling with or without film coolant extraction as well as rotating pin-fin cooling with or without trailing edge ejection in order to guide the efficient rotor blade internal cooling designs. Highly accurate and highly detailed local heat transfer data is needed to aid engineers in their design of blades for advanced gas turbines.
346
Je-Chin Han and Lesley M. Wright 4.2.2.2-5 Notes ________________________________________
347
1. J.C. Han, S. Dutta, and S.V. Ekkad, “Gas Turbine Heat Transfer and Cooling Technology,” Taylor & Francis, Inc., New York, New York, December 2000, ISBN # 1-56032-841-X, 646 pages. 2. B. Lakshminryana, “Turbine Cooling and Heat Transfer,” Fluid Dynamics and Heat Transfer of Turbomachinery, John Wiley, New York, 1996, pp. 597-721; M.G. Dunn, “Convection Heat Transfer and Aerodynamics in Axial Flow Turbines,” ASME Journal of Turbomachinery. 123 no.4 (2001):.637-686. 3. R.J. Goldstein, “Heat Transfer in Gas Turbine Systems,” Annuals of The New York Academy of Sciences, New York, New York, Vol. 934, 2001, 2001, 520 pages. 4. D.E. Metzger, L.W. Florschuetz, D.I. Takeuchi, R.D. Behee, and R.A. Berry, “Heat Transfer Characteristics for Inline and Staggered Arrays of Circular Jets with Crossflow of Spent Air,” ASME Journal of Heat Transfer, 101 (1979): 526-531. 5. L.W. Florschuetz, R.A. Berry, and D.E. Metzger, “Periodic Streamwise Variations of Heat Transfer Coefficients for Inline and Staggered Arrays of Circular Jets with Crossflow of Spent Air,” ASME Journal of Heat Transfer 102 (1980): 132-137; R.N. Koopman and E.M. Sparrow, “Local and Average Transfer Coefficients Due to an Impinging Row of Jets,” International Journal of Heat and Mass Transfer 92 (1976): 73-82. 6. L.W. Florschuetz and C.C. Su, “Effects of Crossflow Temperature on Heat Transfer Within an Array of Impinging Jets,” ASME Journal of Heat Transfer 109 (1987): 74-82. 7. Ibid. 8. D.M. Kercher and W. Tabakoff, “Heat Transfer by a Square Array of Round Air Jets Impinging Perpendicular to a Flat Surface Including the Effect of Spent Air,” ASME Journal of Engineering for Power, Vol. 92 (1970): 73-82; L.W. Florschuetz, C.R. Truman, and D.E. Metzger, “Streamwise Flow and Heat Transfer Distributions for Jet Array Impingement with Crossflow,” ASME Journal of Heat Transfer 103 (1981): 337-342. 9. See note 8 (Florschuetz, et al.). 10. J.C. Bailey and R.S. Bunker, “Local Heat Transfer and Flow Distributions for Impinging Jet Arrays of Dense and Sparse Extent,” ASME Paper No. GT-2002-30473 (2002); also see note 8 (Florschuetz, et al.). 11. Y. Huang, S.V. Ekkad, and J.C. Han, “Detailed Heat Transfer Distributions Under an Array of Orthogonal Impinging Jets,” AIAA Journal of Thermophysics and Heat Transfer 12 (1998): pp. 73-79. 12. S.V. Ekkad, Y. Huang, and J.C. Han, “Impingement Heat Transfer on a Target Plate with Film Holes,” AIAA Journal of Thermophysics and Heat Transfer 13 (1999): 522-528. 13. T. Wang, M. Lin, and R.S. Bunker, “Flow and Heat Transfer of Confined Impingement Jets Cooling,” ASME Paper No. 2000-GT-223 (2000). 14. L. Gao, S.V. Ekkad, and R.S. Bunker, “Impingement Heat Transfer Under Linearly Stretched Arrays of Holes,” ASME Paper No. GT2003-38178 (2003). 15. L.W. Florschuetz, D.E. Metzger, and C.C. Su, “Heat Transfer Characteristics for Jet Array Impingement with Initial Crossflow,” ASME Journal of Heat Transfer 106 (1984): pp. 34-41. 16. R.E. Chupp, H.E. Helms, P.W. McFadden, and T.R. Brown, “Evaluation of Internal Heat Transfer Coefficients for Impingement Cooled Turbine Airfoils,” AIAA Journal of Aircraft. 6 (1969): 203-208. 17. R.S Bunker and D.E. Metzger, “Local Heat Transfer in Internally Cooled Turbine Airfoil Leading Edge Regions. Part I: Impingement Cooling Without Film Coolant Extraction,” ASME Journal of Turbomachinery 112 (1990): 451-458. 18. D.E. Metzger and R.S. Bunker, “Local Heat Transfer in Internally Cooled Turbine Airfoil Leading Edge Regions. Part II: Impingement Cooling with Film Coolant Extraction,” ASME Journal of Turbomachinery. 112 (1990): 459-466. 19. D.E. Metzger, R.A. Berry, and J.P. Bronson, “Developing Heat Transfer in Rectangular Ducts With Staggered Arrays of Short Pin Fins,” ASME Journal of Heat Transfer 104 (1982): 700-706. 20. M.K. Chyu, Y.C. Hsing, T.I.P. Shih, and V. Natarajan, “Heat Transfer Contributions of Pins and Endwall in Pin-Fin Arrays: Effects of Thermal Boundary Condition Modeling,” ASME Paper No. 98-GT-175 (1998). 21. VanFossen, G.J., 1982, “Heat-Transfer Coefficients for Staggered Arrays of Short Pin Fins,” ASME Journal of Engineering for Power, Vol. 104, pp. 268-274; also see note 19. 22. S.C. Arora and W. Abdel-Messeh, “Characteristics of Partial Length Circular Pin Fins as Heat Transfer Augmentors for Airfoil Internal Cooling Passages,” ASME Paper No. 89-GT-87 (1989). 23. D.E. Metzger, S.C. Fan, and S.W. Haley, “Effects of Pin Shape and Array Orientation on Heat Transfer and Pressure Loss in Pin Fin Arrays,” ASME Journal of Engineering for Gas Turbines and Power 106 (1984): 252-257. 24. M.K. Chyu,, “Heat Transfer and Pressure Drop for Short Pin-Fin Arrays With Pin-Endwall Fillet,” ASME Journal of Heat Transfer 112 (1990): 926-932. 25. R.J. Goldstein, M.Y. Jabbari, and S.B. Chen, “Convective Mass Transfer and Pressure Loss Characteristics of Staggered Short Pin-Fin Arrays,” International Journal of Heat and Mass Transfe 37 (1994): Suppl. 1, pp. 149-160. 26. M.K. Chyu, Y.C. Hsing, and V. Natarajan, “Convective Heat Transfer of Cubic Fin Arrays in a Narrow Channel,” ASME Journal of Turbomachinery 120 (1998): 362-367.
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes 27. D.E. Metzger, W.B. Shephard, and S.W. Haley, “Row Resolved Heat Transfer Variations in Pin-Fin Arrays Including Effects of Non-Uniform Arrays and Flow Convergence,” ASME Paper No. 86-GT-132 (1986). 28. M.K. Chyu, V. Natarajan, and D.E. Metzger, “Heat/Mass Transfer from Pin-Fin Arrays With Perpendicular Flow Entry,” ASME HTD-Vol. 226 (1992), Fundamentals and Applied Heat Transfer Research for Gas Turbine Engines. 29. T.K. Kumaran, J.C. Han, and S.C. Lau, “Augmented Heat Transfer in a Pin Fin Channel with Short or Long Ejection Holes,” International Journal of Heat Mass Transfer, 34 no. 10 (1991): 2617-2628. 30. Hwang, J.J. and Lu, C.C., 2000, “Lateral-Flow Effect of Endwall Heat Transfer and Pressure Drop in a Pin-Fin Trapezoidal Duct of Various Pin Shapes,” ASME Paper No. 2000-GT-0232. 31. M.K. Chyu, Y. Yu, H. Ding, J.P. Downs, and O. Soechting, “Concavity Enhanced Heat Transfer in an Internal Cooling Passage,” ASME Paper No. 97-GT-437 (1997); H.K. Moon, T. O’Connell, and B. Glezer, “Channel Height Effect on Heat Transfer and Friction in a Dimpled Passage,” ASME Paper No. 99-GT-163 (1999); G.I. Mahmood, M.L. Hill, D.L. Nelson, P.M. Ligrani, H.K. Moon, and B. Glezer,, “Local Heat Transfer and Flow Structure on and above a Dimpled Surface in a Channel,” ASME Journal of Turbomachinery 123 (2001): 115-123; S.W. Moon, and S.C. Lau, “Turbulent Heat Transfer Measurements on a Wall with Concave and Cylindrical Dimples in a Square Channel,” ASME Paper No.GT-2002-30208 (2002). 32. R.S. Bunker and K.F. Donnellan, “Heat Transfer and Friction Factors for Flows Inside Circular Tubes with Concavity Surfaces,” ASME Paper No. GT-2003-38053 (2003). 33. N. Syred, A. Khalatov, A. Kozlov, A. Shchukin, and R. Agachev, “Effect of Surface Curvature on Heat Transfer and Hydrodynamics within a Single Hemispherical Dimple,” ASME Paper No. 2000-GT-236 (2000). 34. G.M.S. Azad,Y. Huang, and J.C. Han, “Jet Impingement Heat Transfer on Dimpled Surfaces Using a Transient Liquid Crystal Technique,” AIAA Journal of Thermophysics and Heat Transfer 14 no. 2 (2000): 186-193; G.M.S. Azad, Y. Huang, and J.C. Han, “Impingement Heat Transfer on Pinned Surfaces Using a Transient Liquid Crystal Technique,” Proceedings of the 8th International Symposium on Transport Phenomena and Dynamics of Rotating Machinery 2 (2000): 731-738. 35. M.E. Taslim, L. Setayeshgar, and S.D. Spring, “An Experimental Evaluation of Advanced Leading Edge Impingement Cooling Concepts,” ASME Paper No. 2000-GT-222 (2000). 36. M.E. Taslim, Y. Pan, and S.D. Spring, “An Experimental Study of Impingement on Roughened Airfoil Leading-Edge Wall with Film Holes,” ASME Paper No. 2001-GT-0152 (2001). 37. M.E. Taslim and L. Setayeshgar, “Experimental Leading-Edge Impingement Cooling Through Racetrack Crossover Holes,” ASME Paper No. 2001-GT-153 (2001); M.E. Taslim,Y. Pan, and K. Bakhtari, “Experimental Racetrack Shaped Jet Impingement on a Roughened Leading-Edge Wall with Film Holes,” ASME Paper No. GT-2002-30477(2002). 38. B. Glezer, H.K. Moon, J. Kerrebrock, J. Bons, and G. Guenette, “Heat Transfer in a Rotating Radial Channel With Swirling Internal Flow,” presented at the International Gas Turbine & Aeroengine Congress and Exhibition, Stockholm, Sweden, June 2-5, ASME Paper No. 98-GT-214 (1998). 39. G. Pamula, S.V. Ekkad, and S. Acharya, “Influence of Cross-Flow Induced Swirl and Impingement on Heat Transfer in a Two-Pass Channel Connected by Two Rows of Holes,” ASME Paper No. 2000-GT-235 (2000). 40. S. Kieda, T. Torii, and K. Fujie, “Heat Transfer Enhancement in a Twisted Tube Having a Rectangular Cross Section With or Without Internal Ribs,” ASME Paper No. 84-HT-75 (1984). 41. Y.M. Zhang, G.M.S. Azad, J.C. Han, and C.P. Lee, “Heat Transfer and Friction Characteristics of Turbulent Flow in Square Ducts with Wavy, and Twisted Tape Inserts and Axial Interrupted Ribs,” Journal of Enhanced Heat Transfer 7 (2000): 35-49. 42. Y.M. Zhang, W.Z. Gu, and J.C. Han, “Heat Transfer and Friction in Rectangular Channels With Ribbed or RibbedGrooved Walls,” ASME Journal of Heat Transfer 116 no. 1 (1994): 58-65. 43. D.E. Metzger and C.S. Fan, “Heat Transfer in Pin-Fin Arrays With Jet Supply and Large Alternating Wall Roughness Ribs,” HTD-Vol. 226 (1992) Fundamental and Applied Heat Transfer Research for Gas Turbine Engines, pp. 23-30. 44. Z.J. Zuo, A. Faghri, and L. Langston, “A Parametric Study of Heat Pipe Turbine Vane Cooling,” presented at the International Gas Turbine and Aeroengine Congress and Exhibition, Orlando, Florida, June 2-5, ASME Paper No. 97-GT443 (1997). 45. S. Yamawaki, T. Yoshida, M. Taki, and F. Mimura, “Fundamental Heat Transfer Experiments of Heat Pipes for Turbine Cooling,” ASME Paper No. 97-GT-438 (1997). 46. J.L. Kerrebrock and D.B. Stickler, “Vaporization Cooling for Gas Turbines, the Return-Flow Cascade,” ASME Paper No. 98-GT-177 (1998). 47. G.B. Bruening and W.C. Chang, “Cooled Cooling Air Systems for Turbine Thermal Management,” ASME Paper No. 99GT-14 (1999). 48. J.C. Corman and T.C. Paul, “Power Systems for the 21st Century “H” Gas Turbine Combined Cycles,” GE Power Systems, Schenectady, New York, GER-3935 (1995): 1-12. 49. T. Guo,T. Wang, and J.L. Gaddis, “Mist/Steam Cooling in a Heated Horizontal Tube, Part 2: Results and Modeling,” ASME Paper No. 99-GT-145 (1999). 50. X. Li, J.L. Gaddis, and T. Wang, “Mist/Steam Heat Transfer in Confined Slot Jet Impingement,” ASME Paper No. 2000-
348
Je-Chin Han and Lesley M. Wright
349
GT-0221(2000); X. Li, J.L. Gaddis, and T. Wang, “Mist/Steam Cooling by a Row of Impinging Jets,” ASME Paper No. 2001-GT-0151(2001); X. Li, J.L. Gaddis, and T. Wang, “Mist/Steam Heat Transfer with Jet Impingement onto a Concave Surface,” ASME Paper No. GT-2002-30475(2002). 51. J.C. Han, L.R. Glicksman, and W.M. Rohsenow, “An Investigation of Heat Transfer and Friction for Rib-Roughened Surfaces,” International Journal of Heat and Mass Transfer 21 (1978): 1143-1156; J.C. Han, J.S. Park, and C.K. Lei, “Heat Transfer Enhancement in Channels with Turbulence Promoters,” ASME Journal of Engineering for Gas Turbines and Power 107 (1985): 628-635; J.C. Han, “Heat Transfer and Friction Characteristics in Rectangular Channels with Rib Turbulators,” ASME Journal of Heat Transfer 110 (1988): 321-328; J.C. Han and J.S. Park, “Developing Heat Transfer in Rectangular Channels with Rib Turbulators,” International Journal of Heat and Mass Transfer 31(1988): 183-195. 52. See note 51 (Han [1988]) and (Han & Park [1988]). 53. J.C. Han, Y.M. Zhang, and C.P. Lee, “Augmented Heat Transfer in Square Channels with Parallel, Crossed, and V-shaped Angled Ribs,” Journal of Heat Transfer, Transactions ASME 113 (1991): 590-596; J.C. Han and Y.M. Zhang, “High Performance Heat Transfer Ducts with Parallel and V-Shaped Broken Ribs,” International Journal of Heat and Mass Transfer 35 (1992): 513-523; S.V. Ekkad and J.C. Han, “Detailed Heat Transfer Distributions in Two-Pass Square Channels with Rib Turbulators,” International Journal of Heat and Mass Transfer 40 (1997): 2525-2537; S.V. Ekkad,Y. Yuang, and J.C. Han, “Detailed Heat Transfer Distributions in Two-Pass Smooth and Turbulated Square Channels with Bleed Holes,” International Journal of Heat and Mass Transfer 41 (1998):, 3781-3791. 54. S.C. Lau, R.T. Kukreja, and R.D. McMillin, “Effects of V-shaped Rib Arrays on Turbulent Heat Transfer and Friction of Fully Developed Flow in a Square Channel,” International Journal of Heat and Mass Transfer 34 (1991): 1605-1616.; M.E. Taslim, T. Li, and D.M. Kercher, “Experimental Heat Transfer and Friction in Channels Roughened with Angled, VShaped, and Discrete Ribs on Two Opposite Walls,” ASME Journal of Turbomachinery, 118 (1996): 20-28; X. Gao and B. Suden, “Heat Transfer and Pressure Drop Measurements in Rib-roughened Rectangular Ducts,” Experimental Thermal and Fluid Science 24 (2001): 25-34; D.H. Rhee, D.H. Lee, H.H. Cho, and H.K. Moon, “Effects of Duct Aspect Ratios on Heat /Mass Transfer with Discrete V-Shaped Ribs,” ASME Paper No. GT2003-38622(2003). 55. H.H. Cho, S.J. Wu, and H.J. Kwon, “Local Heat/Mass Transfer Measurements in a Rectangular Duct with Discrete Ribs,” ASME Journal of Turbomachinery 122 (2000): 579-586; also see note 54 (Rhee, et al.). 56. See note 53 (Han & Zhang). 57. J.C. Han, J.J. Huang, and C.P. Lee, “Augmented Heat Transfer in Square Channels with Wedge-Shaped and Delta-Shaped Turbulence Promoters,” Journal of Enhanced Heat Transfer 1 no.1 (1993): 37-52. 58. R.S Bunker and S.J. Osgood, “The Effect of Turbulator Lean on Heat Transfer and Friction in a Square Channel,” ASME Paper No. GT-2003-38137 (2003). 59. M.E. Taslim and S.D. Spring, “Effects of Turbulator Profile and Spacing on Heat Transfer and Friction in a Channel,” AIAA Journal of Thermophysics and Heat Transfer 8 no. 3 (1994): 555-562. 60. J.C. Bailey and R.S. Bunker, “Heat Transfer and Friction in Channels with Very High Blockage 45° Staggered Turbulators,” ASME Paper No. GT-2003-38611(2003); M.E. Taslim and S.D. Spring, “Experimental Heat Transfer and Friction Factors in Turbulated Cooling Passages of Different Aspect Ratios where Turbulators are Staggered,” AIAA Paper No. 88-3014, 24th Joint Propulsion Conference (1988). 61. J.R. Shen, Z. Wang, P.T. Ireland, T.V. Jones, and A.R. Byerley, “Heat Transfer Enhancement Within a Turbine Blade Cooling Passage Using Ribs and Combinations of Ribs With Film Cooling Holes,” ASME Journal of Turbomachinery 188 (1996): 428-433; D. Thurman and P. Poinsette, “Experimental Heat Transfer and Bulk Air Temperature Measurements for a Multipass Internal Cooling Model with Ribs and Bleed,” ASME Paper No. 2000-GT-233 (2000); also see note 53 (Ekkad, Huang, and Han). 62. See note 53 (Ekkad, Huang, and Han). 63. See note 61 (Shen, et al.). 64. J.H. Wagner, B.V. Johnson, and F.C. Kopper, “Heat Transfer in Rotating Serpentine Passages With Smooth Walls,” ASME Journal of Turbomachinery 113 (1991): 321-330; S. Dutta and J.C. Han, “Rotational Effects on the Turbine Blade Coolant Passage Heat Transfer,” Annual Review of Heat Transfer 9 (1997): 269-314. 65. J.C. Han, Y.M. Zhang, and K. Kalkuehler, “Uneven Wall Temperature Effect on Local Heat Transfer in a Rotating TwoPass Square Channel with Smooth Walls,” ASME Journal of Heat Transfer 114 (1993): 850-858. 66. Ibid. 67. J.A. Parsons, J.C. Han, and Y.M. Zhang, “Wall Heating Effect on Local Heat Transfer in a Rotating Two-Pass Square Channel with 90-Degree Rib Turbulators,” International Journal of Heat and Mass Transfer 37 no. 9 (1994): 1411-1420; Y.M. Zhang, J.C. Han, J.A. Parsons, and C.P. Lee, “Surface Heating Effect on Local Heat Transfer in a Rotating Two-Pass Square Channel with 60-Degree Angled Rib Turbulators,” ASME Journal of Turbomachinery 117 (1995): 272-278. 68. S. Acharya, V. Eliades, and D.E. Nikitopoulos, “Heat Transfer Enhancements in Rotating Two-Pass Coolant Channels with Profiled Ribs: Part 1 – Average Results,” ASME Paper No. 2000-GT-0227 (2000). 69. J.H. Wagner, B.V. Johnson, R.A. Graziani, and F.C. Yeh, “Heat Transfer in Rotating Serpentine Passages With Trips Normal to the Flow,” ASME Journal of Turbomachinery 114 (1992): 847-857. 70. B.V. Johnson, J.H. Wagner, G.D. Steuber, and F.C. Yeh, “Heat Transfer in Rotating Serpentine Passages with Trips Skewed to the Flow,” ASME Paper No. 92-GT-191, ASME Journal of Turbomachinery. 116 (1992): 113-123.
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes 71. C.W. Park, S.C. Lau, R.T. Kukreja, “Heat/Mass Transfer in a Rotating Two-Pass Square Channel with Transverse Ribs.” AIAA Journal of Thermophysics and Heat Transfer 12 (1998): 80-86; C.W. Park, C. Yoon, S.C. Lau, “Heat (Mass) Transfer in a Diagonally Oriented Rotating Two-Pass Channels with Rib-Roughened Walls,” ASME Journal of Heat Transfer 122 (2000): 208-211. 72. B.V. Johnson, J.H. Wagner,G.D. Steuber, and F.C. Yeh, “Heat Transfer in Rotating Serpentine Passages with Selected Model Orientations for Smooth or Skewed Trip Walls,” ASME Journal of Turbomachinery 116 (1994): 738-744. 73. J.A. Parsons, J.C. Han, and Y.M. Zhang, “Wall Heating Effect on Local Heat Transfer in a Rotating Two-Pass Square Channel with 90-Degree Rib Turbulators,” International Journal of Heat and Mass Transfer 37 (1994): 1411-1420. 74. S. Dutta and J.C. Han, “Local Heat Transfer in Rotating Smooth and Ribbed Two-Pass Square Channels With Three Channel Orientations,” ASME Journal of Heat Transfer 118 (1996): 578-584. 75. L. Al-Hadhrami and J.C. Han, “ Effect of Rotation in Two-Pass Square Channels with Parallel and Crossed 45 angled Rib Turbulators,” International Journal of Heat and Mass Transfer 46 (2003): 653-669. 76. J. Guidez, “Study of the Convective Heat Transfer in a Rotating Coolant Channel,” ASME Journal of Turbomachinery 111(1989): 43-50. 77. C.Y. Soong,, S.T. Lin, and G.J. Hwang, “An Experimental Study of Convective Heat Transfer in Radially Rotating Rectangular Ducts,” ASME Journal of Heat Transfer 113 (1991): 604-611. 78. M.E. Taslim, L.A. Bondi, and D.M. Kercher, D.M., “An Experimental Investigation of Heat Transfer in an Orthogonally Rotating Channel Roughened with 45 Deg Criss-Cross Ribs on Two Opposite Walls,” ASME Journal of Turbomachinery 113 (1991): 346-353. 79. G.M.S. Azad, J.M. Uddin, J.C. Han, H.K. Moon, and B. Glezer, “Heat Transfer in a Two-Pass Rectangular Rotating Channel with 45-Degree Angled Rib Turbulators,” ASME Journal of Turbomachinery 124 (2001): 251-259; L. AlHadhrami, T.S. Griffith, and J.C. Han, “Heat Transfer in Two-Pass Rotating Rectangular Channels (AR=2) with Parallel and Crossed 45° V-shaped Rib Turbulators,” ASME Journal of Heat Transfer 125 (2002): 232-242. 80. T.S. Griffith, L. Al-Hadhrami, and J.C. Han, “Heat Transfer in Rotating Rectangular Cooling Channels (AR=4) with Angled Ribs,” ASME Journal of Heat Transfer 124 (2001): 617-625. 81. E. Lee, L.M. Wright, and J.C. Han, “Heat Transfer in Rotating Rectangular Channels (AR = 4:1) with V-Shaped and Angled Rib Turbulators with and without Gaps,” ASME Paper No. GT-2003-38900 (2003). 82. L.M. Wright,W.L. Fu, and J.C. Han, “Thermal Performance of Angled, V-Shaped, and W-Shaped Rib Turbulators in Rotating Rectangular Cooling Channels (AR=4:1),” ASME Paper No. GT 2004-54073 (2004). 83. H.H. Cho,Y.Y. Kim, K.M. Kim, and D.H. Rhee, “Effects of Rib Arrangements and Rotation Speed on Heat Transfer in a Two-Pass Duct,” ASME Paper No. GT 2003-38609 (2003). 84. P. Agarwal, S. Acharya, and D.E. Nikitopoulos, “Heat/Mass Transfer in 1:4 Rectangular Passages with Rotation,” ASME Paper No. GT 2003-38615 (2003). 85. W.L. Fu, L.M. Wright, and J.C. Han, “Heat Transfer in Two-Pass Rotating Rectangular Channels (A=1:2 and AR=1:4) with Smooth Walls,” ASME Journal of Heat Transfer 127 (2005): 209-356; W.L. Fu, L.M. Wright, and J.C. Han, “Heat Transfer in Two-Pass Rotating Rectangular Channels (AR=1:2 and AR=1:4) with 45° Angled Rib Turbulators,” ASME Paper No. GT 2004-53261 (2004); W.L. Fu, L.M. Wright, and J.C. Han, “Buoyancy Effects on Heat Transfer in Five Different Aspect-Ratio Rectangular Channels with Smooth Walls and 45-Degree Ribbed Walls,” ASME Paper No. GT 2005-68493 (2005). 86. R.J. Clifford, W.D. Morris, and S.P. Harasgama, “An Experimental Study of Local and Mean Heat Transfer in a Triangular-Sectioned Duct Rotating in the Orthogonal Mode,” ASME Journal of Engineering for Gas Turbines and Power 106 (1984): 661-667. 87. S.P. Harasgama and W.D. Morris, “The Influence of Rotation on the Heat Transfer Characteristics of Circular, Triangular, and Square-Sectioned Coolant Passages of Gas Turbine Rotor Blades,” ASME Journal of Turbomachinery 110 (1988): 44-50. 88. S. Dutta, J.C. Han, Y.M. Zhang, and C.P. Lee, “Local Heat Transfer in a Rotating Two-Pass Triangular Duct with Smooth Walls,” ASME Journal of Turbomachinery 118 (1996): 435-443; S. Dutta, J.C. Han, and C.P. Lee, “Local Heat Transfer in a Rotating Two-Pass Ribbed Triangular Duct with Two Model Orientations,” International Journal of Heat and Mass Transfer 39 (1996): 707-715. 89. L. Rathjen, D.K. Hennecke, S. Bock, S., and R. Kleinstuck, “Detailed Heat/Mass Transfer Distributions in a Rotating Two Pass Coolant Channel with Engine-Near Cross section and Smooth Walls,” in Heat Transfer in Gas Turbine Systems, edited by Richard J. Goldstein, Annals of the New York Academy of Sciences 934 (2001): 432-439. 90. See note 95. 91. See note 81. 92. See note 85 (ASME Paper No. GT 2005-68493). 93. G. Su, H.C. Chen, J.C. Han, and D. Heidmann, “Computation of Flow and Heat Transfer in Two-Pass Rotating Rectangular Channels (AR=1:1, AR=1:2, AR=1:4) with 45-Deg Angled Ribs by a Reynolds Stress Turbulence Model,” ASME Paper No. GT2004-53662 (2004). 94. See note 85. 95. Ibid.
350
Je-Chin Han and Lesley M. Wright 96. Ibid. 97. C. Prakash and R. Zerkle, “Prediction of Turbulent Flow and Heat Transfer in a Radially Rotating Square Duct,” ASME Journal of Turbomachinery 117 (1992): 255-261. 98. Y.L. Lin, T.I-P Shih, M.A. Stephens, and M.K. Chyu, “A Numerical Study of Flow and Heat Transfer in a Smooth and Ribbed U-Duct With and Without Rotation,” ASME Journal of Heat Transfer 123 no. 2 (2001): 219-232. 99. H.C. Chen, Y.J. Jang, and J.C. Han, “Computation of Flow and Heat Transfer in Rotating Two-Pass Square Channels by a Reynolds Stress Model,” International Journal of Heat and Mass Transfer 43 no. 9 (2000): 1603-1616; H.C. Chen, Y.J. Jang, and J.C. Han, “Near-Wall Second-Moment Closure for Rotating Multiple-Pass Cooling Channels,” AIAA Journal of Thermophysics and Heat Transfer 14 no. 2 (2000): 201-209. 100. Y.J. Jang, H.C. Chen, and J.C. Han, “Flow and Heat Transfer in a Rotating Square Channel with 45-Degree Angled Ribs by Reynolds Stress Turbulence Model,” ASME Journal of Turbomachinery 123 no. 1 (2001): 124-132. 101. M. Al-Qahtani,Y.J. Jang, H.C. Chen, and J.C. Han, “Prediction of Flow and Heat Transfer in Rotating Two-Pass Rectangular Channels with 45-Degree Rib turbulators,” ASME Journal of Turbomachinery 124 ( April, 2002): 242-250. 102. G. Su, S. Tang, H.C. Chen, and J.C. Han, “Flow and Heat Transfer Computations in Rotating Rectangular Channels with V-Shaped Ribs,” AIAA Journal of Thermophysics and Heat Transfer 18 no. 4 (2004): 534-547. 103. See note 101. 104. See note 93. 105. D.K. Lezius and J.P. Johnston, “Roll-Cell Instabilities in Rotating Laminar and Turbulent Channel Flows,” Journal of Fluid Mechanics 77 (1976): 153-175. 106. M. Elfert, “The Effect of Rotation and Buoyancy on Flow Development in a Rotating Circular Coolant Channel,” 2nd International Symposium on Engineering Turbulence Modeling and Measurements, May 31-June 2, 1993, Florence, Italy. 107. D.G.N. Tse and D.B. McGrath, “A Combined Experimental/Computational Study of Flow in Turbine Blade Cooling Passage. Part I: Experimental Study,” ASME Paper No. 95-GT-355 (1995). 108. D.G.N. Tse and G.D. Steuber, “Flow in a Rotating Square Serpentine Coolant Passage with Skewed Trips,” ASME Paper No. 97-GT-529 (1997). 109. S.C. Cheah, H. Iacovides, D.C. Jackson, H. Ji,,and B.E. Launder, “LDA Investigation of the Flow Development through Rotating U-Ducts,” ASME Journal of Turbomachinery 118 (1996): 590-595. 110. J.P. Bons and J.L. Kerrebrock,, “Complementary Velocity and Heat Transfer Measurements in a Rotating Cooling Passage with Smooth Walls,” ASME Paper No. 98-GT-464 (1998). 111. J. Schabacker, A. Bolcs, and B.V. Johnson, “PIV Investigation of the Flow Characteristics in an Internal Coolant Passage with 45° Rib Arrangement,” ASME Paper No. GT 99-120 (1999); D. Chanteloup, Y. Yuaneda, and A. Bolcs, “Combined 3D Flow and Heat Transfer Measurements in a 2-pass Internal Coolant Passage of Gas Turbine Airfoil,” ASME Paper No. GT 2002-30214 (2002). 112. S.Y. Son, K.D. Kihm, and J.C. Han, “PIV Flow Measurements for Heat Transfer Characterization in Two-Pass Square Channels with Smooth and 90-degree Ribbed Walls,” International Journal of Heat and Mass Transfer 45 no. 24 (2002): 4809-4822. 113. T.M. Liou and C.C. Chen, “LDV Study of Developing Flows through a Smooth Duct with a 180 Deg Straight-Corner Turn,” ASME Journal of Turbomachinery 121 (1999): 167-174. 114. T.M. Liou, M.Y. Chen, and M.H. Tsai, “Fluid Flow and Heat Transfer in a Rotating Two-Pass Square Duct with In-Line 90° Ribs,” ASME Paper No. GT 2001-0185 (2001). 115. T.M. Liou, M.Y. Chen, and Y.M. Wang, “Heat Transfer, Fluid Flow, and Pressure Measurements inside a Rotating Duct with Detached 90° Ribs,” ASME Paper No. GT 2002-30201(2002). 116. Liou, T. M., and Dai, G. Y., 2003, “Pressure and Flow Characteristics in a Rotating Two-Pass Square Duct with 45 Deg Angled Ribs,” ASME Paper No. GT 2003-38346. 117. T.M. Liou, Y.S. Hwang, and Y.C. Li, “Flowfield and Pressure Measurements in a Rotating Two-Pass Duct with Staggered Rounded Ribs Skewed 45° to the Flow,” ASME Paper No. GT 2004-53173 (2004). 118. See notes 116 and 117. 119. A.H. Epstein, J.L. Kerrebrock, J.J. Koo, and U.Z. Preiser, “Rotational Effects on Impingement Cooling,” GTL Report No. 184 (1985). 120. C.H. Mattern and D.K. Hennecke, “The Influence of Rotation on Impingement Cooling,” ASME Paper No. 96-GT-161 (1996). 121. B. Glezer, H.K. Moon, J. Kerrebrock, J.Bons, and G. Guenette, “Heat Transfer in a Rotating Radial Channel with Swirling Internal Flow,” ASME Paper No. 98-GT-214 (1998). 122. J.A. Parsons, J.C. Han, and C.P. Lee, “Rotation Effect on Jet Impingement Heat Transfer in Smooth Rectangular Channels with Four Heated Walls and Radially Outward Crossflow,” ASME Journal of Turbomachinery 120 (1998): 7985. 123. K.V. Akella and J.C. Han, “Impingement Cooling in Rotating Two-Pass Rectangular Channels,” AIAA Journal of Thermophysics and Heat Transfer 12 no. 4 (1998): 582-588.
351
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes 124. J.A. Parsons and J.C. Han, “Rotation Effect on Jet Impingement Heat Transfer in Smooth Rectangular Channels with Heated Target Walls and Film Coolant Extraction,” ASME Paper No. 96-WA/HT-9 (1996). 125. K.V. Akella and J.C. Han, “Impingement Cooling in Rotating Two-Pass Rectangular Channels with Ribbed Walls,” AIAA Journal of Thermophysics and Heat Transfer 13 no. 3 (1999): 364-371. 126. J.A. Parsons, J.C. Han, and C.P. Lee, “Rotation Effect on Jet Impingement Heat Transfer in Smooth Rectangular Channels with Four Heated Walls and Film Coolant Extraction,” ASME Paper No. GT-2003-28905 (2003). 127. F.T. Willett and A.E. Bergles, “Heat Transfer in Rotating Narrow Rectangular Ducts with Heated Sides Oriented at 60Degree to the R-Z Plane,” ASME Paper No. 2000-GT-224 (2000); F.T. Willett and A.E. Bergles, “Heat Transfer in Rotating Narrow Rectangular Pin-Fin Ducts,” Experimental Thermal and Fluid Science 25 (2002): 573-582. 128. L.M. Wright, E. Lee, and J.C. Han,, “Effect of Rotation on Heat Transfer in Narrow Rectangular Cooling Channels (AR = 8:1 and 4:1) with Pin-Fins,” ASME Paper No. GT 2003-38340 (2003). 129. F. Zhou and S. Acharya,, “Mass/Heat Transfer in Dimpled Two-Pass Coolant Passages with Rotation,” Heat Transfer in Gas Turbine Systems, ed., R. J. Goldstein, Ann. N. Y. Acad. Sci., 934 (2001): 424-431. 130. T.S. Griffith, L. Al-Hadhrami, and J.C. Han, “Heat Transfer in Rotating Rectangular Cooling Channels (AR = 4) with Dimples,” ASME Journal of Turbomachinery 125 (2003): 555-563.
352
BIOGRAPHY
4.2.2.2 Enhanced Internal Cooling of Turbine Blades and Vanes
Je-Chin Han Turbine Heat Transfer Laboratory Department of Mechanical Engineering Texas A&M University College Station, Texas 77843-3123 phone: (979) 845-3738 email: [email protected]
Dr. Han is presently the Marcus Easterling Endowed Chair Professor at Texas A & M University. He has been working on many turbine blade cooling projects since 1980, including innovative cooling concepts, rotor coolant passage heat transfer, unsteady high turbulance effects on blade film cooling and turbine edge cooling (leading edge, trailing edge, blade tip, and end-wall regions). This research has been on both aircraft engines and industrial gas turbine applications funded by NASA- Glenn, GE - aircraft engines, DOE - AGTSR + UTSR, and Siemens Westinghouse. He is the co-author of more than 150 referreed journal papers and one book in gas turbine heat transfer and cooling technology. He has been doing research on gas turbine blade cooling for more than 25 years.
Lesley M. Wright Turbine Heat Transfer Laboratory Department of Mechanical Engineering Texas A&M University College Station, Texas 77843-3123 phone: (979) 845-1767 email: [email protected]
Lesley Wright is a Ph.D. candidate in the Department of Mechanical Engineering at Texas A&M University. She joined the Turbine Heat Transfer Laboratory in September of 2001 after obtaining her B.S. in Engineering from Arkansas State University. While experimentally investigating the effect of rotation on internal gas turbine blade heat transfer, she earned her M.S. in Mechanical Engineering from Texas A&M in May of 2003. Lesley continues to study both internal and external turbine blade heat transfer; in addition, she is considering new experimental methods and their application to better understand gas turbine heat transfer and cooling techniques.
4.2.3
Airfoil Endwall Heat Transfer
4.2.3-1 Introduction The flow in a gas turbine influenced by the inner hub and outer casings of the airfoils is defined as secondary or endwall flows. These flows often contain vortices that give rise to velocity components that are orthogonal to the primary flow direction as depicted in figure 1 by the ribbon arrows, which is specific to vane endwalls. These flows constitute one of the most commonplace and widespread three-dimensional flows arising in axial flow turbomachinery. In typical modern day turbine designs, endwall flows for first stage vanes are responsible for over 30% of the total pressure loss through a turbine stage leading to a reduction in turbine efficiencies on the order of 3%. While overall airfoil losses have been reduced through the use of threedimensional geometries that make use of bowed or leaned airfoils, for example, the endwalls have remained fairly conventional and the source of much of the remaining pressure losses. The heat transfer consequences are immense because of the increased convective coefficients and mixing out of film-coolant near the surface. It is clear from a thermodynamic analysis of a turbine engine, that to improve performance, there is a need to increase the aspect ratio for turbine airfoils and to increase turbine inlet temperatures. To improve a turbine’s performance, these trends require that endwall flows be carefully considered in turbine designs. The endwall flow through an airfoil cascade under isothermal conditions with an approaching two-dimensional boundary layer agrees well with that depicted in figure 1. The flow model shows that the inlet boundary layer separates from the approaching endwall to form what is known as a horseshoe vortex. One leg of the horseshoe vortex, present on the pressure side of the airfoil (concave side), is convected into the passage and is promoted by the inherent pressure gradient between the two airfoil surfaces. This pressure side leg of the horseshoe vortex develops into what is known as the passage vortex. The other leg of the horseshoe vortex, present on the suction side of the airfoil (convex side), has an opposite sense of rotation to the larger passage vortex and develops into what is known as a counter vortex. The counter vortex can be thought of as a planet rotating about the axis of the passage vortex (sun). The actual rotation of the vortices depicted in figure 1 were drawn to exaggerate the vortex motion and for the passage vortex is generally about two rotations before exiting the airfoil passage. While this picture represents a timeaveraged representation, measured data indicates that the vortex is not steady. While the development of the vortical structures originates in the endwall regions, the growth can be such that the passage vortex occupies a large portion of the airfoil exit. This
Karen Thole Mechanical Eng. Dept, Penn State Univ. University Park, PA 16802-1412 phone: (814) 865-2519 email: [email protected]
Fig. 1. Classic secondary flow pattern for a turbine airfoil passage.(reproduced with permission from American Society of Mechanical Engineers [ASME]). Source: Langston, L. S. “Crossflows in a Turbine Cascade Passage,” ASME J of Engineering for Power 102 (1980): 866 - 874.
353
Fig. 2. Illustration of the near wall flows as taken through oil and dye surface flow visualization (reproduced with permission of the publisher from ASME). Source: Friedrichs, S., Hodson, H. P. and Dawes, W. N., “Distribution of Film-Cooling Effectiveness on a Turbine Endwall Measured Using the Ammonia and Diazo Technique,” J of Turbomachinery 118 (1996): 613-621.
vortical structure growth extends up to 30-40% of the total span for older vane designs and has been reduced to approximately 10-15% of the total span in the last 15 years. The flow patterns previously described make it difficult to cool the endwall, particularly when considering the near endwall flow, as illustrated in figure 2. The surface flow visualization in figure 2, achieved through an oil and surface dye technique, illustrates 1 0.8 0.6 0.4 0.2 -0.2 -0.4 -0.6 the strong cross flows that occur. The cross flows are driven by the inherent pressure gradients from the pressure0to the suction side of adjacent airfoils. For example, these cross flows influence how the film-cooling jets exit from the holes as well as influence the endwall heat transfer coefficients. This section briefly describes a simplified theoretical approach to understanding vane endwall flows. Following this brief discussion the following will be provided: heat transfer coefficients experienced along a turbine vane endwall with varying effects, filmcooling designs, and geometrical means for reducing endwall heat transfer. Outer casing flows of vanes are similar to inner hub flows; however, for blades the outer casing flows encompass yet another difficult region which is the blade tip region. Generally, there is a gap between the blade tip and outer casing that provides for a leakage flow that is inherently driven by the pressure difference between the pressure and suction sides of the blades. A brief section about tip flows will be provided at the end of this section.
4.2.3-2 Theoretical Development of Endwall Flows As the endwall boundary layer approaches a turbine airfoil, the flow stagnates, whereby the total pressure becomes the static pressure along the span of the vane. Given that the fluid nearer to the endwall has a lower velocity, a stronger deceleration in the boundary layer occurs for the higher speed fluid than for the lower speed fluid. As a result of these differences in the deceleration, a transverse static pressure gradient occurs along the vane span causing the higher speed fluid to turn toward the endwall plate. Subsequently, the formation of a horseshoe vortex occurs just upstream of the turbine vane. One of the legs of the horseshoe vortex wraps around the pressure side of the vane and the other leg wraps around the suction side of the vane. Figure 3a shows measurements of the horseshoe vortex upstream of the turbine vane and figure 3b shows extreme damage from the effects of this vortex on actual hardware. These measurements were made for a simple case with an isothermal flow and an approaching boundary layer that was 9 % of the vane span (Z/S = 0.09). From the combined contour/vector plot in figure 3a, one can see where the flow separates from the upstream endwall and how the flow is then rolled into a vortex. It is also clear to see that, if a film-cooling jet were injected into this region, it would be difficult to maintain the coolant along the endwall. The flow fields discussed thus far have described the condition for a uniform, isothermal flow field with an approaching two-dimensional boundary layer along the endwall. In practice, this idealized flow situation rarely happens since an upstream combustor is present whereby there can be large variations in the exiting flow. Non-uniformity of inlet profiles in addition to the viscous boundary layers along the endwall are caused by temperature gradients at the combustor exit. The development of crossflow and vortical motions in a curved passage, such as an airfoil passage, can be understood by considering flow along two streamlines, as shown in figure 4. Two idealized cases are considered: a) a gradient of velocity due to a turbulent inlet boundary layer with an isothermal flow, b) a linear temperature gradient typical at the exit of a combustor with a uniform velocity field. Assuming steady, incompressible, inviscid flow with negligible variation of velocity in the n-direction, the centripetal acceleration for the streamlines A and B must be balanced by the pressure gradient across the pitch:
U/U
(U2+W2)1/2/Uh = 1
0.15
h 1.0
0.8
0.6
0.10
0.4
Z/S
0.2
0.05
Horseshoe Vortex Damage
0.00 -0.20 1
-0.15 0.8
0.6
-0.10 X/C 0.4
0.2
-0.05 0
0.0
-0.2 0.00
-0.4
-0.6
-0.2 -0.4 -0.6
Fig. 3a. Measurements of the horseshoe vortex just upstream of the vane at the vane-endwall juncture (reproduced with permission from ASME). Horseshoe Source: M. Kang, A. Kohli, and K.A. Thole, “Heat Vortex Transfer and Flowfield Measurements in the Leading Damage Edge Region of a Stator Vane Endwall,” J of Turbomachinery 121 (1999): 558-568.
Fig. 3b. Actual hardware showing effects of the horseshoe vortex on a first vane
(1)
354
Karen Thole
From boundary layer theory, for both streamlines from the developing boundary layer on the sidewalls should be equal. Therefore
.
(2)
Due to the viscous turbulent boundary layer at the endwall, it is evident that VsB < VsA. So for the boundary layer assumption to hold, the radius of curvature of the streamline at B must be reduced. This creates a crossflow in the boundary layer from the pressure surface towards the suction surface of the blade, and thus generates secondary flow (flow that is not aligned with the streamwise direction) as depicted in figure 4a. Now consider figure 4b, a constant velocity profile with linear temperature profile. The same physics hold for this case; however, the resulting vortex is reversed in direction. In this instance the temperature at A is greater than at B, therefore ρA < ρB. Now RB must be greater than RA for the normal pressure gradient to balance and the cross flow is generated towards the pressure side of the adjacent vane row. The change in streamline curvature would be less severe in this case compared to figure 4a since the velocity term is squared in the relationship of Equation 1. As one can see from these simple, idealized flow situations, there can be large variation in the expected secondary flow pattern that can be derived in a turbine vane passage. The important driver for how the flow develops in a turbine vane passage is the total pressure profile entering the passage. As this total pressure profile becomes the static pressure along the vane stagnation, the flow will be driven from a high pressure region to a low pressure region. In most turbine designs there is a flow leakage slot between the combustor and the turbine whereby cooler fluid is injected into the main hot gas path. This leakage can also have an effect on the secondary flow patterns that develop. The bottom line when considering endwall flows is that the profile exiting the combustor, which is often referred to as the combustor pattern factor that ultimately enters the turbine, should be known to fully predict the secondary flows that will develop in the turbine passage. In practice, the combustor pattern factor is one of the parameters used in designing cooling schemes for the airfoils and their associated platforms.
4.2.3-3 Endwall Heat Transfer The heat transfer coefficients given in figure 5 are represented in terms of a non-dimensional Stanton number based on exit velocity. In the region upstream of the vanes, there is a high heat transfer region that occurs between the stagnation point and the reattachment of the flow on the suction side of the airfoil. This is the area which experiences very high acceleration. As the flow moves through the passage, it is apparent that the location of the peak Stanton numbers (peak heat transfer) is being swept from the outer pressure surface towards the suction side of the central vane. Figure 6 presents the Stanton numbers for two freestream turbulence levels (Tu = 0.6% and 19.5%) approaching the stagnation location along a line parallel to the incoming velocity vector. Figure 6 illustrates that at a location of approximately 15% of the chord (∆X/C = 0.15) upstream of the vane, the Stanton numbers begin to increase dramatically. This increase can be related to the position where the flow separates from the endwall, as shown by Figure 3a, which occurs at approximately X/C = 0.12. The heat transfer continues to increase as the flow approaches the vane. Depending upon the particular combustor design, the turbulence levels exiting a combustor that enter into the turbine can be relatively high. In general, high levels of turbulence lead to higher levels of heat transfer. Figure 6 illustrates this increase through both the Stanton numbers (left axis) and the augmentation of Stanton numbers (right axis) whereby augmentation refers to the Stanton number (convective heat transfer coefficients) at high turbulence divided by the Stanton number at low turbulence levels. Figure 6 illustrates that there is an approximate 20% increase in heat transfer resulting from a turbulence level of 19.5%.
355
Fig. 4. Illustration of different vortical patterns that are possible for two idealized flow conditions: a) isothermal with an inlet boundary layer and b) inviscid flow with a temperature profile. Source: B. Lakshminarayana, Fluid Dynamics and Heat Transfer of Turbomachinery (NY: Wiley Interscience, 1996). St =
h [ x103 ] � C p U ex
1.00 0.75 0.50 0.25 Y/C 0.00 -0.25 -0.50 -0.75
20 18 16 14 12 10 8 6 4
-1.00 -0.25 0.00 0.25 0.50 0.75 X/C Fig. 5. Contours of non-dimensional heat transfer coefficients (reproduced with permission from the publisher of ASME). Source: See fig. 3a above.
4.2.3 Airfoil Endwall Heat Transfer 10 14
12 8 10
2 1.8 Tu =19.5% Tu = 19.5% Tu =1%
o
1.61.4
St
St/StSt/Sto 1.41.2
4 6
o
1.0 1.2
4 2 00.2
0.1
0.1 0.2
0.08 0.3 0.06
x/C | ∆ X| / C
0.4 0.04
1 0.5
2
Tu = 19.5% Tu = 0.6% Augmentation, St/St
12
1.81.6
Tu = 0.6% o Augmentation, St/St
Augmentation, St/St
St 6 (x 10e-03) St 3 8 (x10 )
2 0.3 -0.1
14
0.8
1.8
o
10
1.6
(x 10e-03) 8
St/St
6
1.2
4 2 -0.1
0
0.1
0.2
x/C
0.3
0.4
1 0.5
Fig. 6. Non-dimensional heat transfer coefficients (left axis) along the endwall approaching the vane leading edge for low (Sto) and high freestream turbulence (St); and augmentation factors (right axis) for heat transfer coefficients (reproduced with permission from the publisher of ASME).
Fig. 7. Pitchwise averaged non-dimensional heat transfer coefficients (left axis) for low and high freestream turbulence; and augmentation factors (right axis) for heat transfer coefficients (reproduced with permission from the publisher of ASME).
Source: R. Radomsky and K. Thole, “High Freestream Turbulence Tu = 19.5% 12 Effects in theTuEndwall J. of = 0.6% Leading Edge Region,” ASME 1.8 (2000): 699-708. Turbomachinery 122 Augmentation, St/St
Source: See fig. 6 above.
14
10
2
o
1.4
o
Figure 7 compares the heat transfer coefficients averaged over the pitch (vane-to-vane) for a range of axial positions. As the St 1.6 (x 10e-03) flow accelerates through the passage, the heat transfer coefficients also increase, which is to be expected. For most of the axial distance, o the8 augmentation due to freestream turbulence is relativelySt/St constant at 25% above the low freestream turbulence case. Only near the end of the vane where the passage vortex dominates, beyond 1.4 X/C of 0.38, is there a decrease in the augmentation. 6 Note that the heat transfer data shown in this paper were taken with a two-dimensional inlet boundary layer under low speed conditions, to allow for highly-resolved data, with matched Reynolds number conditions. There is evidence in the literature that the 1.2 4 secondary flow patterns remain the same at both low and high Mach number conditions. Rather, these secondary flows are a stronger function of the airfoil geometry and inlet profile conditions that of Mach number. There is no data in the literature that discusses how the 2 transfer on the endwall is altered depending upon an1inlet flow condition that is relevant to that exiting a combustor. As was stated heat -0.1 0 0.1 0.2 0.3 0.4 0.5 previously, it is important tox/C consider that the profile exiting the combustor can vary greatly from that of a two-dimensional boundary layer assumption.
4.2.3-4 Endwall Film-Cooling One method of combating the high heat transfer coefficients along the endwall is through the use of film-cooling holes whereby cooler air is injected through discrete holes in the endwall. Film-cooling hole placement, particularly in the endwall region, has traditionally been based upon designer experience whereby a number of design variables are considered. For example, one should take into account leakage from a slot at the combustorturbine interface whereby cooler gases leak into the main gas path (combustor film-cooling carryover). Most turbines are designed such that pressures outside the main gas path are higher than those found in the main gas path to insure that the hot combustion gases are not ingested below the platform. If designed properly, this leakage flow could be relied upon as a source of coolant. It is also important to remember that the secondary flows that develop are affected by the leakage flows at the vane inlet, as indicated by the previous discussion in this section. Other design variables include roughness effects, film-cooling migration (as will be discussed), film-cooling limitations resulting from internal cooling schemes, and cooling hole manufacturing. Manufacturing of cooling is generally done through the use of electro-discharge machining (EDM) or laser drilling with both being subjected to access limitations. Generally, EDM is more expensive than laser drilling and conversely laser drilling can tend to be more limited in terms of hole geometries that can be manufactured. Consider a leakage slot flow between the combustor and turbine whereby the coolant is 0.75% of the core flow. The cooling to the endwall that can be provided is shown in figure 8. Coolant exits across much of the width of the slot albeit in a very nonuniform manner. This non-uniform slot flow arises from the static pressure distribution and secondary flow development along the endwall. Although there is a large uncooled region,
Fig 8. Measured adiabatic wall temperatures for coolant exiting a combustor/vane leakage slot (reproduced with permission from the publisher of ASME). Source: D.G. Knost and K. Thole, “Adiabatic Effectiveness Measurements of Endwall FilmCooling for a First Stage Vane,” J. of Turbomachinery 127 (2005): 297305.
356
Karen Thole often referred to as a bow wake, around much of the vane at the leading edge and along the pressure side of the airfoil, there is also a well-cooled region in the center of the passage. As the slot flow increases, the cooling potential also increases to a point after which the benefit is small. As was previously mentioned, endwall film-cooling has largely been based on designer experience. One difficulty in designing the film-cooling hole pattern is knowing beforehand the local static pressure along the endwall, which varies greatly along the endwall as the flow accelerates through the passage. If one considers that a single supply feeds all the film-cooling holes and inviscid flow through the holes, it can be shown that a global, ideal (loss-free) blowing ratio for all the film-cooling holes is given by the following equation, (3) where Po,in and ps,in are defined at the inlet to the turbine. If one then wants to compute the local, ideal blowing ratio for each cooling hole (now taking into consideration that the local velocity varies), the following equation can be used (4)
where ps,∞ is the local static pressure defined at the exit of the cooling hole. Equation 4 indicates that the same blowing ratio will occur for each hole placed along a constant static pressure line. Designing an endwall cooling pattern to achieve a uniform blowing ratio is one methodology that some companies have used.
T − Taw η= ∞ T∞ − Tc
9a
9b
9c
Fig. 9. Contours of adiabatic effectiveness for two film-cooling hole patterns (left and center) with a mid-passage gutter for the cooling hole pattern in the center (right) (reproduced with permission from the publisher of ASME). Source: See figure 8 above.
357
Figure 9a shows a cooling hole pattern whereby the holes in the passage were placed along a constant static pressure line. Figure 9b shows an endwall cooling hole pattern with holes that were placed on lines parallel to the incoming flow direction. The endwall in Figure 9b also provides the space for including a mid-passage gap that occurs between vanes as the vanes are mated on the turbine disk. Vanes are generally cast in either doublets or singlets and then placed in a turbine disk with some type of seal between the vane platforms. Generally, relatively low levels of coolant leak from this mid-passage gap (less that 0.3%). Figure 9c shows the same film-cooling hole pattern as that in figure 9b with the exception being the mid-passage gap is present in figure 9c. The coolant flow conditions for all three vanes shown in figure 9 are the same with 0.75% of the core flow exiting in the form of coolant from the upstream slot, 0.5% coolant exiting the film-cooling holes and no net coolant flow through the mid-passage gap for figure 9c. The contours plotted in figure 9 represent the non-dimensional adiabatic wall temperatures. As discussed in the previous section, adiabatic surface temperatures represent the local fluid temperature. From all three contour plots in figure 9, it is clear that the upstream slot can definitely be an important part of cooling mid portion of the endwall. In comparing figure 9c to the rest, however, it
4.2.3 Airfoil Endwall Heat Transfer is also seen that the mid-passage gap limits the area of the cooling present. The midpassage gap provides a convective coolant trough through which coolant from the upstream slot enters and passes through the gap along the mid-passage before exiting the gap near the trailing edge of the airfoil. The streamlines superimposed upon the experimental measurements of the adiabatic wall temperatures presented in figure 9, were computed using computational fluid dynamics (CFD). These streamlines were extracted from the CFD simulations at a location very near to the endwall (less than 2% of the span). The streamlines, for the most part, can be used to predict the trajectory of the coolant flow exiting the cooling holes. There are some regions, however, where the streamlines differ from the jet trajectories, such as at the entrance to the passage closer to the pressure side. In comparing all of the cooling hole patterns in figure 9, it is clear that more uniform cooling can be achieved by placing cooling holes along constant static pressure lines as in figure 9a. All three patterns illustrate the difficulty of cooling the endwall along the pressure side of the vane and along the leading edge-endwall juncture. In general these are very difficult areas to cool because of the secondary flows that were described previously. The horseshoe vortex in the stagnation region can lead to coolant being swept upstream of the holes at low blowing ratios, whereas at high blowing ratios the coolant separates from the endwall and impacts the vane surface rather than the endwall surface. Despite the cooling holes injecting coolant towards pressure side of the vane, the sweeping motion of the passage vortex prevents cooling at the juncture between the pressure side of the vane and the endwall.
Fig.10. Example of contoured endwall for a first vane design Source: W. Colban, K. Thole, and M. Haendler, “Heat Transfer and Film-Cooling Measurements on a Stator Vane with FanShaped Cooling Holes,” International Gas Turbine and Aeroengine Congress and Exposition, Reno, GT2005-68258.
4.2.3-5 Leading Edge Modifications Because industry is concerned with problems at the vane leading edgeendwall juncture, a number of more recent studies have begun to evaluate geometric modifications to airfoils in this region. Both overall vane endwall contouring, as shown in figure 10, as well as more specific geometric features particular to the leading edge region have been considered. At this point there have been three different geometric concepts tested for an asymmetric airfoil geometry that include the following classifications: fences, fillets, and bulbs. Methods using flow control such as suction combined with injection have also been reported but are generally not feasible for gas turbines where gas temperatures exceed airfoil melting temperatures. Chung and Simon first presented their concept for secondary flow control in 1993 that encompassed using a fence placed in mid-passage between two turbine airfoils1. While their tests indicated a reduction in strength of the passage vortex, industry’s concern was in cooling the fence and that it acted as a fin conducting heat to the platform as it was exposed to the hotter main gas path fluid. While the fillet and bulb appear to be similar, as shown in figure 11a, the physical concepts behind why there are these two different modifications are quite different. Sauer et al. and Becz et al. studied the effects of various sized leading edge endwall bulbs as shown in figure 11a.2 The objective of this modification was to intensify the suction side branch of the horseshoe vortex and, through an interaction of the stronger suction side branch with the passage vortex, weaken the passage vortex. Becz et al. found that when compared with a fillet geometry, such as that shown in figure 11b, there were slightly lower aerodynamic losses for a fillet rather than a bulb3. The philosophy behind the Zess and Thole modification, which was a fillet as shown in figure 11b, was to eliminate the leading edge vortex thereby delaying the development of the passage vortex4. They presented a direct comparison between measured and predicted flow fields for an airfoil with no fillet and an airfoil with a fillet design. Their results indicated no presence of a leading edge vortex for a filleted airfoil and a delayed passage vortex. Moreover, their experimental results indicated a large reduction (by more than an order of magnitude in some locations) of turbulence levels for a filleted airfoil. The only papers to have considered the endwall heat transfer effects has been that of Shih and Lin, Lethander and Thole, and Han and Goldstein (who both performed computational studies to evaluate different fillet designs)5. Shih
Filleted Vane (Measured) Filleted Vane (Measured) 0.25 0.20
z/S
0.15 0.25 0.10 0.20 0.05 0.15
z/S
0 0.10 -0.25 -0.20 -0.15 -0.10 -0.05 0.05
x/C
0
Fig.11. Various leading edge geometries: a) fillet and0 bulb designs as shown by Becz -0.20 -0.15 -0.10 -0.05 0 et al. and -0.25 b) CFD predictions and flowfield x/C vane by Zess and measurements for a filleted Thole (reproduced with permission from the publisher of ASME). Sources: See notes 2 and 4.
358
Karen Thole and Lin, investigated not only the use of a leading edge modification but also of inlet swirl to control secondary flows in an airfoil passage. Two different modifications along with no modification were considered for three different inlet swirl configurations: no swirl and swirls of opposite rotation with linearly varying swirl angle from endwall to endwall. The two modifications differed in how they blended into the vane geometry with one fading into the vane surface and the other fading into the endwall. While no single case provided maximum reduction for every metric, a case with inlet swirl and no fillet came very close to doing so. In comparison, the aerodynamic and heat transfer benefits realized for cases with fillets and no inlet swirl were not as great. The computational work by Lethander and Thole combined a three-dimensional computational fluid dynamics (CFD) package along with an optimization package to design an optimal fillet with a rather limited number of independent parameters. Given their specified constraints, the results from their optimization study indicated a successful fillet design was one that was relatively large. While their aerodynamic losses were predicted to slightly increase with their optimal fillet, the thermal environment for the airfoil platform was significantly reduced. Han and Goldstein used the fillet design of Zess and Thole and found that there was a reduction in heat transfer but only near the leading edge-endwall juncture6. In the downstream portion of the passage there was little reduction. Three-dimensional endwall contouring, which includes a more comprehensive geometric modification than simply a modification to the endwall-airfoil leading edge juncture, has also been investigated computationally by Harvey, et al. and experimentally verified by Hartland, et al.7. To design the endwall contour, they used a linear sensitivity matrix in conjunction with superposition methods prior to applying an inverse design algorithm. The results of the experimental verification confirmed a predicted reduction in exit flow angle deviations. Moreover, the experiments indicated a 30% reduction in loss, which was higher than predicted. In a later study, Brennan, et al. and Rose, et al. applied similar computational and experimental (respectively) methodologies as Harvey et al. and Hartland, et al.8. They applied these methods to a high pressure turbine for a single stage in both the vane and blade passages. They reported stage efficiency improvements of 0.59%, which exceeded their predicted improvement of 0.4%. Using endwall contouring and leading edge modifications show promise in reducing secondary flows; however, there are numerous effects that need to be considered. Because this modification must be practically feasible, required manufacturing, space limitations, and cooling are all practical issues that must be addressed. Moreover, some of these designs may be sensitive to the inlet flow conditions which need to be considered.
4.2.3-6 Other Relevant Vane Endwall Studies
359
As discussed in the previous section, a number of studies have been completed with regards to vane endwalls. To form a more complete list of studies presented in the literature on this topic, this section describes additional studies of interest. Blair measured adiabatic effectiveness levels and heat transfer coefficients for a range of blowing ratios through a flush slot placed just upstream of the leading edges of a single passage channel9. One of the key findings was that the endwall adiabatic effectiveness distributions showed extreme variations across the vane gap with much of the coolant being swept across the endwall toward the suction side corner. Granser and Schulenberg reported similar adiabatic effectiveness results in that higher values occurred near the suction side of the vane10. One of the more detailed endwall film cooling studies have been conducted by Friedrichs et al.11. The results of their studies indicated that the endwall cross-flow was altered so that the cross-flow was turned toward the inviscid streamlines as a result of filmcooling injection. There have been a few studies that have measured endwall heat transfer as a result of injection from a two-dimensional, flush slot just upstream of the vane. A series of experiments have been reported for various injection schemes upstream of a nozzle guide vane with a contoured endwall by Burd and Simon, Burd et al., and Oke et al.12. In the studies presented by Burd and Simon, Burd et al. and Oke et al., coolant was injected from an interrupted, flush slot that was inclined at 45° just upstream of their vane13. Similar to others, they found that most of the slot coolant was directed toward the suction side at low slot flow conditions. As they increased the percentage of slot flow to 3.2% of the exit flow, their measurements indicated better coverage occurred between the airfoils. Colban et al. reported flow field and endwall effectiveness contours for a backward-facing slot with several different coolant exit conditions14. Their results indicated the presence of a tertiary vortex that developed in the vane passage due to a peaked total pressure profile in the near-wall region. For all of the conditions simulated, the effectiveness contours indicated the coolant from the slot was swept towards the suction surface. While this study was completed for the same vane geometry as that reported in our paper, the slot geometry had been altered to be flush with the endwall surface. In addition to the previously mentioned studies, the only other studies to have combined an upstream slot with film-cooling holes in the passage of the vane were those of Kost and Nicklas and Nicklas15. One of the most interesting results from the Kost and Nicklas, and Nicklas studies was that they found for the upstream leakage slot flow alone, which was 1.3% of the passage mass flow, the horseshoe vortex became more intense. This increase in intensity resulted in the slot coolant being moved off of the endwall surface and heat transfer coefficients that were over three times that measured for no slot flow injection. They attributed the strengthening of the horseshoe vortex to the fact that for the no slot injection the boundary layer was already separated with fluid being turned away from the endwall at the injection location. Given that the slot had a normal component of velocity, injection at this location promoted the separation and enhanced the vortex. Their adiabatic effectiveness measurements indicated higher values near the suction side of the vane due to the slot coolant migration. CFD studies have also been presented to characterize the endwall film-cooling. Results presented by Knost and Thole indicated, similar to the experiments, the presence of a warm ring on the endwall (bow wave) around the vane where no coolant was present despite the combined slot cooling and film-cooling16.
4.2.3 Airfoil Endwall Heat Transfer 4.2.3-7 Blade Tip Heat Transfer The performance of a turbine engine is a strong function of the maximum gas temperature at the rotor inlet. Turbine blade designers concentrate on finding adequate cooling schemes for high pressure turbine blades, particularly the tip region where heat transfer is quite high. The clearance between a blade tip and its associated outer casing, also known as the blade outer air seal, provides a flow path across the tip that leads to aerodynamic losses and high heat transfer rates along the blade tip. The flow within this clearance gap is driven by a pressure differential between the pressure and suction side of the blade, and is also affected by the viscous forces as the fluid comes into contact with the walls of the gap. A complete description of blade tips and the associated problems is given by Bunker and is well described by figure 1217. Gap size, rotational effects, blade geometries and Reynolds numbers were all highly influence the heat transfer coefficients. One method for improving the thermal environment along the blade tip is to inject coolant into the tip region. In a review paper on tip heat transfer, Bunker states that for blade tips there have been very few film-cooling studies reported in the literature even though film-cooling is widely used18. The discussion given in this section is relevant to tip film-cooling since that is what is typically used in industry. Many tip heat transfer and film-cooling studies have been completed without rotational effects. In general, there is evidence in the literature that supports a widely variable effect of rotation, which warrants further studies. One of the first pioneers in this region is D. Metzger whereby Kim et al.presents a summary of his work19. In Fig. 12. CFD prediction of streamlines across a addition to concluding that there is only a weak effect of the relative motion between blade tip (reproduced with permission from the a simulated blade and shroud on tip heat transfer coefficient, they stated that there publisher of ASME). is a strong dependency of cooling effectiveness for a tip on the shape of the hole and injection locations. Four hole configurations were discussed by Kim et al. that Source: E.M. Hohlfeld, J.R. Christophel, E. included the following: discrete slots located along the blade tip, round holes located L. Couch, and K.A. Thole, “Predictions of Cooling from Dirt Purge Holes Along the Tip of along the blade tip, angled slots positioned along the pressure side and round holes a Turbine Blade,” GT2003-38251. located within the cavity of a squealer tip. The studies reported by Kim et al. were performed in a channel that simulated a tip gap, whereby no blade with its associated flow field was simulated. In comparing the discrete slots to the holes, their data indicated a substantial increase in cooling effectiveness using the discrete slots for all blowing ratios tested. Injection from the pressure side holes provided cooling levels of similar magnitude to the holes placed on the tip with better spreading occurring in the case of the pressure side injection. Kim et al. also reported that an increase in coolant mass flow generally yielded improved cooling with tip surface holes, but for pressure side holes, increased coolant flow yielded decreased cooling effectiveness. Kwak and Han reported measurements for varying tip gaps with cooling holes placed along the pressure surface at a 30˚ breakout angle and on the tip surface at a 90˚ angle for both flat and squealer tip geometries20. They found a substantial improvement in effectiveness with the addition of a squealer tip. The coolant circulated within the squealer tip providing a better distribution of the coolant along much of the tip compared with no squealer cases. Only along parts of the suction side was the cooling effectiveness poor. They found that for the flat tip, good cooling was provided to the trailing edge due to the accumulation of coolant that exited in this area. Their results also indicated that more coolant resulted in improved effectiveness. Results by Christophel et al. compared adiabatic effectiveness levels measured along a blade tip with injection along the pressure side of the blade21. Their results indicated better cooling can be achieved for a small tip gap compared with a large tip gap with different flow phenomena occurring for each tip gap setting. They also found that increased blowing leads to increased augmentations in tip convective heat transfer coefficients, particularly at the entrance region to the gap. When combined with adiabatic effectiveness measurements it was found that the coolant provided an overall net heat flux reduction to the blade tip when using film-cooling and was nearly independent of coolant flow levels.
360
Karen Thole 4.2.3-8 Notes _________________________
361
1. J. T. Chung and T. W. Simon, “Effectiveness of the Gas Turbine Endwall Fences in Secondary Flow Control at Elevated Freestream Turbulence Levels,” ASME 93-GT-51(1993). 2. H. Sauer, R. Műller, and K. Vogeler, “Reduction of Secondary Flow Losses in Turbine Cascades by Leading Edge Modifications at the Endwall,” ASME 2000-GT-0473 (2000);S. Becz, M. S. Majewski, and L. S. Langston, “Leading Edge Modifications Effects on Turbine Cascade Endwall Loss,” GT2003-38898 (2003). 3. See note 2 above. 4. G. A. Zess and K. A. Thole, “Computational Design and Experimental Evaluation of Using a Leading Edge Fillet on a Gas Turbine Vane,” J of Turbomachinery 124 (2002): 167-175. 5. Shih, T. I-P. and Lin, Y.-L., 2002, “Controlling Secondary-Flow Structure by Leading-Edge Airfoil Fillet and Inlet Swirl to Reduce Aerodynamic Loss and Surface Heat Transfer,” ASME GT-2002-30529;Lethander, A. and Thole, K. A., 2003, “Optimizing the Vane-Endwall Junction to Reduce Adiabatic Wall Temperatures in a Turbine Vane Passage,” ASME Paper GT2003-38940;Han, S. and Goldstein, R., 2005, “Influence of Blade Leading Edge Geometry on Turbine Endwall Heat (Mass) Transfer,” GT2005-68590. 6. See notes 4 and 5 above. 7. J. C. Hartland, P. G. Gregory-Smith, N. W. Harvey, and M. G. Rose, “Nonaxisymmetric Turbine End Wall Design: Part II – Experimental Validation,” J of Turbomachinery 122 (2000): 286-293;Neil W. Harvey, Martin G. Rose, Mark D. Taylor, Jonathan Shahrokh and David G. Groegory-Smith, “Nonaxisymmetric Turbine End Wall Design: Part I – Three-Dimensional Linear Design System,” J of Turbomachinery 122 (2000): 278-285. 8. G. Brennan, N. W. Harvey, M. G. Rose, N. Fomison, and M. Taylor, “Improving the Efficiency of the Trent 500 HP Turbine Using Non-Axisymmetric End Walls: Part 1: Turbine Design,” ASME 2001-GT-0444 (2001); M. G. Rose, N. W. Harvey, P. Seaman, D. A. Newman, and D. McManus, “Improving the Efficiency of the TRENT 500 HP Turbine Using Non-Axisymmetric End Walls: Part 2: Experimental Validation,” ASME 2001-GT-0505(2001); also see note 7 above. 9. M. F. Blair, “An Experimental Study of Heat Transfer and Film Cooling on Large-Scale Turbine Endwalls,” J of Heat Transfer (November 1974): 524-529. 10. D. Granser and T. Schulenberg, “Prediction and Measurement of Film Cooling Effectiveness for a First-Stage Turbine Vane Shroud,” ASME Paper Number 90-GT-95 (1990). 11. S. Friedrichs, H. P. Hodson, and W. N. Dawes, “Aerodynamic Aspects of Endwall Film-Cooling,” J of Turbomachinery 119 (1997): 786-793;.S. Friedrichs, H. P. Hodson, and W. N. Dawes, “The Design of an Improved Endwall Film-Cooling Configuration,” J of Turbomachinery 121 (1999):772-780. 12. S. W. Burd, and T. W. Simon, “Effects of Slot Bleed Ijection over a Contoured Endwall on Nozzle Guide Vane Cooling Performance: Part I: Flow Field Measurements,” ASME Paper No. 2000-GT-199; S. W. Burd, C. J. Satterness, and T. W. Simon, “Effects of Slot Bleed Ijection over a Contoured Endwall on Nozzle Guide Vane Cooling Performance: Part II Thermal Measurements,” ASME Paper No. 2000-GT-200; R. Oke, T. Simon, S. W. Burd, and R. Vahlberg, “Measurements in a Turbine Cascade Over a Contoured Endwall: Discrete Hole Injection of Bleed Flow,” ASME Paper Number 2000-GT-214; R. Oke, T. Simon, T. Shih, B. Zhu, Y.L. Lin, and M. Chyu, “Measurements Over a Film-Cooled, Contoured Endwall with Various Coolant Injection Rates,” ASME Paper Number 2001-GT-140. 13. Ibid. 14. W. F. Colban, K. A. Thole, and G. Zess, “Combustor-Turbine Interface Studies: Part 1: Endwall Measurements,” J of Turbomachinery 125 (2002):.193-202; W. F. Colban, A. T. Lethander, K. A. Thole, and G. Zess, “Combustor-Turbine Interface Studies: Part 2: Flow and Thermal Field Measurements,” J of Turbomachinery 125 (2002):.203-209. 15. F. Kost and M. Nicklas, “Film-Cooled Turbine Endwall in a Transonic Flow Field: Part I – Aerodynamic Measurements,” ASME Paper Number 2001-GT-0145 (2001);M. Nicklas, “Film-Cooled Turbine Endwall in a Transonic Flow Field: Part II – Heat Transfer and Film-Cooling Effectiveness Measurements,” ASME Paper Number 2001-GT-0146 (2001). 16. D. K. Knost, and K. A. Thole, “Computational Predictions of Endwall Film-Cooling for a First Stage Vane,” GT-2003-38252. 17. R. Bunker, “Turbine Blade Tip Design and Tip Clearance Treatment,” von Karman Lecture Series 2004-02 (2003). 18. R. S. Bunker, “A Review of Turbine Blade Tip Heat Transfer,” Turbine 2000 Symposium on Heat Transfer in Gas Turbine Systems, Cesme, Turkey, 2000. 19. Y. W. Kim and D. E. Metzger, “Heat Transfer and Effectiveness on Film Cooled Turbine Blade Tip Models,” J of Turbomachinery 117 (1995);Y. W. Kim, J. P. Downs, F. O. Soechting, W. Abdel-Messeh, G. Steuber, and S. Tanrikut, “A Summary of the Cooled Turbine Blade Tip Heat Transfer and Film Effectiveness Investigations Performed by Dr. D. E. Metzger,” J of Turbomachinery 117(1995): 1-11. 20. J. S. Kwak and J. C. Han, “Heat Transfer Coefficient and Film-Cooling Effectiveness on a Gas Turbine Blade Tip,” GT200230194;J. S. Kwak and J. C. Han, “Heat Transfer Coefficient and Film-Cooling Effectiveness on the Squealer Tip of a Gas Turbine Blade,” GT2002-30555.
4.2.3 Airfoil Endwall Heat Transfer 21. J. R. Christophel, K. A. Thole, and F. Cunha, “Cooling the Tip of a Turbine Blade Using Pressure Side Holes—Part 1: Film Effectiveness Measurements,” J of Turbomachinery 127 (2005): 270-277; J. R. Christophel, K. Thole, and F. Cunha, “Cooling the Tip of a Turbine Blade Using Pressure Side Holes—Part 2: Heat Transfer Measurements,” J of Turbomachinery 127 (2005): 278-286.
362
BIOGRAPHY
4.1 Introduction 4.2.3 Airfoil Endwall Heat Transfer
Karen Thole William S. Cross Professor Mechanical Eng. Dept, Penn State Univ. University Park, PA 16802-1412 phone: (814) 865-2519 email: [email protected]
Dr. Karen Thole holds a Professorship in Mechanical Engineering at Penn State University. Her B.S. and M.S. degrees are in Mechanical Engineering from the University of Illinois, and her Ph.D. in Mechanical Engineering is from the University of Texas at Austin. After receiving her Ph.D. in 1992, she spent two years as a post-doctoral researcher at the University of Karslruhe in Karslruhe Germany. After her post-doctoral position, she accepted an assistant professor position at the University of Wisconsin-Madison where she taught and performed research in the Mechanical Engineering Department. In 1999 she accepted a position in the Mechanical Engineering Department at Virginia Tech where she was promoted to full professor in 2003. Dr. Thole’s areas of expertise are heat transfer and fluid mechanics specializing in understanding high freestream turbulance effects. Over the past few years she has developed a number of unique testing facilities directed towards gas turbine heat transfer issues including a combustor simulator that simulates the flow field effects relevant to those entering the turbine section of an engine. Dr. Thole has been responsible for attracting research funding amounting to over $4 million from such agencies at the Department of Energy, US Air Force, Pratt & Whitney, Modine Manufacturing, Siemens-Westinghouse and the National Science Foundation. She has published over 100 peer reviewed papers with a number of these presentations given to international audiences.
4.3
Turbine Blade Aerodynamics
Sumanta Acharya
Gazi Mahmood
Louisiana State University CEBA 1419B, Mechanical Engineering Department Baton Rouge, LA 70803 phone: (225) 578-5809 email: [email protected]
363
4.3-1 Introduction The aerodynamics of the flow in a turbine stage (stator/rotor) is rather complex and is still the subject of many ongoing research activities in the gas turbine community. The flow is inherently three dimensional due to the vane/blade passage geometry with features such as twisting of the vane/blade along the span, clearance between the blade tip and the shroud, film cooling holes, and end wall contouring1. The passage flow is characterized by boundary layer effects, secondary flows generated by the passage pressure gradients, and vortical flow structures such as the leading edge horse-shoe vortices, tip-leakage flow vortices, and corner vortices2. The effects of centrifugal-buoyancy, shockboundary layer interaction, and flow interactions between the stator and rotor rows complicate the passage flow field even further. Along the end walls, the flow structure is strongly threedimensional with the passage vortex and coolant injection on the hub side and the tip-leakage vortex on the tip side. In the midspan regions located away from the passage walls and outside the viscous shear layer, the radial flow is almost negligible and the flow is effectively two dimensional. The fluid dynamics in this region can then be based on two dimensional planar cascade flow studies without any significant loss of information. The three dimensional complex flow structures near the hub endwall region and in the blade tip-shroud clearance have been simulated in annular vane/blade passages with and without rotating blade row3. Studies of the complex end-wall flows have also been performed in stationary cascades with three dimensional airfoil shapes4. The qualitative features of the passage flows, which comprise mainly of the passage crossflow (flow from the pressure side of vane/blade to suction side of adjacent vane/blade) and vortical flows induced by the leading edge, the corners, and the injected coolant flows have been studied in detail in stationary cascades and are considered to be similar in both stationary and rotating blade rows. The primary difference in the secondary flow structure between the blade passage and vane passage is that the vortical flows and cross flows in the blade passage are stronger because of higher turning of the flows along the blade passage. Secondary flows are the major source of aerodynamic losses, which account for 35%-40% of all losses5 and thermal loading in the turbine passage, and thus require special considerations by the turbine designers. The primary objectives of this chapter are to present and analyze the features of the flow field in the turbine vane/blade passage near the hub endwall and mid-span locations of the blade. Toward this effort, reported measurements and computations of pressure, velocity distributions, flow turning angles, turbulence intensity, and vorticity distributions in the cascade test section are presented. Recent efforts to reduce the secondary flows by structural modifications in the passage are discussed. In this chapter, basic fluid dynamic principles and mathematical models of the flow in the passage are not discussed, and the reader is referred to notes 1, 2, and 6 for additional details6. Also details on the aerodynamic design methodology for the vane/blade passage are not presented.
4.3-2 Flow Field in the Mid-span Region
Fig. 1. Streamlines and static pressure distribution in the mid-span plane along blade passage.
Fig. 2. Flow yaw angle (deg) contours in mid-span plane along blade passage.
Source: See Note 56 (Acharya).
Source: See Note 56 (Acharya).
Figure 1 shows the streamlines and static pressure distribution along the mid-span plane of the blade passage. Flow along the blade passage at the mid-span locations turns with the passage contour and essentially follows the ideal flow behavior except very close to the blade walls. At zero degree angle of incidence, the streamline splits at the stagnation point corresponding to the blade leading edge with one leg moving along the pressure side and the other leg moving along the suction side of the blade. The pressure gradient from the pressure side to the suction side leads to the development of secondary flows. These secondary flows and the endwall boundary layer produce deviations to the nearly-inviscid mid-span streamlines shown in figure 2. The flow turning angle, known as the yaw angle relative to the axial +X direction, at the mid-span plane through the blade passage is shown in figure 2. The yaw angle is nearly uniform along a constant pitch line from the pressure side to the suction side, and also changes uniformly along the axial length of the passage. The high yaw angle near the leading edge occurs because of the stagnation region where the streamlines sharply turn around the blade suction side. Figure 3 shows the distribution of the static pressure coefficient, Cp, which is determined from the difference of blade surface pressure and reference pressure at the passage inlet normalized by the passage inlet dynamic pressure. The lowest Cp on the suction surface corresponds to the location at the passage throat area where the flow velocity is the highest. The highest Cp is the stagnation point location on the blade section at the mid-span height. The pressure distribution does not change along most of the blade span or height except near the hub or tip region. The blade loading or lift that provides work on the turbine shaft is determined based on the area circumscribed by such pressure curves as shown in figure 3. The pressure side velocity increases steadily as the Cp decreases on the pressure side from the leading edge to the trailing edge. Along the suction surface, the velocity initially increases toward the throat, but starts to decline when it encounters the adverse pressure gradients downstream of the throat in a subsonic flow. The peak velocity in figure 3 corresponds to the location of the minimum Cp on the suction surface. Due to the adverse pressure gradient on the suction surface downstream of the minimum Cp, there is the potential of boundary layer separation from the suction-side blade surface near the trailing edge and this represents a major source of profile losses in the blade passage. Boundary layer separation at the blade trailing edge can also occur due to a finite trailing-edge thickness and can lead to a distinct wake region. For blade profiles with high loading, flow separation is a major issue. With increased loading on the blade surface, suction surface pressures are reduced, and the velocity and Mach number over the suction surface increases with the local Mach number reaching supersonic values. This leads to local shocks as schematically depicted in figure 3, and creates additional aerodynamic losses such as shock losses or wave drag7. Downstream of the shock, suction surface pressure rises in the adverse pressure gradient region and
364
Sumanta Acharya boundary layer separation can occur earlier leading to increased profile losses for the highly loaded blade. Refer to notes 5 and 8 for discussion on the limiting pressure and velocity distributions on the blade surfaces and provide guidelines to limit the wake region over a very small region at the trailing edge in the blade design8.
4.3-3 Flow Field in the Endwall Region
Fig. 3. Pressure and velocity distribution on blade surface at different spanwise locations. Source: See Note 56 (Acharya).
Fig. 4. Streamlines showing separation lines in a near endwall plane of a linear blade passage.
365
Source: See Note 56 (Acharya).
The flow field near the hub endwall region of the blade passage is dominated by the boundary layer, strong pressure gradients, and cross flow in the pitchwise direction from the pressure side to the suction side. The resulting near-wall flow field is complex and consists of strong secondary flows and vortex roll-up9. When the endwall boundary layer approaches the blade row, a vortex is formed near the junction of the blade leading edge and the endwall. This vortex is termed as the leading edge horseshoe vortex. The horse-shoe vortex splits at the leading edge, and propagates downstream into the passage on both the pressure side and the suction side of the blade passage forming two legs of the early passage vortex flows. Corner vortices are also induced in the corner formed by the blade and the hub endwall. The streamlines slightly above the endwall in figure 4 show some distinct features of the endwall boundary layer flow. These features are identified by the separation lines in the figure. The streamlines along the blade leading edge bifurcate as they approach the saddle point. The saddle point is the location on the endwall where the zero degree incidence line meets the separation line and corresponds to the lowest friction velocity. The incoming endwall boundary layer detaches along the separation line, and secondary vortical flows are formed in the regions immediately downstream and adjacent to the separation line. This is indicated by the high concentration of the streamlines adjacent to the separation line. The strong reverse flow in the vortex regions counter the boundary layer streamlines causing them to be concentrated more densely near the separation line. The leading edge horse-shoe vortex immediately downstream of the saddle point is clearly evident in figure 4. The region between the separation line and the blade suction side in figure 4 represents the suction-side leg of the horse shoe vortex10. The region along the separation line directed from the pressure side to the suction side represents the pressure-side leg of the horse-shoe vortex11, and is driven by the passage pressure gradients. The suction side leg vortex and pressure side leg vortex meet together in the mid-passage region where the two separation lines in the passage merge. This location occurs close to the suction surface, and the merger of the two vortices forms a stronger vortex known as the passage vortex. The passage vortex then travels along the blade suction surface toward the passage exit. The axial development and structures of these vortex flows will be analyzed further in Section 4.3-4. Downstream of the pressure side separation line, the endwall boundary layer region is very thin and skewed toward the suction side. This is evidenced by the streamline concentration being sparse in this region as they turn from the pressure side to the suction side. The strong vortical motions of the pressure side leg vortex entrain most of the fluid from the incoming boundary layer and a new boundary layer forms downstream. Comparing the streamlines in figure 4 with those in the mid-span regions in figure 1, it is clear that the turning of the streamlines inside the blade passage and around the leading edge is much greater near the
4.3 Turbine Blade Aerodynamics endwall region which causes the cross flow here to be stronger. The flow yaw angle contours along the passage in figure 5 show the higher magnitudes of the flow turning near the endwall compared to the flow turning in the mid-span regions away from the endwall. The endwall pressure gradients are shown in figure 6. The suction side pressure magnitudes along the endwall are higher compared to the suction side pressure values in the freestream region (see figure 1). This results in smaller pressure gradients in the endwall from the pressure side to the suction side as shown in the line plot of figure 6. The magnitude of ΔP in the figure is the pitchwise pressure difference between two points at the same axial chord location, one located on the pressure side and the other on the suction side. As mentioned earlier, the turning of the boundary layer fluid in the endwall region is much higher compared to the turning of the free-stream in the passage, which seems to contradict the results in figure 6. According to Niehuis, et al., in the free-stream flows away from the endwall, equilibrium exists between the pitchwise pressure gradient and centrifugal force on the fluid elements at the curved streamlines12. This equilibrium breaks down in the endwall region because the centrifugal force on the fluid elements in the low velocity boundary layer reduces. As such, the weak endwall region streamlines easily turn to a greater degree with relatively smaller pressure gradients as those in figure 6.
Fig. 5. Flow yaw angle (deg) contours in a plane near endwall along a linear blade passage. Source: See Note 56 (Acharya).
Fig. 6. Surface pressure distribution at endwall of a blade passage in linear cascade. Source: See Note 56 (Acharya).
366
Sumanta Acharya
367
The axial pressure distributions along the blade surface (figure 3) change with the span location. The region below 14% span can be considered to be the boundary layer region in the figure. The difference between the free stream static pressure coefficient and endwall region static pressure coefficient Cp on the pressure surface is small and almost uniform. This suggests that pressure gradient on the pressure surface occurs mostly in the axial direction rather than in the spanwise direction. The differences in the suction surface Cp between the 50% span and below 14% span in figure 3 are significant and occur because of the strong cross flow from the pressure side to the suction side in the endwall boundary layer and the vortex leg along the suction surface13. The locations of the lowest Cp within the boundary layer occurs further downstream of the lowest Cp location in the mid-span free-stream region. The endwall cross flow drives the low momentum boundary layer fluid toward the suction surface-endwall junction and causes these observed differences in the Cp distributions. This is also the reason why the axial location of the lowest Cp in the boundary layer and the lowest Ps on the endwall are different. Also note that Cp magnitude decreases significantly from the endwall (i.e. 4.4% span) to the boundary layer edge (i.e. 13.5% span) all along the axial direction. Such spanwise pressure gradient drives the boundary layer fluid and the endwall region secondary flows toward the mid-span direction near the suction surface. The implications of such migrations are realized in the Heat Transfer Analysis section. The velocity distribution on the blade surface near the endwall is also shown in figure 3. The suction surface velocity in the endwall region is lower compared to the free-stream velocity in mid-span because of the influence of strong secondary flows and higher Cp around the suction side edge (i.e. at 4.4% span). As Fig. 7. Surface oil-flow visualization on a linear blade surface and end-wall with the lowest Cp location, the associated peak suction in a linear cascade. LE= leading edge, TE= trailing edge, and BL= boundry surface velocity in the endwall region also moves down layer the axial direction relative to the peak velocity location at the mid-span. The difference in velocity distribution Source: See Note 14. on the pressures side is opposite to what is observed on the suction surface between the mid-span location and endwall region. This can be attributed to the smaller pressure surface Cp near endwall as well as the thin boundary layer, which is also skewed and thicker toward the suction surface, downstream of the endwall separation line. With the knowledge of velocity and pressure distribution on the blade surface at the mid-span and near endwall, it is now appropriate to discuss the three dimensional flow on the blade surface as a whole. It is apparent by now that the blade suction surface flow near the endwall wall region becomes skewed and three dimensional due to the interaction of the boundary layer and flow separation on the suction surface and endwall. The flow visualization on the suction surface of a two dimensional linear blade in figure 7, as observed in Hodson and Dominy, clearly shows the near surface flow behavior14. The flow on the pressure surface is two dimensional for most part of the span as the oil streaklines indicating the surface streamlines are parallel to the endwall in the upper flow visualization of figure 7. The uniform pressure distribution along the pressure surface span (see figure 3) and very weak interaction of the boundary layers between the pressure surface and endwall are responsible for such flow behavior. However, the laminar boundary layer near the pressure surface leading edge may diffuse with a rise in surface pressure when the incoming flow is at high speed. In this case, the boundary layer separates along the line S6 and re-attaches along the line R6 creating a closed separation bubble along most of the span near the leading edge. The laminar boundary layer accelerates following the re-attachment and continues to grow along the pressure surface toward the trailing edge. The two-dimensional separation bubble has no apparent influence on the secondary flows on the endwall.
4.3 Turbine Blade Aerodynamics Near the leading edge at the endwall, the pressure surface streamlines are inclined toward the endwall indicating the flow is driven by the horse-shoe vortex. The separation lines created by the oil streaklines on the suction surface of figure 7 reveals some interesting features of the boundary layer behavior. The separation lines divide the flow on the suction surface into three regimes: (i) two dimensional laminar boundary layer regime, (ii) turbulent boundary layer regime, and (iii) three dimensional flow regime. (i) Two dimensional laminar regime: This regime extends from the leading edge to the lowest suction pressure on the suction surface and between the S2s separation lines near the two endwalls in figure 7. The surface streamlines are seen to be nearly parallel to the endwall in this regime. The laminar boundary layer starting at the leading edge undergoes a high acceleration on the suction surface. According to Hodson and Dominy, the overacceleration in the boundary layer causes a two dimensional separation bubble near the blend point of the circular leading edge and the suction surface15. This separation bubble extends across most of the span, but it is not apparent in the bottom surface flow visualization of figure 7. The suction surface leading edge separation bubble is shown by the flow visualization in Gregory-Smith et al.16. Following the re-attachment behind the separation bubble, the laminar boundary layer accelerates along the suction surface and continues to grow until the separation line S3s. (ii) Turbulent regime: This regime is limited by the re-attachment line following the separation at S3s and trailing edge and between the S2s lines. The laminar boundary layer separates at the lowest suction pressure located at axial distance at S3s because of the adverse pressure gradient (see figure 3) and forms another closed separation bubble. The boundary layer undergoes transition and becomes turbulent as it re-attaches behind the separation bubble on the suction surface. The turbulent boundary layer grows along the suction surface and may separate again due to the adverse pressure gradient near the trailing edge to form the trailing edge wake. (iii) Three dimensional flow regime: This regime is indicated by the region between the separation line S2s and endwall. The regime begins at the location where the suction side leg of the leading edge horse-shoe vortex and pressure side leg vortex from the adjacent blade meet on the suction surface. The pair then emerges as the passage vortex which then moves toward the mid-span as it follows the suction surface toward the passage exit. The suction surface boundary layer separates along the S1s and S2s lines near endwalls in figure 7 as the passage vortex and suction side leg vortex climbs up the suction surface. The distinct appearance of the separation line S2s indicates that the suction side leg vortex maintains its existence in the axial development of the passage vortex which will also be shown in further detail in the next section. The inclination of the surface streamlines toward the mid-span in this regime is caused by the entrainment of the boundary layer fluids (both at the endwall and the suction surface) by the passage vortex. Note that the surface streamlines are symmetric about the mid-span of the blade surface in figure 7. The patterns become asymmetric in threedimensional cascade by the influence of radial forces as will be shown in further sections. The locations of the separation bubbles and separation lines on the blade surface are strongly influenced by the inlet flow angle and Reynolds number or Mach number of the incoming flow. For the high speed compressible flow (with the Mach number>0.70), the flow expands and accelerates along the passage creating local supersonic region at the passage throat17. As a result, a series of weak compression fans are developed at the suction surface near the throat. Detemple-Laake also shows that at transonic and supersonic flow, shocks are formed across the span at the trailing edge of the blade surface18. The shock at the suction surface trailing edge is deflected by the wake from the adjacent blade trailing edge. The shock at the pressure side trailing edge is reflected at the adjacent blade suction surface as a sequence of compression-expansion-compression waves. At all Mach numbers tested (exit Mach number ranges between 0.70 and 1.3), schlieren photographs show that flow separates locally from the blade pressure surface and suction surface forming separation bubbles similar to the subsonic flow pattern19. The separation lines for the suction side leg vortex and the passage vortex on the suction surface move nearer to the mid-span as the Mach number is increased. The suction side leg vortex is deflected by the shock from the adjacent blade pressure side trailing edge and moves closer to the passage vortex at supersonic flow. The endwall pressure distributions for high speed compressible flows show the same behavior as that at the low speed flows. Static pressure on the endwall increases slightly at the trailing edge due to the expansion at the trailing edge.
4.3-4 Development and Structure of Secondary Flows in the Passage We have shown in the earlier section that the secondary flows in the turbine vane/blade passage are dominated by the vortex flows located in the hub endwall region. So far, these vortex flows have been deduced from pressure distributions, near-wall streamlines and saddle points or surface oil-flow visualizations. The vortex flows have been identified as the suction side leg and the pressure side leg originating from the leading edge horse-shoe vortex that eventually merge in a complex way to form the passage vortex. The three vortex structures (horse-shoe, pressure side leg, suction side leg) are the primary sources of the vortex flows in the passage. In addition, smaller corner vortices are induced at the corner of blade edge at the endwall. Vortices are also induced on the suction side near the meeting point of the pressure side leg vortex and suction side leg vortex flows and are advected with these legs along the suction surface toward the passage exit. This section will discuss the structure and development of the three primary vortex flows along the passage at different axial locations by presenting the flow visualization images, streamlines, pressure losses, vorticity, turbulence intensity, and flow turning angles. The induced vortex flows will be identified later in the section. Leading edge horse-shoe vortex: The leading edge horse-shoe vortex is formed at the junction of an endwall and the blunt leading edge of the blade. As the flow approaches the leading edge stagnation line, static pressure rises across the flow from the endwall. The static pressure increases more in the free-stream region above the boundary layer since the free-stream velocity is higher compared to the velocity in the boundary layer20. This spanwise pressure gradient in the vicinity of the leading edge causes a vortex roll-up, known as the leading edge horse-shoe
368
Sumanta Acharya vortex. The vortex is confined in a region which is smaller than one boundary layer thickness from the endwall21. The endwall location where the reverse flow in the horse-shoe vortex meets the incoming boundary layer flow is termed as the saddle point which is generally located along the line corresponding to the zero degree incidence-angle as shown in figure 4. The center of the horse-shoe vortex is located in between the saddle point and the blade leading edge. The exact locations of the saddle point and horse-shoe vortex center depends on the radius of curvature of the blade leading edge and the oncoming boundary layer thickness. The flow visualization in figure 8 shows the typical formation of an instantaneous horse-shoe vortex at the blade leading edge in a linear cascade22. In the figure, the roll-up of the vortex is in the counter-clockwise direction. Periodically, a pair of horse-shoe vortices form that have the same relative size and the same sense of rotation. The time-averaged structure of the horse-shoe vortex obtained from a numerical simulation at the same blade leading edge is also shown in figure 8. The vortex structure is similar to the leading edge horse-shoe vortex in front of a cylinder23. The vortex center in figure 8 is located at the point of the maximum kinetic energy. At the corner of the leading edge, a small counter rotating vortex is induced by the large horse-shoe vortex. This clockwise rotating vortex is known as the leading edge corner vortex. Pressure side leg and suction side leg vortices:
369
Fig. 8. Typical horse-shoe vortex and corner vortex at a blade leading edge. HS=horse-shoe, LE=leading edge, and Tke=turbulent kinetic energy.
The path of the leading edge horse-shoe vortex is along the separation line shown in figure 4, and is driven by the endwall Source: See Notes 22, 56 (Acharya). region pressure gradient and cross flow. The horse-shoe vortex essentially splits near the leading edge with one leg moving along the pressure side and the other leg moving along the suction side. These two primary legs of the horse-shoe vortex represent the origin of the major secondary flow system that develops in the blade passage. The two legs rotate opposite to each other. The pressure side leg of the horse-shoe vortex is driven along the separation line across the passage from the leading edge to the adjacent blade suction side. The suction side vortex leg wraps around the suction side from the leading edge along the separation line in figure 4. The vortices are termed as the right side clockwise and left side counter-clockwise vortices, respectively, in the flow visualization images in figure 924. The vortices in the figure are being viewed in the flow direction in pitchwise planes at different axial locations. The pressure side and suction side of the blade passage in the study are located on the right and left hand side, respectively, in the images. At location A, which represents the pitchwise plane going through the leading edge, a pair of vortex structures appears along both the pressure side leg and suction side leg of the vortex system. As with the leading edge horseshoe vortex pair, these pairs form periodically from a single pressure side and suction side leg vortex. The vortices are very close to the passage pressure side and suction side at plane A. As the vortex legs advect to farther downstream locations at planes B, C, and D in the passage in figure 9, the endwall cross flow and pressure gradient from the pressure side to the suction side sweeps the pressure side leg of the vortex toward the suction side of the passage. As a result, the pressure side vortex approaches Fig. 9. Flow visualization of pressure side leg and suction closer to the suction side leg vortex from the adjacent blade at side leg vortices from. Vph= pressure side leg vortex and the downstream locations. Nearly half way down the passage at Vsh=suction side leg vortex. location D, the two vortex legs merge to form a single structure Source: See Note 9.
4.3 Turbine Blade Aerodynamics of counter-rotating vortices. At the merging point, the clockwise pressure side leg vortex is larger and stronger compared to the smaller suction side leg vortex. Note that the vortices remain close to the endwall from the leading edge location to their merger location half way through the passage. The time-averaged vortex structures, as shown in figure 10, are obtained at a pitchwise plane near the leading edge for a much higher Reynolds number than that used for the flow visualization25. Even though the blade geometries are different for the data in figures 9 and 10, the basic structure of the vortices such as the sense of rotation and relative locations are similar. The positive axial vorticity in figure 10 signifies a clockwise rotation of the pressure side leg vortex, while the negative axial vorticity indicates a counter-clockwise rotation of the suction side leg vortex. The small vortex located at the edge of the pressure side and rotating counter-clockwise is termed as the pressure side leading edge corner vortex in Wang, et al. 26. This vortex is induced by the leading edge horse-shoe vortex and driven along the pressure side leg vortex. Passage vortex and induced vortices: The passage vortex is comprised of the pressure side leg vortex and suction side leg vortex as these vortex legs approach and merge with each other nearly half-way into the passage and along the suction side of the blade. This is shown as a counterrotating vortex pair at location D in Fig. 9. As the pressure side leg vortex is much larger and stronger compared to the suction side leg vortex when they meet, the sense of rotation of the passage is considered be the same as the rotation of the pressure side leg vortex. The passage vortex remain close the suction surface of the blade as it is driven toward the passage exit by the cross flow and pressure gradient along the passage. Figure 11 shows the passage vortex formation in the same passage as in figure. 9 at locations downstream of plane D. The suction side is located on the left hand side in the flow visualization images27 in figure 11. At locations E and F, the passage vortex Vp is rotating counter-clockwise while the suction side leg vortex Vsh is rotating clockwise as they are viewed opposite from the main flow direction. As both the vortices funnel boundary layer fluid and main flow toward their centers, the passage vortex gradually grows larger in size and lifts above the endwall as it travels along the suction surface toward the passage exit. The passage vortex movement toward the mid-span is caused primarily by the spanwise pressure gradient in the boundary layer on the suction surface. Because of this movement of the passage vortex and as the endwall boundary layer fluid is wrapped around the passage vortex, the endwall boundary layer becomes skewed and thicker toward the suction side. The suction side leg vortex wraps around the passage vortex as it travels along with it. This is evidenced in the flow visualization in figure 11. At location E, the suction side leg vortex Vsh is located at the top right hand corner of the Vp. At the downstream F location, Vsh moves to the left side of the Vp. At the passage exit, a study in Wang, et al. shows that Vsh moves around further and is located at the passage vortex bottom part near the suction surface28. In the plane F of figure 11, the small clockwise rotating vortex Vwip located adjacent to the suction surface and above the passage vortex is termed as the wall vortex. This vortex is formed by the strong induction of the passage vortex and
Fig. 10. Velocity vectors and axial vorticity representing pressure side leg vortex and suction side leg vortex near LE. VpLc=pressure side LE corner vortex and ωx=axial vorticity. Source: See Note 25.
Fig. 11. Flow visualization of passage vortex and induced vortices. Source: See Note 9.
370
Sumanta Acharya
371
originates at the same location where the passage vortex starts to form (merging point of the counter-rotating vortex legs). The wall vortex stays above the passage vortex and is driven along with the passage vortex. The small corner vortex induced at the junction of blade leading edge and endwall (see figure 8) by the horse-shoe vortex is driven along the pressure side and suction side edge with the two main vortex legs. In Wang, et al., they are identified as the pressure side leading edge vortex and suction side leading edge vortex, respectively29. They remain small as they travel inside the passage and their sense of rotation are opposite to the rotation of the main vortex legs they are associated with. The pressure side leading edge corner vortex sticks along the pressure surface corner near the leading edge as shown in figure 10. When it enters the passage farther downstream, it is also driven toward the suction side of the adjacent blade along with the pressure side leg vortex by the endwall cross-flow and pressure gradient. The suction side leading edge corner vortex on the other hand remains adjacent to the suction side edge until it meets with pressure side leg vortex from the neighboring blade. The corner vortices are not visible in the smoke flow visualization as the large vortex legs entrain most of the smoke in the flow and enough smoke is not available to generate their small patterns clearly. The threedimensional structures of different vortex flows in a blade passage sketched in figure 11 are adopted directly from Wang, et al.30. The sketch shows two additional small vortex flows located along the pressure surface corner and suction surface corner. They originate about half way downstream in the passage. The existence of such corner vortices is indicated by the high local mass transfer results at the blade surface-endwall corner in Goldstein et al.31. These corner vortices rotate in the same direction as the rotation of the suction side leg vortex. The time-averaged structures of the passage vortex in the same flow as in figure 10 are shown at two axial locations in figure 12. The data presented is measured with a five-hole pressure probe32. The velocity vectors are determined based on the resolved components33. The plane H in figure 12 is located about half way down the passage and the plane I is located near the exit of the passage. Unlike figure 11, the flow is being viewed in the axial direction and therefore, the suction surface is on the left hand side in the plots of figure 12. Thus, the rotation of the passage vortex is in the clockwise direction in the velocity vector plots. The existence of the suction side leg vortex is not apparent in the vector plots as it becomes weak in the downstream locations. In a small region just above the passage vortex and adjacent to suction surface at plane I, the vectors seem to turn counter-clockwise indicating the presence of the wall vortex or suction side leg vortex. However, the vortex is clearly apparent in the vorticity plots in figure 12. The positive axial vorticity indicates the passage vortex while the negative axial vorticity located above the positive region indicates the suction side leg vortex or the wall vortex. The same arrangement of the vortex systems in the downstream locations have been observed in the flow visualization. The vortex center is located at the location of the maximum vorticity in congruence with the forced vortex motion.
Fig. 12. Velocity vectors and axial vorticity representing passage vortex at suction side. SS=suction side, Z=pitchwise distance from pressure side, and ωx=axial vorticity.
Fig. 13. Axial vorticity downstream of passage exit in a linear blade cascade. VP= passage vortex, Vsh=suction side leg vortex, and Vwip= wall vortex. Source: See Note 56 (Acharya).
4.3 Turbine Blade Aerodynamics Note that the passage vortex center in both the vector and vorticity plots moves farther away from the endwall (Y/S=0.0 location) as the passage vortex travels from location H to I in figure 12. This is also consistent with the flow visualization in figure 11. The axial vorticity magnitudes of the passage vortex in plane I are somewhat smaller than those in plane H. The flow turns further away from the axial direction in plane I resulting in smaller component of vorticity in the axial direction. However, the axial vorticity of the wall vortex increases in plane I as the intensity of this vortex grows as it is driven along with the passage vortex. The high negative axial vorticity at the bottom left corner in plane I in figure 12 indicates the suction side corner vortex34. The corner vortices are less likely to develop if the blade surface-endwall corner is filleted. The induced wall vortex, corner vortices, and additional vortices due to the trailing edge wake can be clearly identified in the flow downstream of a blade passage as in figure 13. The data are obtained from the same passage flow as in figure 12, but the location of the data is in a pitchwise plane, K, slightly downstream of the passage exit. The projection of the trailing edge at this plane is located at Z/P=0.0 and the axial direction is into the plane of the data. The positive axial vorticity and the negative axial vorticity just above the positive region in figure 13 also indicate the passage vortex and wall vortex, respectively. The locations of these vortices are even higher above the endwall compared to those in figure 12. In figure 13, the negative trailing edge wake vorticities on the left of the passage vortex form due to the wake in the adverse pressure gradient region at the trailing edge (see figure. 3). The corner Fig. 14. Flow turning angles in a plane near exit of a blade passage. vortices indicated by the negative vorticities about Z/P=0.0 and just above the endwall are enhanced by the trailing wake flows at this location. Figure 14 shows the influences of the various vortex structures on the flow orientation near the exit plane of a blade passage35. The uniformity of the flow angles near the suction side is severely affected by the vortex flow. Both the pitch and yaw angles in figure 14 are referenced to the axial +X direction. The blade turning angle at this location is about 65 degree. The positive pitch angles in the figure indicate that the flow is directed away from the endwall while the negative pitch angles indicate that the flow is directed toward the endwall. The under-turning of the flow yaw angles, which is less than 59 deg near Y/S=0.20, is very high near the passage vortex center. The high over-turning of the flow yaw angles, which is greater than 69 degree near Y/S=0.25, in the vicinity of the suction side occur because of the wall vortex or suction side leg vortex. The over-turning of the yaw angles also occur in the endwall boundary layer region at Y/S<0.12 where the cross-flow is very strong (see figure.4). Such under- and over-turning of the exit flows affect the blade loading and aerodynamic losses in the next row of blades in the turbine stage.
4.3-5 Pressure Loss The vortex structures are a significant source of pressure or aerodynamic losses across the blade passage. They entrain fluid from the free stream flow and enhance convective turbulent transport in the endwall region as well
Fig. 15. Turbulent kinetic energy (Tke) and total pressure loss coefficient Cpt along a blade passage. SS=suction side, PS=pressure side, and ∆Pt=Pt-Pref. Source: See Note 25 (Saha).
372
Sumanta Acharya as on the blade surface. This also results in high local thermal loading on the turbine passage walls. Figure 15 shows the turbulent kinetic energy and total pressure loss generated by the secondary vortices at two axial locations of the blade passage of figure 12. The total pressure loss is determined from the difference of local total pressure and the reference total pressure at the passage inlet. The axial locations of figure 15 correspond to the pitchwise planes H and I in figure 12. At Z/P<-0.30, in plane H, the magnitudes of Tke/V2 larger than 0.020 and Cpt larger than 0.30 can be considered to be located within the secondary vortex flow region. In plane I, the secondary vortex flow region is represented by the contours of Tke/V2 and Cpt for Z/P<-0.50. Turbulent kinetic energy decreases from the axial location H to the location I in the secondary vortex flow region because the flow accelerates as it travels downstream. Because of its size and the magnitude of the vorticity, the passage vortex is primarily responsible for the turbulent kinetic energy and the total pressure losses in the secondary flow region near the suction side. Note that in both the axial locations of figure 15 the high turbulent kinetic energy just above the endwall and outside the secondary flow region occurs because of the boundary layer flow. Also it is important to realize that in the axial location I near the exit plane, the total pressure losses in the passage vortex core region are more than five times the total pressure losses in the free stream region with Cpt<0.20.
4.3-6 Aerodynamics of 2-D Vane Cascade
373
The results presented so far on the flow structure in a blade passage are very typical of other linear blade cascade studies reported36. The magnitudes of the flow quantities are different from one study to the other, but the pattern and arrangement of the secondary flows are similar. The secondary flow patterns in a two-dimensional vane cascade are also expected to be similar to that in a linear blade cascade. This is because the formation of the leading edge horse-shoe vortex is inevitable for the vane passage as well. The horse-shoe vortex is then driven by the endwall cross flow and pressure gradient in the passage forming the suction side leg and pressure side leg vortices. The secondary flows are weaker and smaller in a linear vane passage than in a linear blade passage because of the smaller flow turning and weaker endwall cross flow in the vane passage37. However, the same components of the secondary flows are present in both the vane and blade passages. Due to the weaker secondary flows in the vane passage, the pressure losses are smaller relative to those across a blade passage. Figure 16 presents the data measured in a linear vane cascade38. The pressure side leg vortex (clockwise rotating) and the suction side leg vortex (counter-clockwise rotating) at plane SS-1 in figure 16 are located near the pressure side and adjacent to the suction side of the vane, respectively. At a farther downstream location SS-3, the pressure side leg vortex is driven close to the suction side and pair up with counterrotating suction side leg vortex to form the passage vortex. Like the blade passage flow, the suction side leg
Fig. 16. Secondary velocity vectors and streamwise velocity contours showing vortex flow formations in a linear vane cascade. SS= suction surface and PS= pressure surface. Source: See Note 33.
Fig. 17. Static pressure and secondary velocity vectors across a plane in a vane passage of annular cascade. PS=pressure side and SS=suction side. Source: See Note 3. (Sieverding)
4.3 Turbine Blade Aerodynamics vortex clearly appears much weaker and smaller than the passage vortex at this location in the vane passage. However, the relative arrangement of the passage vortex and suction side leg vortex at SS-3 in figure 16 is different than that observed at plane H in figure 12. The evolution of the passage vortex structure along a linear vane cascade and downstream of a linear vane passage show qualitative similarity with those found in figures 12, 13, and 1639.
4.3-7 Aerodynamics of 3-D Cascade A three-dimensional cascade is usually formed when the linear or twisted blade or vane profiles are stacked in an annular passage. Thus, the flow area in the cascade passage increases from the hub side to the tip side of the adjacent blades. Because of this passage structure, a radial pressure gradient is added to the flow everywhere in the annular passage40. This radial pressure gradient directing towards the hub neutralize the radially outward centrifugal force experienced by the free-stream fluid which is in equilibrium across the annular passage. The radial pressure gradient increases from the pressure side to suction side41. The data in figure 17, adopted from Sieverding et al., show the static pressure coefficients and secondary velocity vectors across an axial plane slightly upstream of the passage exit of an annular vane cascade. In the figure, the difference in static pressure coefficient between the hub and casing is higher on the suction side than on the pressure side. The effects of such pressure gradients are greater at the passage exit than within the passage. The vectors show the formation of the passage vortex near the hub and casing side. The radial displacement of the vortex centers are mostly the consequence of the pitchwise cross flow rather than the radial pressures as observed in a linear cascade. The total pressure losses near the hub and casing walls due to the two passage vortex structures are also nearly identical. However, in the same study, the total pressure losses at the passage exit are found to be asymmetric with respect to the spanwise meridian unlike that observed across a linear cascade. The total pressure losses downstream of the cascade are considerably higher near the hub wall than those near the casing42. This is attributed to the influence of the radial pressure gradient. Near the endwall in the passage, the streamwise velocity decreases in the boundary layer and the secondary flow region. As a consequence the centrifugal force generated from the circumferential component of the streamwise velocity decreases near the endwalls. But, the radial pressure gradient remains unchanged. Thus, the non-equilibrium behavior near the casing results in the reduced total pressure loss there.
*The original version of this material was published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in AGARD Conference Proceedings CP469 ”Secondary Flows in Turbomachines” in 1990.
Fig. 18. Total pressure loss coefficients at exit plane of a linear and an annular cascade. Source: See Note 13.* (Moustapha)
Fig. 19. Spanwise distribution of pitchwise averaged total pressure loss coefficient for linear and annular cascades in Fig. 18. Source: See Note 13.* (Moustapha)
374
Sumanta Acharya A comparative study of a linear cascade and an annular cascade with the same two dimensional blade geometry employed is presented in figure 1843. The total pressure losses in figure 18 are obtained at the same exit location relative to the blade trailing edge in the two passages. It is clearly evident from the plots that the radial pressure gradient in the annular cascade plays a significant role and alters the pressure loss distribution when compared to the linear passage which experiences no such pressure gradient. The pressure losses in the linear cascade have two distinct high loss regions, located symmetrically about 50% blade span, near the suction surface. These are the signatures of passage vortices from the top and bottom endwalls. In contrast, the annular cascade has one high total pressure loss region between 48% and 77% span. In this loss region, the peak losses are 0.70 and 0.80 at 53% and 70% span, respectively, indicating that the cores of the hub side and casing side passage vortices are very close to each other. The plots in figure 19 determined from the local data in figure 18 compares the pitchwise averaged total pressure loss magnitudes between the linear cascade and the annular cascade. The two peaks in figure 19 for the planar cascade are a consequence of the distinct passage vortex pair observed in the local data. For the annular cascade, one peak in the average pressure loss distribution occurs because of a single high pressure loss region in the local data. In general, the average total pressure losses are much higher for the annular cascade than for the linear cascade along most of the span. Between 20% and 40% span at the inner endwall side the average losses are higher for the linear rig as the passage vortex from this endwall is located in this region. The static pressure distributions on the annular endwalls as shown in figure 20 indicate different distributions for the casing wall and hub wall. The pitchwise pressure gradient extends all the way down to the trailing edge for the casing wall. While the cross pitch pressure gradient for the hub wall is high in the first half of the passage, the gradient decreases significantly in the latter half of the passage compared to that for the casing. Such pressure distributions provide the radial pressure gradients between the two endwalls in the annular cascade which is responsible for the radial movement of the secondary flows as explained previously. The magnitudes of the static pressure coefficient near the suction side in figure 20 are higher for the hub wall than for the casing in the first half of the passage. In the second half of the passage, these magnitudes near the suction surface are higher for the casing wall. Thus, the endwall cross flow covers most of the casing
Fig. 20. Static pressure coefficients on endwalls in a blade passage in an annular cascade. Source: See Note 13.* (Moustapha)
2
wall, while it covers only the first 3 rd of the hub wall. Note the qualitative similarity of the contour distributions between the linear passage endwall (figure. 6) and annular passage casing wall (figure 20). Fig. 21. Surface flow visualization on a vane suction surface in an annular cascade. Source: See Note 45. *The original version of this material was published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in AGARD Conference Proceedings CP469 ”Secondary Flows in Turbomachines” in 1990.
375
4.3 Turbine Blade Aerodynamics Blade surface pressure distributions on the pressure surface are about the same for the annular cascade and linear cascade44. The radial pressure gradient on the pressure surface (annular cascade) is present only near at the hub side. On the contrary, the radial pressure gradient is present along the entire span on the blade surface and is larger in the first onethird of the surface in axial direction compared to the rest of the surface. When compared at the same corresponding radial location, the magnitudes of static pressure coefficient on the suction surface are always higher in the annular cascade than in the linear cascade due to the radial pressure gradient. However, the suction surface pressure coefficient distributions in the axial direction follow similar patterns in both types of cascade employing same blade geometry. Thus, the two dimensional local separation bubbles also appear on the blade suction surface in an annular passage. The vane suction surface flow pattern in an annular cascade is presented in figure 2145. The visualization reveals the asymmetric surface streamlines relative to the spanwise meridian unlike the streamline patterns on the suction surface in a linear cascade. In figure 21 also, the separation lines divide Fig. 22. Measured secondary velocity vectors at rotor exit in an anthe near surface flow behavior into three regions as has been nular passage. observed in case of a linear passage. The first region is the two dimensional laminar region extending from the leading edge Source: See Note 3. (Gallus) to the inclined separation line (spanwise) between passage vortex separation lines. In the radial/spanwise direction this region is limited between the separation lines for passage vortex. Near surface flow simply follows the suction surface in the first region. The second region is the turbulent flow region that extends behind the inclined re-attachment line to the trailing edge in figure 21. Unlike that in the linear cascade (figure 7), the suction surface separation bubble formed here (figure 21) by these inclined separation and re-attachment lines is asymmetric. The third flow region, which is the three dimensional boundary layer region in figure 21 limited by the passage vortex separation line and endwall, is larger on the casing side than on the hub side. The separation lines extend all the way to the trailing edge. The separation line for the passage vortex at the casing is farther away from the casing than the separation line for the hub-end passage vortex line is from the hub wall. This is in accordance with the passage vortex movement observed in figure 18 for the annular cascade and as described, is caused by the radial imbalance of the radial forces. The inclination of the surface streamlines in the third region caused by the passage vortex funneling and entrapping fluid is an indication of the vortex strength. The three dimensional vortex flows near the hub and casing walls are enhanced in an annular rotor passage where the blades rotate relative to a stator passage. In the blade stage, there is a gap between the blade tip and casing to allow for rotation. Due to the pressure gradient from the blade pressure side to the suction side there is a leakage flow from the pressure side to the suction side in the tip gap. The tipgap flows generate an additional vortex flows near the casing wall which develop and grow along with the casing side passage vortex in the rotor passage. The tip vortex influences the passage vortex from the casing wall. Figure 22 shows the tip-clearance vortex and passage vortex structures near
Fig. 23. Relative total pressure contours in an annular rotor passage showing vortex flows. PS= pressure side and SS= suction side. Source: See Note 47.
376
Sumanta Acharya the casing wall and hub wall at the exit plane of a rotating rotor passage. As indicated in the figure, the counter-rotating secondary flows in the passage vortex near the casing wall are caused by the influence of the tip vortex. According to Gallus et al., the flow interactions between the rotor-stator rows change the static pressure distributions on the rotor surface periodically46. The isobar contour lines in figure 23 show that the tip-clearance vortex strengthens the radial inward movement of the casing passage vortex in a rotor passage47. This occurs as the tip vortex grows and intensifies along the passage and pushes the passage vortex in the radial direction. The size and strength of the passage vortices formed in the rotor passage fluctuate depending on the position of the trailing edge wake and passage vortices from the stator row. When the stator wake and passage vortices hit the rotor blade leading edge, the passage vortices in the rotor passage grow larger and stronger. At the rotor exit, the locations of the passage vortices and trailing edge wake also fluctuate depending on the wake and vortices from the upstream stator row48. The tip-clearance vortex increases the suction side static pressure and decreases the pressure side static pressure at the tip region of the rotor blade. This reduces the blade loading at the tip region in the rotor stage.
4.3-8 Aerodynamics With Passage Modifications Recently there have been a number of studies directed Fig. 24. Different leading edge fillet profiles employed in aerodynamic at structural modifications of the blade passage with the aim of loss reduction studies reducing the secondary flows in the passage. The secondary Source: See Notes 22, 25, 49. flows are the significant sources of aerodynamic losses and increased thermal loading in the passage walls. The large passage vortex structure also makes the exit flow turning non-uniform across the entire passage exit plane. This subsequently increases the noise level, secondary losses, and fluctuations of the blade loading on the following blade row. Cold air is injected through tiny holes in the endwall to provide a protective film on the endwall from the hot gas in the passage main flow. The effectiveness of film cooling is adversely affected by the secondary flows in the endwall. The coolant air injected from the holes located upstream of and adjacent to the separation line (figure 4) is lifted up from the endwall by the passage/pressure side leg vortex and suction side leg vortex. This exposes a large part of the endwall immediately downstream of the separation line to the hot gas. The wake and exit passage vortices also affect the coolant flow injected from the holes located in the platform between the two stages. The non-uniformity in the exit flow angles alters the expected trajectories of these coolant paths. The structural modifications of the passage are undertaken at or near the endwall that only affects the flow in the boundary layer and, beneficially alters the secondary flow behavior. Therefore the blade profile remains unchanged for most of the blade span, and only the minimum change occurs in the blade loading. The geometrical modifications are still the subject of ongoing investigations and include leading edge fillet additions and endwall profiling. In Fig. 25. Leading edge horse-shoe vortex in a linear blade passage the discussion below, attention will be focused on these two. with leading edge fillet. Tke= turbulent kinetic energy.
377
Leading Edge Fillet: This modification is also termed as the leading edge contouring near the endwall. Fillets are placed at the junction of the leading edge and endwall. Several forms
Source: See Notes 22, 25.
4.3 Turbine Blade Aerodynamics of the fillet profiles have been tested and are shown in figure 24. As can be seen in the figure, two types of basic construction of fillet profiles can be identified: (i) profile with varying height from the blade surface to the endwall and (ii) profile of bulb with surface thickness at the outer periphery49. The thickness of the type (i) fillet profiles reduces to zero as they extend out from the blade surface to the endwall. These fillets may blend with either the endwall or blade wall or with both the endwall and blade surface as they wrap around the leading edge extending inside the passage. The fillet profiles of the type (ii) blend with the blade surface as they wrap around the leading edge, but meet the endwall with a finite thickness. Type (ii) fillets simply thicken the blade profile near the leading edge at the endwall. All types of fillets studied until now have asymmetric profile with respect to the leading edge and have their highest point located at the leading edge. The height of this highest point from the endwall i.e. the maximum height of the fillets is typically one boundary layer thickness of the incoming flow. The studies mentioned above show that type (i) fillets are the most effective in reducing the secondary flows in the blade passage. These fillets reduce the size and strength of the leading edge horse-shoe vortex. Consequently, the strength of the passage vortex is reduced. The high total pressure losses due the passage vortex then also decrease across the blade passage. Figure 25 shows the horse-shoe vortex structure at the leading edge with a fillet profile of type (i) employed at the leading edge. The profile height varies linearly to zero from the blade surface to the endwall and blends with the endwall and blade wall inside the passage on the pressure side and suction side (Fillet 150). The blade passage is the same as that in figure 8. The wedge shaped object on the left of the flow visualization image of figure 25 is the fillet profile. The size of the horse-shoe vortex is about half in the flow visualization and about one-fifth in velocity vector plot with the fillet compared to the case without any fillet. Note that the flow visualization is observed at a low speed to avoid any smearing and diffusion of smoke. The flow area at the leading edge is reduced in the passage with the fillet. For the incompressible flow, this will cause the boundary layer fluid to be displaced from the leading edge plane. Also, the adverse pressure gradient along the leading edge plane (due to the stagnation) is reduced by the fillet slope. All these factors are responsible in reducing the size of the horse-shoe vortex with the fillet. The turbulent kinetic energy is also reduced significantly in figure 25 compared to what is observed without the fillet. This indicates that the strength of the horse-shoe vortex is also reduced by the fillet. There is also no apparent structure of the leading edge corner vortex in the secondary velocity vectors with the fillet. As the horse-shoe vortex is reduced, the Fillet 1 is expected to reduce the passage vortex size and strength downstream in the blade passage. Figure 26 shows the passage vortex at a plane 92% axial chord (near the exit) with and without fillet in the same blade passage. Comparing the velocity vectors in figure 12 (Plane I) and figure 26, it can be seen that the location of the passage vortex center with the Fillet 1 moves little higher above the endwall than without the fillet. In an upstream location near the suction side in the blade passage, the suction side leg vortex is reduced in size and weakens with the Fillet 1 compared to that without the fillet51. The significant differences are observed in the total pressure loss contours of figure 26. The high total pressure loss region (Cpt>0.45) can be considered as the signature of the passage vortex. The Cpt contours presented here are measured
Fig. 26. Passage vortex and total pressure loss at 92% axial chord with and without fillet. Source: See Note 25.
Fig. 27. Secondary velocity vectors and turbulent kinetic energy (k) at pressure side (Plane PS1) of a linear vane cascade with and without fillet. PS= pressure side. Source: See Note 49. (Zess)
378
Sumanta Acharya with a five-hole pneumatic probe unlike the magnitudes in figure 15 which are obtained from computations in the same passage with an incoming boundary layer of smaller thickness. The total pressure losses in figure 26 are much lower at the bottom part of the passage vortex with the fillet than without the fillet. This indicates that the Fillet 1 has reduced the passage vortex both in size and strength. Also revealed, the under-turning (yaw angle) of the exit flow with the Fillet 1 occurs over a larger region in the passage vortex core52. Similar results about the passage vortex and associated total pressure losses are observed with other fillet profiles of type (i)53. Figure 27 shows the effectiveness of another fillet profile of type (i) in reducing the passage vortex in a linear vane cascade. The quantities in figure 27 are measured in a plane normal to the vane pressure surface54. The velocity vectors in the figure show the structure of the pressure side vortex near the pressure surface. As can be clearly seen, the pressure side leg vortex is not complete for the filleted vane unlike the vortex for the unfilleted case. The spiral of the vortex is not complete in the same location for the filleted case as the passage vortex is weakened. This will eventually make the passage vortex weaker down the passage for the filleted vane. Also, the location of the passage vortex center appears to shift farther away from the pressure side for the filleted case compared to that for the unfilleted case. The turbulent kinetic energy magnitudes in figure 27 are much smaller for the filleted vane than those for the unfilleted vane. The k-contours indicate a well-defined vortex core for the unfilleted case while the k-contours for the filleted case are much uniform in the y/P direction. The fillet causes the passage vortex in this plane to fluctuate along y/P as the velocity component in this direction has the largest fluctuations with the fillet. On the contrast, the large fluctuations in the w-velocity component cause the passage vortex to fluctuate in the z/S direction for the unfilleted vane.
Fig. 28. Axisymmetric axial profiling of endwall: (a) endwall profile through blade passage, (b) endwall profile upstream of blade passage.
Endwall Profiling: Endwall profiling is achieved in two waysaxial profiling along the passage with no pitchwise variation and nonaxisymmetric profiling along the passage with profile variations in both the axial and pitchwise directions. The profiling is aimed either to accelerate the boundary layer fluid at the endwall or to reduce the pitchwise pressure gradient at the endwall. (i) Axial Profiling of the Endwall: Since there is no variation of the profile in the pitch direction, this profiling is also termed as the two dimensional axisymmetric contouring. The profiling is employed on either of the endwalls in the passage, but not on the both endwalls. The height of the profile increases over a smooth curve from the leading edge to the trailing edge such that aspect ratio of the exit plane or the throat area is unaffected as shown in figure 28. This type of endwall profile was studied in linear vane passages55. The axisymmetric profiles of the endwall upstream of the blade/vane passage such as the profile (b) of figure 28 are also studied56. Upstream profiles in the first stage nozzle guide vane are used for the gas path transition from the combustor chamber to the turbine inlet. In any profile shown in figure 28, the inlet velocity to the blade/vane passage decreases (due to increased passage area) and the flow acceleration through the passage increases (due to decreased passage area). This leads to a reduction in the boundary layer thickness and suppresses the growth of secondary flows on the endwalls. Also, the exit flow angle is expected to undergo less under-turning and over-turning due Fig. 29. Streamwise velocity and secondary velocity vectors at 0.90Cax in a linear vane passage with endwall profiling through to the higher flow acceleration downstream with the endwall profiling passage aft. extending through the passage.
379
Source: See Note 55. (Burd)
4.3 Turbine Blade Aerodynamics Figure 29 shows the effects of the endwall profiling through the passage aft of a vane cascade on the secondary flows57. The data are presented in a plane near the passage exit where the endwall profile has become almost flat. The low streamwise velocity magnitudes adjacent to the pressure side in figure 29 are located within the boundary layer on this side. The high velocities near the contour endwall are the results of flow accelerations along the endwall. The two concentrated low velocity regions (U/U2<0.90) adjacent to the suction side near the flat and contour endwalls are located in the passage vortex at this plane. The passage vortex region near the contour endwall is about half the size of that near the flat endwall. The velocity vector plot in figure 29 reveals that fluids are being displaced from the contour endwall region toward the mid-span (y/S=0.50) as all the vectors in this region are pointing toward the midspan. The vectors pointing toward the suction side near y/S=1.0 indicate that the cross-pitch flow is much stronger at the flat endwall than at the contour endwall. As this component of the flow is one of the major contributors for the growth of the passage vortex, the passage vortex is smaller near the contour endwall. The passage vortices can be identified by the small turning vectors creating an apparent clockwise and counterclockwise motion near the contour endwall and flat endwall, respectively. The static pressure distributions near the endwalls along the passage in figure 30 illustrate further the effects of the profiled endwall on the crosspitch and secondary flows. The contoured profile for the data in figure 30 is similar to that in figure 29 except the profile now extends across the entire passage length. The other endwall of the passage is flat without any contouring. In figure 30, the pressure distribution at the flat wall is similar to what is observed on the non-profiled endwalls in a linear vane passage. At the contoured endwall side, the contour lines of constant pressure near the leading edge are aligned more in the pitch direction than in the axial direction. In contrast, the constant pressure lines at the flat wall side near the leading edge are aligned more in the axial direction. The pressures at the contoured endwall are higher than those at the flat endwall for the first 40% axial chord. The pressures then are lower at the contoured endwall than at the flat endwall for the latter 60% axial chord. Thus, the pressure gradient at the contoured endwall is more parallel to the vane surface than to the pitch direction. The pressure gradient at the flat endwall is more parallel to the pitch direction than to the vane surface. As a consequence, the cross flow in the pitch direction is stronger on the flat endwall than on the contoured endwall. The reduced strength of the endwall cross flow then suppresses the growth of the passage vortex as mentioned earlier. The total pressure loss at the passage exit is reduced when the passage vortex near the contoured endwall is weakened and reduced in size. This is illustrated in the total pressure measurements in figure 3158. The planar vane cascade in the figure employs flat endwalls at both the hub and tip while the tip wall is axially contoured and the hub wall is flat for the contoured vane cascade. The contouring here extends across the entire passage length. The data in figure 31 are presented in a plane located 10% axial chord downstream of the passage exit. Hence, the endwall profile is flat at this location and the total passage height (z/S) is same for both the planar cascade and contoured cascade. The passage vortex regions in the figure can be identified by the highly concentrated circular contour lines near the endwalls. The parallel contour lines about y/P=0.50 indicate the wake region. The loss distributions are almost symmetric about the mid-span location z/S=0.50 for the planar cascade and the passage vortices are located away from the endwall regions as expected. While the loss distributions are asymmetric in the contoured cascade, the passage vortex loss region about z/S=0.10 from the flat wall side is similar to the passage vortex loss region in the planar cascade. However, the core loss region of this passage vortex has shifted closer to the flat wall side and further away from the suction side trailing edge.
Fig. 30. Static pressure distributions near endwalls in a linear vane passage with endwall profiling extending from leading edge to trailing. Source: See Note 55. (Shih)
Fig. 31. Total pressure loss, Cpt distributions at 1.10Cax for a linear vane cascade with and without endwall profiling. Source: See Note 55. (Dossena)
380
Sumanta Acharya
381
The loss contours due to the passage vortex from the contoured endwall side in the contoured cascade are located just adjacent to the contoured wall side at z/S=0.95. Thus, the spanwise (z/ S) extent of the passage vortex at the contoured endwall side is smaller compared to those in the planar cascade. Another important difference between the two cascades is identified by comparing the wake regions. The pitchwise width of the wake region for the contoured cascade is smaller than that for the planar cascade. This can be attributed to the velocity and pressure distributions on the vane suction surface in the contoured cascade. The lowest pressure and consequently the peak velocity on the vane suction surface shifts toward the trailing edge in the presence of the contoured endwall. Thus, the diffusion rate of velocity is lower over the suction surface in the contoured cascade59. As a result the extent of the adverse pressure gradient region near the suction surface trailing edge decreases reducing the extent of the trailing edge wake. Thus, the profile loss in the contoured cascade also decreases. The effect of axial profiling of endwall in an annular vane passage was measured60. The endwall contouring is employed at the tip wall in the last half of the passage. The contouring affects the pressure distributions significantly on the vane suction surface at the tip region. The suction surface pressure at the tip region rises in the first 50% of axial chord. In the latter 50% of axial chord, the pressure decreases on the suction surface and the maximum suction side velocity shifts toward the trailing edge at the tip region. As a result, the adverse pressure gradient on the blade suction surface is reduced at the vane tip region. In the annular flow area, the static pressure and total pressure loss distributions are affected in the latter part of the passage where the tip contouring is located. In the aft part of the passage, the radial pressure gradient directed toward the hub endwall in an annular passage with no tip wall contouring is inverted in the upper half span near the suction side when the tip contouring is employed. Unlike the annular passage without tip contouring, the radial flow angle at the exit plane is negative across most of the plane with the tip wall contouring. Downstream of the passage exit the total pressure loss region due to wake reduces in the pitch/circumferential direction significantly for the contoured annular passage (tip wall profiling) compared to that for the plain annular passage. This occurs as the adverse pressure gradient at the vane surface trailing edge reduces for the contoured tip/casing endwall. The same relative behavior for the wake region has been observed for the linear vane cascade with and without tip wall contouring61. At the same downstream location as above, Boletis (1985) shows that the high total pressure loss region near the contoured casing wall (annular) is also reduced compared to that near the plain annular casing wall. However, the magnitudes of the total pressure loss in this region are about the same for both type of casing walls. (ii) Non-axisymmetric Profiling of Endwall: In this case, the endwall profile variations can be achieved by varying the height of the profile over a smooth curve in the axial direction and over another smooth curve in the pitch direction. The objective is to increase the endwall height near the passage pressure side and decrease the endwall height near the suction side with respect to a baseline flat endwall. The endwall region static pressure on such profile is expected to decrease near the pressure side and increase near the suction side, thus reducing the pitchwise pressure gradient and the strength of the cross-pitch flows at
Blade LE Pres s side ure
Endwall
on cti Su e sid
Pitchwise distance
Endwall profile height (m)
Fig. 32. Non-axisymmetric profile of endwall employed in a linear blade cascade. Source: See Note 62.
Fig. 33. Measured static pressure and computed surface streamlines at a flat endwall and at a non-axisymmetric contoured endwall in a linear blade passage. Ps= wall static pressure.
4.3 Turbine Blade Aerodynamics the endwall. Figure 32 shows the profile of such an endwall that is employed62. The figure also includes the profile height variations across the passage. Harvey et al and Hartland et al. provide guidelines for designing the non-axisymmetric contour profiles for the linear cascade63. The measured static pressure distributions and computed surface streamlines at a flat endwall and at the contoured endwall of figure 32 are presented in figure 33 for a linear blade passage. The other endwall profile of the passage is always flat in this case. The surface static pressure Ps on the contoured endwall in figure 33 increases near both the pressure side and suction side compared to the Ps at the same locations on the flat endwall. But, the pitchwise pressure gradient, that drives the cross-pitch flow, in the first 40% axial chord decreases for the contoured endwall compared to that for the flat endwall. This clearly affects the cross-pitch flow on the endwall as shown in figure 33. The turning of the streamlines near the leading edge is much lower on the contoured endwall than on the flat endwall. The distance of the saddle point from the leading edge is also smaller for the contoured endwall than for the flat endwall. This indicates that the leading edge horse-shoe vortex is smaller in size above the contoured endwall. Inside the passage, the streamlines are also turning less toward the suction surface and appear to be more parallel to the blade surface on the contoured endwall. This occurs as the strength of the cross-pitch flow near the contoured endwall is decreased. The consequences of the results in figure 33 are weaker passage vortex and lower total pressure loss across the blade passage with the non-axisymmetric contoured endwall. These will be shown next. Figure 34 shows the streamlines in a pitchwise plane located 9% axial chord down the passage. The blade profile and contoured endwall profile are identified as solid objects in the figure. The structure of the pressure side leg vortex at this location is very clear near the pressure side of the flat endwall case. On the other hand, the streamlines near the pressure side for the contoured endwall case have not completed the full revolution to create a vortex structure. This happens as the pressure side leg vortex is weakened and reduced in size by the contoured endwall. As the pressure side leg vortex is driven from the horse-shoe vortex, this also validates the assertion that the horse-shoe vortex reduces with the contoured endwall. The passage vortex can be identified in figure 34 at the suction side where the total pressure loss coefficients, Cpt are very high. The extent of the passage vortex can be considered for Cpt>0.40 in this case. Then, clearly the passage vortex size for the contoured endwall passage is about half of that for the flat endwall linear cascade. The magnitudes of Cpt also indicate that the passage vortex is much weaker for the contoured endwall as the Cpt are lower for the contoured endwall at the passage vortex location than for the flat endwall. As such, the mass-averaged total pressure loss across the passage reduces significantly with the contoured endwall. Several other profiles of the non-axisymmetric contoured endwall have been tested successfully in blade and vane passages64. The results are similar to what we have discussed so far. These endwall profiles reduce the total pressure loss across the blade passage by weakening the endwall cross flows and passage vortex. Endwall Film Injection: Coolant air injected through tiny holes in the endwall covers the endwall with a layer of film of cold air and protects the endwall from the hot gas streak in the blade passage (figure 35). Wall static pressure changes in the vicinity of the coolant injection holes as the coolant jet blocks the boundary
Fig. 34. Streamlines and total pressure loss coefficients with and without non-axisymmetric contoured endwall showing pressure side leg vortex and passage vortex. Source: See Note 62.
Fig. 35. Coolant injection through holes in endwall for film cooling. L= hole length and D= characteristic scale of hole shape.
382
Sumanta Acharya layer and interacts with the boundary layer fluid downstream of the hole. Thus, the coolant jets significantly influence the endwall pressure field and the cross-pitch flow in the blade passage. As the coolant jets also interact with the vortex flows, especially with the pressure side leg vortex and suction side leg vortex along the separation line, the secondary flow dynamics changes along with the total pressure losses across the passage. As mentioned in the beginning of Section 4.3-7 that the coolant film is lifted away from the endwall by these vortices, the separation lines in the passage is sometimes termed as the “liftoff” line. The physics of mixing and diffusion of the jets in the boundary layer and interactions between a coolant jet and the boundary layer or the vortices are frequently complicated by the action of neighboring coolant jets. Such physics are studied under the subject of “Jets in Cross Flow” and hence, will not be discussed here. The primary objective of this section is to discuss the effects of the film injection on the secondary flow field and not the dynamics of the jets. The arrangement of the coolant holes in figure 35 is expected to provide coverage for the entire passage endwall and hence, sometimes is termed as the full-coverage film cooling. The effective coverage of the endwall by the coolant depends on various factors like injection angle, coolant hole orientation, coolant hole shape, hole size, L/D ratio, relative locations of the holes, and mass flux or local blowing ratio from individual hole. These are also the fundamental characteristics of the coolant holes and must all play the role together when the coolant holes are employed. The local blowing ratio is defined as the ratio of the mass flux of the coolant to the mass flux of the passage flow. It is not always easy to measure the mass flux of individual holes with accuracy. Thus, an inlet blowing ratio, Minlet is defined based on coolant flow through an idealized, loss free hole at the passage inlet condition65.
M inlet =
383
Po , plenum − Po ,inlet P0,inlet − Pinlet
Here, Po,plenum is the stagnation pressure of the coolant supply plenum, Po,inlet is the stagnation pressure at the passage inlet, and Pinlet is the static pressure at the passage inlet. The boundary layer is energized and strengthened with the properly ejected Fig. 36. Coolant jet locations downstream of injection holes in a linear coolant jets. This enables the boundary layer fluid to withstand vane passage. the pitchwise pressure gradient in the passage and cross-pitch flow is weakened as a consequence. Thus, with proper design Source: See Note 66. and configuration the endwall film injection can also provide an effective structural modification that reduces the secondary flows and aerodynamic losses. It is difficult to generalize the flow field at the endwall when the coolant jets are ejected. Each geometric configuration and flow parameter associated with the coolant holes just mentioned can alter the endwall boundary layer uniquely. On the other hand, the secondary and cross flows affect the coolant jets. The illustrations that are going to be presented next do not represent a typical behavior of the near wall flow. The readers will have some understanding about the relative importance of the coolant jet configuration and secondary flows. Figure 36 shows the measured locations of the coolant jets as they travel downstream from their ejection points in a linear vane passage66. The configuration of the coolant holes is also shown in the figure. The holes are ejecting at 35 degree with respect to the endwall surface. Data are presented near
4.3 Turbine Blade Aerodynamics the endwall where the quantities are affected the most. Location P1 is located along a pitchwise plane just downstream of the first row of holes and location P2 is located along a pitch plane just downstream of the 3rd and last row of holes. The coolant concentration is defined as the ratio of coolant density to the free-stream density67. The coolant concentration is the highest at the core of jet. Also, the coolant jets have forced vortex motions at the core68. Thus, the locations of the coolant jets in figure 36 can be readily identified at the locations of high turbulence intensity and coolant concentration. At location P1, ten distinct jets with high magnitudes from the ten holes upstream are clearly identifiable. The four jets near the suction side are attached to the endwall while the jets nearest the pressure side appear to be slightly lifted up from the endwall. Local blowing ratio is high near the suction side because the wall static pressure is low there and local blowing ratio is low near the pressure side because the wall static pressure is high there. Thus, the jets have higher momentum near the suction side than the jets near the pressure side. High momentum reduces mixing of the jets with the surrounding flow. Also, the suction side main flow has higher kinetic energy than the pressure side main flow. The high kinetic energy bends the suction side jets toward the endwall and high momentum aids the process. On the other hand, the low momentum jets near the pressure side easily penetrate the low energy main flow and lifts up from the endwall. At location P2 of figure 36, turbulence intensity Tu>11% near the endwall across the pitch is caused by the combined effects of all coolant jets as Tu is maximum 11% at this location without any coolant injection. Besides the signatures of three jets (Tu≥19%) from the last row of holes near the pressure side, no other jets are distinct at this location. The higher coolant concentration near the pressure side is the result of large number of jets near the pressure side compared to the number of jets near the suction side. In addition, some jets in the first and second rows are directly lifted away from the endwall by the up-wash flows of the pressure side leg vortex and suction side leg vortex. This action mixes the jets easily with the main stream and the coolant concentration from these jets reduces significantly. Some jets on the pressure side may also have been swept toward the middle of the passage by the cross flow and pressure side leg vortex. Thus, the coolant concentration is the highest near y/P=0.45 at location P2. Similar results about the locations of the coolant jets are reported in a linear vane passage69. The effectiveness of individual coolant jet is largely dependent upon its ability to stick persistently to the endwall to provide the maximum coverage. The location chosen for a coolant hole is therefore very important in this respect. Figure 37 provides evidence by how strongly the secondary flows deflect and block some coolant jets simply because of their location70. The coolant holes shown in figure 37 are arranged in four pitchwise rows at upstream of leading edge, 30% axial chord, 60% axial chord, and 90% axial chord. Four individual holes are also located at the pressure side of the blade passage. All the holes have same shape and geometry. The dark traces on the endwall are produced by the ejected coolant jets as they travel along the endwall. The length, level of darkness, and lateral spreading of the traces indicate the distance traveled by the jets, level of consistency of coolant, and lateral coverage by the jets, respectively, before they are mixed with the main fluid. Ammonia gas mixed with the coolant air stream reacts with the Diazo coating on the endwall and produces such traces71. Surface flow visualization as the coolant jets ejecting indicates the separation or lift-off lines of the pressure side leg/passage vortex and suction side leg vortex in figure 37. The five holes from the pressure side at 30% axial chord and the holes at the last two rows are located downstream of the lift-off line for the pressure side leg vortex. Friedrichs et al. (1996) shows that this line has moved downstream compared to that without coolant injection. The 4th and 5th
Fig. 37. Visualization of surface flow and coolant jet trajectories along endwall in a linear blade passage at Minlet=1.0. Source: See Note 70.
Fig. 38. Passage vortex and total pressure losses at exit flow with (Minlet=2.0) and without coolant injection in a linear blade passage. Source: See Note 65.
384
Sumanta Acharya coolant jets from the pressure side at 30% axial chord seem to dislocate the lift-off line slightly downstream from their positions. The jet traces from the holes located upstream of the passage in figure 37 shows virtually very little or no traces near the pressure side and all jets near the suction side are deflected around the lift-off line. The strong leading edge horse-vortex is either lifting the jets or deflecting the jets even if they are shooting directly toward the leading edge. The jets from the middle holes in the upstream row have low momentum and are aimed toward the pressure side while the cross flow here is directed in the axial direction have low kinetic energy. Thus, these jets have small trajectories and are easily swept in the main flow as soon as they are ejected. The traces of the jets from the holes inside the passage, except those near the pressure side, are swept toward the suction side by the cross flow. Similar behavior of the jets inside a blade passage is also observed72. The 3rd to 5th jet from the pressure side at 30% axial chord and 2nd to 4th jet from the suction side at 60% axial chord are additionally pulled in by the passage vortex (see the lift-off line) and in cases, are entrained into the passage vortex at the hole location itself. The traces nearest the pressure side are almost parallel to the pressure side as the boundary layer is very thin here and flow behaves as inviscid. A single jet from one of the four holes at the pressure side corner interacts with and strengthens the jet downstream and the combined jet trajectory is very long at this location. The jets at 90% axial chord and nearest the suction side in the previous two rows eject with high momentum due to the low wall static pressure. The main flow kinetic energy is also high at these locations because of its high speed. These keep these jet traces narrow and stick to the endwall for a longer distance. Also note that the last row of jets are covering a large area on the endwall as they are swept toward the suction side by the cross flow. These jets affect the cross flow as seen in the surface flow visualization of figure 37 (the top image). The streamlines downstream of the holes at 90% axial chord are parallel to the passage rather than being turned toward the suction side as compared to the streamlines upstream of these holes. Thus, these jets have weakened the cross flow near the passage exit. The effects of the same cooling holes as in figure 37 on the passage vortex structure and total pressure losses at the exit flow are shown in figure 3873. But, the inlet blowing ratio for the data with the endwall coolant injection is now 2.0. The plots with no coolant injection are included in figure 38 for comparison. The dashed lines in the figure indicate the spanwise locations of the passage vortex cores. The passage vortex is identified in the vector plot at the location of the clockwise rotation and in the total pressure loss contour at the location of circular region with high loss magnitudes. As noted in both the vector and contour plots, the passage vortex with coolant flow is located much nearer the endwall than with no coolant flow. The momentum of the ejected coolant adds energy to the boundary layer fluid. Therefore, when the passage vortex entrains these boundary layer fluids, the total pressure losses near the bottom part of the passage vortex are reduced. Coolant jets can be injected from continuous slots located in the upstream endwall/platform of the blade passage inlet. This type of coolant flow is often termed as the slot-bleed injection. The readers are referred to note 74 for information on the secondary flow field behavior with the slot-bleed74.
4.3-9 Notes ______________________________ 1. B. Lakshminarayana, Fluid Mechanics and Heat Transfer of Turbomachinery (New York:John Wiley & Sons Inc., 1996). 2. S. L. Dixon, Fluid Mechanics, Thermodynamics of Turbomachinery, 3rd ed. (Oxford: Butterworth-Heinemann Ltd., 1995). 3. H. E. Gallus, J. Zeschky, and C. Hah, “Endwall and Unsteady Flow Phenomena in an Axial Turbine Stage,” ASME Tran. J. Turbomachinery 117 (1995): 562-570; E. Boletis, “Effects of Tip Endwall Contouring on the Three-Dimensional Flow Field in an Annular Turbine Nozzle Guide Vane: Part 1- Experimental Investigation,” ASME Tran. J. Engr for Gas Turbines and Power 107 (1985): 983-990; C. H. Sieverding, W. Van-Hove, and E. Boletis, “Experimental Study of the Three-Dimensional Flow Field in an Annular Turbine Nozzle Guidevane,” ASME Tran. J. Engr for Gas Turbines and Power 106 (1984): 437-444. 4. A. Duden, I. Raab, and L. Fottner, “Controlling the Secondary Flow in a Turbine Cascade by Three-Dimensional Airfoil Design and Endwall Contouring,” ASME Tran. J. Turbomachinery 121(1999): 191-199; S. P. Harasgama and C. D. Burton, “Film Cooling Research on the Endwall of a Turbine Nozzle Guide Vane in a Short Duration Annular Cascade: Part 1- Experimental Technique and Results,” ASME Tran. J. Turbomachinery, Vol. 114 (1992): 734-740. 5. R. P. Dring and W. H. Heiser, Turbine Aerodynamics, Chap.4 in Aerothermodynamics of Aircraft Engine Components, AIAA education series (New York: AIAA Inc., 1985). 6. L. Fielding, Turbine Design- The Effect of an Axial Flow Turbine Performance of Parameter Variation, (New York: ASME Press, 2000); J.P. Gostelow, Cascade Aerodynamics, (OxfordPergamon Press Ltd., Oxford, U.K., 1984).
385
4.3 Turbine Blade Aerodynamics 7. J. L. Kerrebrock, Aircraft Engines and Gas Turbines, 2nd ed. (Massachusetts: The MIT Press, 1992). 8. H. Cohen, G. F. C. Rogers, G.F.C., and H.I.H.Saravanamuttoo, Gas Turbine Theory, 4th ed. (Essex, U.K.: Longman Group Ltd., 1996); also see note 5 above. 9. H. P. Wang, S. J. Olson, R. J. Goldstein, and E.R.G. Eckert, “Flow Visualization in a Linear Turbine Cascade of High Performance Turbine Blades,” ASME Tran. J. Turbomachinery 119 (1997): 1-8. 10. Ibid. 11. Ibid. 12 R. Niehuis, P. Lücking, and B. Stubert, “Experimental and Numerical Study on Basic Phenomena of Secondary Flows in Turbines,” AGARD Conf. Proc. No. 469, Secondary Flows in Turbomachines (1990): 5.1-5.17. The original version of this material was published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in AGARD Conference Proceedings CP-469 ”Secondary Flows in Turbomachines” in 1990. 13. L.S. Langston, “Research on Cascade Secondary and Tip-Leakage Flows- Periodicity and Surface Flow Visualization,” AGARD Conf. Proc. No. 469, Secondary Flows in Turbomachines (1990): 19.1-19.15; S.H. Moustapha, G.J.Paron, and J.H.T. Wade, “Secondary Flows in Cascades of Highly Loaded Turbine Blades,” ASME Tran. J. Engr. Gas Turbines and Power 107 (1985): 1031-1038. The original version of this material was published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in AGARD Conference Proceedings CP-469 ”Secondary Flows in Turbomachines” in 1990. 14. H.P. Hodson and R. G. Dominy, “Three-Dimensional flow in a Low-Pressure Turbine Cascade at Its Design Condition,” ASME Tran. J. Turbomachinery 109 (1987): 177-185. 15. Ibid. 16. D.G. Gregory-Smith, C.P. Graves, and J.A. Walsh , “Growth of Secondary Losses and Vorticity in an Axial Turbine Cascade,” ASME Tran. J. Turbomachinery 110 (1988): 1-8. 17. E. Detemple-Laake, “Measurement of the Flow Field in the Blade Passage and Side-wall Region of a Plane Turbine Cascade,” AGARD Conf. Proc. No. 469, Secondary Flows in Turbomachines (1990): 10.1-10.13. The original version of this material was published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in AGARD Conference Proceedings CP-469 ”Secondary Flows in Turbomachines” in 1990. 18. Ibid. 19. Ibid. 20. W.A. Eckerle and L.S. Langston, “Horseshoe Vortex Formation Around a Cylinder,” ASME Tran. J. Turbomachinery 109 (1987): 278-285. 21. See notes 13 and 20 above. 22. G.I. Mahmood, R. Gustafson, and S. Acharya, “Experimental Investigation of Flow Structure and Nusselt Number in a Low Speed Linear Blade Passage With and Without Leading Edge Fillets,” ASME Tran. J. Heat Transfer 127 (2005): 499-512. 23. See note 20 above. 24. See note 9 above. 25. G.I. Mahmood and S. Acharya, “Experimental Investigation of Secondary Flow Structure in a Blade Passage With and Without Leading Edge Fillets”, in review, ASME Trans. J. Fluids Engineering (2005);A.K. Saha, G.I. Mahmood, R. Gustafson, and S. Acharya, “Predicted and Measured Flow Field and Heat Transfer in a Linear Blade Cascade Employing Fillets,” in preparation for the ASME Tran. J. Turbomachinery. 26. See note 9 above. 27. See note 9 above. 28. See note 9 above. 29. See note 9 above. 30. See note 9 above. 31. R. J. Goldstein, H.P. Wang, and M.Y. Jabbari, “The Influence of Secondary Flows Near the Endwall and Boundary Layer Disturbance on Convective Transport From a Turbine Blade,” ASME Tran. J. Turbomachinery 117(1995): 657-665. 32. See note 25 above. 33. M.B. Kang and K.A. Thole, “Flowfield Measurements in the Endwall region of a Stator Vane,” ASME Tran. J. Turbomachinery 122 (2000): 458-466. 34. See note 14 above. 35. See note 25 above. 36. D. G. Gregory-Smith and J.G.E. Cleak, “Secondary Flow Measurements in a Turbine Cascade With High Inlet Turbulence,” ASME Tran. J. Turbomachinery 114 (1992): 173-183; C. Hah, “A Navier-Stokes Analysis of ThreeDimensional Turbulent Flows Inside Turbine Blade Rows at Design and Off-Design Conditions,” ASME Tran. J. Engr. Gas Turbines and Power 106 (1984): 421-429; L.S. Langston, “Crossflows in a Turbine Cascade Passage,” ASME Tran. J. Engr. for Power 102 (1980): 866-874.
386
Sumanta Acharya
387
37. K. Hermanson, S. Kern, G. Picker, and S. Parneix, “Predictions of External Heat Transfer For Turbine Vanes and Blades With Secondary Flowfields,” ASME Proc. Turbo Expo, GT-2002-30206, 2002. 38. See note 33 above. 39. A. Yamamoto, “Production and Development of Secondary Flows and Losses in Two Types of Straight Turbine Cascades: Part 1- A Stator Case,” ASME Tran. J. Turbomachinery 109 (1987): 186-193;R.P. Roy, K.D. Squires, M. Gerendas, S. Song, W.J. Howe and A. Ansari, “Flow and Heat Transfer at the Hub Endwall of Inlet Vane PassagesExperiments and Simulations,” ASME Proc. Turbo Expo, 2000-GT-198, 2000. 40. See note 3 above. 41. See note 3 above. 42. See note 3 above. 43. See note 13 above. 44. See note 13 above. 45. R.G. Dominy and S.C. Harding, “An Investigation of Secondary Flows in Nozzle Guide Vanes,” AGARD Conf. Proc. No. 469, Secondary Flows in Turbomachines (1990): 7.1-7.15. 46. See note 3 above (Gallus). 47. M.V. Hoyningen-Huene, W. Frank, and A.R. Jung, “Three-Dimensional Time-Resolved Flow Field in the First and Last Turbine Stage of a Heavy Duty Gas Turbine, Part I: Secondary Flow Field,” ASME Proc. Turbo Expo, 2000-GT-0438, 2000. 48. J. Zeschky and H.E. Gallus, “Effects of Stator Wakes and Spanwise Nonuniform Inlet Conditions on the Rotor Flow of an Axial Turbine Stage,” ASME Tran. J. Turbomachinery 115 (1993): 128-136. 49. H. Sauer, R. Müller, and K. Vogeler, “Reduction of Secondary Flow Losses in Turbine Cascades by Leading Edge Modifications at the Endwall,” ASME Tran. J. Turbomachinery 123 (2001): 207-213; G. A. Zess and K.A. Thole, “Computational Design and Experimental Evaluation of Using a Leading Edge Fillet on a Gas Turbine Vane,” ASME Proc. Turbo Expo, GT-2001-0404, 2001; A.T. Lethander, K.A. Thole, G. Zess, and J. Wagner, “Ortimizing the VaneEndwall Junction to Reduce Adiabatic Wall Temperatures in a Turbine Vane Passage,” ASME Proc. Turbo Expo, GT2003-38939, 2003; S. Becz, M.S. Majewski, and L.S. Langston, “Leading Edge Modification Effects on Turbine Cascade Endwall Loss,” ASME Proc. Turbo Expo, GT-2003-38898, 2003;S. Becz, M.S. Majewski, and L.S. Langston, “An Experimental Investigation of Contoured Leading Edges for Secondary Flow Loss Reduction,” ASME Proc. Turbo Expo, GT-2004-53964, 2004; also see notes 22 and 25 above. 50. See notes 22 and 25 above. 51. See note 21 above. 52. See note 25 above. 53. See note 49 above. 54. See note 49 above (Zess). 55. S.W. Burd and T.W. Simon, “Flow Measurements in a Nozzle Guide Vane Passage With a Low Aspect Ratio and Endwall Contouring,” ASME Proc. Turbo Expo, 2000-GT-0213, 2000; T.I-P Shih, Y.-L Lin, and T.W. Simon, “Control of Secondary Flows in a Turbine Nozzle Guide Vane by Endwall Contouring,” ASME Proc. Turbo Expo, 2000-GT-0556, 2000; V. Dossena, A. Perdichizzi, and M. Savini, “The Influence of Endwall Contouring on the Performance of a Turbine Nozzle Guide Vane,” ASME Tran. J. Turbomachinery 121(1999): 200-208; F.C. Kopper, R. Milano, and M. Vanco, “Experimental Investigation of Endwall Profiling in a Turbine Vane Cascade,” AIAA Journal, AIAA 80-1089R 19, No. 8 ( August 1981). 56. S. Acharya, “Eendwall Cooling With Endwall Contouring and Leading Edge Fillet,” Smi-annual Report Submiited to UTSR, South Carolina, Project No. 02-01-SR098, June 2003-December 2003; D.E. Bohn, K. Kusterer, N. Sürken, and F. Kreitmeler, “Influence of Endwall Contouring in Axial Gaps on the Flow Field in a Four-Stage Turbine,” ASME Proc. Turbo Expo, 2000-GT-472, 2000; L.P. Timko, “Energy Efficient Engine High Pressure Turbine Component Test Performance Report,” Contract Report for NASA, Report No. NASA CR-168289; also see note 55 above (Shih). 57. See note 55 above (Burd). 58. See note 55 above (Dossena). 59. Ibid. 60. See note 3 above (Boletis). 61. See note 55 above (Dossena). 62. R. Gustafson, G.I. Mahmood, and S. Acharya, “Control of Secondary Flows in a Low Speed Blade Cascade with a Nonaxisymmetric 3-D Endwall,” in preparation for the ASME Trans J. Turbomachinery. 63. N. W. Harvey, M.G. Rose, M.D. Taylor, S. Shahpar, J. Hartland, and D.G. Gregory-Smith, “Nonaxisymmetric Turbine End Wall Design: Part I- Three-Dimensional Linear Design System,” ASME Tran. J. Turbomachinery 122 (2000): 278-285; J.C. Hartland, D.G. Gregory-Smith, and M.G. Rose, “Non-axisymmetric Endwall Profiling in a Turbine Rotor Blade,” ASME Proc. Turbo Expo, 98-GT-525, 1998. 64. G. Ingram, D.G. Gregory-Smith, M. Rose, N. Harvey, and G. Brennan, “The Effect of End-wall Profiling on Secondary Flow and Loss Development in a Turbine Cascade,” ASME Proc. Turbo Expo, GT-2002-30339, 2002; J.C. Hartland, D.G. Gregory-Smith, N.W. Harvey, and M.G. Rose, “Nonaxisymmetric Turbine End Wall Design: Part II- Experimental
4.3 Turbine Blade Aerodynamics Validation,” ASME Tran. J. Turbomachinery 122 (2000): 286-293; J. Yan, D.G. Gregory-Smith, and P.J. Walker, “Secondary Flow Reduction in a Nozzle Guide Vane Cascade by Non-axisymmetric End-wall Profiling,” ASME Proc. Turbo Expo, 99-GT-339, 1999. 65. S. Friedrichs, H.P. Hodson, and W.N. Dawes, “Aerodynamic Aspects of Endwall Film-Cooling,” ASME Tran. J. Turbomachinery 119 (1997): 786-793. 66. F. Kost, and F. Nicklas, “Film-Cooled Turbine Endwall in a Transonic Flow Field: Part I- Aerodynamic Measurements,” ASME Proc. Turbo Expo, 2001-GT-0145, 2001. 67. Ibid. 68. F. Bario, F. Leboeuf, A. Onvani, and A. Seddini, “Aerodynamics of Cooling Jets Introduced in the Secondary Flow of a Low-Speed Turbine Cascade,” ASME Tran. J. Turbomachinery 112 (1990): 539-546. 69. Ibid. 70. S. Friedrichs, H.P. Hodson, and W.N. Dawes, “Distribution of Film-Cooling Effectiveness on a Turbine Endwall Measured Using the Ammonia and Diazo Technique,” ASME Tran. J. Turbomachinery 118 (1996): 613-621. 71. M.Y. Jabbari, K.C. Marston, E.R.G. Eckert, and R.J. Goldstein, “Film Cooling of the Gas Turbine Endwall by DiscreteHole Injection,” ASME Tran. J. Turbomachinery .118 (1996):. 278-284; also see note 70 above. 72. Ibid. 73. See note 65 above. 74. H.D. Pasinato, Z. Liu, R.P. Roy, W.J. Howe, and K.D. Squires, “Prediction and Measurement of the Flow and Heat Transfer Along the Endwall and Within an Inlet vane Passage,” ASME Proc. Turbo Expo, GT-2002-30189, 2002; S.W. Burd and T.W. Simon, “Effects of Slot Bleed Injection Over a Contoured Endwall on Nozzle Guide Vane Cooling Performance: Part I- Flow Field Measurements,” ASME Proc. Turbo Expo, 2000-GT-0199, 2000; R. Oke, T.W. Simon, T. Shih, B. Zhu, Y.L. Lin, and M. Chyu, “Measurements Over a Film-Cooled, Contoured Endwall with Various Coolant Injection Rates,” ASME Proc. Turbo Expo, 2001-GT-0140, 2001.
388
BIOGRAPHY
4.3 Turbine Blade Aerodynamics
Sumanta Acharya Louisiana State University CEBA 1419B, Mechanical Engineering Department Baton Rouge, LA 70803 phone: (225) 578-5809 email: [email protected]
Sumanta Acharya is currently the L. R. Daniel Professor in Mechanical Engineering, and the Director of the Turbine Innovation and Energy Research (TIER) Center at Louisiana State University (LSU), Baton Rouge . He received his B.S. in Mechanical Engineering from IITKharagpur, India, in 1978, his Ph.D. in Mechanical Engineering from the University of Minnesota in 1982, and has been on the faculty at LSU since 1982. He has worked in several areas pertaining to turbine blade cooling and aerodynamics including internal cooling, film cooling, blade tip leakage flows, and 3D endwall flows (hub). In the gas turbine area he has received funding from Pratt & Whitney, General Electric, the University Turbine Systems Research Program (UTSR) of the Department of Energy, the U. S Navy, the Air Force Office of Scientific Research, the National Science Foundation, and the State of Louisiana. He has published 130 journal articles and book chapters, and presented over 160 conference papers.
Gazi Mahmood Louisiana State University CEBA 1419B, Mechanical Engineering Department Baton Rouge, LA 70803
Dr. Gazi Mahmood is currently working as a postdoctoral research associate at the Turbine Innovation and Energy Research (TIER) Center of the Louisiana State University. He received his B.Tech. in Mechanical Engineering from the Indian Institute of Technology in Bombay in 1995 and his Ph.D. in Mechanical Engineering from the University of Utah in 2001. The research interests of Dr. Gazi Mahmood include turbulent channel flows, fluid dynamics and convective heat transfer on special surface structures, aerodynamics, gas turbine cooling, pumps, infrared imaging techniques, flow control on wing-shaped bodies, fluidic actuators, and miniature contact probes for flow measurements. Presently, he has been working in a research project investigating the aerodynamic performances and endwall cooling in the gas turbine passages with endwall contouring and leading edge fillets. His work experiences also include research on dimples, pin-fins, and vortex passage employed in heat exchangers, wing bodies, bearing cooling, and turbine blade cooling. He was a lecturer at the Mechanical Engineering Department in Columbia University (New York) in 2001.
4.4.1
Buckets and Nozzles
4.4.1-1 Introduction Gas turbine engines, both aircraft and industrial power generation, represent one of the most aggressive applications for structural materials. With ever growing demands for increasing performance and efficiencies, all classes of materials are being pushed to higher temperature capabilities. These materials must also satisfy stringent durability and reliability criteria. As materials are developed to meet these demanding requirements, the processing of these materials often becomes very complicated and expensive. As a result, the cost of materials and processes has become a much larger consideration in the design and application of high performance materials. Both the aircraft engine and power generation industries are highly cost competitive, and market advantage today relies on reducing cost as well as increasing performance and efficiency. The firing temperatures of all gas turbines, both industrial power generation and aircraft engines, have increased over the past ~30 years. More recently, the rate of temperature increase has slowed for aircraft engines but not for industrial gas turbines (IGT). As a result, the materials temperature capability requirements for these two classes of gas turbines are converging. For many years, the high performance requirements of military and commercial aircraft engines fueled the development of advanced materials and processes. Many of these high temperature materials are now being used in industrial gas turbines as output, efficiency, and reliability requirements continue to grow. Directionally solidified and single-crystal nickel-base superalloys have been developed for investment casting of hot gas path components and have been scaled up to the part sizes required for IGT components but not without significant challenges in producibility, defect allowances, and repair. The application of nickelbase superalloys in industrial gas turbines has required particular emphasis on technology development for the production of buckets and nozzles in large IGT sizes. Processing scale-up from aircraft engine-sized parts to large IGT-sized parts has presented unique materials development and processing challenges. There has been much synergy in the development of these materials for both aircraft engines and industrial power generation turbines, and this synergy is likely to continue to grow as we strive to push materials capability to the limit while providing robust designs for reliable, long-life service.
4.4.1-2 Background
Stephen J. Balsone GE Gas Turbines LLC P.O. Box 648; GTTC 174D Greenville, SC 29602 Phone: 864 254-5294 Email: [email protected]
411
Higher operating temperatures are historically the primary means of improving aircraft engine thrust or industrial power generation gas turbine output. Higher operating temperatures require higher temperature capability materials and associated technologies such as improved oxidation and environmental coatings. For industrial power generation gas turbines, the firing temperature (as defined by the gas temperature that enters the first rotating stage of buckets or blades) has a profound effect on the performance of the turbine. Since the early 1970’s there has been a continuous increase in the output and efficiency of large industrial gas turbines (IGT) for electrical power generation. This increase is due in large part to the introduction of high temperature structural materials. The use of these advanced materials has resulted in an increase in gas turbine firing temperature from 982ºC (1800ºF) to greater than 1427ºC (2600ºF) over the past 30 years. For every 10ºC (50ºF) increase in the firing temperature, the gas turbine combined-cycle efficiency improves by approximately 1%. A 1% improvement in efficiency means millions of dollars in savings to an electrical power producer looking to deliver electricity at the lowest cost to its customers. Nickel-base (Ni-base) superalloys are the alloys of choice for high temperature, high strength structural applications, and they have become the standard for IGT hot gas path components such as buckets, nozzles, and shrouds. Many of these investment cast Ni-base superalloys were derived from aircraft engine alloys developed for use in both commercial and military aircraft gas turbines. In addition to investment cast Ni-base superalloys, other
high temperature materials are in production or being developed for IGT applications. High temperature coatings such as metallic coatings for oxidation and corrosion resistance and ceramic coatings for thermal protection are becoming standard for hot gas path and combustion hardware. Ceramic-matrix composites are also being developed for high temperature applications such as turbine shrouds, combustion liners, and turbine nozzles. In order to reduce the amount of cooling air required to keep materials within their high temperature capability, several options have been pursued. The obvious is the use of higher temperature capability materials. The second is the use of thermal barrier coatings (TBCs). The third is the incorporation of more efficient cooling schemes. These approaches are complementary, and system approaches have been developed that take advantage of all three. Advanced industrial gas turbines firing at temperatures of 1425ºC (2600ºF) and above require advanced cooling schemes to keep the metal temperatures within design limits. While aircraft engines have used advanced air cooling techniques including extensive use of convection and film cooling to maintain part temperatures at required levels, they are limited to these techniques because these are the only approaches that are available for turbines that must be flown on aircraft. However, industrial gas turbines do not have these weight limitations and therefore can use other cooling techniques such as steam cooling for buckets and nozzles. The first benefit of steam cooling is that it allows higher firing temperatures for a given set of parts lives or reliability goals. The higher thermal transport properties of steam enable designers to maintain metal temperatures within acceptable limits in higher gas temperature environments. Secondly, film cooling requires that relatively cool air be mixed with the high temperature gas as it passes through the first stage nozzles. This film lowers the temperature of the gas thus increasing the difference between combustor discharge temperature and the firing temperature for a set combustor discharge temperature. Therefore for a given NOx production rate, steam cooling permits higher firing temperatures and hence maximizes output and efficiency. Thus the use of steam cooling allows NOx emissions and efficiency goals to be met simultaneously.
4.4.1-3 Process Development – Investment Casting of DS and SX Alloys The investment cast Ni-base superalloys currently in IGT use have been principally derived from aircraft engine alloys developed to meet stringent high performance requirements of both commercial and military aircraft gas turbines. However, the introduction of these aircraft engine alloys in IGT hot gas path components has posed significant development challenges1. Figure 1 shows a General Electric F-class gas turbine first stage bucket in relation to a typical aircraft engine turbine blade2. This first stage bucket is a directionally solidified (DS) Ni-base superalloy made from GTD-111TM, an alloy derived from Rene’ 80 and developed specifically to meet property requirements for long-life operation in IGT’s. This first stage bucket demonstrates a greater than 10X increase in part size and a greater than 20X increase in part weight as compared with the aircraft engine blade. These DS Ni-base superalloy buckets are up to 76 cm (30 in) in length and weigh up to 18 kg (40 lbs). In the most advanced IGT’s, they are designed and manufactured with complex internal serpentine cooling passages. They are manufactured to very tight dimensional tolerances, and inspection and acceptance standards are today approaching aircraft engine requirements. The earliest industrial gas turbines used forged turbine buckets with cast nozzles. This changed in the late 60’s when castings began to be used in bucket applications. Many nozzle and bucket castings used in industrial gas turbines are made by using the conventional equiaxed investment casting process. Vacuum is used in most cases, except for some of the cobalt alloys, to prevent the highly reactive elements in the superalloys from reacting with the oxygen and nitrogen in the air. With proper control of metal and mold thermal conditions, the molten metal solidifies from the surface toward the center of the mold, creating an equiaxed structure. To prevent shrinkage porosity, care is taken to allow proper feeding of molten metal to the casting while it solidifies. Directional solidification (DS) was first introduced into industrial gas turbines in the late 1980’s. Although it has been in aircraft engines for more than 25 years, considerable process development was necessary to scale it up to the sizes of buckets used in industrial gas turbines. By exercising careful control over temperature gradients, a planar solidification front can be developed in these large buckets. The result is a bucket with an oriented grain structure that runs parallel to the major axis of the part and contains no transverse grain boundaries. The Fig. 1. Size comparison of an IGT first elimination of these transverse grain boundaries confers additional creep stage bucket with a typical aircraft enand rupture strength on the alloy, and the orientation of the grain structure gine turbine blade. (Courtesy of General provides a favorable modulus of elasticity in the longitudinal direction to Electric Company). TM enhance fatigue life. By directionally solidifying the alloy GTD-111 , an increase of approximately 23ºC (40ºF) in creep strength and an increase Source: See Note 2. of approximately 10X in fatigue life can be realized3.
412
Stephen J. Balsone Figure 2 shows General Electric’s H SystemTM gas turbine third stage and fourth stage buckets made of DS GTD-444TM Ni-base superalloy4. The alloy development of DS GTD-444TM serves as a good example of the alloy modification conducted to adapt an aircraft engine alloy for IGT application. DS GTD-444TM is a directionally solidified version of the first generation SX Ni-base superalloy Rene’ N4. When design requirements exceeded the material capability of DS GTD-111TM, the SX Rene’ N4 alloy was selected because of the alloy’s very good high temperature strength. Grain boundary element modifications were made to the alloy to produce a SX Rene’ N4 derivative as a DS alloy with improved large-part castability for IGT application. The result was a castable, high temperature DS Ni-base superalloy with improved creep and fatigue properties as compared with DS GTD-111TM. DS GTD-444TM was first introduced in production in 1999. More recently, industrial gas turbine OEM’s have worked with suppliers and universities to develop large, single-crystal (SX) castings that offer additional creep and fatigue benefits through the elimination of grain boundaries. While directionally solidifying GTD-111TM as a single crystal offers some improvements in creep and fatigue, much more significant improvements can be achieved by adopting alloys developed as single-crystal compositions for aircraft engine applications. The use of one such alloy, Rene’ N5, can produce an increase of more than 35ºC (~60ºF) in creep strength and 2X to 3X increase in fatigue life compared to DS GTD-111TM. However, to successfully produce large single crystals weighing more than 11 kg. (~25 lbs) and measuring more than 450mm (~18 inches) in length, modifications had to be made to the alloy. These modifications were necessary to reduce the reaction between the metal and mold materials to enable reasonable yields to be obtained when these buckets and nozzles are solidified for long times at high temperatures5. Figure 3 shows General Electric’s H SystemTM gas turbine first stage bucket made of SX Rene’ N56. This Ni-base superalloy was adopted directly from aircraft engine application and is currently used in production for first stage buckets, as well as first stage nozzles and shrouds, in General Electric’s H SystemTM and FB-class gas turbine product lines. These applications are among the first use of SX Ni-base superalloys in IGT’s. Significant material and processes development was required for the introduction of SX Ni-base superalloys in IGT applications, including SX investment casting equipment and technology, ceramic mold and core development, and post-cast joining, machining, coating, and inspection technology to fabricate these very complex, high performance airfoils. Critical processing technology developments were required to scale up the investment casting of Ni-base superalloys for IGT applications. DS and SX investment casting furnaces were scaled up to handle the size and weight of IGT buckets. Advances in mold materials and construction were required to hold the large volumes of molten metal during the DS and SX casting withdrawal process. Ceramic cores were improved to increase high temperature strength in order to minimize core deformation and hold critical dimensional tolerances. Processing considerations were also important to alloy selection and chemistry modifications, as Ni-base superalloys were adapted from aircraft engines to IGT applications. Alloy chemistries were modified to prevent formation of melt-related defects such as freckles, porosity, and hot tearing in the large IGT parts. In addition, minor alloying elements were adjusted to control grain boundary strength. These alloy chemistry changes to improve the castability of alloys in large sizes and to increase casting yields had to be balanced with the alloy mechanical property and environmental resistance requirements for IGT hardware for robust, long-life service. These materials and processes technology developments are often conducted in joint partnership between the gas turbine OEM and the investment casting suppliers.
413
Fig. 2. Third and fourth stage IGT buckets made from DS GTD-444TM. (Courtesy of General Electric Company). Source: See Note 2.
Fig. 3. First stage IGT bucket made from SX Rene’N5. (Courtesy of General Electric Company). Source: See Note 2.
4.4.1 Buckets and Nozzles The main drivers for yield in large DS and SX airfoil castings are dimensional and grain defects. Dimensions and associated tolerances result from the aerodynamic, mechanical, and heat transfer design of the part that must meet the gas turbine cycle requirements. Grain defects are metallurgical in nature and are controlled by local solidification, chemistry, heat treatment, and geometry. Typical grain defects encountered by developers of DS and SX IGT airfoils are shown in figure 47. Grain defects are typically dealt with by adjusting processing conditions, mold configurations, local geometry, and alloy chemistry8.
4.4.1-4 Process Development – High Gradient Casting High gradient is a term used to describe any one of a number of investment casting techniques used to solidify castings under well-controlled, large thermal gradients. This results in cast material have a much-refined microstructure with improved chemical homogeneity. In the conventional DS or SX investment casting process, heat from the mold is removed by conduction to a chill plate and by radiation to the surroundings, including heat loss from the open baffle in the bottom of the furnace during the mold withdrawal process. In a high gradient casting process, heat from the mold is removed by immersing the mold into a cooling media directly from the mold withdrawal from the furnace. The high gradient cooling media is often a bath of liquid metal such as aluminum or tin9. High gradient casting technology promises to increase casting yields and throughput by eliminating melt-related defects such as freckle formation in complex airfoil geometries. In addition, high gradient casting produces a significant refinement of the cast alloy microstructure, as often measured by the primary dendrite arm spacing. This microstructural refinement is also accompanied by a reduction in casting microporosity, a reduction in the volume fraction of eutectic phase, and an improvement in the morphology and distribution of carbides. The chemical homogeneity of the cast alloy is improved, and this allows Ni-base superalloys to be more fully solutioned during heat treatment, resulting in a significant increase in high temperature creep and fatigue properties10. This microstructural refinement and improved chemical homogeneity achievable by a high gradient casting process can provide both incremental gains and big leaps in investment casting technology. The improved materials capability provided by high gradient casting provides important design benefits such as component upgrades without redesign, the flexibility to balance higher component performance with extended component life, and the ability to cast more exotic, high performance, hard-to-cast Ni-base superalloys. In addition, a direct material substitution may provide cost reduction by replacing a more costly DS or SX Ni-base superalloy with a less expensive alloy at equivalent performance by improving the capability of the less costly alloy via a high gradient casting process.
4.4.1-5 Alloy Development – Buckets Over the past several decades, advanced airfoil alloy development for industrial gas turbines has progressed from poly-crystalline Ni-base superalloys such as U500, U700, and Alloy 738, to poly-crystalline and then directionally solidified superalloys such as (DS) GTD-111TM. GE’s most advanced DS Ni-base superalloy in production use is GTD-444TM, a DS version of the single-crystal (SX) Rene’ N4 alloy used in aircraft engine applications. Figure 5 shows a schematic of IGT airfoil alloy development, plotting application temperature as a function of first introduction. Note that figure 5 shows two paths for advanced IGT airfoil alloy development. One path has been followed for the evolution of cast Ni-base superalloys for latter stage bucket applications that require simple or no internal cooling. The second path shows the introduction of SX Rene’ N5, a second generation (containing rhenium) single-crystal Ni-base superalloy developed for aircraft engine applications. SX Rene’ N5 has been adopted for IGT use in first stage buckets having complex internal cooling schemes as well as first stage nozzle and shroud applications in General Electric’s most advanced IGT’s. Table I lists the alloy compositions of some important IGT investment cast Ni-base superalloys.
Temperature
Early Stages Complex Cooling
Rene N5 (SX) GTD-111TM (DS)
Rene 77 U500
1960 Fig. 4. Typical grain defects in SX and DS airfoil alloys that negatively impact casting yields. (Courtesy of General Electric Company).
GTD-111TM
738 GTD-111TM
1970
1980
Latter Stages Simple or No Cooling
GTD-444TM (DS)
(DS)
1990
2000
Fig. 5. IGT bucket alloy evolution showing increase in temperature capability. Source: See Note 2.
Source: See Note 7.
414
Stephen J. Balsone The higher firing temperatures of advanced industrial gas turbines place increased demands on protective coatings for hot gas path components. Coatings are required to protect the components from corrosion, oxidation, and mechanical property degradation. As nickel-base superalloys have become more complex, it has been increasingly difficult to obtain both the higher strength levels that are required and a satisfactory level of corrosion and oxidation resistance without the use of coatings. The function of all coatings is to provide a surface reservoir of elements that will form very protective and adherent oxide layers, thus protecting the underlying base material from oxidation and corrosion attack and degradation. Since the mid-1980s, an increase in the oxidation resistance of IGT airfoil alloys has been required to meet the higher firing temperatures of advanced industrial gas turbines. This demand was met with a series of M-CrAlY overlay coatings with increasing oxidation resistance. In addition to these simple overlay coatings, duplex coatings that consisted of an overlay M-CrAlY with a diffusion aluminide were also developed. The enhanced oxidation protection was achieved by an increased aluminum content in the outer region of the coating. As the internal temperatures of these hot section parts increased, it also became necessary to coat the internal surfaces and cooling holes to provide protection against oxidation and embrittlement that would otherwise occur. Air plasma sprayed yttria-stabilized zirconia thermal barrier coatings (TBCs) have been successfully used to extend life of superalloy components in aircraft engines and industrial gas turbines. A TBC is a multilayer thermal and environmental protection system consisting of an insulating layer of ceramic over a metallic bond coat for substrate oxidation protection and improved metal-ceramic interfacial properties. A typical air plasma sprayed (APS) TBC consists of a low thermal conductivity yttria stabilized zirconia (YSZ) ceramic top coat layer over a metallic bond coat layer. The top ceramic layer may be from 250 to 1250 microns (0.010 to 0.050 inch) thick while the metallic bond coat is typically from 200 to 300 microns (0.008 to 0.012 inch) thick. The bond coat is usually an M-CrAlY chosen to provide oxidation protection to the substrate metal and a rough surface for mechanical adhesion of the ceramic top coat. In conventional APS TBCs, the internal microcracks are “random” or non-directional. Advanced APS TBCs have been developed that mimic the oriented microstructure of vapor phase deposited TBCs. This advanced APS TBC can accommodate the mismatch in thermal strains between the ceramic top layer and the metallic bond coat and substrate. This allows the superior properties of this coating to be achieved by a coating process that can be used to apply TBCs to large industrial gas turbine parts.
4.4.1-6 Alloy Development – Nozzles Nozzles, also called vanes, are the stationary airfoils that direct the hot gas path flow for proper impingement on the rotating buckets. They must be able to withstand high temperatures. For example, the stage 1 nozzles are subjected to the highest gas path temperatures in the turbine but to lower mechanical stresses than rotating buckets. Much of the stress experience by the nozzles is a result of high thermal stresses and to a lesser degree, mechanical stresses such as the aerodynamic loading. As a result, the nozzles must have excellent resistance to thermal fatigue, as well as high temperature oxidation and corrosion resistance. Good creep resistance is also an important design consideration, especially for large, multi-airfoil latter stage nozzles whose size and weight will lead to creep deformation at temperature as the nozzles support their own weight (and the attached diaphragm structure) from the turbine case. Since the 1960’s, the cobalt-base alloy, FSX-414, has been the stage 1 nozzle alloy in General Electric’s E-class and F-class gas turbines11. FSX-414 has been the alloy of choice for nozzle applications for many years because of the superior highest temperature strength of cobalt alloys as compared with nickel-base alloys. In addition, cobalt-base alloys exhibit excellent corrosion resistance due to the level of Cr in the alloys. The oxidation resistance of FSX-414 is such that it is currently used without a protective oxidation coating. For the larger latter stage nozzles, GTD-222, a nickel-base superalloy, was developed to satisfy the requirement for higher creep resistance at relatively lower temperatures12. GTD-222 offered a creep strength improvement of approximately 60ºC (~150ºF) over FSX-414 at the typical operating temperatures for stage 2 and stage 3 nozzle applications. GTD-222 is often coated by an aluminiding process to enhance the oxidation resistance of the alloy. Both FSX-414 and GTD-222 possess good weldability, an important alloy characteristic for newmake nozzle salvage to remove casting defects, nozzle hardware fabrication, and nozzle repair after field service. As the firing temperature of advanced industrial gas turbines increases, nickel-base superalloys with improved high temperature strength and high temperature oxidation resistance are being introduced for nozzle applications. For example, the bucket alloy GTD-111TM has been used for the stage 1 nozzle on General Electric’s FB-class gas turbines. Also, SX Rene N5 has been used for the stage 1 nozzle on General Electric’s H SystemTM gas turbine. In these applications, both GTD-111TM and SX Rene N5 are coated with a TBC to allow for higher hot gas path temperatures. As hot gas path temperatures increase for more turbine output and cooling flows decrease to improve turbine efficiencies, alloy development will continue for nozzle applications. Development of new alloy compositions that possess high temperature creep strength, i.e. high gamma prime content and excellent microstructural stability, along with good weldability (difficult with high gamma prime alloys) will be critical for long-life nozzle applications. Excellent oxidation and corrosion resistance are always important, especially for fuel flexibility and future syn-gas or other alternative fuel usage.
4.4.1-7 Materials Performance
415
As has been mentioned previously, nickel-base superalloys are the highest temperature class of metallic materials currently used in aircraft engines and industrial gas turbines. Among these, the highest temperature subset are the turbine airfoil alloys. The drive to increase output and efficiency in industrial gas turbines results in the need for increased capability materials for both creep and fatigue. It also puts increased demand on thermal and environmental coatings to provide protection for long times in higher temperature environments. Hot section buckets and nozzles must not only have sufficient strength to withstand the mechanical and thermal loading, but must also have coatings that protect the substrate material from damaging effects of exposure to the hot gases. These effects include oxidation, hot corrosion, and embrittlement. Thus these components must be treated as a substrate / coating materials system.
4.4.1 Buckets and Nozzles Industrial gas turbine materials must withstand much longer time at operating temperatures but many fewer loading and unloading cycles than aircraft engine materials. Also, industrial gas turbine materials are not subject to the rapid start-up requirements that aircraft engines must tolerate. However, industrial gas turbines must operate for most of their life at maximum power output at a constant 3,600 or 3,000 RPM (depending on whether the power frequency is 60 or 50 Hz) as opposed to aircraft engines that see short times at maximum “take-off” power levels, drop back to cruise conditions, and have varying RPM shaft speeds. Industrial gas turbines must be capable of burning both clean fuels (e.g. natural gas) and distillate fuels with higher alkali metal levels. They are also exposed to much more air borne contaminates. Industrial gas turbine customers have come to expect long-term durability for airfoils and 30 year lifetimes for major structures (casings, rotors, etc.). Typically, this translates into hot section turbine overhaul intervals of no less/no earlier than 24,000 hours and expected parts lives of greater than 48,000 hours with repair. Reliability of modern combined-cycle power plants is expected to be >98% with an availability of about 90%. These values can easily be achieved with the proper maintenance plan and operation practices. New, more efficient industrial gas turbines will be expected to meet similar standards, despite higher operating temperatures. Improved materials and coatings have allowed significant gains in the durability of critical gas turbine components, even under today’s conditions of higher operating temperatures. The current direction in the industrial gas turbine industry is to provide still greater customer value through the availability of improved component repair techniques that permit refurbishment rather than replacement of expensive components such as turbine buckets and nozzles. Given the drive to reduce life-cycle costs of gas turbine operation, increased use of component repair and refurbishment will become a key for future activity in the industry.
Creep
Turbine airfoil alloys are limited in service temperature primarily by creep strength. The improvements in creep strength originate from changes both in alloy composition and in processing history13. Increasing the concentration of refractory metals such as W, Nb, Ta, and Mo (all of which diffuse slowly because of their high atomic weight) increases the creep strength of nickel-base alloys. The limits to which these additions can be successfully added are determined largely by their solubility and by the tendency of more concentrated alloys to form detrimental second phases. At the service temperatures of turbine airfoils (> two-thirds of the melting temperature), the principal creep deformation mode is grain boundary sliding. Advanced solidification processing techniques have been developed to alter the grain structure of these alloys during casting. These processes either create columnar grains with long axes parallel to the turbine blade axis (directional solidification or DS) or result in castings that have no grain boundaries at all (single crystals or SX). These structures lead to considerable increases in creep strength. Both of these solidification processing technologies have taken a long time to reduce to practice. Today, however, DS and SX buckets and nozzles are being used in advanced industrial gas turbines14. The service life requirements for industrial gas turbines are significantly longer than for aircraft engines. This places more emphasis on time dependent phenomenon such as creep and creep-fatigue interactions. Creep-fatigue interactions become more pronounced when materials are thermally cycled while operating at higher temperatures for long times. The effect of creep-fatigue interactions can significantly reduce the fatigue strength of the material. This phenomenon places an additional demand for fatigue data with hold times to simulate the effect of this interaction.
Fatigue
A consequence of the increased firing temperatures of industrial gas turbines is the increased severity in the thermal cyclic loading of the hot gas path buckets and nozzles15. Increased cooling effectiveness achieved with advanced air cooling and steam cooling schemes produces higher thermal gradients in these parts. While the designer can achieve the same or even lower bulk metal temperatures compared to lower firing temperature machines, the thermal gradient increases in more advanced gas turbines. This increases the thermal strain associated with these parts. The increased severity of these thermal cycles is being addressed by moving toward more thermal strain-tolerant materials such as DS and SX hot gas path components. Additionally, thermal barrier coatings are effective in reducing the thermal load into the cooled component and thus reducing the thermal gradient and thermal strains.
Environmental
Environmental degradation involves oxidation or corrosion of the alloy in the hot gas path. Oxidation involves the reaction between oxygen and the metal alloy to form various oxides. These chemical reactions remove material or deplete the material of strength. At high temperatures, these reactions can occur rapidly and create the potential for failure if an excessive amount of the alloy is consumed. The oxidation behavior of an alloy depends on its chemistry, casting segregation, and exposure conditions. At high temperatures, rapid oxidation attack can occur unless there is a barrier to oxygen diffusion and reaction on the exposed alloy surfaces. Ni-base superalloys containing a sufficient amount of Al will form a protective, adherent, and slow growing alumina (Al2O3) scale to prevent extensive oxidation damage16. Alloy chemistries can be furthered modified to improve oxidation behavior by adding Y or reducing S. Hot corrosion is another environmental damage mode. It is a rapid form of attack that is generally associated with alkali metal contaminants, such as sodium and potassium, which react with sulfur in the fuel to form molten sulfates. Sodium at levels of only 2ppm (parts per million) or less in the fuel or in the air can lead to hot corrosion damage17. In general, uncoated, cooler areas of a hot gas path component are susceptible when fuel is contaminated, synthetic fuel is used, or there is a lot of debris taken into the turbine from the environment. Basically, molten deposits on the component break down the protective oxide scale, and rapid, unpredictable degradation proceeds. The temperature range where this phenomenon occurs is between 650-925oC (~1200-1700oF). At these temperatures, substrate alloys that form faster growing chromia scales show better resistance in corrosion tests. Generally these systems have over 10% Cr.
416
Stephen J. Balsone As Ni-base superalloy development has evolved, the ability to obtain high temperature strength and creep resistance along with satisfactory oxidation and corrosion resistance has become increasingly difficult without the introduction of protective coatings. Many high strength Ni-base superalloys are not capable of forming a sufficiently protective oxide scale because the chemistry of the alloy is dictated by other requirements such as high strength, creep resistance, and microstructural stability. Thus, the alloy chemistry can not be optimized for oxidation or corrosion resistance. In today’s advanced industrial gas turbines, coatings are required to provide protection from oxidation, corrosion, and mechanical property degradation with service. The function of these coatings is to provide a surface reservoir of elements such as Al and Cr that will form stable, adherent oxide layers that will protect the substrate alloy from environmental attack. Table I. Nominal Composition of IGT Cast Ni-Base Superalloys wt % Buckets
U500 U700 (Rene 77) Alloy 738 MAR M247 GTD-111TM GTD-444TM PWA 1483 Rene N5 CMSX-4® PWA 1484
Nozzles
FSX414 GTD-222TM GTD-111TM Rene N5
Ni
Cr
Co
bal bal bal bal bal bal bal bal bal bal
18.50 15.00 16.00 8.25 14.00 9.80 12.80 7.00 6.50 5.00
18.50 17.00 8.30 10.00 9.50 7.50 9.00 7.50 9.00 10.00
10.00 bal bal bal
28.00 22.50 14.00 7.00
bal 19.00 9.50 7.50
Fe
0.20
1.00
Mo
W
Al
Ti 3.00 3.35 3.40 1.00 4.90 3.50 4.00
1.20 4.90
4.00 5.30 1.75 0.80 1.50 1.50 1.90 1.50 0.60 2.00
2.60 10.00 3.80 6.00 3.80 5.00 6.00 6.00
3.00 4.25 3.40 5.50 3.00 4.20 3.60 6.20 5.60 5.60
2.30 1.50 1.50
7.00 2.00 3.80 5.00
0.80 3.00 6.20
1.00
Nb
0.90 0.50
Ta
1.75 2.80 2.80 4.80 4.00 6.50 6.50 9.00
1.00 2.80 6.50
Mn
V
C 0.07 0.07 0.10
B
other
0.10 0.08
0.006 0.020 0.001 0.015 0.010 0.009
Hf 0.15
0.05
0.004
Re 3.0, Hf 0.15, Y 0.01
Hf 0.15
Re 3.0, Hf 0.10 Re 3.0, Hf 0.10
0.25 0.10 0.10 0.05
0.010 0.008 0.010 0.004
Re 3.0, Hf 0.15, Y 0.01
4.4.1-8 Conclusions To increase the output and efficiency of IGT’s, the firing temperature continues to increase, placing higher demands on the temperature capability of gas turbine materials. Over the past ten years, firing temperature has increased by approximately 93ºC (200ºF), and combined-cycle efficiency has increased by approximately 4%. Materials and processes improvements have enabled these performance increases along with improving the durability and reliability of advanced IGT’s. Continued growth in IGT firing temperature and efficiency will require continued materials and processes technology development. Ni-base superalloys have been key to past progress and are key to future IGT growth. Cast Ni-base superalloys for hot gas path applications will require higher temperature strength with improved oxidation/corrosion resistance. Larger component sizes and complex geometries with sophisticated internal cooling schemes will require new investment casting technology to produce defect-free, high performance DS and SX Ni-base superalloys. Materials and processes technology development for Ni-base superalloys continues today and into the future to assure that when new design requirements demand the world’s best materials, Ni-base superalloys will be ready to meet the most challenging high temperature applications.
4.4.1-9 Notes ________________________
417
1. B.B. Seth, “Superalloys - The Utility Gas Turbine Perspective,” Superalloys 2000, ed. T.M. Pollock, R.D. Kissinger, R.R. Bowman, K.A. Green, M. McLean, S.L. Olson, and J.J. Schirra,Warrendale, PA: TMS, 2000, 3-16; M.B. Henderson, J. Hannis, G. Mccolvin, and G. Ogle, “Materials Issues for the Design of Industrial Gas Turbines,” Advanced Materials and Processes for Gas Turbines, ed. G. Fuchs, A. James, T. Gabb, M. McLean, and H. Harada,Warrendale, PA: TMS, 2003, 3-13. 2. S.J. Balsone, “Nickel-Base Superalloy Materials Technology for Advanced IGT Applications,” Niobium for High Temperature Applications, ed. Y-W. Kim and T. Carneiro,Warrendale PA: TMS, 2004, 3-9. 3. P.W. Schilke, “Advanced Gas Turbine Materials and Coatings,” GER-3569G, General Electric Company, 2004. 4. See Note 2 above. 5. See Note 1 above. 6. See Note 2 above. 7. J.C. Schaeffer, “Single Crystal Materials Technology for Advanced Gas Turbine Applications,” presentation at Power-Gen Europe 2005, Milan, Italy, June 28-30, 2005. 8. M. Konter and M. Thumann, “Materials and Manufacturing of Advanced Industrial Gas Turbine Components,” Journal of Materials Processing Technology, 117 (3), 2001, 386-390. 9. R.F. Singer, “Advanced Materials and Processes for Land-Based Gas Turbines,” Materials for Advanced Power
4.4.1 Buckets and Nozzles Engineering, ed. D. Coutsouradis, J.H. Davidson, J. Ewald, P. Greenfield, T. Khan, M. Malik, D.B. Meadowcroft, V. Regis, R.B. Scarlin, F. Schubert, and D.V. Thornton (Dordrecht, Netherlands: Kluwer Academic Publishers Group, 1994, 1707-1729; A.J. Elliott, S. Tin, W.T. King, S.-C. Huang, M.F.X. Gigliotti, and T.M. Pollock, “Directional Solidification of Large Superalloy Castings with Radiation and Liquid-Metal Cooling: A Comparative Assessment,” Metall. Trans. A, 35A (10), 2004, 3221-3231. 10. S. Balsone, G. Feng, L. Peterson, and J. Schaeffer, “Microstructure and Mechanical Behavior of Liquid Metal Cooled Directionally Solidified GTD-444TM,” Solidification Processes and Microstructures: A Symposium in Honor of Wilfried Kurz, ed. M. Rappaz, C. Beckermann, and R. Trivedi, Warrendale, PA: TMS, 2004, 77-83. 11. See Note 3. 12. Ibid. 13. See Note 7. 14. R. Viswanathan and S.T. Scheirer, “Materials Technology for Advanced Land Based Gas Turbines,” Turbomachinery International, 42 (1), 2001. 15. See Note 7. 16. See Note 3. 17. Ibid.
418
BIOGRAPHY
4.4.1 Buckets and Nozzles
Stephen J. Balsone GE Gas Turbines LLC P.O. Box 648; GTTC 174D Greenville, SC 29602 phone: (864) 254-5294 email: [email protected]
Mr. Stephen J. Balsone is the Manager for High Temperature Materials Development in the Materials & Processes Engineering organization at GE Energy, Greenville SC. The group is responsible for high temperature materials and processes development for industrial power generation gas turbines, principally hot gas path and combustor components. Before joining GE Energy in 2000, Mr. Balsone was Program Manager for Structural Materials Development & Reliability at GE’s Global Research and Development Center, Schenectady NY, responsible for the development of high temperature structural metallic materials, principally supporting GE’s aircraft engine and industrial gas turbine businesses. Prior to joining GE in 1996, Mr. Balsone worked for 14 years at the Materials & Manufacturing Directorate, Air Force Research Laboratory, Wright-Patterson Air Force Base, Dayton OH. His work included structural materials development for military aircraft engines and hypersonic vehicles. Mr. Balsone received his MS degree in Aeronautical Engineering from the Air Force Institute of Technology and his BS degree in Metallurgical Engineering & Materials Science from Carnegie Mellon University. He is currently a PhD candidate at the University of Michigan.
4.4.2
Protective Coatings for Gas Turbines
Kang N Lee Cleveland State University NASA Glenn Research Center Cleveland, OH 44135 Current Address:
Rolls-Royce Corpation P.O. Box 420, Speed Code W-08 Indianapolis, IN 46206 phone: 317-230-4469 email: [email protected]
419
4.4.2-1 Introduction Economical and environmental concerns, i.e. improving efficiency and reducing emissions, are the main driving force behind the ever increasing demand for higher gas turbine engine inlet temperatures. Technology improvements in cooling, materials and coatings are required to achieve higher inlet temperatures1. Advances in the development of airfoil cooling designs have been achieved by combining high convective cooling efficiencies with film cooling. Material improvements have been dramatic during the past several decades. The improvement in alloy composition and the development of directional and single crystal casting technologies have allowed increased alloy operation temperatures, and hence increased turbine inlet temperatures2. Improved high temperature mechanical properties of alloys, however, have been made typically at the expense of environmental resistance. This trend, combined with higher operating temperatures, has resulted in environmental degradation of materials, deteriorating the mechanical properties and shortening the service life of components3. The need to protect alloys from environmental degradation motivated the development of protective coatings. The idea to apply a layer with protective properties on the surface of Ni-based superalloys was first practiced in the 1960s4. Two types of protective coatings have been most widely used: diffusion aluminide coatings based on β-NiAl phase and MCrAlY (M = Ni, Co, or NiCo) overlay coatings based on a mixture of β-NiAl and γ’-Ni3Al or γ phases5. As the temperature capability of Ni-based superalloys approaches their intrinsic limit, further improvements in their temperature capability have become increasingly difficult6. Therefore, during the past two decades, the emphasis in gas turbine materials developments has shifted to thermal barrier coatings (TBC), which are ceramic coatings with a very low thermal conductivity that reduce the alloy surface temperature by insulating it from the hot gas. Current state-of-the-art thermal barrier coatings comprise two layers: a diffusion aluminide or MCrAlY bond coat and a low thermal conductivity partially stabilized zirconia (YSZ: 7 to 8 wt% Y2O3-ZrO2) top coat. Thermal barrier coatings were first successfully tested in a research turbine engine in mid 70s. By the early 80s they entered revenue service on the vane platforms of aircraft engines, and today they are flying in revenue service on vane and blade surfaces7. Thermal barrier coatings are expected to play an increasingly significant role in advanced gas turbine engines both in aero and industrial applications in the future. Major improvements in turbine inlet temperatures can be achieved by replacing Ni-based superalloy hot section components with silicon-based ceramic matrix composite (CMC) and silicon nitride (Si3N4) ceramics8. These materials have superior high temperature mechanical properties, such as strength and creep resistance, compared to Ni-based superalloys. They are also light and possess excellent high temperature oxidation resistance in clean, dry air, due to the formation of slow-growing, protective silica scale9. One major disadvantage of these materials is the lack of environmental durability in combustion environments. Water vapor, a combustion reaction product, reacts with the protective silica scale, forming gaseous reaction products, such as Si(OH)4 10. In high pressure, high gas velocity combustion environments, this reaction results in rapid recession of these materials. These materials also suffer from severe hot corrosion in environments contaminated by molten salt11. A new class of coatings, environmental barrier coating (EBC), has been developed in the 90s to protect Si-based ceramics and ceramic composites from the degradation by water vapor12. The current state-of-the-art environmental barrier coating comprises three layers: a silicon bond coat, a mullite-based intermediate coat, and a barium-strontium-aluminosilicate (1-xBaO·xSrO·Al2O3·2SiO2, 0 ≤ x ≤ 1) top coat13. CMC combustor liners coated with the current state of the art EBC were retrofitted in a Solar Turbines’ industrial gas turbine engine and successfully completed a 14,000 h field test in the late 90s14. This paper will discuss the status of current thermal barrier coatings and environmental barrier coatings, with the focus on key factors affecting their performance.
4.4.2-2a Coatings for Superalloy Components Current thermal barrier coatings consist of two layers: a metallic bond coat and a ceramic top coat15. The bond coat has two key functions: It provides the bonding between the ceramic top coat and the superalloy substrate and protects the superalloy from the environmental degradation. The key function of the ceramic top coat is to reduce the alloy surface temperature by insulating it from the hot gas. Current bond coats are diffusion aluminide coatings based on β-NiAl phase and MCrAlY (M = Ni, Co, or NiCo) coatings based on a mixture of β-NiAl and γ’-Ni3Al or γ phases16. Bond coats oxidize upon thermal exposure, even in the presence of a ceramic top coat, forming an oxide scale, known as TGO (thermally grown oxide). Current top coat is yttria-stabilized zirconia (YSZ: ZrO2 doped with 7~8 wt% Y2O3). YSZ has several important characteristics for a successful top coat17. It has a high melting point, a low thermal conductivity and a high thermal expansion coefficient and is thermodynamically stable in contact with alumina that grows on bond coat. The ZrO2-7~8 wt% Y2O3 composition also has good erosion resistance compared with other ceramics and good phase stability at temperatures <1200oC. The durability of thermal barrier coatings is governed by a sequence of crack nucleation, propagation and coalescence events along the bond coat/TGO or top coat/TGO interfaces that accumulate prior to final failure by large scale buckling or edge lifting18. Stresses in TBC play an important role in crack nucleation. TBC stresses arise from two sources: stresses due to TGO growth and stresses due to the coefficient of thermal expansion (CTE) mismatch between the various layers. Therefore, the bond coat must form the most protective oxide scale possible which, in practice, means an alumina scale that is slow-growing, and adherent19. The nature of stresses in TBC is closely related to the surface roughness of bond coat. Both growth and CTE mismatch stresses are compressive on flat bond coat surfaces20. The thermal mismatch stress (σt) is given by, σt = (αc - αsubstrate) ∆T Ec/(1−νc)
(1)
where αc and αsubstrate are coefficients of thermal expansion for the coating and the substrate, respectively, Ec is the Young’s modulus of the coating, and νc is the Possion’s ratio of the coating. Therefore, YSZ having a lower CTE than the bond coat and superalloy substrate is expected to be in compression on cooling, assuming stresses are relaxed at high temperatures. However, the surface of an initially flat bond coat gradually roughens with thermal exposures, forming TGO with a convoluted morphology and major imperfections21. Rough bond coat surfaces produce out-of-plane stresses along the bond coat/TGO or top coat/TGO interfaces22. These local out-of-plane stresses are responsible for the nucleation of cracks which ultimately lead to the failure of TBC. There are two degradation mechanisms for YSZ that have significant effects on TBC stresses: Phase transformation23 and sintering of YSZ24. These changes result in higher thermal stresses and a decrease in thermal fatigue life. Sintering also causes an increase in thermal conductivity. Two approaches have been investigated to alleviate the high-temperature durability problems of YSZ25: the first approach is alternative stabilizers for ZrO2 and the second approach is alternative materials to ZrO2. Key factors affecting TBC performance to be discussed in subsequent sections are bond coat surface finish, bond coat oxidation, bond coat surface imperfections, thermal conductivity of YSZ, sintering of YSZ, and phase transformation of YSZ.
4.4.2-2b Bond Coat Processing Pack cementation and chemical vapor deposition are widely used methods to form diffusion aluminide coatings on turbine blades. Pack cementation is a chemical vapor deposition process in which component surface is saturated with aluminum in a powder mixture containing aluminum, aluminum oxide (as an inert filler) and a halide activator (usually NH4Cl)26. When the reactor containing the components to be coated and the powder mixture is heated, aluminum halides (AlCl3, AlCl2, AlCl) form which diffuse through the powder mixture and react with the components, resulting in the formation of an aluminide coating. In practice, platinum is added in β-NiAl bond coat to form platinum modified diffusion aluminide coating, (Ni,Pt)Al, which significantly improves the alumina scale adherence27. In this process, components are electroplated with a thin Pt layer prior to the aluminization. The aluminizing process can be divided into a high, medium, and low activity process28. The main phases of the outer zone of the coatings in high, medium, and low activity process are NiAl3+Ni2Al3 (>40 wt.% Al), NiAl (32-38 wt.% Al) and NiAl or NiAl+Ni3Al (<31 wt.% Al), respectively29. Lower activity process produces coatings with higher ductility. A typical as-aluminized (Ni,Pt)Al bond coat surface exhibits large grains of Pt-modified β-NiAl with a cellular network of grain-boundary ridges, whose geometry is very similar to that of underlying bond coat grain boundary structure30. Low Pressure Plasma Spaying (LPPS) and Electron Beam Physical Vapor Deposition (EB-PVD) are widely used methods to deposit overlay MCrAlY bond coat on turbine components31. NiCrAlY bond coat consists of the following main phases: γ-Ni-based solid solution, γ’-Ni3Al phase, β-NiAl phase, and α-Cr-based solid solution32. Alloying NiCrAlY with Co reduces the thermal stability of γ’-phase, decreases its quantity, and converts NiCoCrAlY into β+γ33. It is this phase condition that makes NiCoCrAlY bond coat highly ductile. For Example, EB-PVD-processed Ni-20Co-20Cr-8Al-0.5Y (wt%) has Al-rich β phase and Ni-solid solution γ phase34.
420
Kang N Lee Surface Finish Bond coat surface finish depends on the type of process employed for ceramic top coat. Air plasma sprayed (APS) YSZ top coat employs a MCrAlY bond coat processed typically by low pressure plasma spraying (LPPS). Rough surface of as-processed MCrAlY bond coat facilitates good mechanical bonding to plasma-sprayed YSZ by providing anchoring points. EB-PVD YSZ top coat employs either a diffusion aluminide or a MCrAlY bond coat. In contrast to APS YSZ, the bond coat surface is treated to make it flat prior to the deposition of EB-PVD YSZ. This is to eliminate as-processed surface roughness which generates out-of-plane stresses. It may be that EB-PVD YSZ due to the nature of the process, i.e. vapor phase deposition in high temperature and vacuum, possesses good chemical bonding, eliminating the need for a rough surface for mechanical bonding. Typical surface treatment includes grit blasting, grinding, and shot peening. Beneficial effects of grit blasting the bond coat surface with Al2O3 on TBC performance are discussed in Haynes, et al.35. The intensity of θ-Al2O3 was much lower on grit-blasted specimens compared to the as-processed bond coat surfaces. Grit blasting also eliminated the detrimental process of interfacial void growth during EB-PVD processing. It was suggested that grit blasting either suppressed θ-Al2O3 nucleation or triggered more rapid transformation to or nucleation of α-Al2O3 during EB-PVD process at ~1000oC. The elimination of void growth in the grit-blasted specimens was attributed to the removal of sulfur-contaminated zone at the surface of the bond coat. Lightly polishing the surface of as-deposited (Ni,Pt)Al bond coats prior to YSZ deposition also dramatically increased EB-PVD TBC life36, presumably by the same mechanism by which grit blasting improved the TBC life.
Oxidation
421
The approach to achieving an ‘ideal’ α-alumina scale includes the addition of reactive element (RE), such as Y, Zr or Hf, the addition of precious metals such as Pt, and the desulfurization of the coating and superalloy substrate37. The addition of Pt to β-NiAl is known to improve alumina scale adhesion38. The Pt addition has the additional benefit of reducing the critical Al concentration necessary to form a protective α-Al2O3 scale in β-NiAl, suggesting that Pt improves selective oxidation to from alumina39. The addition of a reactive element, such as Zr, Hf, or Zr+Hf, to β-NiAl is more effective in improving the scale adhesion than the addition of Pt, and has the additional benefit of reducing the growth rate of α-Al2O3 scale40. Figure 1 shows that, at 1200oC, the addition of Hf reduces the scale growth rate by a factor of 40. Comparison of YSZ-coated Y-doped Rene N5 having a (Ni,Pt)Al bond coat and YSZ-coated β-NiAl+Zr showed that the change to a more oxidationresistant bond coat composition increased the TBC lifetime Fig. 1. Isothermal weight gains at 1200°C plotted versus the by more than a factor of five at both 1150 and 1200oC41. The square root of time to show the parabolic reaction kinetics improvement was attributed to the formation of an adherent (Reprinted with permission of Elsevian, Copyright 1998) TGO on β-NiAl+Zr which was superior to that formed on (Ni,Pt)Al-coated Rene N5. Source: See Note 19. (Pint, et al.) Sulfur is known to segregate to metal/scale interface, causing the deterioration of the scale adhesion by forming voids at the metal/scale interface42. Interfacial voids degrade the scale adhesion by limiting the contact between the metal and the scale and also by acting as a stress concentrator or crack initiator43. Consequently, lowering the level of S below ~1 ppm significantly improves the scale adhesion44. The beneficial effect of reactive elements (RE) in improving the scale adhesion has been attributed to the segregation of RE at the metal/scale interface which inhibits the interfacial segregation of sulfur impurity45. They also segregate to the oxide grain boundaries where they significantly reduce the outward transport of Al, thereby reducing the rate of oxide growth which is now mostly by oxygen transport46. The change in predominant growth mechanism also drastically reduces interfacial void formation since the flux of cation vacancies is reduced47. In practice, Hf is most effective in improving the oxidation resistance of β-NiAl and yet it does not have the same strong effect in superalloys or MCrAlY, in which Y is more effective48. Pt appears to improve the resistance to scale spallation through the inhibition of interfacial void formation49. It has been suggested that Pt either inhibits S segregation or changes diffusivities in the substrate which, in turn, inhibits the interfacial void formation50. The composition of the substrate also appears to affect the performance of the TBC even when covered by a bond coat. Cyclic oxidation of (Ni,Pt)Al/YSZ TBC on Rene N5 at 1150oC showed that both desulfurizing (to 0.7 ppma S) of Rene N5 and the addition of Y in Rene N5 increased the TBC life time51. S and Y, presumably, diffuse through the bond coat and affect the scale adhesion. Haynes et al. reported a similar trend of slightly longer TBC life on a desulfurized Re N5 (to 0.9 ppma S), however, Y in Re N5 did not improve the TBC life52.
4.4.2 Protective Coatings for Gas Turbines There are two sources for the loss Al from the bond coat: the formation of alumina scale and the interdiffusion between the bond coat and the substrate. In terms of bond coat life, the concern is that the Al content in bond coat will prematurely fall below the minimum level required to avoid spinel formation. Equally important, however, is the effect of the influx of Al to the superalloy substrate on the solubility of refractory elements used for strengthening, resulting in precipitation of embrittling phases53. Phase transformations due to the Al loss from β-NiAl bond coat on Ni and Ni-16Cr-11Al-Y bond coat on a superalloy substrate at 950oC are illustrated in Figures 2a and 2b, respectively54.
Fig. 2. Phase transformations in (a) β-NiAl bond coat on Ni and (b) Ni-16Cr-11Al-Y bond coat on a superalloy substrate at 950°C (Reprinted with permission of ASM International®. All rights reserved. www.asminternational.org) Source: See Note 3.
Surface Imperfections The surface of an initially flat, single phase β-NiAl and (Ni,Pt)Al bond coat on a single crystal superalloy is shown to progressively roughen with thermal exposures55. Roughening has also been observed during thermal cycling of MCrAlY coatings on Ni-based superalloys56. The most prominent roughening comprises the undulations of the original TBC/bond coat interface57. Adherent TGO follows the roughness contour of the bond coat. The surface roughening is far more severe in cyclic exposures compared to isothermal exposures58. Phase transformations in bond coat have been suggested as a source for the formation of surface roughening known as ‘rumpling’59. Two types of phase transformations occur in β-NiAl and (Ni,Pt)Al bond coat during thermal exposures: β to γ’ phase and martensite transformation of β phase. The β to γ’ phase transformation is due to the depletion of Al from the β phase60, while the martensite transformation is non-diffusional and shear-dominated transformation61. Similar phase transformations are expected to occur in MCrAlY bond coat. The martensite transformation was observed in bond coats thermally cycled at 1150oC to 5 and 100% of TBC life, suggesting that the transformation accompanied the thermal cycling for most of the TBC life62. The volume change for the β to martensite transformation is approximately -2%63. Alloys with martensite + γ’ structure at room temperature undergo a reversible transformation to single phase β when heated to 1100oC 64. Bond coat surface rumpling was attributed to the plastic deformation of bond coat caused by the repeated volume changes accompanying the reversible β - martensite phase transformation during thermal cycling65. Surface rumpling was also attributed to the volume reduction accompanying the β to γ’ phase transformation66. The volume reduction can be accommodated by the development of surface recession or by the formation of internal cavities. In the outer part of the bond coat both Ni and Al diffuse toward the surface, therefore there should be a compensating vacancy flux in the opposite direction. The result is expected to be either the formation of Kirkendall porosity or, if the pores collapse, a decrease of the coating volume. It is suggested that the pores tend to collapse during cycling because of thermal stresses in the bond coat, whereas they coalesce into large cavities in isothermal exposures. Darzens et al. suggest that both the β to martensite and the β to γ’ phase transformations contribute to the bond coat and TGO roughness (instability). Figure 3 is a trace of the bond coat/TGO interface at f = 0.76 (fraction of life)67. About 70% of the instabilities have the β-phase located at the base. Moreover, there is a greater likelihood that β exists at the base of the most prominent (deepest) instabilities. γ’ is present on at least one side of instability with 75% probability, implying that this phase tends to locally impede the TGO displacement. Based on these observations the following mechanism has been suggested67. The amplitude of the instability of the TGO has an isothermal and cyclic components, but the latter is appreciably larger. The isothermal component is believed to be associated with the stress caused by the β to γ’ Fig. 3. A trace of the bond coat/TGO interface at f = transformation. The influence of the transformation on the larger displacements 0.76 (fraction of life). The arrows designate the base upon cycling appears to involve two effects: (a) the martensite transformation of all large TGO instabilities (Reprinted with permisin the β phase results in a volume reduction on cooling larger than in either the sion of TMS. All rights reserved.) γ’ or the substrate; (b) γ’ has greater strength than the surrounding β at high Source: See Note 21. (Darzens, et al.) temperature, impeding the TGO displacement.
422
Kang N Lee Other mechanisms suggested for bond coat imperfections include a thermal ‘ratcheting’ phenomenon associated with the elastic/ plastic mismatch between the bond coat and the growing oxide68, pre-existing ridge on bond coat grain boundaries and preferential intergranualr oxidation and cracking69, and Y2O3/YAG precipitate phases in MCrAlY bond coat with a columnar morphology acting as preferred channels for rapid inward diffusion of oxygen, causing locally thick regions of TGO70. Bond coat imperfections have important implications for TBC performance and failure. Bond coat imperfections cause TGO to displace into the bond coat with each thermal cycle, forming concave regions71. The stiffness of the ceramic top coat should constrain the displacement of TGO into the ceramic top cot, preventing the formation of convex regions on the metal surface72. The local separations will gradually accumulate on thermal cycling, link together and eventually form the critical sized flaw required for TBC buckling73.
4.4.2-2c Top Coat ZrO2-Based Ceramics Processing APS and EB-PVD are the most widely used methods to deposit YSZ top coat on turbine components. Figures 4a and 4b show asdeposited APS and EB-PVD YSZ, respectively. In plasma spraying process, powders of coating material are melted and propelled to the substrate. Upon impingement on the substrate the molten drops are quenched and solidified. A coating is produced by the successive impingement of the drops, referred to as splats, on the substrate. In EB-PVD process, coatings are produced by condensation of vapor on the substrate. A focused electron beam is used to evaporate the coating material. Multiple beams are employed to produce coatings containing components with differing vapor pressures.
a)
b)
50 µm
Fig. 4. As-deposited YSZ (7 wt.% Y2O3): (a) APS; (b) EB-PVD Source: J.I. Eldridge, NASA Glenn Research Center, Cleveland, OH.
APS YSZ has splat structure with inter-splat porosity. Figure 5 is thermal spray coating microstructure showing common features74. Two prominent crack morphologies found within as-plasma-sprayed YSZ are75: (1) Elongated crack-like separations between flattened splats that melted during the spray deposition. These are oriented nominally parallel to the interface. They form because of thermal contraction as the splats cool; (2) Large, more equi-axed, voids contiguous with equi-axed zirconia particles. These are presumably the particles that did not fully melt during the deposition. The separations between splats are efficient in lowering thermal conductivity, but this is at the expense of surface finish, strain tolerance and erosion resistance76. EB-PVD YSZ has a columnar microstructure, which imparts excellent strain tolerance77, and thus longer cyclic life than APS YSZ. Figure 6 illustrates the dependence of the coating microstructure on the substrate temperature and the rotation speed78. Other advantages of EB-PVD YSZ include aerodynamically favorable smooth surface finish and good erosion resistance79. But the columnar structure with open porosity parallel to the direction of heat conduction results in a higher thermal conductivity compared to APS YSZ80.
423
4.4.2 Protective Coatings for Gas Turbines
Fig. 5. Thermal spray coating microstructure showing common features (Reprinted with permission of ASM International®. All rights reserved. www.asminternational.org) Source: See Note 74.
Fig. 6. A schematic illustrating the dependence of the coating microstructure on substrate temperature and rotation speed (Reprinted with permission of Trans Tech Publications. All rights reserved) Source: See Note 78.
Thermal Conductivity In thermal barrier coatings, heat is conducted by lattice waves (phonons) as well as by electromagnetic radiation (photons). Lattice waves are elastic or ultrasonic waves, but their spectrum extends to the very high frequencies, where their waves, λ is of atomic dimensions81. There are two radiation sources in gas turbine engines: far-field radiation and near-field radiation82. Far-field radiation is the radiation from the combustion gas, which is at high temperatures of around 2000oC. Near-field radiation is the radiation from the layer of cooler gas, at around 1200oC, adjacent to the TBC. Radiation can pass through the partially transparent ceramic to the metallic bond coat and substrate. The radiative component of the heat conduction can become a significant portion of the overall thermal conductivity at elevated temperatures. The thermal conductivity by mobile carriers, whether waves or particles, can be expressed in general in the form83, κ = 1/3 Cvl
(2)
where C is the specific heat per unit volume, v is their speed and l is their mean free path. The thermal conductivity is limited by various interaction processes, which transfer energy between the waves84. Based on models for thermal conductivity, a low intrinsic thermal conductivity requires weak binding, a large mean atomic weight, a complex crystal structure, non-directional binding and a large number of different atoms per molecule85. Lattice imperfections reduce the thermal conductivity by scattering phonons and thereby reducing the mean free path86. Point defects, such as solute cations and oxygen vacancies, reduce the lattice thermal conductivity by scattering high-frequency lattice waves, while grain boundaries scatter lattice waves at the low-frequency part of the spectrum. Significant theoretical reductions in the thermal conductivity of YSZ are expected due to grain boundary scattering when the grain sizes are reduced below 100 – 10 nm87. These reductions by point defects and grain boundaries are almost independent of each other, since they scatter lattice waves in different frequency ranges. The effect of yttria dopant level on the thermal conductivity of YSZ was investigated88. Thermal conductivity decreased with increasing yttria contents up to 4.5 - 8 mol%. The decrease in thermal conductivity was attributed to a reduced mean free path in zirconia by an increasing phonon scattering, which was likely due to the combined effect of local elastic strain fields generated by incorporating a larger dopant atoms and the introduction of oxygen vacancies into the lattice. The little improvement in the thermal conductivity beyond 4.5 – 8 mol% yttria level was attributed to vacancy clustering89. The imperfections which scatter lattice waves have little influence on the radiative component. To reduce this component, one needs larger imperfections, such as porosity and inclusions with almost a micron size90. Pores may be preferable, since they present a larger contrast in the index of refraction. Calculations show that the optimum pore diameter is about 0.5 µm for YSZ91. An ideal structure for low thermal conductivity has a very fine grain size of nanometer scale, while the matrix contains relatively large inclusions, of the order of 0.5 µm in diameter92. Besides photon scattering, porosity decreases the thermal conductivity of a solid by reducing the net-section area through which heat can be transported by phonons and so the reduction in thermal conductivity depends on not only the volume fraction of pores but also their aspect ratio and their spatial distribution93. Ideally, flat-pancake shaped pores perpendicular to the temperature gradient, as are formed at splat-boundaries in plasma-sprayed coatings, are most efficient in decreasing the thermal conductivity94 . The intra-columnar fine porosity in EB-PVD YSZ accounts for the reduced thermal conductivity95. This is generally perceived to be much less effective as its distribution generally aligns perpendicular to the coating surface, i.e. parallel to the direction of the primary heat flux96.
424
Kang N Lee Approaches to Reducing Thermal Conductivity Alternative Dopants: Doping with ions heavier than yttrium can theoretically decrease the thermal conductivity by increasing the mean atomic weight. Five dopant additions, Er, NiO, Nd, Gd and Yb, were examined with the aim of maximizing lattice strains and lattice anharmonicity97. The most effective additions by EB-PVD were Gd, Nd and Yb which resulted in a thermal conductivity of 0.88, 1.00 and 1.02 W/m-K, respectively, calculated for a 4 mol% addition at a coating thickness of 150 µm at 500oC. The thermal conductivity of EBPVD reference YSZ was ~1.6 W/m-K. Multi-component dopants, ZrO2-Y2O3-Nd2O3(Gd2O3, Sm2O3)-Yb2O3(Sc2O3), were investigated using a laser rig at 1316oC surface temperature and 950 – 1100oC ceramic/metal interface temperature98. The thermal conductivity of APS YSZ (8 wt% or 4.55 mol% Y2O3) was about 1 W/m-K, which gradually increased to about 1.4 W/m-K after a 20-h test. In contrast, some ZrO2-Y2O3-Nd2O3-Yb2O3 or ZrO2-Y2O3-Gd2O3-Yb2O3 showed thermal conductivity as low as ~0.6 W/m-K, which did not change much after a 20-h test. The thermal conductivity and the rate of thermal conductivity increase were lowest at the total dopant level of 6-13 mol%. The thermal conductivity of EB-PVD ZrO2-(4-6 mol%)Y2O3-Nd2O3-Yb2O3 was as low as 0.85 W/m-K, while the thermal conductivity of EB-PVD YSZ (8 wt% Y2O3) was 1.85-1.9 W/m-K, after a 20-h test. It was proposed that the differing ionic sizes in the solid solution produced distortion, which facilitated the creation of thermodynamically stable, defect clusters that reduced the thermal conductivity and improved sintering resistance99. Varying the coating nano-structure: For EB-PVD thermal barrier coatings, the thermal conductivity has been observed to vary with coating thickness100. For the early stages of deposition (< 100 µm) a value of 0.8 – 1 W/m-K was reported; however, by the time the coating was 250 µm thick, the mean thermal conductivity was between 1.5 – 1.9 W/m-K at room temperature. Figure 7 shows a two-layer coating approximation consisting of an inner (100 µm) fine-structured zone, overcoated with a layer with thermal conductivity close to bulk YSZ (2.2 W/m-K)101. Measurements of the inner, fine-structured grain size gave a value of 3-4 µm, and within each grain was an ultrafine structure of nanometer dimensions. Therefore, renucleation of the EB-PVD coating growth at periodicities less than 100 µm would lower the thermal conductivity102. Reduction by layering: Layering offers a promising route to lower the thermal conductivity of an EB-PVD YSZ, which involves the introduction of interfaces/density changes parallel to the YSZ/ bond coat interface103. A glow discharge plasma was employed to vary the density of YSZ during deposition. The layers were produced by switching the d.c. bias applied to the substrate between high and low levels during deposition. This has the effect of periodically changing the degree of ion bombardment and thus altering the density of the layers produced. Thermal conductivity reduction of the order of 37 – 45% compared to state-of-the-art EB-PVD YSZ has been measured for these layered structures, approaching the values of APS YSZ. The combination of layering at micron dimensions and the introduction of density change from layer to layer work in combination to reduce the thermal conductivity. The layering periodicity is selected to significantly reduce photon transport, while local changes in layer density act to scatter phonons.
Fig. 7. Schematic of a two-layer coating model, consisting of low conductivity inner layer and a high conductivity outer layer (Reprinted with permission of Trans Tech Publications. All rights reserved) Source: See Note 101.
Sintering
425
Sintering leads to an increase in thermal conductivity and thermal stresses. The thermal conductivity of APS YSZ (8 wt.% Y2O3) was investigated as a function of time and temperature using a laser rig104. The thermal conductivity increased from 1.0 W/m-K to 1.15, 1.19, and 1.5 W/m-K after 30 h at the surface temperature of 990, 1100, and 1320oC, respectively. The increase was attributed to sintering as was evidenced by the decrease in the microporosity. Sintering also increased the hardness and modulus, which increases thermal stresses. The Knoop hardness on the coating surface increased from 4 GPa to 7.5 GPa and the surface modulus increased from 70 GPa to 125 GPa after 120 h at 1100oC. EB-PVD coatings are more resistant to an increase in thermal conductivity compared to APS coatings. The effect of bond coat surface imperfections on the sintering of EB-PVD YSZ/(Ni,Pt)Al-coated Rene N5 was investigated105. On thermal exposure, necks
4.4.2 Protective Coatings for Gas Turbines form and clusters of individual columns sintered together to from large-scale dense regions with large gaps in between, resulting in a ‘mud-cracking’ pattern. The pattern of the gaps appeared to reproduce the roughness of the underlying bond-coat alloy. The origin of the ‘mud-cracking’ pattern was attributed to the local undulations in the substrate surface, causing individual columns to converge and narrowing the gap between them. It is thus expected that sintering will be more pronounced on the high-pressure side of blade and on the leading edges than on the low-pressure sides because the surface curvature will tend to cant columns together on the high-pressure side106. In practice, the high-pressure side of a blade also tends to be hotter, further likely to promote the sintering. The influence of the concentration (4-20 mol%) of an alternative dopant Gd2O3 on the sintering and grain growth of ZrO2 solid solutions was investigated107. The onset of measurable shrinkage for the ZrO2-4 mol% Y2O3 and ZrO2-4 mol% Gd2O3 occurred at ~1100 and ~1175oC, respectively. The shrinkage at a given temperature increased as the Gd2O3 content increased to 8 mol% but then decreased for higher Gd2O3 concentration. The grain size data showed a trend similar to the sintering data: the average grain size showed a maximum at a Gd2O3 concentration of 8 mol%. The slower thermal conductivity increase on thermal aging for ZrO2 alloyed with multicomponent dopants Y2O3-Nd2O3(Gd2O3, Sm2O3)-Yb2O3(Sc2O3)108 was presumably due to reduced sintering. It appears that heavy rare earth oxide dopants are effective in reducing the sintering of ZrO2 solid solutions.
Phase Transformation The excellent phase stability of YSZ is due to the formation of the metastable, nontransformable t’ phase which is very stable at temperatures below ~1200oC109. The phase stability of plasmasprayed YSZ (8.6 mol% Y2O3) was examined by XRD110. The assprayed YSZ was primarily a nontransformable tetragonal phase (t’), having about the same composition as the starting powder. The formation of the nontransformable tetragonal phase (t’) is due to the rapid quench in the process. Figure 8 shows the mole fraction of phases versus temperature after 100-h aging111. Aging at 1200 and 1400oC progressively increased the amounts of equilibrium cubic (c) and transformable tetragonal (t) phases at the high temperature and this resulted in cubic (c) and monoclinic (m) phases at room temperature. The yttria content of the remaining t’ phase was lowered to ~ 5 mol% after 100 h at 1400oC. The t to m phase transformation on cooling with its accompanying volume expansion can lead to disintegration of the coating112. In laboratory torch tests, regions of optimal TBC lives were found to correlate with regions having high amounts of the t’ phase, small but nonzero amounts of the m phase, and little or no c phase113. EB-PVD YSZ (7 wt.% Y2O3) was investigated in regard to phase transformation after annealing114. Free-standing YSZ was heat-treated in air, for up to 200 h, in the temperature range 1200 – 1400oC. For 6-8 wt.% EB-PVD YSZ, the equilibrium phase diagram115 predicts a two-phase mixture, consisting of t phase containing 4 wt.% Y2O3 and c phase containing 16 wt.% Y2O3, at the deposition temperature of ~1000oC. The t’ phase, however, forms due to the rapid quench in the process. Annealing for short times (10 h at 1400oC and 30 h at 1300oC) and longer times at 1200oC produced significant amount of m phase, reaching ~ 20 mol% after 100 h at 1400oC. The proportion of c phase increased rapidly for all annealing conditions (1200 – 1400oC); at 1400oC, the content of c phase increased to ~55 mol% as soon as the temperature was reached and remained at that level up to 100 h. In general, the transformation to the c phase was more rapid than the transformation to m phase. Schultz et al. report that EB-PVD TBCs were stable up to 100 h at 1150oC, but transformed to a mixture of t+c+m phases after annealing for 100 h at 1400oC116. A slower cooling rate increased the amount of m phase at room temperature117. For instance, after 100 h at 1371oC, the amount of m phase increased from < 10 mol% to > 30 mol% when slow furnace cooling was used instead of quenching. Overall, the amount of m phase was in the same range for EB-PVD and APS coatings when aged in a similar condition118.
Fig. 8. Plot of the mole fraction of phases versus temperature for a 100-h aging (Reprinted with permission of The American Ceramic Society, copyright 1981, www.ceramics.org) Source: See Note 23.
426
Kang N Lee Alternative Coating Materials Zirconates Rare earth zirconates are being explored as alternate TBC materials. Rare earth zirconates, M2Zr2O7 (M = rare earth element), crystallize in the ordered pyrochlore structure over a composition range119. At elevated temperatures and outside the composition range, the disordered fluorite structure is the stable phase120. Maloney proposed the use of M2D2O7 (M = Gd, La, Y; D = Hf, Ti, Zr) as TBCs, while Suresh et al. independently proposed the use of a much broader family of compounds, where M represents all elements from the lanthanide series (La to Yb)121. The thermal conductivity of hot-pressed Gd2Zr2O7, Nd2Zr2O7, and Sm2Zr2O7 was investigated at 25-700oC122. The thermal conductivity of Gd2Zr2O7, Nd2Zr2O7 and Sm2Zr2O7 at 700oC was 1.6, 1.6 and 1.5 W/m-K, respectively, which was about 30% lower than the thermal conductivity of the reference YSZ (2.3 W/m-K). The decrease was attributed to phonon scattering by point defects. The two types of point defects expected in these materials are substitutional rare-earth solute cations (Gd, Nd, and Sm) replacing zirconium and the corresponding oxygen vacancies created by the substitution of tetravalent zirconium by trivalent rare-earth element123. The scattering strength of oxygen vacancies is larger than that of substitutional solutes, because of the missing anion mass and the missing interatomic linkages associated with the vacancies. The lower thermal conductivity compared to YSZ was attributed to two factors124: (i) higher concentration of oxygen vacancies present, and (ii) more effective phonon scattering by solute cations as a result of the significant atomic weight difference between the cations and zirconium. Two perovskites (SrZrO3 and BaZrO3) and one pyrochlore (La2Zr2O7) were investigated as TBC candidates125. Sintered SrZrO3 cracked ~600oC which was attributed to the volume change due to a phase transformation, making it unsuitable for a TBC. At 1000oC, the thermal conductivity of sintered BaZrO3 and La2Zr2O7 was 3.4 and 1.6 W/m-K, respectively, while the thermal conductivity of sintered YSZ was ~2.2 W/m-K. The Young’s modulus and hardness of sintered BaZrO3 and La2Zr2O7 were ~15% lower than those of sintered YSZ, while the CTEs were slightly lower than that of YSZ up to 1400oC. Low Young’s modulus is beneficial to thermal stresses, while low CTE is detrimental to thermal stresses. Plasma-sprayed La2Zr2O7 performed better than BaZrO3 in thermal cycling126. However, the La2Zr2O7 coating had a significantly shorter life than the YSZ coating, by roughly an order of magnitude in thermal cycling at 1240-1360oC127 . Layered or graded coatings with YSZ as the first ceramic coating of the TBC system and La2Zr2O7 as the final topcoat showed much improved performance, showing lives similar to YSZ128.
Yttria-Stabilized HfO2 Replacing the Zr in YSZ with Hf can theoretically decrease the thermal conductivity by increasing the mean atomic weight. Plasmasprayed hafnia-yttria coatings (HfO2-8.4 wt.% Y2O3, HfO2-11.4 wt.% Y2O3, HfO2-15.0 wt.% Y2O3, HfO2- 27.2 wt.% Y2O3) were evaluated with respect to plasma-sprayed YSZ (6-9 wt.% Y2O3) in a burner rig129. The hafnia-yttria coatings were very sensitive to plasma-spray parameters and high-quality coatings were obtained only when specific parameters were used. In contrast, YSZ coatings were in general relatively insensitive to spray parameter variations. In contrast to zirconia-yttria compositions, the hafnia-yttria compositions with the fully stabilized cubic phase (HfO2- 27.2 wt.% Y2O3) outperformed the partially stabilized compositions. The fully stabilized hafnia-yttria performed about as well as the YSZ coating when sprayed with certain parameter sets. It is possible that the fully stabilized hafniayttria may be more stable at higher temperatures (>1200oC) than the partially stabilized YSZ. The HfO2-27.2 wt.% Y2O3 also showed significantly superior resistance to sintering compared to YSZ (8 wt.% Y2O3) after 15 h at 1400oC.
YAG The thermal conductivity of dense, polycrystalline yttrium-based garnets in the temperature range 23 – 1000oC was investigated130. The high-temperature thermal conductivity of these garnets was in the range 2.4 – 3.2 W/m-K, which is somewhat higher than the thermal conductivity of dense YSZ (~2.4 W/m-K). It was demonstrated that the thermal conductivity of these garnets could be tailored by forming substitutional solid solution alloys. The oxygen diffusivity of YAG is about 10 orders of magnitude lower than that in zirconia, suggesting its potential as an alternative to zirconia in future TBCs for improved durability.
4.4.2-2d Failure Mechanisms of TBC While there are several different ways in which TBCs can fail, the mechanisms of greatest concern are those that intimately involve the growth of the TGO131. The failure is governed by a sequence of crack nucleation, propagation and coalescence events along the bond coat/TGO or top coat/TGO interfaces132. Figure 9 schematically illustrates the cracking sequence by growth misfit, followed by cooling to ambient133. Eventual failure occurs by either buckle, or edge-driven delamination134, with a failure plane located at (or near) the interface between the TBC and the bond coat. The life of specific system appears to correlate with the average thickness of the TGO135.
427
4.4.2 Protective Coatings for Gas Turbines
Fig. 9. Schematic illustration of the cracking sequence by growth misfit, followed by cooling to ambient (Reprinted with permission of Elsevier, copyright 2000.)
APS
Source: See Note 34.
The separation that nucleates the failure sequence appears to be associated with imperfections at (or near) the interface between the TBC and bond coat136. The most prominent are undulations in the original bond coat surface. The trajectory of the delamination that causes final failure resides primarily within the TBC itself and connects the peaks of the undulations. Rabiei & Evans suggest the following failure mechanism137. The sources of stress are those formed upon TGO growth, followed by the changes that happen because of thermal expansion misfit on cooling to ambient. The zones that experience out-of-plane tensile stress are most important, since these are the stresses responsible for nucleating and propagating cracks along delamination planes in the system. See Figure 9 for the cracking sequence. Radial cracks form in the TBC as the TGO thickens because of the out-of-plane stress in the TBC normal to the interface. The TBC cracks do not penetrate to TGO because the interface between the TGO and bond coat is in compression. On cooling to ambient, the CTE misfit causes appreciable tension to develop normal to the interface between the TGO and bond coat and this interface separates. The coalescence of this separation with the radial cracks in the TBC, by rupturing the intervening TGO, is a key event. Once coalescence happens, the energy density available in the TGO attached to the upper portion of the crack becomes available for outward growth of the crack in the TBC.
428
Kang N Lee EB-PVD The most crucial constituents for the failure of EB-PVD TBC are138: (i) Imperfections in the TGO affect the TBC life; (ii) The failure occurs by large scale buckling (LSB), subject to the development of large separated domains at the interface. Life is governed by the evolution of these separations up to the critical size needed for LSB; (iii) The adhesion of the TGO/bond coat interface degrades upon thermal exposure, both because of embrittlement by segregants and the growth of separations around the imperfections in the TGO. Mumm and Evans suggest the following mechanism139. As the TGO thickens with extended elevated temperature exposure, imperfections develop and enlarge. Accordingly, the ambient temperature energy release rate around the imperfections becomes larger the longer the exposure. There may be simultaneous, time-dependent embrittlement of the TGO/bond coat interface, perhaps by S segregation. When the imperfections become large enough, interface separations nucleate in their vicinity. An appreciable energy release rate develops only on cooling, caused by the CTE misfit between the TGO and the superalloy. Separations should from only on cooling. After longer exposures, when the imperfections further enlarge, it is surmised that some of the separations coalesce. The trajectory of the delamination primarily occurs at the TGO/bond coat interface.
4.4.2-3a Coatings for Ceramic Components Key requirements for a successful EBC include140: i) environmental stability, especially in water vapor; ii) coefficient of thermal expansion (CTE) match; iii) chemical compatibility; and iv) phase stability. It is also desirable to have a low thermal conductivity for maximum thermal insulation capability. Table I compares the CTE of SiC, Si3N4, and current EBC materials141. Note the good CTE match between Si-based ceramics and EBC materials . Table I. CTE of Si-based ceramics and current EBC materials
Material CTE (106 /C)
SiC
Si3N4
4.5 ~ 5.5
3~4
* 3Al2O3·2SiO2 ** 1-xBaO·xSrO·Al2O3·2SiO2, 0 ≤ x ≤ 1
Si 3.5 ~ 4.5
Mullite*
BSAS**
Y2SiO5 Sc2SiO5 Er2SiO5 Yb2SiO5
5~6
4~5
5~6
5~6
7~8
7~8
Source: See Note 141.
Table II lists the thermal conductivity of hot-pressed EBC materials determined by a high heat flux laser rig at 200oC - 1400oC142. Mullite (3Al2O3·2SiO2) and BSAS (1-xBaO·xSrO·Al2O3·2SiO2, 0 ≤ x ≤ 1) have thermal conductivities similar to that of YSZ (8 wt.% Y2O3), while rare earth monosilicates, except for Sc2SiO5, have lower thermal conductivities than YSZ. Table II. Thermal conductivity of hot-pressed current EBC materials at 200oC - 1400oC determined by a high heat flux laser rig
Mullite +BSAS
Material
YSZ
Mullite
BSAS
Thermal conductivity (W/m-k)
2.2 ~ 2.9
2.2 ~ 2.8
2.5 ~ 3.0 2.0 ~ 2.3 1.6 ~ 1.9
Y2SiO5
Sc2SiO5
Yb2SiO5
Er2SiO5
2.3 ~ 3.5
1.3 ~ 1.4 1.4 ~ 1.5
Source: See Note 142.
Current EBCs have multi layers designed in such a way that the system satisfies all the key requirements for a successful EBC. They consist of a silicon bond coat and a ceramic top coat. The ceramic top coat typically comprises at least two ceramic layers. The bond coat facilitates the adherence of the ceramic top coat to the substrate and the ceramic top coat provides protection from water vapor and thermal insulation.
4.4.2-3b Processing APS is the most successful and widely used process to apply EBCs142. With EB-PVD process, the low vapor pressure of silica compared to alumina and rare earth oxides makes it difficult to produce coatings with the desired stoichiometry. Other coating processes being explored include chemical vapor deposition (CVD)143, sol-gel, and slurry coatings144, which have the benefit of being none-lineof-sight processes.
429
4.4.2 Protective Coatings for Gas Turbines 4.4.2-3c Testing The key component in EBC testing is water vapor. Laboratory scale high steam rigs145 are used to simulate the high water vapor pressure, while high pressure/high velocity burner rigs146 are used to simulate both the high water vapor pressure and high gas velocity. A low gas velocity (a few cm/sec) is typically employed in laboratory scale high steam rigs due to the simplicity. Since water vapor is the most critical factor that affects the performance of EBCs, this test is suitable for the initial screening of EBC candidates and for the evaluation of long-term EBC performance. The high pressure/high velocity burner rig test is suitable for proof tests of mature EBCs and for subcomponent tests since it closely simulates actual gas turbine environments. Burner rigs can be set up to generate a temperature gradient through the EBC, simulating the temperature profile of cooled components. Laser rigs are used for thermal conductivity measurements as well as for the evaluation of EBC performance under a temperature gradient147.
4.4.2-3d Bond Coat Silicon is the current bond coat. It provides excellent durability to EBC by facilitating adherence to the ceramic top coat and the oxidation resistance148. The excellent performance of Si bond coat is attributed to its close CTE match with Si-based ceramics, oxidation resistance due to the formation of slow-growing silica scale and chemical compatibility with the substrate and the ceramic top coat. The use of silicon bond coat is limited by its melting point (~1416oC). For higher temperature applications where Si bond coat melts, the next ceramic layer, such as a mullite-based coating, becomes the bond coat. The life of EBC is significantly reduced without Si bond coat.
4.4.2-3e Top Coat Current top coat can be divided into two groups, i.e. top coat with a mullite-based layer and top coat without a mullite-based layer. Table III lists the current top coats. Table III. List of current top coats
With a mullite-based layer
Without a mullite-based layer
mullite mullite + BSAS mullite or mullite + BSAS / YSZ mullite or mullite + BSAS / BSAS mullite or mullite + BSAS / RE2SiO5 or RE2Si2O7*
BSAS RE2SiO5* RE2Si2O7* Ta2O5
* RE: rare earth element
Top Coat with a Mullite-Based Layer Mullite has attracted the most interest as a protective coating for Si-based ceramics because of its good CTE match and chemical compatibility with SiC and Si3N4 ceramics149. With the emergence of Si bond coat, mullite and mullite-based coatings became an intermediate layer bridging the Si bond coat and the water vapor-resistant top layer150. Key functions of mullite-based coatings when used as an intermediate coat in current EBCs include chemical compatibility, strain tolerance, and barrier to water vapor transport.
Mullite and Mullite+BSAS Conventionally plasma-sprayed mullite coatings contain a significant amount of metastable amorphous phase due to the rapid cooling of molten mullite during the solidification on a cold substrate151. A subsequent exposure of the mullite coating to a temperature above ~1000oC causes the crystallization of the amorphous phase. Shrinkage accompanies the crystallization, leading to cracking and delamination of the mullite coating. A modified plasma-spraying process enables the deposition of crystalline mullite coating, dramatically improving the crack resistance and adherence152. Plasma-sprayed mullite coating on SiC remained virtually intact at the mullite/SiC interface as well as on the mullite surface after a 50-h exposure in a high pressure (6 atm) hot corrosion rig containing Na2SO4153. CVD mullite also displayed little evidence of damage in Na2SO4 environments154. Further improvement in the crack resistance of plasma-sprayed mullite coatings is achieved by adding a second phase (BSAS)155. The improved crack resistance of the mullite+BSAS composite coating is attributed to the reduced coating tensile stress due to the low modulus BSAS phase156.
430
Glass
Kang N Lee Mullite/YSZ The relatively high silica activity of mullite (0.3 ~ 0.4) causes the selective volatilization of silica and the recession of mullite in high velocity combustion environments157. Therefore, a water-vapor resistant overlay coating is needed on top of the mullite coating, to provide the stability in water vapor. YSZ is a logical candidate for a top coat because it has been successfully used as a TBC for superalloy components in gas turbine engines, signifying its stability in water vapor. The mullite/YSZ system is an effective EBC for SiC/SiC short term exposures. One critical disadvantage of YSZ is the high CTE and sintering. The stresses due to the CTE mismatch and sintering cause severe 100 µm cracking. These cracks provide easy paths for water vapor penetration, resulting 158 in rapid oxidation and premature coating delamination .
BSAS
Mullite/BSAS and Mullite+BSAS/BSAS BSAS is a top coat material developed in the NASA’s High Speed ResearchEnabling Propulsions Materials (HSR-EPM) program in joint research by NASA, General Electric, and Pratt and Whitney159. It has the key attributes for a successful EBC, such as a low silica activity, a low CTE, and a low modulus (~100 GPa for dense BSAS). The low silica activity provides stability in water vapor, while the low CTE and low modulus provide low thermal stresses. The EPM EBCs exhibit dramatically improved durability compared to the mullite/ YSZ EBC. Figure 10 shows plasma-sprayed Si/mullite+20 wt.% BSAS/ BSAS on SiC/SiC composite after 1000 h at 1316oC (1-h cycles) in 90% H2Obalance O2160. The EBC maintained excellent adherence and crack resistance. Pockets of glasses developed within the BSAS top coat. The EPM EBC was applied on SiC/SiC composite combustor liners in Solar Turbines (San Diego, CA) Centaur 50s gas turbine engines under DOE Ceramic Stationary Gas Turbines (CSGT) Program161. One engine used by Texaco in Bakersfield, CA, successfully completed a 14,000-h field test (~1,250oC maximum combustor liner temperature). The higher operating temperature resulted in emissions consistently below 15 ppmv nitrogen oxides (NOx) and below 10 ppmv carbon monoxide (CO) throughout, roughly reducing the NOx and CO loads on the environment by factors of about 2 and 5, respectively. The EPM EBCs have some durability issues that limit their upper use temperature and life162. One key issue is the volatilization of the BSAS top coat in high velocity combustion environments. A projection based on a silica volatility model in conjunction with BSAS volatility data indicates a BSAS recession of ~70 µm after 1000 h at 1400oC, 6 atm total pressure and 24 m/s gas velocity. Actual gas turbines operate at significantly higher pressures and gas velocities, which increases the projected recession to much higher levels. The EBC in Solar Turbines engines suffered significant BSAS recession in some areas after the 14,000-h test163. Another key issue is the chemical reaction between BSAS and the thermally grown silica on Si bond coat. The BSAS-silica reaction produces a low-melting (~1300oC) glass that causes EBC degradation and a premature failure at temperatures above ~1300oC. The pockets of glasses in Figure 10 are due to the BSAS-silica reaction. Therefore, it is desirable to avoid the BSAS second phase in the mullite layer for applications requiring long-term exposures at temperatures above 1300°C ~ 1350°C.
Mullite + BSAS
Fig. 10. Plasma-sprayed Si/mullite+20 wt.% BSAS/ BSAS on a SiC/SiC composite coupon after 1000 h at 1316°C (1-h cycles) in 90% H2O-balance O2 (Reprinted with permission of Elsevier, copyright 2005) Source: See Note 141.
Yb2SiO5 Mullite
Si
SiC/SiC
100 µm Fig. 11. Plasma-sprayed Si/mullite/Yb2SiO5 on SiC/ SiC composite after 1000 h at 1380°C (1-h cycles) in 90% H2O-balance O2. (Reprinted with permission of ASM International, All rights reserved, www.asminternational.org) Source: See Note 142.
Mullite/RE2SiO5 (or RE2Si2O7) and Mullite+BSAS/RE2SiO5 (or RE2Si2O7)
431
Some rare earth silicates have a low CTE, phase stability, and a low silica activity, making them excellent EBC top coat materials164. Volatility data165 and thermodynamic calculations166 indicate that rare earth monosilicates (RE2SiO5; RE = rare earth element) are significantly less volatile than BSAS in water vapor, at least by an order of magnitude, while the volatilities of rare earth disilicates (RE2Si2O7) are similar to that of BSAS. Disilicates of Y, Yb, and Lu exposed to a high velocity and high steam environment at 1450oC ~ 1500oC gradually decomposed to Y2SiO5, Yb2SiO5, and Lu2SiO5, respectively, indicating the selective volatilization of silica167. Figures 11 and 12 show plasma-sprayed Si/mullite/Yb2SiO5 on SiC/SiC composite (1000 h) and on Si3N4 (400 h) at 1380oC (1-h cycles) in 90% H2O-balance O2168. The EBC maintained superb adherence and crack resistance on both substrates. In contrast, plasmasprayed Si/mullite/YSZ on Si3N4 suffered severe cracking and delamination after 280 h at 1380oC (1-h cycles) in 90% H2O-balance O2 (Figure 13)169. This demonstrates the detrimental effect of a large CTE mismatch between EBC and substrate. Figure 14 shows plasmasprayed Si/mullite/Yb2SiO5 on SiC/SiC composite after 100 h in a high pressure/high velocity burner rig (1400oC, 100 h, 6 atm, gas velocity = 24 m/s)169. Note the excellent adherence, crack resistance, and oxidation resistance of the rare earth silicate EBC in a simulated gas turbine environment.
4.4.2 Protective Coatings for Gas Turbines Top Coat without a Mullite-Based Layer BSAS and RE2SiO5 possess key properties desirable for a successful EBC, such as low CTE, low silica volatility, and phase stability170. Ta2O5 has a low CTE, but relatively high volatility in water vapor, higher than the volatility of BSAS by an order of magnitude171, and a phase transformation at ~ 1370oC. The performance of these materials without a mullite-based intermediate layer was investigated.
Mullite
Si Si3N4
BSAS BSAS is reactive with silica, thermally grown on SiC or Si3N4, forming a low melting eutectic (mp ~ 1300oC). The chemical reaction causes the build-up of a thick reaction zone and porosity at the BSAS/substrate interface. Plasma-sprayed BSAS-coated SiC/SiC composite after 100 h (2-h cycles) at 1300oC in 90% H2O - balance O2 developed a thick (~20 µm) reaction zone and large pores at the BSAS/ substrate interface172. Pores are attributed to the bubbling of gaseous species through the low viscosity eutectic glass. In long-term exposures, the pores continue to grow and coalesce, leading to complete coating spallation. The low viscosity reaction zone can lead to the EBC spallation under a high shear stress. Similar behavior is observed in the presence of Si bond coat.
Fig. 12. Plasma-sprayed Si/mullite/Yb2SiO5 on Si3N4 after 400 h at 1380°C (1-h cycles) in 90% H2O-balance O2. (Reprinted with permission of Elsevier, YSZ copyright 2005.)
Mullite
Source: See Note 141.
Si
RE2SiO5, RE2Si2O7, and Ta2O5
100 µm
Plasma-sprayed RE2SiO and Ta2O5-based EBCs showed good adherence on Si-based ceramics under thermal exposures in air. However, these coatings on SiC/SiC composite did not maintain the adherence in water vapor environments175. Consequently, the substrate suffered rapid oxidation, forming thick and porous scale. A premature EBC spallation occurred along the scale since the thick scale constituted a weak mechanical link. Possible explanations for the rapid oxidation include lack of chemical bonding and EBC cracking under thermal cycling. Both can provide an easy access for water vapor into the interface. The fact that the Si/mullite EBC shows far superior oxidation resistance and longer life, although mullite develops similar cracks, suggests that inadequate chemical bonding may be responsible for the lack of oxidation resistance. Other low CTE rare earth monosilicates, such as Y2SiO5, Er2SiO5, Sc2SiO5, and Lu2SiO5, exhibited similar poor oxidation resistance176. CVD Ta2O5 was unstable in an environment containing Na2SO4, rapidly reacting to form NaTaO3 which subsequently interacted destructively with the underlying Si3N4 substrate to form a molten phase177. Rare earth disilicates (RE2Si2O7) were applied on Si3N4 ceramics by a slurry process followed by sintering178. Short-term exposures at relatively low temperatures showed promising results, indicating their merits for further research. 173 5
174
Si3N4
Fig. 13. Plasma-sprayed Si/mullite/YSZ on Si3N4 after 280 h at 1380°C (1-h cycles) in 90% H2O-balance O2. Source: See Note 169.
Yb2SiO5 Mullite Si SiC/SiC
4.4.2-4 Conclusions 100 µm
Thermal barrier coatings for superalloys and environmental barrier coatings for ceramic matrix composites are important technologies to achieve higher gas turbine Fig. 14. Plasma-sprayed Si/mullite/Yb2SiO5 on inlet temperatures. a SiC/SiC composite after 100 h in a high presIn thermal barrier coatings, stresses play the major role in coating failure, which sure/high velocity burner rig (1400°C, 100 h, 6 are influenced by various factors such as bond coat oxidation, bond coat surface atm, gas velocity = 24 m/s). roughening, YSZ phase transformation, and YSZ sintering. Current approaches for improvements include adding reactive element (RE) or precious metals in the bond Source: See Note 169. coat for enhanced oxidation resistance, and alternative oxide stabilizers for ZrO2 and alternative materials to ZrO2 for enhanced phase stability and sinter resistance. Thermal conductivity is another key element for the TBC performance. A lower thermal conductivity TBC lowers the substrate temperature and/or reduces the TBC thickness, which improves the performance and life of gas turbine components. Key future research areas include the development of new bond coat alloy compositions resistant to surface imperfections and the development of low thermal conductivity TBC, without compromising the quality of TGO and TBC life. In environmental barrier coatings, chemical reactions, rather than stresses, appear to play the major role in the coating degradation. Key life-limiting reactions are water vapor volatility of the surface layer, chemical reactions between various EBC layers, including silica TGO, and the oxidation of silicon bond coat. Therefore, it is important to design EBC in such a way to minimize these chemical
432
Kang N Lee reactions. The selection of an EBC for a particular system depends on application requirements, such as the EBC surface temperature, the substrate temperature, and the life goal. Alternative top layer materials to BSAS are being investigated for applications at T>1300oC. Rare earth silicates have shown promising results, while ZrO2- and HfO2-based materials require further research to alleviate the thermal expansion mismatch stress. As advanced gas turbines rely on these coatings for environmental/thermal protection, life models to predict the remaining coating life and non-destructive technologies to monitor the coating’s health are important future research areas.
4.4.2-5 Notes _________________________
433
1. F.O. Soechting, Thermal Barrier Coating Workshop, NASA Lewis Research Center, Cleveland, OH, NASA Conference Publication 3312, (1995): 1-15. 2. M. J. Donachie and S. J. Donachie, Superalloys – Second Edition, ASM International, Materials Park, Ohio (2000). 3. Y. Tamarin, Protective Coatings for Turbine Blades, ASM International, Materials Park, OH (2002). 4. Ibid. 5. Ibid. 6. See note 1 above. 7. R. A. Miller, Thermal Barrier Coating Workshop, NASA Lewis Research Center, Cleveland, OH, NASA Conference Publication 3312 (1995): 17-34. 8. D. Anson and D. W. Richerson, Progress in Ceramic Gas Turbine Development, Vol. 2. Edited by M. van Roode, M. Ferber, and D. W. Richerson, ASME PRESS, New York, NY, 1-10 (2003). 9. N. S. Jacobson, J. Am. Ceram. Soc., 76[1] (1993): 3-28. 10. E. J. Opila and R. Hann, J. Am. Ceram. Soc., 80[1], 197-205 (1997); J. L. Smialek, R. C. Robinson, E. J. Opila, D. S. Fox, and N. S. Jacobson, Adv. Composite Mater., 8[1] (1999): 33-45. 11. N. S. Jacobson, J. L. Smialek, and D. S. Fox, Handbook of Ceramics and Composites, Vol. 1. Edited by N. S. Cheremisinoff, Marcel Dekker, New York, NY, USA, 99-135 (1990). 12. K. N. Lee, H. Fritze, and Y. Ogura, Progress in Ceramic Gas Turbine Development, Vol. 2. Edited by M. van Roode, M. Ferber, and D. W. Richerson, ASME PRESS, New York, NY, 641-664 (2003). 13. K. N. Lee, D. S. Fox, J. I. Eldridge, D. Zhu, R. C. Robinson, N. P. Bansal, and R. A. Miller, J. Am. Ceram. Soc. 86 [8] (2003): 1299-1306. 14. H. E. Eaton, G. D. Linsey, K. L. More, J. B. Kimmel, J. R. Price, and N. Miriyala, ASME paper 2000-GT-0631, International Gas Turbine and Aeroengine Congress and Exposition, Munich, Germany, May 8-11, 2000. 15. See note 7 above. 16. See note 3 above. 17. D. R. Clarke, Surface and Coatings Technology, 163-164, (2003): 67-74. 18. A.G. Evans, D. R. Mumm, J. W. Hutchinson, G. H. Meier and F. S. Pettit, Progress in Materials Science 46 (2001): 505553. 19. M. K. Ferber, J. P. Singh, J. A. Haynes, M. Lance, I. G. Wright, H. Wang, and G. Romanoski, Advanced Turbine Systems Program 1998 Annual Report, Department of Energy (1998); B. A. Pint, I. G. Wright, W. Y. Lee, Y. Zhang, K. Pruβner, and K. B. Alexander, Mater. Sci. Eng., A145, 201-211 (1998); I. G. Wright and B. A. Pint, 1st Int. Conf. on Industrial Gas Turbine Technology (CAME-GT), Brussels, Belgium, July 10-11, 2003. 20. J. Cheng, E. H. Jordan, B. Barber, and M. Gell, Acta Mater., 46 [16] (1998): 5839-5850; also see note 18 above. 21. Y. Zhang, J. A. Haynes, B. A. Pint, I. G. Wright, and W. Y. Lee, Surface and Coatings Technology, 163-164, 19-24 (2003); M. W. Chen, R. T. Ott, T. C. Hufnagel, P. K. Wright, and K. J. Hemker, Surface and Coatings Technology, 163-164, 2530 (2003);V. K. Tolpygo and D. R. Clarke, Acta Mater. 48, 3283-3293 (2000); S. Darzens, D. R. Munn, D. R. Clarke, and A. G. Evans, Met. and Mater. Trans. A, 34A, 511-522 (2003);D. R. Mumm, A. G. Evans, and I. T. Spitsberg, Acta. Mater. 49, 2329-2340 (2001);M. Gell, K. Vaidyanathan, B. Barber, J. Cheng, and E. Jordan, Met. and Mater. Trans. A, 30A, 427-435 (1999);Y. H. Sohn, J. H. Kim, E. H. Jordan, and M Gell, Surface and Coatings Technology, 146-147, 70-78 (2001). 22. A. G. Evans, G. B. Crumley, and R. E. Demaray, Oxid. Met., 20 [5] (1983): 193;G. C. Chang and W. Phucharoen, Surface and Coatings Technology 30 (1987): 13-28; A. M. Freborg, B. L. Ferguson, W. J. Brindley, and G. J. Petrus, Mat. Sci. and Eng., A245 (1998): 182-190;E. P. Busso, J. Lin, S. Sakura, and M. Nakayama, Acta. Mater 49 (2001): 1515-1528; E. P. Busso, J. Lin, and S. Sakura, Acta. Mater 49 (2001): 1529-1536; see also notes 18 and 20. 23. R. A. Miller, J. L. Smialek, and R. G. Garlick, Advances in Ceramics, Vol. 3, Science and Technology of Zirconia. Edited by A. H. Heuer and L. W. Hobbs, American Ceramic Society, Columbus, OH, 241-251 (1981). 24. D. Zhu and R. Miller, Journal of Thermal Spray Technology, 9[2] (2000): 175-180. 25. J. R. Gross, M. N. Rahaman, and R. E. Dutton, Ceram. Trans 154 (2003): 311-320. 26. See note 3. 27. See note 19.
4.4.2 Protective Coatings for Gas Turbines 28. See note 3. 29. Ibid. 30. See note 21 (Tolpygo) and (Gell). 31. See note 3. 32. Ibid. 33. Ibid. 34. A. Rabiei and A. G. Evans, Acta. Mater., 48, (2000): 3963-3976. 35. J. A. Haynes, M. J. Lance, B. A. Pint, and I. G. Wright, Surface and Coatings Technology 146-147 (2001): 140-146. 36. M. Gell, E. Jordan, and K. Vaidyanathan, Surface and Coatings Technology, 120-121, (1999): 53. 37. See note 19. 38. E. J. Felten, Oxid. Metals 10 (1976) 23-28; E. J. Felten and F. S. Pettit, Oxid. Metals, 10, 189-223 (1976); J. G. Fountain, F. A. Golightly, F. H. Stott, and G. C. Wood, Oxid. Metals 10 (1976): 341-345. 39. See note 19 (Wright & Pint). 40. See note 19 (Pint, et al.). 41. Ibid. 42. P. Y. Hou and J. Stringer, Oxid. Metals 38 (1992): 323-345; H. J. Grabke, G. Kurbatov, and H. J. Schmutzler, Oxid. Metals 43 (1995): 97-114. 43. See note 19 (Pint, et al.). 44. W. P. Allen and N. S. Bornstein, High Temperature Coatings I. Edited by N. Dahotre, J. M. Hampikian, and J. Stiglich, TMS, Warrendale, PA, 193-202 (1995); G. H. Meier, F. S. Pettit, and J. L. Smialek, Mater. Corros. 46 (1995): 232-240; M. A. Smith, W. E. Frazier and B. A. Pregger, Mater. Sci. Eng., A203, 388-398 (1995); J. C. Schaeffer, W. H. Murphy, and J. L. Smialek, Oxid. Metals 43 (1995): 1-23. 45. B. A. Pint, Oxid. Metals 45 (1996): 1-37. 46. Ibid. 47. See note 19 (Pint, et al.). 48. See note 19 (Pint, et al.) and (Wright & Pint). 49. See note 19 (Wright & Pint). 50. See note 19 (Pint, et al.). 51. Ibid. 52. See note 35. 53. See notes 3 and 19 (Wright & Pint). 54. See note 3. 55. See note 21 (Zhang, et al.), (Chen, et al.), and (Tolpygo & Clarke). 56. See notes 21 (Sohn, et al.) and 34. 57. See note 34. 58. See note 21 (Tolpygo & Clarke). 59. See note 21 (Zhang, et al.), (Chen, et al.), (Tolpygo & Clarke), and (Darzens, et al.). 60. Ibid. 61. See note 21(Chen et al.). 62. Ibid. 63. See note 21 (Zhang, et al.) and (Chen, et al.). 64. See note 21 (Zhang, et al.). 65. See note 21 (Zhang, et al.) and (Chen, et al.). 66. See note 21 (Tolpygo & Clarke). 67. See note 21 (Darzens et al.). 68. See note 21 (Mumm et al.). 69. See note 21 (Gell et al.). 70. D. R. Mumm and A. G. Evans, Acta Mater. 48, (2000): 1815. 71. See note 21 (Tolpygo & Clarke) and (Mumm et al.). 72. See note 21 (Mumm et al.). 73. See note 21 (Tolpygo & Clarke) and (Mumm et al.). 74. Handbook of Thermal Spray Technology. Edited by J. R. Davis, ASM International, Metals Park, OH, 2004. 75. A. H. Bartlett and R. Dal Maschino, J. Am. Ceram. Soc. 78 (1995): 1018. 76. J. R. Nicholls, K. J. Lawson, A. Johnstone, and D. S. Rickerby, Surface and Coatings Technology 151-152 (2002): 383391. 77. D. S. Duvall and D. L. Ruckler, ASME Paper 82-GT-327 (1982). 78. U. Schulz, K. Fritscher, H.-J. Ratzer-Scheibe, et al., High Temperature Corrosion 4, Edited by R. Streiff, J. Stringer, R. Krutenat, M. Caillet, and R. Rapp, Trans Tech Publication, 957-964 (1997). 79. J. R. Nicholls, Y. Jaslier, and D. S. Rickerby, 4th Int. Symp. On High Temperature Corrosion, Les Embiez, France, May, 1996. 80. See note 78.
434
Kang N Lee
435
81. P. G. Klemens and M. Gell, Mater. Sci. Eng, A245 (1998): 143-149. 82. Ibid. 83. Ibid. 84. Ibid. 85. See notes 17 and 81. 86. See notes 76 and 81. 87. See note 81. 88. S. Raghavan, H. Wang, R. B. Dinwiddie, W. D. Porter, and M. J. Mayo, Scripta Materialia, 39[8] (1998): 1119-1125; J. –F. Bisson, D. Fournier, M. Poulain, O. Lavigne, and R. Mevrel, J. Am. Ceram. Soc. 83, (2000): 1993. 89. See note 17. 90. See note 81. 91. Ibid. 92. Ibid. 93. See note 17. 94. See notes 17 and 76. 95. See note 76. 96. Ibid. 97. D. S. Rickerby, European Patent EP 0 825 271 A1. 98. D. Zhu and R. A. Miller, Ceram. Eng. Sci. Proc. 23[4] (2002): 457-468. 99. Ibid. 100. J. R. Nicholls, K. J. Lawson, A. Johnston, and D. S. Rickerby, High temperature corrosion 5. Edited by R. Streiff, I. J. Wright, R. Krutenat, M. Caillet, and A. Cailerie, Trans Tech Publication, 595-606 (2001). 101. Ibid. 102. See note 76. 103. Ibid. 104. See note 24. 105. V. Lughi, V. K. Tolpygo, and D. R. Clarke, Mater. Sci. Eng. A368 (2004): 212-221. 106. Ibid. 107. See note 25. 108. See note 98. 109. See note 23. 110. Ibid. 111.Ibid.. 112. Ibid. 113. S. Stecura, NASA TM 78976 (1978). 114. U. Schulz, , J. Am. Ceram. Soc. 83 (2000): 904-910. 115. H. G. Scott, J. Mater. Sci. 10 (1975): 1527-1535. 116. U. Schulz, K. Fritscher, and M. Peters, J. Eng. Gas Turbines Power 119 (1997): 817-21. 117. J. R. Van Valzah and H. E. Eaton, Surf. Coat. Technol. 46 (1991): 289-300. 118. See note 114. 119. D. J. M. Bevan and E. Summerville, Handbook on the Physics and Chemistry of Rare Earths: Non-Metallic Compounds I. Edited by K. A. Gschneider and L. R. Eyring, North-Holland Physics Publishing. New York, 412-515 (1979); M. A. Subramanian and A. W. Sleight, Handbook on the Physics and Chemistry of Rare Earths: Non-Metallic Compounds I. Edited by K. A. Gschneider and L. R. Eyring, Elsevier Science Publishers, Oxford, UK, 225-248 (1993). 120. Ibid. 121. M. J. Maloney, U.S. Patent No. 6 117 560, 2000; M. J. Maloney, U.S. Patent No. 6 284 323, 2001; G. Suresh, G. Seenivasan, M. V. Krishnaiah, and P. S. Murti. J. Nucl. Mater., 249, 259-61 (1997); G. Suresh, G. Seenivasan, M. V. Krishnaiah, and P. S. Murti. J. Alloys Compd., 269, L9-L12 (1998). 122. J. Wu, X. Wei, N. P. Padture, P. G. Klemens, M. Gell, E. Garcia, P. Miranzo, and M. I. Osendi, J. Am. Ceram. Soc. 85 (2002): 3031-3035. 123. Ibid. 124. Ibid. 125. R. Vassen, X. Cao, F. Tietz, D. Basu, and D. Stover, J. Am. Ceram. Soc. 83 (2000): 2023-2028. 126. Ibid. 127. R. Vassen, X. Cao, F. Tietz, and D. Stover, Ceram. Eng. Sci. Proc. 22 [4] (2001): 435-442. 128. Ibid. 129. R. A. Miller and G. W. Leissler, NASA TP 3296 (1993). 130. N. P. Padture and P. G. Klemens, J. Am. Ceram. Soc., 80, (1997): 1018-1020. 131. P. K. Wright, Mater. Sci. Eng. A245 (1998): 191; P. K. Wright and A. G. Evans, Curr. Opin. Solid State Mater. Sci. 4 (1999): 255. 132. See note 18.
4.4.2 Protective Coatings for Gas Turbines 133. See note 34. 134. S. R. Choi, J. W Hutchinson, and A. G. Evans, Mech. Mater. 31 (1999): 431; also see note 18. 135. See note 131 (Wright, 1998). 136. See note 34. 137. Ibid. 138. D. R. Mumm and A. G. Evans, Acta Mater 48 (2000): 1815-1827. 139. Ibid. 140. K. N. Lee, Surface and Coatings Technology, 133-134, 1-7 (2000); also see note 12. 141. K. N. Lee, D. S. Fox, and N. P. Bansal, J. Euro. Ceram. Soc. 25 (2005): 1705-1715. 142. N.S. Jacobson, D.S. Fox, J.L. Smialek, E.J. Opila C. Dellacorte, and K.N. Lee, ASM Handbook, Vol. 13B. Edited by S.D. Cramer and B.S. Covino, Jr., ASM International, Materials Park, OH, 565-578 (2005). 143. M. L. Auger and V. K. Sarin, Surface and Coatings Technology 94-95 (1997): 46-52 ; S. N. Basu, P. Hou, and V. K. Sarin, International Journal of Refractory Metals & Hard Materials 16 (1998): 343-52; J. A. Haynes, K. M. Cooly, D. P. Stinton, R. A. Lowden, and W. Y. Lee, Ceram. Eng. and Sci. Proceedings, 20 [4] (1999): 355-362 ; J. A. Haynes, M. J. Lance, K. M. Cooley, M. K. Ferber, R. A. Lowden, and D. P. Stinton, J. Am. Ceram. Soc. 83 [3] (2000): 657-659; Y. W. Bae, W. Y. Lee, and D. Stinton, J. Am. Ceram. Soc., 78 [5] (1995): 1297-300; W. Y. Lee, Y. W. Bae, and D. P. Stinton, J. Am. Ceram. Soc., 78 [7] (1995): 1927-30. 144. See note 12. 145. See note 13. 146. R. C. Robinson and J. L. Smialek, J. Am. Ceram. Soc. 82 [7] (1999): 1817-25 . 147. D. Zhu, K. N. Lee, and R. A. Miller, Ceram. Eng. Sci. Proc., 22 [4] (2001): 443-452. 148. See note 140. 149. J. I. Federer, J. Mater. Eng. 12 (1990): 141-149; D. W. Richerson and J. L. Schienle, Proceedings of the Twenty-Second Automotive Technology Development Contractors’ Coordination Meeting (October 29-November 2, 1984), SAE P-155, 453-461, March 1985; J. R. Price, M. van Roode, and C. Stala, Key Engineering Materials 72-74 (1992): 71-84; K. N. Lee, R. A. Miller, and N. S. Jacobson, J. Am. Ceram. Soc. 78 [3] (1995): 705-710. 150. See note 140. 151. See note 149 (Lee et al.). 152. Ibid. 153. K. N. Lee, Transactions of the ASME 122 (2000): 632-636. 154. W. Y. Lee, K. L. More and Y. W. Bae, J. Am. Ceram. Soc. 79 [9] (1996): 2489-92. 155. See notes 13 and 140. 156. K. N. Lee, J. I. Eldridge, and R. C. Robinson, J. Am. Ceram. Soc., 88 [12] (2005): 3483-3488. 157. See note 153. 158. See note 156. 159. See notes 13 and 140. 160. See note 141. 161. H. E. Eaton, G. D. Linsey, E. Y. Sun, K. L. More, J. B. Kimmel, J. R. Price, and N. Miriyala, ASME paper 2001-GT0513 ASME TURBOEXPO 2001, New Orleans, Louisiana, June 4-7, 2001; K. L. More, P. F. Tortorelli, L. R. Walker, J. B. Kimmel, N. Miriyala, J. R. Price, H. E. Eaton, E. Y. Sun, and G. D. Lindsey, ASME paper 2002-GT-30630, Proceedings of ASME Turbo Expo 2002, Amsterdam, Netherlands, June 3-6, 2002, 162. See note 13. 163. See note 161 (More et al.). 164. K.N. Lee, US patent 6,759,151 (2004); also see note 141. 165. See note 141. 166. P. Meschter, GE Global Research Center, private communication. 167. T. Ohji, The 28th Int. Conf. & Exp. On Adv. Ceram. & Composites, Cocoa Beach, FL (2004); H. Klemm, M. Fritsch, and B. Schenk, The 28th Int. Conf. & Exp. On Adv. Ceram. & Composites, Cocoa Beach, FL (2004); S. Ueno, D. D. Jayaseelan, N. Kondo, T. Ohji, and S. Kanzaki, The 28th Int. Conf. & Exp. On Adv. Ceram. & Composites, Cocoa Beach, FL (2004). 168. See notes 141 and 142. 169. K. N. Lee, unpublished research, NASA Glenn Research Center. 170. See notes 13 and 141. 171. See note 169. 172. See note 13. 173. See note 169. 174. C. M. Weyant, K. T. Faber, J. D. Almer, and J. V. Guibeen, J. Am. Ceram. Soc. 88 [8] (2005): 2169-76. 175. See note 141. 176. Ibid. 177. See note 144 (Lee et al.). 178. See note 167 (Ohji) and (Ueno, et al.).
436
BIOGRAPHY
4.4.2 Protective Coatings for Gas Turbines
Kang N Lee Cleveland State University NASA Glenn Research Center Cleveland, OH 44135 Current Address:
Rolls-Royce Corporation P.O. Box 420, Speed Code W-08 Indianapolis, IN 46206 phone: (317) 230-4469 email: [email protected]
Kang Lee is a pioneer in environmental barrier coatings (EBC) for silicon-based ceramics. He played a key role in the development of the current state-of-the-art EBC. He has also investigated thermal barrier coatings (TBC) based on silicates and developed an emission spectroscopic non-destructive evaluation (NDE) technique for high temperature coatings. His current research focuses on development of advanced EBC for silicon carbide ceramic matrix composite (CMC) turbine hot section components and applying emission spectroscopic NDE to TBC. He has 13 U.S patents and has written two invited book chapters and over 50 technical publications.
4.4
Heat Transfer Analysis
Frank J. Cunha, Ph.D., P.E. Pratt & Whitney United Technologies Corporation 5 Bruce Lane Avon, CT 06001 Phone: (860) 565-8909 Email: [email protected]
389
4.4-1 Introduction The thermal efficiency and specific output of a gas turbine are primarily dependent on two major cycle parameters: the pressure ratio and the turbine inlet temperature1. In an ideal Brayton cycle, thermal efficiency increases up to stoichiometric temperatures and high-pressure ratios, without considering losses, particularly, those associated with turbine cooling. Since turbine airfoil materials melt at temperatures much lower than the stoichiometric temperatures, hot gas-path components, such as turbine airfoils, must be cooled and attention must be given to cycle parasitic losses. The recognition of material temperature limitations has led to the continuous turbine development programs for cooling technologies, material development, and related multi-disciplinary disciplines of fluid dynamics, heat transfer, aerodynamic performance, and structures, all aimed at the durability of turbine hot-gas-path components. The pursuit of improved turbine materials began long ago when the initial temperature limitations were found to be at about 1500°F (800°C)2. Following this initial period, an intensive development period took place when nickel-based alloys were developed and characterized as having high creep resistance characteristics. Material improvements relaxed temperature limitations by about 300°F (167°C)3. Further development of turbine airfoil manufacturing techniques, such directionally-solidification castings and single-crystal castings led to higher metal temperature capability. More recently, numerous testing evaluations have been conducted to characterize new hot-gas path material superalloys in terms of tensile, rupture, fatigue, creep, toughness, corrosion and oxidation resistances, producibility, processing, and other thermophysical properties4. Following extensive laboratory testing, actual operating experience is gained with engine testing subject to real operational environments culminating in mature levels of technology readiness levels for production. Today, many modern turbine airfoils use single crystal superalloys. These are two-phase alloys with a large volume fraction of precipitates, based on the intermetallic compound, Ni3 Al, interspersed in a coherent facecentered cubic matrix comprised of nickel, Ni, with smaller weight percent of various other elements in solid solution5. These elements include: cobalt, Co, aluminum, Al, chromium, Cr, tungsten, W, molybdenum, Mo, tantalum, Ta, hafnium, Hf, rhenium, Re, and ruthenium, Ru. The elements Re and Ru are introduced in the latest generation of single crystal alloys. All these elements have different attributes which can be summarized as follows: Cr, Al, and Hf are used as surface protection elements, Mo, W, and Ta are used in solid solution strengthening, and Re and Ru are used for high creep strength6. The strength of these single crystal alloys is mainly a function of the size and the percentage of precipitates. Experimentally, it has been determined that the peak creep strength is achieved with a volume fraction of of 60%65%7. Much of the behavior of these alloys can be explained on the basis that high volume fraction alloys deformation occurring by shearing of the precipitates. The high volume fraction of precipitates precludes dislocation bypass at low and intermediate temperatures forcing precipitate shearing. However, the energies resisting dislocation shearing of the precipitates are those required to form a local reversal of Al-Ni order or antiphase boundary in stacking fault of the Ni3 Al superlattice. The energies associated with the anti-phase boundary in the superlattice stacking faults determine the strength, fatigue, and fracture characteristics of these alloys8. This is also evident by the increase in yield strength at moderate to high temperatures before a monotonic decrease in yield strength. As a result, the excellent high-temperature creep and fatigue resistance of the superalloys is a result of a combination of solid-solution strengthening, absence of deleterious grain boundaries, and high volume fraction of precipitates that act as barriers to dislocation motion. It should be pointedout, however, that fatigue crack initiation also depends on the microscopic
defects, which can be categorized as intrinsic defects and deviant material defects. Intrinsic defects include carbides, undissolved eutectic pools, and associated micro-porosity, concentrated in interdendritic regions of the casting. Even though the intrinsic defects are normal features of the alloy microstructure, deviant defects, such as crystallographic defects related to low and high angle grain boundaries, freckles and silver grains, and porosity, are considered rejectable defects in the material quality that exceeds specified limits. Modern single crystal superalloys are face-centered single crystal superalloys with cubic symmetry. As a result, there are three independent mechanical properties that characterize the material behavior: the modulus of elasticity, E, the shear modulus, G, and the Poisson ratio, v9. The mechanical and thermal-physical properties are also a function of temperature and change with crystal orientation leading to anisotropic behavior of the material. After the casting process, a series of standard heat treatments, such as, solution annealing, coating heat treatment, and precipitation heat treatment, are used to optimize microstructures of the material10. In parallel to base material development, coatings were also developed to protect the base material from corrosion and oxidation attack11. These coatings were designated as metallic bond coatings characterized as diffusion and overlay type coatings. The substantive difference between the overlay and diffusion type coatings relies on the how the coating constituents are supplied. For diffusion type coatings, the major constituents are supplied by the base metal; whereas, for overlay coatings, the constituents are supplied by external coating sources. The advantage of overlay coatings is that more varied corrosion and oxidation resistant compositions can be applied with increased bonding thickness. The surface roughness of bond coatings allow for the deposition of ceramic type thermal barrier coatings. The addition of these coating systems onto single crystal superalloy substrate has further relaxed the temperature limitations associated with the turbine inlet conditions. As the material development of coatings and base material reach a culminating point, further increases in turbine inlet temperatures can only come from advanced airfoil cooling. Thus, cooling and material technologies have permitted designs with turbine inlet conditions in excess of 3100°F (1700°C) with interface bond-to-metal temperatures topping at 2050°F (1121°C) for durability requirements of oxidation, creep life, and fatigue cycles12. Two of the most relevant parameters for measuring and assessing cooling performance of turbine airfoils are the cooling effectiveness parameter and heat load parameter. By definition, this parameter is a dimensionless temperature ratio of gas-to-metal temperature difference over the gas-to-coolant temperature difference13:
(1)
Clearly, if the cooling effectiveness is non-existent, or zero, there is no cooling effect; whereas, if the cooling effectiveness is equal to unity, the airfoil metal and coolant temperature are the same. These two extreme values of either zero or unity are considered as outer limits for cooling effectiveness parameter. In general, the cooling effectiveness lies in-between these two limits and characterizes the performance of the cooling circuit inside the turbine airfoil. The other parameter denoted as the heat load parameter is defined as the ratio of internal heat to the external heat fluxes as follows14: (2)
Figure 1 illustrates cooling effectiveness for typical cooling configurations as a function of heat load parameter simplified here just by the variable in cooling flow15. The cooling effectiveness is naturally a function of many design variables, that is, the cooling configuration, and the coolant ejection requirements. It also a strong function of the amount of cooling medium used, usually measured as percentage of the mainstream gas. As illustrated in figure 1, the cooling effectiveness increases rapidly with small amounts of coolant. Then, the cooling effectiveness increases monotonically at a lower rate. This implies that for demanding thermal applications, where the thermal load parameter is high, large amounts of coolant may be required. For state-of-the-art engines, turbine cooling air and leakage may be as high as 25-30% of engine mainstream flow. In terms of efficiency, and since the cooling air is drawn from the compressor, which is driven by the turbine, it represents a direct loss of efficiency. In general, a very approximate rule-of-thumb of 1% cooling air may represent a loss of a fraction of that percentage in specific fuel consumption. This leads to the obvious conclusion that turbine cooling needs to be minimized. Clearly, this is not the only loss mechanism in the engine. Other losses may include mixing and aerodynamic losses, such as profile drag, skin-friction, gas diffusion, secondary flows, tip clearance, boundary-layer separation, shocks, losses due to off-design airfoil incidence angles, trailing edge vortex shedding, and blockage losses16. All these losses of engine cycle and turbine efficiencies have to be set against the gains in running the engine at higher turbine inlet temperatures for maintaining required output. Therefore, it is always necessary to reduce and optimize the cooling air requirements for a gas turbine engine design. In figure 1, the simplest cooling configuration is characterized by the lowest cooling effectiveness. For instance, the radial cooling holes in the middle of the airfoil cross-section will not permit air to eject from the airfoil walls into the gas mainstream, avoiding film cooling and corresponding ejection mixing losses. However, forced convection through the radial holes may not be sufficient for high
390
Frank J. Cunha thermal load applications, and even at low thermal loads, this simple cooling arrangement is likely to induce high temperature gradients between the airfoil surface and the cooling hole locations. An improved cooling arrangement over the radial cooling hole arrangement is the multi-pass serpentine cooling configuration. In serpentine cooling arrangements, coolant enters the airfoil through the blade root inlet; passes though multiple circuits, cooling the mid-body of the airfoil before ejecting out of the airfoil through main body film holes or trailing edge slots. The leading edge, with high thermal loads, may be cooled with dedicated cooling circuits. If the leading edge and the mid-body circuits are combined and coupled together, one can refer to these arrangements as cold-bridge or warmbridge designs depending on the feed source. If the feed coolant comes from the mid-body with freshly supplied coolant, we have a cold-bridge cooling design; whereas, if Fig. 1. Cooling effectiveness illustration for different modes of cooling the feed comes from the mid-body with warmed coolant, naturally after heat pick-up from the mid-body serpentine Source: See note 1. arrangement, we have a warm-bridge cooling design. Since the serpentine cooling circuits have large wet perimeters in the turbine airfoil cross-section, it is possible to introduce trips on the airfoil internal wall to promote turbulence and enhance internal coolant heat pick-up. Many trip configurations have been designed throughout the years of cooling technology development. These include normal trips, skewed trips, and angled “chevron” trips with different orientations relative to the flow, different trip heights, different height-to-pitch ratios, different relative positioning with respect to each other, either in a staggered or in-line arrangements. These different arrangements lead to different heat transfer characteristics enhancing heat pick-up by the coolant flowing in the serpentine passages tailored to minimize the effects of local external heat loads. Further cooling improvements are obtained by the introduction of peripherally cooled airfoil designs. This is done to maintain the greatest thermal gradient within a thin outer layer of the airfoil wall. In addition, the coolant can be ejected at many points around the surface of the airfoil. This concept leads to an increase of convective cooling efficiency by means of peripheral cooling circuits. This new parameter can also be regarded as a dimensionless temperature ratio of the difference between the exit coolant to inlet coolant temperatures relative to the difference of metal to coolant inlet temperatures17: (3) If this ratio is non-existent, or zero, it is implied that exit coolant temperature and inlet temperature are the same, and thus, no heat pick-up in the circuit. If the ratio is unity, then the exit coolant temperature and the metal temperature are the same. In this case, it is said that the cooling circuit is 100% efficient convectively. In general, serpentine cooling may be in the order of 15-30% efficient. However, dedicated airfoil peripheral cooling may be 30-60% efficient. This is about 2X the convective efficiency values of typical serpentine cooling. In addition to the improved convective characteristics, film cooling can be used in many points of the airfoil. The film effectiveness is also a dimensionless parameter defined as 18: (4) The overall cooling effectiveness, being a function of both the convective and film cooling effectiveness, can be optimized for any cooling arrangement. If the film exit shape is optimized to cover the airfoil as much as practical, one can obtain overall cooling effectiveness in excess of 75% with film coverage and convective efficiency in excess of 50%. Simply stated, the peripheral airfoil circuits will pick-up heat by convection in a very effective manner and then eject from film openings with high coverage designs to further reduce the thermal load to the part. It should be noted that the thermal load can be reduced externally by the effect of thermal barrier coating. This is a ceramic coating that is applied on top of the metallic bond coating. The net result is the ability to increase the inlet turbine temperature to values in excess of 3100°F (1700°C) while maintaining interface bond-to-metal temperatures at 2050°F (1121°C) for desired life requirements of oxidation, creep life, and thermal-mechanical fatigue cycles19. In the limit, combination of peripheral convective cooling and film may lead to airfoil transpiration cooling which would occur in micro-porous surface to form a continuous film blanketing of the airfoil surface as illustrated in figure 1.
391
4.4 Heat Transfer Analysis Regarding the different cooling schemes, peripheral cooling and conventional serpentine cooling arrangements may include several internal design features for augmenting internal heat transfer. These internal features may be small posts connected to both sides of the coolant passage, as pedestals, or may project to only about half of the coolant passage, as pin-fins20. Fundamentally, both features act in a similar manner, by turbulating coolant flow field and by exposing more surface area to the coolant. These synergistic cooling effects lead to higher internal heat transfer coefficients, which can be regarded as a measure of heat pick-up by the coolant. It should also be noted that there are conduction paths integrated directly with the walls of the airfoil for further cooling. The disadvantage of such augmentation devices is that they increase the coolant flow blockage, which in turn increases the coolant pressure drop. Since the coolant supply pressure may be limited, attention must be given to these blockage effects so as to assure sufficient pressure at the film openings and overcome external gas pressure. The internal-to-external pressure ratio is usually referred to as the back-flow-margin with minimum values established for the leading edge and mid-body of the airfoil. On the other hand, if the internal coolant pressure is much higher than the external pressure at the film exits, then the film jets may actually penetrate through the external boundary layer setting out “blow-off” conditions. Blowing the coolant out nearly normal to the airfoil surface with high velocity is likely to lift the coolant right off the blade and allow the hot mainstream gas to move below the coolant jets making contact with the airfoil surface. This is an adverse film effect that needs to be prevented. To prevent blow-off conditions, the geometrical attributes of film openings need to reduce the ejection velocity by introducing cooling hole shapes, angles, tapering, or diffusion zones while maintaining a high degree of cooling hole filling21. Externally in the main flow field, a mismatch between the gas mainstream and the coolant jet velocities will give rise to mixing losses. In this way, film ejection may trip or further energize the external boundary layer; thus, increasing the external heat transfer coefficient or thermal load to the part. Figure 2 and 3 illustrates typical airfoil cooling schemes. Typically, for the first vane, the coolant is fed into the casing plenum chamber, and from there into the leading edge impingement baffle. From the leading edge baffle, the coolant impinges on the leading edge target surface and is ejected into the external gas-path through several close-spaced rows of holes known generally as showerhead holes. For this component, the highest temperature occurs at the leading edge where the stagnation point is located22. The balance of the coolant flow in this vane will pass through a small gap between the baffle and the internal wall of the airfoil leading to other film holes or slots in the main body or trailing edge. Note that internal cooling is done by convective and impingement cooling. At the trailing edge pedestals may be placed to enhance internal cooling, heat pick-up, and decrease the metal temperature23. The first blade shown in figure 3 is only an example of a wide variety of serpentine cooling configurations. Stators and rotor blades of downstream stages will often have multi-pass cooling of the type shown in figure 2. The coolant flow Fig. 2. Typical airfoil cooling schemes for a two-stage high is led through a series of serpentine passages and makes several pressure turbine illustrating different modes of cooling. passes along the span before finally ejecting at the trailing edge. The purpose of turbine cooling is solely justified Source: See Note 1. by the need to have turbine components withstand adverse environments while maintaining life targets. Lifing and durability of airfoils require a synthesis of mission cycle, range of operating conditions and time spent at each condition, including the number of transients between conditions. Overall cooling effectiveness is a function of the design, which is reflected by two other relevant parameters: the convective efficiency and the film effectiveness. The overall cooling effectiveness should balance the external heat load to the part. In turn, all cooling parameters need to be balanced with their effects of aerodynamic and cycle performance. For blades, the mechanical load, due to centrifugal effects, need to be determined in conjunction with the blade bulk metal temperature distributions to evaluate its creep capability. The next sections will describe turbine airfoil requirements for the purposes of converting energy to produce shaft power output, or specific core power in a manner where the figure of merit associated with thrust specific fuel consumption is optimized with the thrust-to-weight ratio in the case of aircraft engine applications. In this context, external Fig. 3. Typical airfoil cooling schemes for high-pressure and internal heat transfer details are explicitly listed and turbine blade. discussed. A section on durability is presented to describe most of the design problems and possible solutions. Conclusions are Source: See Note 30. then derived based on current assessment methods for airfoil cooling technology advancements.
392
Frank J. Cunha 4.4-2 Heat Transfer Requirements The primary components of gas turbine engines used in aircraft operations comprise the inlet duct, fan, compressor(s), combustors, turbine(s), and nozzle24. Many engines have two turbines: the high and the low-pressure turbine. The high-pressure turbine is required to drive the high-pressure compressor. Together with the combustor, these high-pressure, high-temperature, and high rotatingspeed components make up the gas generator portion of the engine. The high-pressure turbine is the most challenging component of the engine design. Most of the attention given in this section covers the turbine limits, and the cooling required to withstand extremely hot, corrosive, and unsteady environments. The low-pressure turbine drives the fan and the low-pressure compressor. The low-pressure turbine is subjected to much cooler gas temperatures. In power applications, another turbine may be added to drive a rotor, a generator, a pump, or a gas compressor. In land-based gas turbine applications, the compressor work is extracted from the first rotor stages, while the other stages produce useful shaft work transformed into electrical power by the electrical generator connected in tandem. In traditional power generating plants, there are three fundamental elements: (1) the heat source, (2) the heat utilizer, and (3) the waste heat reservoir25. To produce useful work heat is supplied by a working fluid from the heat source. The utilizer is required to convert portion of the heat supplied by the working fluid into useful power. Since not all heat can be converted to useful power, a waste reservoir is required to dispose of the remainder of heat. These are the elements of a modern steam plant which can be divided into two major sub-parts: One half consists of the boiler or nuclear reactor as the heat source with its auxiliaries; and the other, the turbine cycle, consisting of turbine generator, condenser, pumps and feed-water heaters. The major components in the turbine cycle are the heat utilizer and the refrigerator or water heat reservoir. In general steam comes from the boiler drum or nuclear reactor, passes through the turbine, which produces mechanical power on the turbine shaft to drive a generator of electrical energy for distribution. Thermodynamic performance is reflected in the overall cycle efficiency, often times designated as the plant heat balance, which is a measure of thermal energy conversion to electrical generation output in BTU per KW-hr. In land-based power-producing applications, recent improvements in cycle heat balance have being achieved by combining the operations of steam and gas turbines together in the same plant26. In these combined cycle plants, the elevated exhaust gas temperature is used in heat recovery steam-generator to create steam before expanding through a train of steam turbines. Intermediate pressure steam is then used as cooling medium in the gas turbine topping cycle. This is done in a closed loop circuit with reheated steam from the gas turbine being returned to the steam turbine cycle for further expansion. In this layout, all gas turbine stationary and rotating airfoils were designed with steam cooling in a closed-loop system. The main reason for this closed cooling system is to avoid losses associated with coolant ejection onto the main gas stream from airfoil locations. Combined power plants operate at thermodynamic cycle efficiencies as high as 60 percent or a heat balance as low as 5700 BTU/KW-hr. One such unit is the GE H-class machine operating at just under 60 percent at Baglan Bay power plant in Wales, UK27. This is remarkable, since current advanced nuclear generating plants operate at an overall cycle efficiency of 32.5 percent or at a heat balance of about 10,500 BTU/KW-hr.
4.4-3 Gas Heat Transfer Regardless of the type of power plant cycle, it is common knowledge that a design of the turbine engine is very challenging, and it requires a mix of technical disciplines: cycle analysis, aerodynamic, heat transfer, combustion, fluid dynamics, and turbine mechanical design and manufacturing. The basic cycle parameters define the engine capability for all operating modes. The overall design is selected in concert with performance and mechanical design constraints. For instance, in gas turbine design, the average heat transfer design point parameters are dependent upon a set of performance parameters defined at different points in the gas turbine thermodynamic cycle, such as inlet conditions, at cycle point no 2, engine core flow, at cycle point no. 2.5, compressor discharge conditions of temperature and pressure, at cycle point no. 3; turbine inlet conditions of temperature and pressure, at cycle point no. 4; high pressure rotor inlet conditions of temperature and pressure, at cycle point no. 4.1; low pressure inlet conditions of temperature and pressure, at cycle point no. 4.5; and high and low rotor speeds for different ambient conditions.
4.4-4 Gas Temperature Transverse Quality The gas turbine combustor exit gas temperatures are further refined with pattern and profile factors consistent with the burner characteristics28. The combustion process that takes place in the burner requires that incoming air be separated into primary and secondary streams for adequate combustion. The primary air is mixed with the fuel at nearly stoichiometric proportions at the center of the burner, and the combustion produces extremely hot core gaseous products. Secondary air is injected from the sides of the combustion chamber and separates the very hot gases from the structure. The secondary air dilutes and cools the products of combustion to acceptable levels before leaving the burner. Because of the size limitations, the burner may not be sufficiently large to give gases time to mix homogeneously, and as a result, the turbine may experience gas temperature variations which give rise to pattern and profile factors29:
393
4.4 Heat Transfer Analysis Pattern Factor =
(5)
Profile Factor =
(6)
. For the first vane, the peak temperature is obtained from pattern factor and this information is used for thermal-mechanical design. It should be noted current combustor exit temperatures are in excess of 3000°F (1650°C) with expected growth towards 3800°F (2100°C). As the flow passes through the turbine components, the temperature profile retains the same basic form of the high temperature in the center region and lower temperature at the outside region, but it is also affected by the processes that take place during turbine expansion, such as mixing, tip leakage, and development of secondary flows. Figure 4 illustrates the thermal load distribution for the high-pressure turbine components30. The primary gas temperature of interest in blade design is the total temperature relative to the blade31:
T gas, relative = Tgas
V2 W2 − +r 2Jg c c p 2Jg c c p
.
(7)
At the inlet to the rotor, the relative velocity W is smaller than the absolute velocity V, and the relative gas temperature should reflect the 32 . effect of Mach number through the recovery factor,
Fig. 4. Illustration of thermal load distribution for the high-pressure turbine components. Source: See Note 30.
The effect of secondary flows on streamline migration and the effect of film flow should be reflected in the stage Mach number distributions. In addition to these considerations, there are margins or adders that lead to increments imposed on the gas temperature values. These are used to account for the following design tolerances: engine-to-engine variation, deterioration, overshoot, and production scattered tolerance to sustain thrust guarantee33. It should be noted that film, or external surface to gas, thermophysical properties required to determine heat transfer coefficients, for instance, should be calculated using Eckert’s reference temperature34. Other dimensionless ratios used to characterize stage local conditions, such as relative-to-total temperatures and pressures can be assumed nearly constant for operating conditions at idle and above in high-pressure turbine design35.
394
Frank J. Cunha 4.4-5 Airfoil Thermal Load Externally to the turbine airfoil, gas static pressure and gas stream velocity distributions are used for determining the external heat transfer coefficients. The external heat transfer depends on the development of the boundary layers on the airfoil. The external boundary layer is subject to several factors, which make this calculation difficult. These factors include, pressure and temperature gradients, turbulence level, surface curvature, surface roughness, laminar-to-turbulent transition, and flow unsteadiness. Separation bubbles near the airfoil leading edge could also have a major influence on the development of the boundary layer. These could be caused by rapidly changing the curvature or during off-design conditions when incidence angles may be high. If, however, pressure gradient is favorable and strong enough, a turbulent boundary may re-laminarize with subsequent reduction in the heat load. It should be noted that the presence of film cooling will invariably trip or energize the boundary layer. Standard relations for flow over flat plate with suitable modifying factors to account for these effects, can be used to calculate external heat transfer rate to the airfoil as follows36:
(8)
. The empirical correlation factor is obtained by matching test data.
Despite their simplicity, the flat plate correlations provide a good estimate of heat transfer. Figure 5(a) and 5(b) are examples for the pressure and suction sides37. The Nusselt number calculated are based on measurements of static pressure, from which the distribution of the Reynolds number is obtained. On the downstream region of the suction side, with fully turbulent flow, the heat transfer coefficient is predicted well. On the pressure side, the turbulent flat plate correlation predicts heat transfer well on downstream region towards the trailing edge; but on the upstream region, the heat transfer prediction is low. In this upstream region the flow is very unsteady causing very large heat transfer coefficients. Fortunately, heat transfer coefficients are predicted using boundary layer codes, such as those of reference38, which account for curvature effects, transition and surface roughness on the airfoil external wall.
(a)
(b)
Fig. 5. Flat plate correlations estimate of heat transfer on the pressure (a) and suction (b) sides of the high-pressure turbine airfoil in terms of Nusselt numbers versus Reynolds numbers. Source: See Note 37.
Besides, main-body airfoil thermal load calculations, airfoil tip and blade platform thermal loads need to be determined as well. Tip heat transfer is difficult to determine due to the complexities associated with leakage flows over a relatively small gap between rotating blades and stationary shrouds. However, references listed in note 39 can be consulted for determining the external heat transfer coefficients of different tip configurations39. Similarly, platform heat transfer is also complex due to the secondary flow and leakage effects associated with inter-segment gaps. In this regard, note 40 can be consulted for determining the external heat transfer coefficients associated with these difficult situations40.
395
4.4 Heat Transfer Analysis 4.4-6 Leading Edge Thermal Load At the leading edge, effects related to unsteadiness due to the row of upstream airfoils lead to high heat transfer coefficients and empirical correction factors are usually applied as follows41: (9)
where all variables are defined in the nomenclature with
0 <θ < 80.
Turbine airfoils are optimized for aerodynamic performance and heat load to improve airfoil life. The ideal heat transfer objectives, during the conceptual airfoil design, include the following considerations: (1) blunt leading edges to minimize heat transfer coefficients at the airfoil stagnation points, (2) minimize blade count to minimize cooling surface. (3) minimal pressure diffusion to minimize separation and high external heat transfer coefficients, (4) increase mid-body airfoil thickness to reduce losses in internal cooling serpentine cooling arrangements, (5) avoid low trailing edge wedge angles that lead to large distances from the trailing edge and the pressure side coolant ejection (film) bleeds. Some of these considerations may be in conflict with aerodynamic objectives, and, in general, a compromise is reached between cooling and aerodynamic design groups.
4.4-7 Coolant Heat Transfer The overall cooling effectiveness, , as a dimensionless metal temperature is a function of the other cooling parameters, such as the convective efficiency, , and the film effectiveness, , for a given heat load, , to the part . The relationship between these cooling parameters can be expressed in a simple relationship, derived from one-dimensional heat transfer analysis as follows42: (10) If a metal temperature at a point in the airfoils is limited to a required value for a specified thermal load, ; then, the overall cooling effectiveness, , as a measure of metal temperature, is set for the design, and the design process begins with a search of the required internal convective efficiency and film effectiveness.
4.4-8 Film Cooling Film cooling takes the cooling medium and discharges it in a carefully designed pattern of openings in the blade surface to form a protective film over the airfoil surface. This film blanket reduces the external heat load to the airfoil by reducing the driving temperature from the local gas temperature to the film temperature, as shown in Figure 6. The film cooling effectiveness is a function of injection geometry on the airfoil including the hole diameter, dh , hole spacing, x n , effective slot height, , surface distance from the film row, x, inclination of the film hole axis to the surface in terms of in-plane and out-of-plane angles, blowing parameter, , density ratio, the ratio of x/(Ms), and film coverage, Cov43. Film coverage is a fraction of the length along the row of film holes which is covered by the footprint of the breakout of the holes and is usually given by the ratio of film hole breakout, b, to the film hole pitch, p, or Cov=b/p. Film effectiveness can then be expressed analytically as:
(11)
where, the empirical constants, C1 and C2 , are determined by fitting the test data for different film cooling configurations. In general, the exponent C2 is about the same order of magnitude for the pressure and suction sides of the airfoil. However, coefficient C1 is about an order of magnitude different between the pressure and suction side of the airfoil with the pressure side film decaying much faster on the pressure side of the airfoil. It should be noted that when the empirical constants of equation (11) are matched with test data, most of the density ratio for the data reported in the literature is about unity, and this data needs adjustment for typical engine environments.
396
Frank J. Cunha
Fig. 6. Film blanket reduces the external heat load to the airfoil by reducing the driving temperature from the local gas temperature to the film temperature.
The use of diffuser shaped holes for film cooling reduce the penalties associated with film of blowing ratios, M. Shaped holes also provide increased coverage, as shown in figure 744. In the past, shaped holes were almost eliminated in favor laser drilled round holes, but it is now apparent that they should be used due to superior cooling advantage over laser drilled holes. For a given coolant flow, laterally averaged film cooling effectiveness from staggered double row of holes is usually higher than from a single row of holes. However, for the same coolant flow, the double row hole diameters may be expensive to make. The injection of coolant from succeeding rows of film holes results in a cooler film temperature than if only one row of holes were blowing. Film temperatures superposition produced by coolant injection from multiple rows of film holes can be evaluated by successive heat balance stations. This procedure is repeated for blowing from all N successive rows of holes upstream of the heat balance station as follows45:
Fig. 7. The use of diffuser shaped holes for film cooling provide increased film coverage in terms of film effectiveness versus hole locations.
(12) Computer programs are developed to superimpose the film cooling effectiveness distribution downstream from several rows of film holes having individual levels of coolant injection, temperature, and flow rate to provide the effective distribution of cumulative film cooling effectiveness; and, hence, film temperature along the airfoil surfaces.
4.4-9 Trip Strips or Turbulence Promoters for Cooling Passages If the film effectiveness, , is known for a particular design, due to the number and arrangement of film cooling holes; then the only unknown in equation (10) is the convective efficiency, . This convective efficiency has a fundamental solution, which can be given, in simple terms, by the relation: (13) In this expression, the convective efficiency, , is related to the internal heat transfer coefficient, wetted surface area, and coolant flow rate. Expression (13) allows for design features to be evaluated in terms of this performance parameter. That is, for a required convective efficiency, , limited by the required cooling effectiveness, ; then a required product of internal heat transfer coefficient with the wetted surface area, hc A c can be evaluated as a fundamental requirement for the cooling design. In a general, it is the goal of the design to obtain the largest hc A c product, for the least amount for coolant flow rate, m , possible. Insofar as the airfoil internal cooling features are concerned, there may be a large number of configurations that can be explored in the design space as noted in the reference provided by Han, et al.46. These would include configurations with internal cooling features, such as pedestals or turbulating trip strips just to mention just a few.
397
4.4 Heat Transfer Analysis Patankar and Spalding observed that in the near the internal wall of the airfoil, the convection associated with the primary flow direction can be neglected, the so-called Couette flow assumption47. However, for the specific case of predicting the boundary layer behavior under the influence of rotation, this assumption may not be applicable. The boundary layer equations for momentum and energy can be written as: (14) (15) The first term of the right-hand-side of equation (14) represents the shear stress at the wall. The second term represents the crosswise convective mass flux, where the velocity is taken as the mean velocity component normal to the wall, in the near wall region. This crosswise component exists due to rotational Coriolis influences on the boundary layer. The third term represents the axial pressure gradient on the boundary layer. The variable, y, is taken as the cross-stream coordinate away from the wall. All other symbols are provided in the nomenclature. The relations given by equations (14) and (15) are written in dimensionless form by using the traditional groups of + + v , y , and p + defined in the nomenclature. When these groups are substituted into (14) and (15), the following new relationships are obtained; (16) (17) Using the diffusion laws, the effective (laminar and turbulent) shear stress and heat flux are written in dimensionless form as
(18) and (19) where
is the dynamic viscosity,
is the effective viscosity, and
is the effective Prandtl number.
Two ordinary differential equations describing momentum and energy transport in the boundary layer can be obtaining by equations (16) and (17) with equation (18) and (19), respectively. If the mixing-length hypothesis is used for the viscosity ratio as , as described by Crawford and Kays48, the resulting set of equations become
(20)
(21) + where the effective sub-layer A =26 was used proposed by White49. The boundary conditions are: Equations (20) and (21) can be solved numerically for given Prandtl numbers to yield the desired profiles for the velocity and temperature near the wall. Typical results are succinctly summarized as the law of the wall correlations:
(22) and
(23)
where, E and P can be regarded as constant factors for a smooth wall condition and given Prandtl numbers. These are used to calculate the Stanton number at the wall, which is given by the relation:
398
Frank J. Cunha
(24) Note that the Nusselt number is related to the Stanton number by the relation; Nu =(Re) (Pr) (St). Solutions (22) through (24) are applied for smooth channels; however, these correlations have been modified to account for non-smooth roughened walls. Typical representations of dimensionless velocity and temperatures have catalogued by Han for a series of configurations50. These relationships can assume the form (25) and (26) The functions R and G are the momentum and thermal roughness functions determined experimentally. All other variables are described in the nomenclature. In blade cooling passages, repeated roughness elements in the form of regularly spaced ribs are used to increase heat transfer rates. As a by-product of the desired increases in heat transfer rates, the pressure losses through the channel also increase. It is usually necessary to determine the best combination of heat transfer rate increases with the lowest possible pressure losses. As an example, several trip strips configurations in a cooling channel are shown in figure 8. These are described as follows: (a) (b) (c) (d)
Parallel 90 degree ribs Parallel inclined ribs Parallel v-shaped (Chevron) ribs Criss-crossed inclined ribs
Fig. 8. Several trip strips configurations in cooling channels.
Fig. 9. Illustrates the geometrical parameters for the trip-strip configurations.
The pertinent parameters to consider include the dimensionless heat transfer coefficient, the Stanton Number, the friction factor, and the thermal performance index. Figure 9 illustrates the geometrical parameters for the trip-strip configurations. The ribbed channel pressure losses were originally modeled using the “law of the wall” similarity by Nikuradse for sand grain roughness51. The correlating parameter became known as the roughness function, R, which, for rectangular channels, is defined as:
(27)
399
4.4 Heat Transfer Analysis where: e is the rib height, f is friction factor, and d h is the channel hydraulic diameter. The friction factors and the roughness function for the rib configurations (a)-(d) of figure 8 are given in Table I. Table I. Friction factor and momentum roughness function for the configurations of Figure 8.
Friction factor
Roughness function
Sources: (a) See Note 52. (b) Han, J.C., J.S. Park, and C.K. Lei, “Heat transfer enhancements in channels with turbulance promoters”, Trans. ASME J. Eng. Gas Turbine and Power, Vol. 107, 628-635,1985. (c) Han, J.C., Y.M. Zhang and V.P. Lee, “The influence of surface heat flux ratio on heat transfer augmentation in square channels with parallel, crossed and V-shaped angled ribs”, ASME Paper 91-GT-3, 1991. (d) Lau, S.C., R.T. Kukreja and R.D. McMillin, “Effects of V-shaped rib arrays on turbulent heat transfer and friction of fully developed flow in square channel”, Int. J. Heat Mass Transfer 34, 7, 1605-1616, 1991.
400
Frank J. Cunha Using the Reynolds analogy between momentum and heat transfer, Dipprey and Sabersky developed a thermal analog to the roughness function to correlate heat transfer data in sand roughened channels52. Webb et al., extended the analysis to rib roughened channels53. The thermal roughness function was given by:
f −1 G = R + 2St f 2
The corresponding heat transfer correlations for geometries (a) – (d) of figure 8 are given in Table II.
(28)
Table II. Stanton Number and thermal roughness function for the configurations of Figure 8
Sources:
(a) See Note 52. (b) Han, J.C., J.S. Park, and C.K. Lei, “Heat transfer enhancements in channels with turbulance promoters”, Trans. ASME J. Eng. Gas Turbine and Power, Vol. 107, 628-635,1985. (c) Han, J.C., Y.M. Zhang and V.P. Lee, “The influence of surface heat flux ratio on heat transfer augmentation in square channels with parallel, crossed and V-shaped angled ribs”, ASME Paper 91-GT-3, 1991. (d) Lau, S.C., R.T. Kukreja and R.D. McMillin, “Effects of V-shaped rib arrays on turbulent heat transfer and friction of fully developed flow in square channel”, Int. J. Heat Mass Transfer 34, 7, 1605-1616, 1991.
401
The motivation for placing the ribs in a channel is to increase the rate of heat transfer from the channel walls to the bulk fluid. This is accomplished by two distinct mechanisms. First, the placement of ribs act as turbulators to break-up the near wall flow increasing the turbulence level and enhancing the exchange of fluid in the near wall region with the core flow by action turbulent diffusion. While this increases the heat transfer, it also increases the wall friction resulting in higher pressure losses. A second heat transfer mechanism is active if the ribs are inclined to the core flow direction. In this case, the ribs induce secondary flows in the core flow, which can circulate fluid from the middle of the channel towards the walls thus increasing the heat transfer along these walls. The secondary flows have a smaller effect on the pressure losses, thus they are more beneficial of increasing the thermal performance of the cooling channel. The flow in a ribbed channel can be divided conceptually in a main core flow and near wall flow. The flow structures that have been observed in the vicinity of the ribbed wall are shown in Figure 1054. This structure varies primarily upon the geometrical ratio of p/e. As an example, when p/e>7, three different separated flows regions are observed: (1) In front of the rib, a separated
4.4 Heat Transfer Analysis recirculation zone exists; (2) on top of each rib a separation bubble forms due to upstream facing step; (3) following the back side of the rib, a separated recirculation region exists which can extend up to 7 rib heights downstream before the flow reattaches. From this point the boundary layer re-develops until the next separated region is encountered upstream of the rib. When p/e<7, the two separated regions which lie between the pair of ribs combine. On top of each rib a separation bubble exists. When p/e=7, the boundary layer between the separation zones and preceding the up and downstream ribs does not have a chance to develop. Separation occurs on the top of each rib. When p/e=10, the separation bubble at the top of the rib may not re-attach on the top surface of the rib but it may extend past the back of the rib and combine with the separation region following the rib. The re-attachment occurs between ribs. This is believed to be the optimum configuration. However, the effects of the Reynolds number may determine the validity of these results. The core flows develops rapidly in a ribbed channel compared to a smooth one. Friction factors become constant after the 3rd rib or 4 diameters downstream. The Nusselt number shows similar behavior. The presence of the ribs increase turbulence and induce secondary flows that tend to move fluid from the main core towards certain walls and away from the others. Similar core fluid flow and heat transfer enhancements are found in rotating channels caused by Coriolis as illustrated in figure 11. This is particularly true for channels with aspect ratio of unity. Sources listed in note 55 can Fig. 10. Observed flow structures in the vicinity of the ribbed wall. be consulted for determining the effects of rotation on internal heat Source: See Note 53. transfer coefficients where rotational and buoyancy numbers are the fundamental dimensionless parameters55. The thermal performance can be defined by the parameter of St/f 1/3 as given by Webb and Eckert56. It relates the heat transfer to friction factors or pressure loss in the channel.
4.4-10 Impingement Cooling for CrossOver Holes and Inserts The impingement configurations have been used extensively in the gas turbine industry. In general, there are two locations in airfoil design where impingement cooled has been used. The first location is used for leading edge cooling. Dedicated cooling air is allowed to pass through rib cross-over openings leading to jet impingement on the inner wall of the airfoil leading edge57. After impingement cooling the air flow passes through a series of film holes forming a shower-head arrangement through the airfoil the leading edge. Internal jet impingement, convective cooling through the film holes, and film cooling become very effective means to cool this section of the airfoil with very high heat loads. The other location is for trailing edge cooling. In this case, the cooling air is allowed to pass through a series of rib cross-over openings58. Jet impingement occurs after each cross-over hole. The air impinges on each subsequent ribs and surrounding walls before discharging at the trailing edge exit slots. Many references can be consulted for obtaining impingement heat transfer coefficients for these two configurations59. For high-pressure vanes, inserts or baffles can be used to cool the airfoil with impingement cooling. There are several references that can be consulted in this regard60. Each impingement correlation for this type cooling will be dependent on the cross flow degradation from upstream impingement jets. A useful correlation has been provided by Floerschuetz et al as follows61:
Fig. 11. Core fluid flow and heat transfer enhancements are found in rotating channels caused by the rotating vortices inside the cooling passages. Source: See Note 30.
402
Frank J. Cunha
m
(29) where
A M B N
C
nx
ny
nz
1.180 0.612 0.437 0.092
-0.944 0.059 -0.095 -0.005
-0.642 0.032 -0.219 0.599
0.169 -0.022 0.275 1.040
The different variables are described in the nomenclature. The validity of correlation is the defined by the following parameter range:
It should be noted that for the calculation of the Reynolds number in correlation (29) needs to be determined for the flow passing through the insert impingement holes. Flow through these openings is a result of pressure differences which can be obtained from an overall flow network analysis.
4.4-11 Pin-Fin or Pedestals for Trailing Edge Cooling In trailing edge designs, typically two rows of cross-over holes are used for cooling. The flow acceleration is high through these cross-over impingement openings. The coolant flow Mach number profiles follows that of the coolant static pressure profile in that it assumes almost stepwise profiles as the cooling flow crosses through these openings. The step-wise profiles are undesirable as they lead to relative high peaks in internal heat transfer coefficients at the walls of the blade as illustrated in figure 12. In other words, there are regions in the airfoil trailing edge walls, which attain relative lower metal temperatures due to high internal heat transfer coefficients. Meanwhile, other areas attain relatively higher metal temperatures due to lower internal heat transfer coefficients. These metal temperature differences can lead to high thermal strains, which in conjunction with transient thermal response during take-off or power loading can, in turn, lead to undesirable thermal-mechanical fatigue damage. It is, therefore, desirable to use a trailing edge cooling design, which improves the internal profiles for Mach number, static pressure drop, and internal heat transfer coefficient distributions along the airfoil trailing edge. In that regard, internal features, such as pedestals provide design advantages. Optimum designs may include internal pedestals with many different cross-sectional areas, such as: circular, oval, racetrack, square, rectangular, diamond cross-sections, just to mention only a few. In the illustrative example that follows, attention is given to the most simplest form of an internal cooling feature: such as round pedestals.
403
Fig. 12. Localized heat transfer enhancement on different segments in trailing edge cavities. Source: See Note 58.
4.4 Heat Transfer Analysis The relevant cooling characteristics of round pedestals could easily be extended to other cross-sectional area pedestals for specific applications. In this context, for a staggered array of round pedestals the internal unobstructed trailing edge channel area will have a cross-sectional area, A=WH, where W is the width and H is the channel height. The corresponding perimeter becomes Per = 2 (W+H) , and the hydraulic diameter is given by d H = 4A/Per . The surface area becomes A duct = 2WL , where L is the length of the channel. The surface of the plate not being covered by the pedestals becomes , where N L and N T are the number of pedestals in the longitudinal and transverse directions of the trailing edge channel. It should be noted that the number of pedestals can be given by N T = W/S n and N L = L/S p , where S n and S p are the transverse and longitudinal pitches, respectively. With these geometrical parameters, the Reynolds number, associated with the pin diameter, is calculated as follows: , where Vmax = V(A/A min ) with . In this case, the coolant velocity, V , is the free stream inlet velocity. With the knowledge of the pin Reynolds number, the following relationships for the Nusselt number from table III can be obtained, following Zukauskas62. Table III. Pedestal Nusselt numbers [17].
For the area not affected by the pedestals, the plate Nusselt number is given by obtained from the relationship Ishida et al [65]:
.The pin fin efficiency is
63
(30) where all variables are described in the nomenclature. The overall internal heat transfer with pedestals is then based on the following two contributions: (1) the plate contribution, and (2) the pin contribution as follows:
(31) The corresponding heat transfer multiplier, HM, normalized by the Dittus-Boelter correlation64, by
, is then given
(32) The pressure drop calculations through a bank of pedestals is given by Ishida et al. and are presented here for completeness as follows 65 .
(33)
404
Frank J. Cunha The corresponding friction multiplier, FM , normalized with Blasius resistance formula,
, becomes
(34) Relationships provided by equations (32) and (34) can be used to design a trailing edge cooling channel with round pin-fins. These were provided here to represent a design procedure; however, many other empirical correlations can be obtained for other type of internal features as described by Han et al.66. The overall procedure for obtaining heat transfer and friction multipliers would follow the steps outlined here. In many instances, plots of FM/HM versus ReH are used to compare the performance of different configurations. One such plot is shown in figure 13, based on mass-transfer augmentation, from Chyu67. By mass-transfer to heat transfer analogy, the results of this plot are directly related to heat transfer assessment. Figure 13 allows for comparison and determination of which configuration would provide the best heat transfer for the least pressure drop. Plots similar to that of figure 13 are usually called performance index plots and become useful early in the design process.
(a)
(b)
Fig. 13. Pedestal trailing edge configuration and test results; (a) transparent view of blade showing pedestal bank at trailing edge (b) effect of Reynolds number on performance index (HM/FM) in straight pins with fillets. Source: See Note 20.
4.4-12 Bulk Temperatures for Cooling Passages To estimate airfoil cavity cooling flow, one needs to consider the inlet coolant flow rate, at the root of the blade passage, denoted here as mRoot, as well as the cooling flow rate at the several film holes, denoted as mfilm, and flow variation through the cavity. The linear function of cooling flow rate distribution from the root as flow exits the airfoil cavity leads to the following flow relationship68:
(35) with LH as the blade span radial height to the film hole location. The temperature of the coolant inside the cavity is obtained by considering an infinitesimal small control volume, dr, in the supply cavity. This consideration leads to the following energy balance69:
405
4.4 Heat Transfer Analysis
(36)
(37) where Q denotes the total heat transfer onto the corresponding section of radial span, LH. The coolant flow rate is then introduced in equation (36). The resulting expression is integrated from root to a radial section of the airfoil. This yields an expression for the coolant temperature increase in terms of heat transfer, and rotational pumping, as a function coolant flow rate. The result is:
4.4-13 Thermal-Mechanical Aspects of Durability Once the internal and external heat transfer loads have been assessed, the gas and coolant temperatures, together with the internal and external heat transfer coefficients, will form a set of boundary conditions for a heat conduction analysis to predict the metal temperatures and stresses in the airfoil. Usually, this analysis begins with the modeling of a 3D section of the blade as a solid model and then proceeds by using a finite element code, such as ANSYS70, for completing the thermal-stress analysis. An example is provided here for illustrative purposes. It was taken from the NASA report71 for the E3 HPT blade. The blade solid model and the corresponding finite element ANSYS model are shown in figure 14. As can be seen from the transparent solid model of figure 14, this blade is cooled by convection with cavity wall trip strips, not shown, and with a set of pedestals in the trailing edge. While external surfaces are locally film cooled from the leading edge with three rows of showerhead holes, there are no film cooling holes on either the pressure or suction surfaces of the airfoil. The trailing edge has a 3 centerline cooling flow ejection. Fig. 14. NASA E blade solid model and the corresponding finite element 3 The NASA E HPT blade is used as a vehicle ANSYS model. to demonstrate the thermal-mechanical analysis for a typical uncoated anisotropic turbine component. The Source: See Note 23. 3D model of figure 14, using ANSYS, is composed of 21567 elements, 13494 of which are 3D-10 node tetrahedral structural solid elements in the trailing edge, while 8073 are 3D-20 node structural solid elements in the rest of the airfoil. This hybrid finite element model was selected to improve the meshing of the pedestals in the trailing edge region of the airfoil. This modeling technique is common in durability assessment of turbine airfoils. The thermal boundary conditions set-up for a blade model is a very laborious process. On the external side, computational fluid dynamic codes can be used to set-up external pressure and Mach number distributions for free-stream conditions with specified turbulence intensity levels. Boundary layer programs are used to establish the external heat transfer coefficients based on expected wall roughness. Gas recovery temperatures and film cooling effectiveness, usually obtained from previous testing, are used to determine the adiabatic wall (film) temperatures. In the end, external heat transfer coefficients and film temperatures are specified as external boundary conditions for the model. Inside the airfoil, flow network models are used to determine internal pressure drops and Mach number distributions.
406
Frank J. Cunha Heat transfer correlations, based on previous testing, can be used to determine the internal heat transfer characteristics of the blade cooling system under the effects of rotation. Thus, the network flow analysis provides the means to determine internal heat transfer coefficients and coolant temperatures as internal boundary conditions for the model. The thermal results are shown in figure 15, which illustrates a region on the suction side with high metal temperatures corresponding to high thermal loads due to flow acceleration. In this example selected for illustrative purposes, the mechanical boundary conditions for the model are such that the bottom plane must remain fixed in the direction of the blade stacking axis or the z-direction. One node near the middle of an aft internal rib remains fixed in space for all three directions, and one other node near the leading edge is fixed in the y-direction. The top plane nodes are constrained to remain coupled and planar. The primary crystal coordinate system has the z-axis coincident with pull direction of the centrifugal force experienced by the blade at rated design speed. In order to simulate this, a uniformly distributed load is applied normally to the top face of the blade section. The planar boundary condition applied this way is a first approximation to boundary conditions that may be experienced by the blade section. The assumption is that there is sufficient constraint of the surrounding blade material to counteract bending moments seen as a result of temperature gradients and non-uniform deformation rates72. The blade deformation will largely be controlled by the conditions at the ends of the blade. At the root, the blade will experience a relatively lower and more uniform temperature. Thus, if the root of the blade experiences a more uniform temperature, any other cross-section will be forced to deform at a uniform rate73. Therefore, to a first approximation, the constraint of planar sections has been applied in the blade model. The results of this structural analysis are shown in figure 16 in terms of von Mises total strains. It should be noted that the analysis provided here could be applied to a more complex blade model as well. The strain results can be used for fatigue assessments. In general, strain range versus cycles to crack initiation are available for the material under consideration leading to assessment of the airfoil fatigue life.
407
3 Fig. 15. NASA E blade thermal results.
Source: See Note 23.
The allowable stress in the blade depends on its temperature 3 Fig. 16. NASA E blade thermal strain results. and the target life. The link between these considerations is the behavior of the blade material in creep. Although the maximum Source: See Note 23. stress occurs at the root, this airfoil section may not be the most critical area for creep life. This is because even though the stress may largest at the root, the metal temperature is the lowest. The critical section is then where combination of stress and temperature will make the blade reach the creep limit first. References listed in note 73 present approaches to take the combination of metal temperature and stress in assessing the creep life74. For aircraft propulsion engine mission cycles are characterized by takeoff, climb, cruise; whereas for power-generation engine the mission cycle is simpler with long periods of running at constant conditions between starts and stops. At each segment of the mission cycle, the engine will run at different speed and temperature conditions, so that the overall life needs to be considered as the cumulative effect of the life expended in each mission segment. For each set of conditions, a thermal and structural analysis needs to be made, and the results, together with material data, will be used to estimate the creep life of the component. When the turbine airfoil is subjected to repeated loads, the airfoil may incur fatigue damage75. Whereas creep occurs through initially elastic and plastic bulk deformation of the material, fatigue starts at localized plastic points due to defects, inclusions, corrosion and oxidation damage. As soon as cracks exist in the component, the stress is transferred to the remaining material through load shakedown. Eventually, if cracks reach critical size sound material will not be sufficient to support the load leading to failure. As a result, crack initiation, and not propagation, is the preferred way to design for fatigue. In general the total strain, similar to the strain distribution of figure15, can be used with material data to determine the number cycles that the component can withstand. This leads to the concept of damage, D, which is defined as the mission actual cycles relative to the material cyclic capability under the same operating conditions. The overall damage is the accumulated sum of the damage from all flight segments or types. The same definition can be applied for creep damage; but instead of cycles, actual hot time becomes the variable. In the end, the total damage of participating modes should be less than unity as described by the following damage summation equation:
4.4 Heat Transfer Analysis Overall Mission Cycle
Fatigue Cycles
k =1
i =1
∑{ ∑
D Fatigue +
Creep Load Cycles
∑
D Creep }
= 1
(38)
(38)
i =1
There are other relevant modes of damage, such as high cycle fatigue, corrosion, and oxidation; but discussion of these topics would be beyond this subject matter. Suffice to say that metal temperature, steady and unsteady stresses for the turbine component are extremely relevant for all damage modes. As a result, one should strive for obtaining representative and accurate turbine component heat transfer analysis and preparation for component life analysis.
4.4-14 Conclusions Aircraft propulsion engines, land-based power generation, and industrial machines have, as a primary component, the turbine as means to produce thrust or generate power. In the turbine section of the engine, airfoil components are subjected to extremely complex and damaging environments. The combination of high gas temperatures and pressures, strong gradients, abrupt geometry changes, viscous forces, rotational forces, and unsteady turbine vane/blade interactions, all combine to offer a formidable challenge in terms of turbine durability. Nevertheless, the ultimate goal is to maintain or even improve the highest level of turbine performance and simultaneously reduce the amount of flow needed to achieve this end. Coolant flow is a penalty to the cycle and thermal efficiency, and requires management by characterizing of the turbine airfoil thermal loads and the requirements for film cooling. Despite being a difficult situation to assess, relatively simple correlations can be used to determine the airfoil overall cooling effectiveness in terms of thermal heat load, convective efficiency and film effectiveness. In this context, existing design approaches were presented to characterize airfoil thermal loads and existing airfoil cooling schemes. Emphasis was given to the general characteristics of turbine cooling including film cooling, impingement cooling, and convective cooling for different parts of the airfoil such as leading edge, mid-body, trailing edge, tip and endwalls. Convective cooling is presented in terms of fundamental cooling enhancements. The heat transfer phenomena associated with turbulating trip strips and pedestals were presented and discussed in some detail. In the end, internal and external heat transfer conditions are used to determine the durability of the airfoils in terms of oxidation, creep life and fatigue cycles. Recent literature dealing with these topics is listed to provide more in-depth overview of the subject matter.
4.4-15 Notes ________________________ 1. J. Moustapha, M.F. Zelesky, N. Baines, and D. Japikse, Axial and Radial Turbines, Concepts NREC, 2003 2. Ibid. 3. Ibid. 4. D.P. DeLuca and C.G. Annis Jr, “Fatigue in Single Crystal Nickel Superalloys,” Office of Naval research (ONR) FR-23800, Aug. 1995; N.J. Arakere, “High-Temperature Properties of Single Crystal Superalloys in Air and Hydrogen,” ASME IGTI 2001-GT-0585, New Orleans, June 2001; R.A. Naik, D.P. DeLuca, and D.M Shah, “Critical Plane Fatigue Modeling and Characterization of Single Crystal Nickel Superalloys,” ASME IGTI 2001-GT-30300, Trans. ASME Journal of Engineering for Gas Turbines and Power, 2004; D.W. Maclachlan and D.M. Knowles, “The Effect of Material on the Analysis of Single Crystal Turbine Blades: Part I – Material Model,” Fatigue and Fracture Engineering Material Science 25 (2002): 385-398;D.W. Maclachlan and D.M. Knowles, “The Effect of Material on the Analysis of Single Crystal Turbine Blades: Part II – Component Analysis,” Fatigue and Fracture Engineering Material Science 25 (2002): 385-398. 5. V. Seetharaman, “Thickness Debit Properties of PW1484,” Pratt and Whitney Materials and Processes Engineering, Interim Report, Nov. 2002. 6. C.T. Sims, N.S. Stoloff, and W.C. Hagel, Superalloys II, (John Wiley & Sons, 1980). 7. T.L. Dame, “Anisotropic Constitute Model for Nickel Base Single Crystal Alloys: Development and Finite Element Implementation,” PhD Thesis, University of Cincinnati, 1985. 8. F.R.N. Nabarro and H.L. deVillers, The Physics of Creep, (Taylor and Francis, 1995). 9. F.J. Cunha, M.T. Dahmer, and M.K. Chyu, ”Thermal-Mechanical Life Prediction System for Anisotropic Turbine Components,” ASME GT2005-68108, Reno, NV, 2005, Trans. ASME Journal of Turbomachinery, 2006. 10. See note 5 above. 11. J.E. Heine, J.R. Warren, and B.A. Cowles, “Thermal Mechanical Fatigue of Coated Blade Materials,” Wright-Patterson Final Report ERDC-TR-89-4027, Sept. 1988. 12. See note 1 above. 13. L. Torbidoni and J.H. Horlock, “ A New Method to Calculate the Coolant Requirements of a High Temperature Gas Turbine Blade,” ASME Paper GT2004-53729, Vienna, Austria, 2004. 14. Ibid.
408
Frank J. Cunha
409
15. See note 1. 16. H. Schlichting, Boundary-Layer Theory, (McGraw-Hill, 7th Edition, 1979). 17. See note 13. 18. Goldstein, R.J., “Film Cooling”, Advances in Heat Transfer, Vol. 7, pp. 357-358. 19. See note 1. 20. M.K. Chyu, “Heat Transfer and Pressure Drop for Short Pin-Fin Arrays with Pin-Endwall Fillet.” Trans. ASME Journal of Heat Transfer 112 (1990): 926-932; A.A. Zukauskas, “Heat Transfer from Tubes in Cross Flow,” Advances in Heat Transfer 8 (1972): 116-133. 21. R.J. Goldstein, E.R.G Eckert, and F. Burggraf, “Effects of Hole Geometry and Density on Three-Dimensional Film Cooling,” Trans. ASME J. Heat Transfer, 1974. 22. J.E. Albert, D.G. Bogard, and F. Cunha, “Adiabatic and Overall Effectiveness for a Film Cooled Blade,” IGTI-ASME GT2004-53998, Vienna, Austria, 2004. 23. F.J. Cunha, M. Dahmer, and M.K. Chyu, ”Analysis of Airfoil Trailing Edge Heat Transfer and Its Significance in Thermal Mechanical Design and Durability,” ASME GT2005-68107, Trans. ASME Journal of Turbomachinery, 2006. 24. See note 1. 25. J.K. Salisbury, Steam Turbines and Their Cycles (Malabar, FL: R.E. Krieger Publishing Co., 1950). 26. F.J. Cunha, “Integrated Steam/Gas Cooling System for Gas Turbines,” US Patent No. 5,340,274, 1994; A. Jacala, R.M. Davis, M.A. Sullivan, R.P. Chyu, and F. Staub, “Closed Circuit Steam Cooled Bucket,” US Patent No. 5,536,143, 1996; F.J. Cunha, D.A. DeAngelis, T.A. Brown, S. Chopra, V.H.S. Correia, and D.R. Predmore, “Turbine Stator Vane Segments Having Combined Air and Steam Cooling Circuits,” US Patent No. 5,634,766, 1996. 27. L. Langston, “A Year of Turbulence”, Power and Energy, ASME Magazine, June, 2004. 28. A.H. Lefebre, Gas Turbine Combustion, Taylor and Francis, 2nd Ed., 1999. 29. Ibid. 30. J.C. Han, “Turbine Blade Cooling Studies at Texas A&M University ~ 1980-2004”, The 2004 AIAA Thermophysics Award Lecture, 2005. 31. See note 16. 32. Ibid. 33. See note 1. 34. See note 16. 35. See note 1. 36. W.M. Kays and M.E. Crawford, Convective Heat and Mass Transfer, McGraw-Hill, 1980, 2nd Ed. 37. F.O. Soechting and O.P. Sharma, “Design Code Verification of External Heat Transfer Coefficients,” Paper No. AIAA-88-3011, 1988. 38. W.M. Kays and M.E. Crawford, “STAN5 - A Program for Numerical Computation of Two-Dimensional Internal and External Flows,” NASA CR-27-42, 1976. 39. R.L. McKnight, J.H. Laflen, and Spamer, “Turbine Blade Tip Durability Analysis”, NASA CR-165268, Feb. 1981; J. Christophel, K.A. Thole, and F.J. Cunha, “Measured Adiabatic Effectiveness and Heat Transfer for Blowing from Tip of a Turbine Blade,” Trans. ASME Journal of Turbomachinery 127, (2005): 251-262; J. Christophel, K.A. Thole, and F.J. Cunha, “Cooling the Tip of a Turbine Blade Using Pressure Side Holes – Part I. Adiabatic Effectiveness Measurements,” Trans. ASME Journal of Turbomachinery 127 (2005): 270-277; J. Christophel, K.A. Thole, and F.J. Cunha, “Cooling the Tip of a Turbine Blade Using Pressure Side Holes – Part II: Heat Transfer Measurements,” Trans. ASME Journal of Turbomachinery 127 (2005): 278-286; E. Couch, J. Christophel, E. Hohlfeld, K.A. Thole, and F.J. Cunha, “Comparison of Measurements and Predictions for Blowing from Tip of a Turbine Blade,” AIAA J. of Propulsion and Power 21 (2005): 335-343. 40. M.F. Blair, “An experimental Study of Heat Transfer and Film Cooling on Large-Scale Turbine Endwall,” Trans. ASME J of Heat Transfer 96 (1974): 524-529; R.A. Graziani, M.F. Blair, J.R. Taylor, and R.E. Mayle, “An Experimental Study of Endwall and Airfoil Surface Heat Transfer in a Large Scale Turbine Blade Cascade,” Trans. ASME J. Eng. Gas Turbine Power 102 (1980): 257-267; O.P. Sharma and T.C. Buttler, “ Reynolds Stresses and Dissipation Mechanisms Downstream of a Turbine Cascade,” Trans. J of Turbomachinery 109 (1987): 229-236; R.J. Goldstein and R.A. Spores, “Turbulent Transport on the Endwall in the Region Between Adjacent Turbine Blades,” Trans. J of Turbomachinery 110 (1988): 862-869; J.T. Chung and T.W. Simon, “Effectiveness of the Gas Turbine Endwall Fences in Secondary Flow Control at Elevated Freestream Turbulence Levels,” ASME Paper 93-GT-51, 1993; Y. Yu and M.K. Chyu, “Influence of Gap Leakage Downstream of the Injection Holes on Film Cooling Performance,” Trans. J of Turbomachinery 120 (1998): 541-548; W.W. Ranson, K.A. Thole, and F.J. Cunha, “Adiabatic Effectiveness Measurements and Predictions of Leakage Flows Along a Blade Endwall, AIAA J. of Propulsion, Nov. 2005. 41. F. Kreith, Principles of Heat Transfer , IEP- A Dun-Donneley Pub., 1976, 3rd Ed. 42. See note 13. 43. See note 18. 44. See note 21. 45. See note 18.
4.4 Heat Transfer Analysis 46. J.C. Han, S. Dutta, and S. Ekkad, Gas Turbine Heat Transfer and Cooling Technology, Taylor and Francis, 1st Ed., Ch. 1, 2000. 47. Patankar, S.V., and Spalding, D.B., Heat and Mass Transfer in Boundary Layers, CRC Press, Cleveland, 1968. 48. See note 36. 49. F.M. White, Viscous Fluid Flow, McGraw-Hill, 2nd Ed, 1991. 50. See note 41. 51. J. Nikuradse, (1933) “ Laws of flow in rough pipes,” VDI Forsch.. 361(1933) English Translation. NACA TM-1292 (1965). 52. J.C. Han, “Heat transfer and friction in channels with two opposite rib-roughnened walls,” Trans. ASME Journal of Heat Transfer 106 (1984): 774-781. 53. R.L. Webb, E.R.G. Eckert, and R.J. Goldstein, “Heat transfer and friction in tubes with repeated-rib roughness,” Int. J. Mass Transfer 14 (1971): 601-617; R.L. Webb and E.R.G. Eckert, “Application of rough surfaces to heat exchanger design,” Int. J. Mass Transfer 15 (1972): 1647-1658. 54. Ibid. 55. F.J. Cunha, “Numerical Prediction of Heat Transfer in Internal Cooling Passages of Gas Turbine Blades,” IGTI- ASME Cogen Turbo Power 8 (1993): 307-316; F.J. Cunha, “A Calculation procedure to Analyze Three Dimensional Parabolic Flow Problems of Heat Transfer with Rotation,” ASME HTD 300 (1994): 123-137; D.G.N. Tse and Steuber, “Flow in a Rotating Square Serpentine Coolant Passage with Skewed Trips,” ASME Paper 97-GT-529, 1997; J.H. Wagner. B.V. Johnson, R.A. Graziani, and F.C. Yeh, “Heat Transfer in Rotating Serpentine Passages with Trips Normal to the Flow,” ASME Paper 91-GT-265, Transactions ASME, 1992. 56. See note 53. 57. U. Uysal, P. Li, M.K. Chyu, and F.J. Cunha, “ Heat Transfer on Internal Surfaces of a Duct Subjected to Impingement of a Jet Array with Varying Jet Hole-Size and Spacing,” ASME GT2005-68944, Trans. ASME Journal of Turbomachinery, 2006; R. Gardon and J. Cohonpue, “Heat Transfer Between a Flat Plate and Jet Air Impinging on It,” ASME/AICHE International Heat Transfer Conference Proceedings, 1962; H. Martin, “Heat and Mass Transfer Between Impingement Gas Jets and Solid Surfaces,” Advances in Heat Transfer 13, 1977. 58. M.K. Chyu, U. Uysal, and P-W Li, “Convective Heat Transfer in a Triple-Cavity Structure Near Turbine Blade Trailing Edge,” Proceedings of IMECE’02, International Mechanical Engineers Congress, New Orleans, Nov. 17-22, 2002. 59. See note 46. 60. D.M. Kercher and W. Tabakoff, “Heat Transfer by Square Air Jets Impinging Perpendicular to a Flat Surface Including the Effect of Spent Air,” ASME Paper 69-GT-4, Trans. ASME Journal of Engineering for Power, 1970; L.M. Florschuetz, C.R. Truman, and D.E. Metzger, “ Streamwise Flow and Heat Transfer Distributions for Jet Array Impingement with Crossflow, Trans. ASME Journal Heat Transfer, March, 1981. 61. Ibid (Florschuetz). 62. See note 20 (Zukauskas). 63. See note 20 (Chyu). 64. F.W. Dittus and L.M.K. Boelter, University of California at Berkely Publications in Engineering 2 (1930): 443. 65. K. Ishida and K. Hamabe, “Effect of Pin-Fin Aspect Ratio and Arrangement on Heat Transfer and Pressure Drop of Pin Fin Duct for Airfoil Internal Cooling Passage,” ASME Paper 85-WA/HT-62, 1985. 66. See note 46. 67. See note 20 (Chyu). 68. See note 23. 69. Ibid. 70. ANSYS User’s Manual, Swanson Analysis Systems, Inc., Volumes I,II,III,IV, Revision 5.0. 71. R.D. Thulin and D.C. Howe, “Energy Efficient Engine High-Pressure Turbine Design Report”, NASA CR-165608, PWA-5594-171, March 1982. 72. See note 4 (Maclachlan, 2002). 73. Ibid. 74. F.R. Larson and J. Miller, “A Time-Temperature Relationship for Rupture and Creep Stress,” Transactions of the ASME (1952): 765-775; F. Garafalo, Fundamentals of Creep and Creep-Rupture in Metals, The Macmellan Co.,1965; D.C. Stouffer and L.T. Dame, Inelastic Deformation of Metals Models, Mechanical Properties and Metallurgy, John Wiley & Sons, 1st Ed.,1996. 75. D. Burgreen, “Structural Growth Induced by Thermal Cycling,” Journal of Basic Engineering, ASME 68-WA/Met-14, 1968; F.O. Soechting, ”Turbine Low Cycle Fatigue Design Program,” Wright-Patterson Final Report AFWAL-TR-862124, 1985; also see note 21.
410
BIOGRAPHY
4.4 Heat Transfer Analysis
Frank J. Cunha, Ph.D., P.E. Pratt & Whitney United Technologies Corporation 5 Bruce Lane Avon, CT 06001 phone: (860) 565-8909 email: [email protected]
Dr. Cunha has worked in the area of turbine cooling design and analysis, heat transfer and mechanical design for a period spanning 25 years. During this time, Dr. Cunha assumed lead design responsibilities at several original equipment manufacturers, including Siemens-Westinghouse, General Electric, and Pratt and Whitney and has received 25 US patents for design of turbine engine hot gas path components. Throughout his career, Dr. Cunha has continued to apply recent advances in cooling technology to both commercial and military engine programs to achieve the highest rotor inlet temperature level in the aerospace industry. Dr. Cunha has been a leader in developing advanced testing cooling programs at a consortium of Universities, and has published numerous journal and symposium technical papers on turbine cooling technologies.
5.0
Turbine System Economics and Reliability, Availability & Maintainability (RAM)
5.0-1 Introduction With the varied and fast changing global power market, the complexity of turbine system economics has increased dramatically. In the past, power plants were primarily government regulated and base loaded. Dispatch and electricity pricing was relatively predictable. In today’s market, with IPPs, there are endless variations in the way power is produced, provided, regulated, and purchased. OEMs and power producers need to understand methods to quantify and compare parameters, and to understand the drivers and uncertainties to properly evaluate decisions and their potential for profitability in this constantly changing marketplace.
5.0-2 The Power Market Drivers To understand the power market, one must keep in mind the key differences between this market and others: Electric power can not be economically stored. Unlike other commodities, electric power cannot be easily or economically stored. For the most part, it must be produced on demand. While there are some efforts to retain energy generated during off peak hours using technologies such as pumped storage, flywheels, and/or superconductors, the cost is high and the efficiency and reliability of these methods is low. The demand for electric power is constantly fluctuating. The fluctuating demand for electric power is clear. Demand varies during the day, with a morning and evening peak and varies over the year with a winter and summer peak. Some of this fluctuation can be predicted based on historical information, such as the typical change in consumption over a day, and the typical seasonal variations, but the fluctuations can shift significantly from the norm due to uncontrollable events like periods of severe weather. Utilities have a high capital investment cost. The initial required investment for a power plant varies based on the type of power plant. Typically, utilities have very high fixed costs, spending almost five times the initial investment per dollar than other manufacturing endeavors1. These fixed costs, which are typically between $475/kW and $1430/kW2, include equipment for generation, transmission, distribution, and permitting. Power plants have a relatively long life cycle.
Bonnie Marini, Ph.D. Manager, Gas Turbine Design Gas Turbine Engineering Siemens Power Generation Orlando, FL phone: 407 736-6428 email: [email protected]
Power plants do require significant initial investment, but they compensate with very long life. Power plants operate for decades, with units operating 30 years or more. Fuel prices are subject to negotiation and electricity prices are constantly varying. market.
These factors impact the ability to compete effectively in the deregulated
There is an unflinching expectation that required electric power is always and immediately available.
445
Electricity has become a critical and integral part of the economy and there is no tolerance for an inadequate supply no matter what the circumstances. This is reflected in the fact that electric consumption is generally accepted as one of the lead economic indicators.
These unique features of the power market create a complex situation to evaluate and choose effective strategies for power generation. Power Market Solutions The nature of this constantly fluctuating demand for a commodity that must be essentially produced on demand drives a varied supplier base, which can be broken into three basic types of operating modes – base loaded plants, intermediate loaded plants, and peakers. Base load plants operate continuously for long periods of time. They are typically large plants (> 200MW), which are economical and reliable to operate. These plants often do not have the ability to change load quickly and take advantage of spot market peak pricing. These units operate year round and all day with an overall economy, which allows them to compete and operate profitably, even at low demand times. Base load plants provide the core of the power grid, and act to regulate and maintain the grid frequency. Intermediate load supplier includes plants that operate to meet normal fluctuating demands in the morning and evening hours and typically operate for 10 to 14 hours per day. Peak load suppliers include plants which can start up quickly and supply power to meet high demands during periods of high or low temperature when the combined base and intermediate load capacity is not adequate. These plants are typically more expensive to operate, but offer operational flexibility. Electricity supplied during a peak demand is sold at a premium, making this the most profitable time to generate.
5.0-3 The Economics of Making Electricity The US market is currently a hybrid market consisting of regulated market regions and deregulated market regions, with each having different economic drivers. The regulated markets have contracts that are cost based. Rates are negotiated, fixed, and renegotiated allowing for a set return on investment for the power producer. The deregulated market is an auction market, with suppliers bidding into the market, offering power to the grid. A controller ranks the bids and purchases power as needed from the lowest cost supplier on up in price until the demand is met. The traditional power market is regulated to meet local demand for power. In this scenario, the cost of electricity is typically locked in by regulation and varies little. A network of base load units and expensive peakers is put in place to enable the supply to meet the fluctuating demand for power. Profits are based on a cost plus regulated model. For a power producer, when choosing an OEM (Original Equipment Manufacturer) or AE (Architect Engineer), the primary economic variables to consider are first time plant cost, service costs, fuel costs, and plant availability and reliability. Fuel costs far outweigh the other factors and the driver in this case for the OEM industry is efficiency. The power market has changed significantly with deregulation and different variables must be considered to appropriately assess the potential from a customer and supplier point of view. In the deregulated market power prices fluctuate drastically over time. In different areas and different countries, the rules vary, but it is common to have power generators bidding to supply to the grid and in some cases, there are penalties for promising power and then not being able to deliver. This market, which seems as harried and volatile as the stock exchange floor, bids the operation of units which vary in their ability to come up to full power, to sustain partial load and to assure delivery of power at a precise time. RAM and Economics When a power plant is off line for maintenance, there is no income stream. When power is offered on the spot market and the plant does not start, there may be penalties. These scenarios drive a need to consider RAM. RAM is an acronym for Reliability, Availability, and Maintainability. Reliability is used to express and quantify the unplanned maintenance needs of a power plant. Reliability is a measure of how often a plant is available in comparison to the total number of hours the plant would be available with no unexpected maintenance. An ideal power plant has a reliability is 100%.
Availability is a measure of how often a unit is capable of providing service. The availability can be quantified as the ratio of the total number of hours the unit is actually available in comparison to the total number of hours. Availability considers both scheduled and unscheduled maintenance and compares that to an ideal situation with no maintenance outages at all. An ideal power plant has an availability that is less than 100%. Units with less frequent and shorter maintenance intervals have higher availabilities.
446
Bonnie Marini
Maintainability is used to express the cost of maintenance. This includes the cost for parts, and the cost of the servicing. Maintainability can be used to compare plants that require frequent, lower cost servicing, with plants that require less frequent, higher cost servicing. Combining these considerations, RAM looks at how often you can use the equipment and how much it costs to keep it in operating condition. The concept of RAM is used to consider the trade off between higher technology immature technologies, and less advanced, but more reliable operation. Modern Power Plant Economics Throughout the US market exists a hybrid market in which the trend has been to move towards deregulation. The change from regulation to deregulation changed the motivation and strategy of power producers. The old descriptions of the power industry as stable and constant were traded in for adjectives like competitive, flexible, dynamic – even risky. This volatile market favors plants that are highly fuel efficient for base load and plants with quick, “push button” starting capability to meet daily peak demands. As a power producer, the ideal situation is to be efficient enough to compete in the base loaded market or flexible enough to change load quickly to compete in the peak market. For an OEM to have access to the largest market share, this means providing engines that can do both. This need to “have it all” has driven the technology of power plants. Equipment manufacturers are competing to deliver plants with the highest efficiency and the most flexibility. This demand for more performance results in pushing the technology envelope. There have been positive and negative consequences to this change in the industry. The driver for the change was to increase competition in the marketplace, and consequently to reduce utility bills. A discussion of the result in that regard is out of the scope of this paper, but significant secondary consequences occurred which are relevant. The technology of gas turbines has advanced very quickly in this market. Efficiencies have increased and emissions have reduced. The introduction of new technologies has increased in volume and scope, resulting in issues in reliability and availability for less mature engines. A power producer must decide between buying the most advanced technology with the highest efficiency, or buying a model that is more mature, and more reliable, but not as efficient. Utilities are businesses. Regulated utilities are mainly concerned with maintaining the lowest life cycle costs for their units. For deregulated enterprises profitability is what is important. Prior to deregulation, sound economic decisions were important, and with regulation, these calculations were fairly simple and reliable. Subsequent to deregulation, sound economic decisions became absolutely critical to survival, while becoming far more complex. Utilities must understand how to make appropriate purchasing decisions, and OEMs must understand how these choices are made to compete effectively. The economics and the technology are intimately tied together. While it may be clear which technical decision advances technology the fastest, it is not always as clear which technical decision makes the best business case. Power plant economics explores the cost of a decision over time. Basic Power Plant Economics Economic evaluation of a power plant can be explored in a number of ways. All methods account for the cost of a decision over time. There are several accepted methods to examine cost over time. A common approach is to evaluate net present value (NPV). The net present value method looks at the value of the project over time by converting all income and expenditures into equivalent values at the current time and subtracting the initial investment. To do this, the future interest rate and the rate of inflation must be estimated and expressed as a discount rate, r. While these estimates are somewhat inaccurate, the sensitivity to the assumption can be explored to understand the implication of one choice over another.
Where; In: initial investment r: discount rate or weighted average capitol cost for a given company t: time CashFlow: Income – expenses
447
5.0 Turbine System Economics and Reliability, Availability & Maintainability (RAM) The basic rule of thumb is that the NPV should be greater than zero for an investment, though most investors will strive to get a hurdle rate that is higher. To calculate the NPV of a power plant, the following parameters are needed: 1) capital investments, 2) projected price of electricity, 3) size of the plant, 4) capacity factors (how many hours the plant will operate in a given year), 5) dispatch payments 6) projected fuel costs, 7) operating and maintenance costs, 8) start up time and costs, 9) regulating costs, and 10) the discount rate (the cost of money). For a power producer, the initial investment is the cost of the plant and initial costs for support equipment and hiring staff. The income is calculated per annum, and is income from the electricity sold to the grid. This is the price of electricity ($/MWh) multiplied by number of hours the facility is producing power (MWh) in a typical year. In a deregulated market, the price of electricity is determined by the market and for calculation, should be evaluated based on historical data and forecasted fuel and electricity prices. In some markets, dispatch payments are another source of income. Dispatch payments are fees paid to producers that are capable of providing power to the grid with a very short lead time, typically in the range of 10 minutes from demand to supply (spinning and nonspinning reserve). These payments are for the assurance of capability and are paid regardless of whether the capability is leveraged. This provides incentive to suppliers to develop and maintain a fast start supply, so the grid can adequately respond to unplanned peak needs. Expenses are determined on a per annum basis and include the cost of fuel, and operating and maintenance costs. The fuel cost is a variable cost and must be estimated. For the calculation, fuel cost must be converted to annual cost in dollars by multiplying the fuel price times cumulative fuel consumed per annum. The operating and maintenance costs include personnel costs, and the costs for scheduled and unscheduled maintenance. Operating and maintenance costs are impacted directly by the mode of operation of a plant. The frequency of maintenance is influenced by both operating hours and number of starts. A single start for a combined cycle facility can result in a significant incremental increase in maintenance costs, with one producer estimating $20,000 in incremental costs for one combined cycle start3. Expenses may also include regulating costs. Regulating costs are government instituted economic consequence to encourage industry to make decisions that have been determined to be for the good of the people. These costs include taxes and the cost of complying with government regulations. Taxes may be implemented to influence companies to choose preferred technologies or to impact the local job market. Technologies may be politically preferable due to environmental or safety issues as viewed by the regulating government. Certain fuel sources may be preferred to enhance energy independence or to promote local industry and employment. Regulating costs also include costs for adhering to regulations that are put in place to assure the safety of a facility, such as OSHA requirements. Regulations and taxes are dependent on the current political climate and are subject to frequent changes, often resulting in the changing position of a certain facility in the market place. An example of this kind of influence is environmental regulations where emissions credits are traded. Over time a facility built to meet a certain regulation may become covered by a regulation with a more aggressive limit. This may mean that additional operating costs are incurrent to purchase additional emissions credits, thus influencing the economics of the power producer. The calculation complexity increases further when looking over the life of the power plant. Power plants have a long life and many changes occur over the life of the plant. Some of these risks can be hedged by investing in futures to fix the future price of commodities such as fuels, or to insure against adverse business conditions, such as long periods of mild weather. Some gas turbines have the capability to operate on alternate fuels. Some units are purchased with the flexibility to switch between gas and oil, and various qualities of oil can be considered. As the price for oils and gases fluctuate, a plant with this capability can be more competitive, but at a price. This option requires more capital investment, both in the unit and in the supporting auxiliaries, and potentially in licensing and permitting fees. Units desire to remain competitive over time by being highly efficient and as a result of this are also driven to increase capital investment over time. New technology is regularly introduced and can be purchased as upgrades to improve efficiency. To develop a symbiotic relationship with customers, some OEMs offer access to upgrades to customers who purchase long term service agreements, thereby integrating the need for consistent high quality service with the need for continually competitive technology. To understand the drivers for profitability, and the importance of RAM, the sensitivity of the calculation can be explored to further understand the uncertainty of the calculation. The variables can be examined in terms of controllable variable and uncontrollable variables.
5.0-4 Operating Strategies and Options For a base loaded plant, low margins are compensated by long operating times. Long, uninterrupted operating times are supported by reliability and maintainability. Gas turbine technology has been in service for many decades, and most units have
448
Bonnie Marini reliabilities of greater than 95%. However, the base line efficiency of a competitive unit is constantly increasing. The old robust engine with learned out technology is simply not efficient enough. At the other end of the operating spectrum are peakers that are looking to leverage the high costs of electricity during peak needs. In June 25 of 1998, the price per megawatt-hour of electricity in parts of the Midwest soared briefly from $40 to $70004. Though the higher end of this scale is the exception and not the norm, the implication is clear that the investment costs and fuel costs pale in the face of this return and the only significant factor is how much the plant can generate. In this market the goal is to be ready to run when the prices increase. Here again RAM is the driver since for the most part, a window of high potential for peak need can be identified, and so owners can schedule planned maintenance outside these windows. However, availability and reliability are extremely important because if an owner pushes the start button and does not get power, a competitor will quickly jump in and take over that share of the market. If a plant is inoperable due to unplanned maintenance (low reliability) then the opportunity to compete during this need will not even be possible. Operators in the intermediate load business are balancing all of these needs. They want to be chosen for operation, so efficiency is important, and they need to ready to operate. Reliability, availability, and maintainability are all equally important. Gas Turbines in the Power Market Gas turbines operate in all three operating regimes. Simple cycle gas turbines are installed to meet peak demands. They are extremely flexible, relatively low in initial investment cost, and quick to install. While internal machine efficiencies are quite high, gas turbine simple cycles exhaust at roughly 1000◦F, wasting a significant amount of energy and resulting in a rather low cycle efficiency, and limiting the application to peak markets. For intermediate and base load applications, gas turbines are used in combined cycles. In this arrangement, the gas turbine waste heat is used in a heat recovery steam generator to power a bottoming steam cycle. The combined efficiency of such power plants are quite competitive (> 55%), easily fitting into the intermediate load market, and capable of competing as a base loaded units. In both scenarios there are economic complexities. The base loaded plants must maintain very low costs. The business cases are based on low margins and high volume. Large plants with high availability and low maintenance needs have the advantage in this market. The business cases for load following plants are based on a low volume high profit model, which requires them to be available when the need is present. The economics improve when an unexpected peak in demand occurs. Those who can meet these peaks quickly and reliably can reap the benefit of selling when the market is at its highest. For these reason, RAM – reliability, availability and maintainability become significant driving factors in the power market. Controlling RAM While Increasing Technology Level To maintain competitiveness, OEMs are aware that new technology must be introduced at a lower risk level. Steps are taken to control risk during design, prior to implementation, and during operation. During design, risk analysis and management methods are used which allow a quantified assessment of the probability of a particular failure and the consequences. Results of these analyses are used to mitigate risks by either changing the design, or altering the consequential impact. Prior to implementation, new technologies are being tested more thoroughly. In the past, most of the testing was done with similar technologies in small engines, or aircraft engine test beds. These approaches are similar, but not the same as the IGT (Industrial Gas Turbine) application. To further reduce risk, some OEMs such as Siemens, have built IGT test beds, so fully scaled new technology can be fully instrumented and tested in a controlled environment. For further reaching changes, an entire test site is constructed through cooperation between an OEM and a power producer to validate a new design. In some cases, new technologies are introduced to a single customer site with an agreement to test out the technology prior to release to a larger fleet. Subsequent to implementation, engines can be fitted with improved monitors and sensors, condition monitoring, and better controls. OEMs are now offering producers the option to purchase a monitoring contract, where the OEM constantly monitors plant operation. Monitoring a fleet of engines allows the OEM to develop probabilistic indicators of potential failures. When symptoms occurs, the owner is informed, and corrective actions can be taken in a controlled manner, at a convenient time, greatly reducing the likelihood of unexpected failure during a peak need. All of these features combine to reduce the risk of increased RAM costs.
5.0-5 Conclusion Since deregulation the focus on efficiency so over shadowed other needs that the technical envelope was pushed very hard, very fast. The result was an improvement in the efficiency of gas turbines, but there was a partnering risk when leveraging immature technologies. Using RAM in business models allows appropriate evaluation of the benefit and risk of immature technologies and allows user to apply these technologies intelligently. Today, OEMs and operators are both cognizant of the need to make sound economic evaluations of technology options and to consider technology maturity and the resultant RAM into their calculations. Additional actions (further testing) are taken and additional products (such as online monitoring) are being offered to reduce and control RAM. With these considerations, good choices can be made by power producers that will provide for profit for the company and reliable power for the communities served.
449
5.0 Turbine System Economics and Reliability, Availability & Maintainability (RAM) 5.0-6 Notes ––––––––––––––––––––––– 1. Bonbright, Danielson and Kamerschen, Principles of Public Utility Rates, Arlington, Virginia, Public Utility Reports, Inc., 1988. 2. “Cost Comparison IGCC and Advanced Coal,” by Stu Dalton, Director Fossil, Emission Control and Distributed Energy Resources, EPRI. Presented at EPRI Roundtable on Deploying Advanced Coal July 29, 2004. 3. Panel of Combined Cycle Users Group & Gas Turbine Users Group, Electric Power Conference, Chicago, April 5-7, 2005. 4. “Exploiting Uncertainty, The “real options revolution in decision-making,” by Peter Coy, Published in Business Week, June 7, 1999.
450
BIOGRAPHY
5.0 Turbine System Economics and Reliability, Availability & Maintainability (RAM)
Bonnie Marini, Ph.D. Manager, Gas Turbine Design Gas Turbine Engineering Siemens Power Generation Orlando, FL phone: (407) 736-6428 email: [email protected]
Dr. Marini has been working in the power industry since 1980 and has a PhD in experimental fluids mechanics. She is currently the Manager of Turbine Technology and Processes for Siemens Power Generation. In this position she is responsible for technical approaches and processes for turbine hot gas path design and for advanced turbine development projects. Prior to this, she led the team developing upgrade products for the Siemens gas turbine service fleet. Other experience includes combined cycle analysis, steam turbine gas path design, and systems design for nuclear power plants while working for various AE firms, OEMs, and utilities
6.0.1 The DOE Turbine Program: Overall Program Description
6.0.1-1 Introduction The focus of the DOE Office of Fossil Energy (FE) Advanced Turbine Program is on the key technologies needed to enable development of advanced turbines that will operate cleanly and efficiently when fueled with coal-derived synthesis gas and hydrogen fuels. Developing turbine technology to operate on these fuels is critical to the development of advanced zero-emission power generation technologies such as FutureGen type plants that will minimize emissions of carbon dioxide. These plants will most likely be based on integrated gasification combined-cycle systems, and consequently should be capable of utilizing coal-derived synthesis gas as well as hydrogen. The Turbine Program is an investment in secure U.S. electric power production that is clean, efficient, affordable, and fuel-flexible, and will make possible the continued use of coal our Nation's largest domestic fossil energy resource — coal. The FE Turbine Program R&D is supporting the adaptation and development of existing and new advanced gas turbines for application to coal-derived hydrogen fuels and synthesis fuels. Studies, both ongoing and completed, have identified concepts for optimization and modification of large frame combustion turbines in integrated gasification combined-cycle (IGCC) applications. These studies have determined the concepts, technologies, and modifications needed to meet the goals for near-zero emissions, higher efficiency, and lower capital cost machines for application to coal-derived fuels such as syngas and hydrogen. Technology base activities will provide the basic underpinning for the Program areas to resolve advanced systems, material, heat transfer, aerodynamic, and combustion technical issues, as machines and systems are modified for high-hydrogen fuels derived from coal. The FE Turbine Program, as administered by DOE’s National Energy Technology Laboratory (NETL), is designed to provide low-cost solutions to Presidential initiatives, and provide technological solutions to high level DOE goals. These initiatives include: 1) Climate Change Initiative (http://www.whitehouse.gov/news/releases/2002/02/climatechange.html) 2) Clear Skies Initiative (http://www.whitehouse.gov/news/releases/2002/02/clearskies.html) 3) FutureGen Initiative (http://www.fe.doe.gov/programs/powersystems/futuregen/) 4) Hydrogen Initiative (http://www.eere.energy.gov/hydrogenandfuelcells/pdfs/review04/2_mill er_philadelphia_04.pdf) Specific goals presented below are written for Advanced Coal-Based Power Systems, and are designed to support these Presidential initiatives. The Advanced Power Systems goals are addressed for the most part by the efforts of the DOE-FE Gasification and Turbine Programs. This is particularly true for the 2010 goal, with improved efficiency and costs. The 2012 goal brings in the additional accomplishments and progress made by the CO2 Sequestration Program.
The Advanced Power Systems goal for 2010 states: By 2010, develop advanced power systems capable of achieving 45–50% thermal efficiency at a capital cost of $1,000/kW or less for coal-based plant utilization. The Sequestration interim goal for 2012 states: By 2012, R&D will be completed to integrate this technology with CO2 capture and sequestration into a zeroemissions configuration(s) that can provide electricity with less than a 10 percent increase in cost. A main objective of the Advanced Turbine Program is to support the FutureGen Initiative. The FutureGen Initiative and the associated project can be described as an effort to “…validate the technical feasibility and the economic viability of “zero” emission energy from coal. By 2012, begin operation of a nominal 275megawatt (MW) prototype plant that will produce electricity and hydrogen with “zero” emissions; and prove the effectiveness, safety, and performance of CO2 sequestration.” The Fossil Energy 2015 goal states: “Create partnerships that provide technology by 2015 for near-zero emission plants (including carbon) that are fuel-flexible and capable of multi-product output, and efficiencies over 60 percent with coal and 75 percent with natural gas. It is FE’s intent that program spending will be completed during the stated goal date, thereby completing the R&D at a full-scale prototype or component scale. It is through this prototype scale testing that the ability to meet these goals will be demonstrated and substantiated. Subsequent testing and deployment of the technology at a demonstration scale will be completed through other programs and is expected to take four years. x x x x
FE Turbine Program contributions to the 2010 Advanced Power Systems goals are planned to be: Efficiency: Demonstrate 2–3 percentage points of improvement in combined-cycle (CC) performance (above base line). Cost: Demonstrate a 20–30% reduction in CC capital cost plus enhanced value for lower COE. Emissions: Demonstrate combustor emissions with 2 ppm NOx (@15% O2) in simple cycle exhaust.
It is expected that these advances to achieve the 2010 goal will contribute to the 2012 goal for IGCC-based power systems that capture carbon. The challenge here is maintaining the 2010 performance advances but now the turbine fuel will be nearly pure hydrogen. Additionally, the Turbine Program plans to contribute to the 2012 Carbon Sequestration goal by providing advanced and highly integrated CO2 compression technology to reduce the compression penalty (auxiliary load) by 25–40%. In supporting the FutureGen Project is a primary goal of the FE Advanced Turbine program. The Turbine Program takes the approach to provide the latest advances made through pursuing the 2010, 2012, and 2015 goals. This will allow installation of the most advanced hydrogen fueled turbine at the FutureGen project. It is envisioned that the FutureGen turbine could be installed with a plan that would allow the machine to be optimized in the field for combustion and firing temperature performance. This approach would allow for a machine fueled with 100 percent hydrogen to operate with the highest efficiency and lowest NOx emissions. Plans for the FE Turbine Program contributions to the 2015 goal are: Efficiency: x Hydrogen turbine CC with 3–5% points improvement (total above base line). x Oxy-fuel turbine based IGCC system > 50% eff. (HHV) with CO2 capture and compression. Cost: x Competitive COE for zero emission systems. Emissions: x H2 Turbine-based IGCC demonstrated with 2 ppm NOx (@15% O2) x Oxy-fuel turbine based IGCC with zero emissions (100% turbine exhaust captured and sequestered, and zero criteria pollutants and CO2)
2
Multiple Products: x H2 turbine-based IGCC with higher capacity gasification. x Oxy-fuel turbine based IGCC with multi-product production. A generalized technical approach to realize the 2010 goals is presented below in Table 1. Table 1. Generalized Technical Approach to 2010 Goal, and Potential Benefits Technical issue to pursue Combustor for 2 ppm NOx More durable catalysis for in combustor NOx formation prevention H2 Premixing with out flash back Higher turbine inlet temperatures (TIT) (~ 210 °F) Better TBC materials Enhanced cooling techniques Increase rotor torque limitation Compressor and air separation unit integration Ceramic parts Enhanced aerodynamics
Benefit to gas turbine or power plant Eliminates SCR and other penalties (NH4 slip, cost back pressure) Reduced O&M, makes catalytic combustion possible Enables low NOx combustion and related robust combustion techniques Approximately 1% pt. improvement to simple cycle per each ~ 70 °F increase Higher TIT, less air extraction, reduced RAM overall improvement in efficiency Higher TIT and less air extraction Higher power output with reduced capital cost (~ 20%) 0.5–1.0% points efficiency improvement Higher TIT Higher throughput & specific power
Table 2 below presents a list of current FE turbine projects that received funding in FY06. Following the table is a summary of individual projects. The University Turbine Systems Research (UTSR) Program is summarized in a separate section of the Handbook. Table 2 Active Turbine Program Projects That Received Funding in FY 06 Turbine Program Key Activities Hydrogen Turbines For FutureGen Advanced IGCC / H2 Gas Turbine Advanced Hydrogen Turbine for FutureGen Catalytic Combustion for Ultra-Low NOx Hydrogen Turbines Micro-mixing for Ultra-low emissions H2 / SYNGAS Combustion Catalytic Combustor for Fuel-Flexible Turbine System Study for Improved Gas Turbine Performance for IGCC Application Fuel Flexible Combustion System for Coproduction Plant Applications Syngas & Hydrogen Combustion Partial Oxidation GT for Coproduction from Coal in Industrial Applications Advanced Oxy-Fuel Turbines for FutureGen Zero Emissions Coal Syngas-Oxygen Turbo machinery Coal-Based Oxy-Fuel System Evaluation and Combustor Development. Advanced Research for FutureGen University Turbines Systems Research Systems Analysis of Advanced Brayton Cycles Turbine Materials & TBC for High Hydrogen Fuels HX in Hydrogen Fuel Turbines Novel Concepts for the Compression of Carbon Dioxide Super Sonic Shock Compression for Compression of CO2 Low-swirl Injectors for Hydrogen Gas Turbines
Contractor
Contract ID
General Electric Siemens Power Generation Precision Combustion Incorporated
42643 42644 42647
Parker Hannifin Siemens Power Generation General Electric General Electric NETL In House Turbine Support GTI
42648 41891 41889 41776 NETL 42649
Siemens Power Generation Clean Energy Systems
42646 42645
SCIES UC Irvine ORNL Ames Lab SwRI Ramgen LBNL
41431 42652 FEAA070 AL05205018 42650 42651 FWP678402
Summary of Active Turbine Program Projects Individual Turbine Program project descriptions provide insight into the breadth and depth of research being supported by DOE. The descriptions provide a detailed overview of turbine activities, and tie individual project goals to the larger national goal of energy security, which is attainable by using the nation’s most abundant fossil energy resource — coal. Project Summary: NETL In-House Combustion R&D in support of Turbine Program NETL’s Energy Systems Dynamics Focus Area is actively involved in a number of projects to support DOE’s Turbine Program. These projects include research in low-emissions combustion, model validation, sensor and controls development, and materials. Research in turbines combustion in focused primarily on development of hydrogen and oxy-fuel combustion approaches to meet Future Gen efficiency and emissions targets. Hydrogen combustion: To meet DOE FutureGen targets for zero carbon and low-NOx emissions, new combustion technologies will be required. To facilitate carbon capture, IGCC systems will remove carbon upstream of the turbine by shifting the syngas composition to produce a high-hydrogen-content fuel. Highhydrogen-content fuels will present some new challenges for combustor development. Lean-premix combustion strategies currently employed for natural gas (NG) fired engines will not be easily retrofit for high-hydrogen-content fuels due to the potential for flashback. Early IGCC systems will likely employ diffusion flame combustion systems for this reason, but diffusion flame combustors are not likely to meet FutureGen NOx targets of 2ppm. Solutions may come from the NETL Combustion Program, which has several projects focusing on assessing the flashback potential for fuels with a hydrogen content ranging from 20–100 percent. In addition, alternative combustion approaches such as trapped vortex combustion are being investigated, which have the potential for good flame stability and higher velocity flows to avoid flashback with premixed hydrogen flames. Research also is under way to assess the potential for dilute diffusion flame systems burning hydrogen. Oxygen-blown gasifiers will have nitrogen available from the air separation unit, and NETL is investigating approaches to using this nitrogen to dilute the hydrogen fuel, enabling both efficient, stable combustion and low-NOx emissions. Hybrid approaches employing dilute diffusion flame and partial premixing are also being investigated. These hydrogen combustion studies combine chemical kinetic and computational fluid dynamic (CFD) simulations, with laboratory and bench scale combustion testing at pressures up to 30 atm using NETL’s High Pressure Combustion research facilities. On a more fundamental level, NETL has an optically accessible, swirl stabilized combustor that is being used to develop validation data for Large Eddy Simulation and other advanced simulation methods. This activity is a multi-agency collaboration involving DOE, Department of Defense, National Aeronautics and Space Administration, and several major research Universities to develop and validate advanced simulation tools that are being used to design fuel-flexible combustion systems. Oxy-Fuel Combustion: NETL also is investigating fundamental issues associated with oxy-fuel combustion. Oxy-Fuel combustion systems are being considered for zero- emission power cycles where all of the carbon from the fuel can be captured. These systems will either use steam or carbon dioxide as a diluent to manage combustion temperatures. Thermodynamic and chemical kinetic modeling of the combustor indicates that CO2 dilution may result in unacceptable CO levels in the combustor effluent. NETL is examining combustion issues associated with high steam-loaded systems to develop the database necessary for oxy-fuel combustor design. Sensors and Controls: NETL is working on development of flame sensors and controls to improve emissions and stability of advanced turbine systems. NETL’s Combustion Control and Diagnostic Sensor (CCADS) is a flame ionization sensor with a demonstrated capability to measure flashback and lean blowoff for gas turbine combustors. Continued development of this sensor system now is focusing on measuring the fuel-air equivalence ratio, which will offer the potential for adjustment of fuel flows and lower NOx emissions.
Richard Dennis
Project Summary: Advanced Hydrogen Turbine for FutureGen (CID: 42644) Participant: Siemens Westinghouse Power Corporation Siemens Westinghouse Power Corporation (SWPC), with support from Florida Turbine Technologies, major universities, and others, intends to advance the state-of-the-art gas turbine for integration into a coal-based IGCC power plant that will be fueled with coal-derived hydrogen fuel and syngas. The project objectives will lead to significant advancements in IGCC plant efficiency, near-zero emissions, and a reduction in plant cost. The project will further develop and optimize integration of this advanced G-class gas turbine into an IGCC plant to ensure that DOE’s FutureGen Program objectives of plant efficiency, NOx emissions, and capability for CO2 sequestration are achieved. The proposed three-phase, 10-year SWPC project will begin with Phase I, which will focus on identifying and down-selecting the advanced technologies needed to achieve the challenging program goals, producing the required new component conceptual designs, and generating an R&D implementation plan. Phase II will entail development of new component detailed designs, and technology validation test programs. Engine and system fabrication, with deployment and testing in an IGCC plant, will be carried out in Phase III. SWPC will focus key development efforts on gas turbine combustion system, performance enhancements, and required materials/coatings advances. Combustion development will concentrate on advanced concepts evaluation, down-selection, and development to produce operational systems for burning coal derived hydrogen and syngas fuels, with natural gas burning capability as a back-up. To implement the new performance enhancing concepts, the program will include evaluating and down-selecting the most promising concepts for improving component efficiencies, enhanced cooling, and maintaining the turbine’s rated inlet temperature while operating on the above range of fuels. Materials/coatings selection and development will support the goal of higher efficiency, while supporting the extended fuel flexibility capability by targeting improvements in component durability and life cycle costs. Advancements in sensors and controls will be carried out to provide a capability for monitoring flame temperature, emissions, individual component metal temperature, coating durability, and turbine blade tip clearance control. Overall plant performance and economic optimization efforts of the SWPC program will lead to the effective integration of advanced G-class gas turbines into future low-emissions, coal-based IGCC plants, thereby ensuring that a cost-effective supply of electricity is available in the United States that uses our domestic coal resources. (DOE award: $45.5 million; plus contractor cost-share; project duration: 56 months, Phase III of this contract was not awarded) Project Summary: Advanced IGCC/Hydrogen Gas Turbine Development (CID: 42643) Participant: GE Energy GE Energy proposes a gas turbine development project entitled Advanced IGCC/Hydrogen Gas Turbine Development. This project will develop gas turbine technology for advanced IGCC and FutureGen power generation plants, to support DOE’s overall coal-based power generation goals of high efficiency (45–50% (HHV), near-zero emissions (<3 ppm NOx @ 15% O2), and competitive capital costs (<$1,000/kW). Gas turbine improvements in this program address DOE’s Turbine Program goals of 3–5 points of improvement in combined-cycle efficiency, less than 2 ppm NOx emissions, and compatibility with either traditional IGCC coal synthesis gas fuel or with high-hydrogen fuel produced from FutureGen type plants. The proposed GE project will leverage existing state-of-the-art gas turbine technology, while developing, validating, and prototype testing the technologies and systems needed to meet DOE’s goals. Emissions reductions for low-Btu and hydrogen fuels will be addressed through combustion technology advancements, with the goal of achieving the same type of emissions improvement DLN (dry-low-NOx) technology accomplished for natural gas fueled gas turbines. Based on previous technology program developments, combustion testing of promising combustion technology platforms and enablers will be performed and evaluated, leading to full-scale design and development of an optimum system. IGCC plant level efficiency improvements will be achieved primarily through increased IGCC gas turbine firing temperature, to the same levels as today’s natural gas fired gas turbines. This will be enabled through application of high-temperature 7FB gas turbine materials/design technology and development of technologies to allow increased turbine
5
6.0.1 The DOE Turbine Program: Overall Program Description
mass flow and output. Environmental testing of advanced high-temperature materials and coatings exposed to IGCC/hydrogen fuels will be performed, along with selective coating development/improvements. The GE technology advancements will start with R&D in Phase I, proceed to design and validation at the component level in Phase II, and result in prototype testing in a GE 7FB IGCC gas turbine in Phase III. (DOE award: $45.6 million; plus contractor cost-share, project duration: 75 months, Phase III of this contract was not awarded) Project Summary: Coal-Based Oxy-Fuel System Evaluation and Combustor Development (CID: 42645) Participant: Clean Energy Systems, Inc. Clean Energy Systems, Inc. (CES) will develop and demonstrate operation of its proprietary oxy-fuel combustor technology on syngas. Currently, CES is demonstrating a 20-MWt combustor on natural gas. CES’ next objective is to demonstrate its combustor technology on coal syngas, enabling zero-emission coal plants with higher heating value efficiencies between 50% and 60%, coupled with 100% carbon dioxide capture. CES will implement a research, design, development, and validation project with three phases, as prescribed by DOE. Phase I will include an R&D Implementation Plan and Conceptual Design. Part of the design process calls for a System Study consisting of definition of subsystem performance parameters, power system modeling, evaluation of alternative system configurations, and fuel variability evaluation. The deliverables from Phase I will be a report of optimized cycles and an oxy-coal syngas fuel combustor conceptual design. CES will team with SWPC, and the design progression will be conducted in collaboration to assure that the final combustor/turbine product is viable, useful, and optimum for both available and next-generation turbine hardware. Similarly, CES will team with other major subsystem suppliers and consultants, including ConocoPhillips for gasification, Air Products and/or Air Liquide for air separation, Kinder Morgan CO2 Company for carbon dioxide systems, G.C. Broach Company for heat recovery steam generators, and Western Research Institute for computer modeling. CES will employ existing assets to the maximum extent possible as it develops the oxy-syngas combustor. In particular, CES will use the previous cycle studies undertaken by DOE/NETL, Lawrence Livermore National Laboratory (LLNL), and numerous other parties; as well as using the existing Kimberlina Power Plant demonstration facility. The Kimberlina facility will be evaluated to confirm that, with reasonable modification, it can be configured to burn simulated coal syngas from tube trailer supplies of mixtures of hydrogen, carbon monoxide, and methane. This approach will allow the introduction of various syngas compositions, which will validate operation of the combustor over a wide fuel range. System studies will include air separation units, turbine types, gasification methods, and syngas cleanup methods and requirements. CES will work with suppliers and subcontractors to obtain the various subsystem parameters. System modeling and subsystem integration opportunities will be evaluated by CES and outside contractor(s) using the Aspen Plus® software program. Phase II of the CES program will entail developing a detailed design and conducting validation testing, to allow a pre-commercial combustor to be fabricated and tested in Phase III. CES will incorporate information and knowledge gained from SWPC, including operating states, combustor size, and combustor configuration (single or multiple combustors). Operating state flexibility is one of the merits of the CES combustor, and its design can be initiated prior to knowing precise desired operating parameters, though nominal guidelines from the turbine manufacturer will be incorporated. Phase III will involve fabricating an appropriately sized pre-commercial prototype combustor and conducting longer-term testing using actual and/or simulated coal syngas. Testing at this stage of the product development will be conducted at the Kimberlina Power Plant using blended syngas, and then transition to a site with actual syngas if an appropriate site is available. (DOE award: $4.5 million; plus contractor cost-share, project duration: 39 months) Project Summary: Zero Emissions Coal Syngas-Oxygen Turbo Machinery (CID: 42646)
6
Richard Dennis
Participant: Siemens Westinghouse Power Corporation Siemens Westinghouse Power Corporation (SWPC), with support of Clean Energy Systems (CES), Florida Turbine Technologies, a major university, and others, intends to propose a multi-phase project for research and development of turbines and related systems to utilize high-hydrogen fuels derived from coal. Activities will include development of a turbine for an Oxy-Fuel Rankine Cycle System that would be integrated into a highly efficient, near-zero emission power plant. The focus on fuel flexibility through combustion with oxygen syngas is seen as key to continued use of coal, our largest domestic fossil energy resource, coupled with capture of CO2 and all of the Clean Air Act criteria pollutants. Phase I will initially review cycle optimization based on previous work on a limited number of system studies to assess likely operating conditions of the turbines. This will require that both the combustion part of the cycle along with the secondary part of the cycle be optimized together to select the most optimal pressure and temperatures in the bottoming cycle. In order to evaluate the range of cycle options, three cycles will be considered. The baseline cycle will be based on the ultra-supercritical steam cycle, in which turbo machinery designs can be developed using materials from current industrial gas turbine frames. A second cycle will be developed by moving beyond the current ultra supercritical designs, emphasizing realistic near-term achievements with an acceptable increase in risk. Finally, a high-efficiency, higher risk cycle will be developed incorporating concepts from the latest advanced gas turbine frames. Conceptual designs of equipment for selected cycles will follow to identify total plant costs and technology challenges. Critical component identification and cycle selection will lead to a more specific cycle with proposed cost and schedule at the end of Phase I. The required R&D that will be conducted in Phase II and Phase III will also be identified as part of Phase I of the program. The Phase II of this SWPC project will involve the detailed design of the components and selected material development, both of which are required to support component development. A significant effort will address the challenges of how a working fluid, composed mainly of H2O and CO2, impacts rotating components, as well as the associated material issues such as stress corrosion, general corrosion/erosion, creep effects, and thermal mechanical fatigue. It is expected that material development will not only be required for major components, but also for surface engineering and innovative cooling schemes so that the turbo machinery can withstand the elevated temperatures required for the coal Syngas-Oxygen turbo machinery. Phase III will involve prototype testing of certain sub-components, and a scale-model test of a steam turbine component. Fabrication and testing will include the specialized components such as rotating blades and other stationary components critical to the overall performance of the power plant. It is expected that by the end of successful completion of Phase III, feasibility of this type of high-efficiency, zero-emission cycle will be demonstrated. DOE award: $14.5 million; plus contractor cost-share, project duration: 56 months, Phase III of this contract was not awarded Project Summary: Catalytic Combustion for Ultra-Low NOx Hydrogen Turbines (CID: 42647) Participant: Precision Combustion, Inc. (PCI) The proposed PCI project will develop and demonstrate an ultra-low-NOx rich catalytic combustion system for fuel-flexible hydrogen combustors in megawatt-scale turbines. This will further develop PCI’s rich catalytic combustion technology for fuel flexible hydrogen application, in collaboration with Solar Turbines, and provide a roadmap to commercialization of the technology across all size ranges of power generation turbines. In a current DOE program, this technology has demonstrated subscale ultra-low-NOx with syngas and with hydrogen diluted with nitrogen (low single-digit NOx corrected to 15% O2 with operation at IGCC base load combustor temperatures and 10 atm. pressure). The technology offers low single-digit NOx emissions, even with hydrogen as the only fuel; fuel flexibility for similar low emissions of syngas or natural gas; and the potential to support increased firing temperature (and efficiency) while maintaining low emissions. The benefits include combustors capable of delivering near-zero NOx without costly postcombustion controls and without the need for added sulfur control. This advances DOE objectives for
7
6.0.1 The DOE Turbine Program: Overall Program Description
achievement of low single-digit NOx emissions, improvement in efficiency versus post-combustion controls, fuel flexibility, a significant net reduction in IGCC system net capital and operating costs, and a route to commercialization across the power generation field. In the proposed project, PCI will develop the technology for fuel flexible use of hydrogen in a megawatt-scale combustor through design and analysis (CFD, Chemkin), and sub-scale, mid-scale, and full-scale testing. The work plan is in three phases. Phase I involves development of conceptual designs for catalytic combustion technology for hydrogen fuel, and an R&D implementation plan including supporting analysis, fabrication, and testing of several small to intermediate scale components. Phase II, Detailed Design and Validation Test Program, concentrates on development of multiple full-scale modules for validation testing, to include full pressure testing and resolution of key issues related to startup, load shifting, turndown, shutdown, module interactions, and system design. In Phase III, the full-scale design will be frozen, and a full combustor system with multiple catalytic combustor modules will be fabricated for an initial engine rig testing. This will be followed by either engine loop test using the full-scale engine hardware at mid-pressure, or full-scale engine testing. Solar Turbines, the world's largest manufacturer of mid-range industrial gas turbines (1–15 MW), will be an active participant, developing combustor design and hardware for its engines as well as design, testing, and system-level interactions with PCI. Solar and PCI have an established interactive relationship, with a current DOE-supported engine trial program now under way for natural gas-fired catalytic combustion. The new focus on the smaller machine will facilitate more rapid yet economic combustor development targeted to an engine 5- to 15-MW size that may be considered a building block for larger turbine applications. Because the catalytic system is scalable and modular in nature, integration to larger engines can be facilitated. (DOE award: $4.9 million; plus contractor cost-share, project duration: 60 months) Project Summary: Micro-mixing Lean Premix System for Ultra-LOW Emission Hydrogen/Syngas Combustion (CID: 42648) Participant: Parker Hannifin Corporation, Gas Turbine Fuel Systems Division The general focus of this project is to develop the next generation of environmentally friendly, hydrogen/syngas, gas turbine fuel injection technologies. Parker intends to prototype and test innovative, multi-point fuel injector technologies that satisfy DOE’s objectives of reducing NOx emissions to 2 ppm. Detailed studies and experimentation with these injectors are proposed to elucidate the effects of various operating parameters on overall turbine performance. The impact of nozzle design and operating conditions on combustion efficiency, emissions, and lean stability will be characterized. Burner technologies will be developed for lean and ultra-lean premixed hydrogen/syngas combustion in combustor geometries similar to those used in gas turbine engines, and with compositions similar to those obtained in coal gasification plants. Parker will investigate the impact that hydrogen content in syngas has on flashback and emission characteristics in lean premixed combustion systems, and will develop strategies to mitigate the impact of flashback and auto-ignition. Through a university partner program, data on flammability limits, stability characteristics, laminar and turbulent flame propagation, as well as the impact of the anchoring mechanism, burning conditions and syngas composition, will be collected and synthesized into models. Diagnostics, corroborated with computational analyses, will be used to determine the role of chemistry, and transport and fluid mechanics in the mechanisms of combustion. Starting from already proven, Macrolamination technology, Parker’s general approach is to adapt the proven designs and concepts to hydrogen/syngas combustion and hydrogen enriched combustion. With Macrolamination technology, elegant and sophisticated multi-point lean-premix nozzles and burners can be developed with exceptional affordability. Parker proposes to develop, build, and test a large number of burners spanning a wide range of sizes, from small-scale single-cup premixers to 1-Megawatt size premixers. The modularity of the macrolamination approach affords the flexibility to build multiple scales of these injectors from a basic building block (a single mixing cup) affordably and expeditiously.
8
Richard Dennis
The combination of lean conditions and multiple point injection (for fast and efficient mixing) will be the primary vehicle for achieving low-NOx emissions in this project. In order to further reduce NOx levels, consideration will be given to operability at ultra-lean conditions, and designs that mitigate lean-combustion instabilities will be developed. Stability augmentation will be achieved through optimization of swirl and other aerodynamic features. Zone staging will also be used to enhance lean operability. Parker has assembled a seasoned team to perform the proposed scope of work, including industrial participants (Parker Hannifin and Solar Turbines), and a University partner (University of California Irvine). Parker Hannifin will lead the project and the other participants will be subcontractors to Parker. The team believes that the nozzles and combustor development tasks defined in this project will serve national interests by helping to maintain U.S. leadership in the gas turbine market for power generation applications. (DOE award: $1.2 million; plus contractor cost-share, project duration: 32 months). Project Summary: Partial Oxidation Gas Turbine for Power and Hydrogen Co-Production from Coal Derived Fuel in Industrial Applications Participant: Gas Technology Institute (GTI) The objective of this project is to provide a detailed assessment and evaluation of the feasibility, opportunities, and challenges of using MW-scale Partial Oxidation Gas Turbines (POGT) for coal-based co-production of electricity and hydrogen or syngas for steel, forest and paper, oil refinery, food, and other industries. The feasibility and performance assessments will be conducted for turbine-based plants that are integrated and optimized to provide high efficiencies, ultra-low emissions of criteria pollutants (2 ppm NOx), and reduced costs. This assessment and evaluation project will build upon the existing POGT technology that has been under development by GTI since 1995, and can be effectively combined with a coal gasifier. The POGT operates under reducing conditions as a combustion gas turbine — generating power through the partial oxidation of the gasifier product gas, and achieving highly efficient extraction of both thermal and pressure energy from the partially oxidized stream. Because of the partial oxidation reactions, the POGT also acts as a fuel reformer to convert hydrocarbons that are present in gasifier product gas into hydrogen-rich syngas. The ultra-low NOx emissions are achieved because the oxygen-deficient atmosphere suppresses NOx formation and converts the NH3 and HCN present in the gasifier product gas into N2. The POGT uses a smaller air compressor than an equivalent conventional gas turbine. All of these factors, combined with system integration benefits, will provide significant cost reductions for industrial applications. In this project, the Team will conduct detailed techno-economic and engineering assessments of a plant consisting of a coal gasifier, a POGT, and a hydrogen purification unit, with emphasis on the POGT. The POGT evaluation will be based on the analytical, experimental, and modeling results from ongoing GTI projects. The engineering evaluation of modifications needed to convert a conventional gas turbine to POGT will be based on, but not limited to, estimates by two leading turbine manufacturers (Solar and SWPC) for converting their own product lines and future planned products. A comprehensive market evaluation will be conducted to define the specifics and applicability of the proposed system in different industrial segments. System configurations will be chosen for specific applications and required co-products. These selected configurations will be optimized to provide the best achievable energy efficiency, and lowest emissions. The major result of the project will be an R&D Implementation Plan, cost, and schedule to bring the technology to commercialization, and an R&D Plan for modification of one or more existing gas turbines to a POGT will be developed and reported to DOE. Large industrial users (in particular, the steel, glass, forest and paper, oil refinery, and food industries) will directly benefit from this project. GTI anticipates that this new technology, providing a single on-site source of co-products from coal (electricity, syngas, or hydrogen), will provide customers with reduced product costs and improved efficiencies. GTI will lead a team that includes Solar Turbines Incorporated (Solar) and Siemens Westinghouse Power Corporation (SWPC). These gas turbine manufacturers command a large share of the current world turbine market, where Solar’s share of the 1- to 30-MW gas turbine market is the largest of all manufacturers. SWPC
9
6.0.1 The DOE Turbine Program: Overall Program Description
is one of the largest worldwide vendors of gas turbine technology in the 30- to 100-MW range. The team will also include Oak Ridge National Lab to assist with material studies for POGT components, and Georgia Tech University as a major U.S. expert in basic combustion science and flame stability. This team has the experience, the resources, and the will to bring this new technology to the industrial marketplace quickly and effectively. The U.S. energy markets, overall U.S. economy, and U.S. taxpayers will benefit from the project due to a wider use of domestically available coal for industrial energy needs, which will replace imported premium fuels. (DOE award: $999,992; plus contractor cost-share, project duration: 22 months) Project Summary: Super Sonic Shock Compression for the Efficient Compression of Large Volumes of Carbon Dioxide (CID: 42651) Participant: Ramgen Power Systems, Inc. Ramgen Power Systems is applying its super sonic shock wave compression technology toward the efficient and cost-effective compression of large quantities of CO2 for sequestration. Ramgen will design, validate, fabricate, and test a 100:1 pressure ratio, two-stage super-sonic compressor through three phases and 5 years of development. The development plan has numerous design reviews, risk assessments and go/no-go decision points. At the conclusion of Phase III, the sub-scale pre-commercial test unit will be the basis for a full-scale optimization and test program. The benefits of Ramgen’s technology approach are: fewer stages, higher efficiency, lower part count and therefore lower capital and maintenance costs, and smaller size for comparable mass flow and pressure ratio. Shock compression technology has the potential to simultaneously develop a very high compression ratio per stage, and very high efficiency. Shock compression is affected by the mole weight of the gas. Since CO2 is heavier than air, Ramgen’s shock compression approach benefits from the low speed of sound characteristic. Conversely, conventional compressors are at a disadvantage with the heavy CO2 gas, because shocks are bad for performance in a conventional compressor. This allows Ramgen to build a 2-stage CO2 shock compressor for a pressure ratio of 100:1, while conventional technology will typically require six stages of compression. The efficiency of the shock compression system will be at least as good as conventional approaches. In addition, development of Ramgen’s compression technology is cross-cutting and capable of delivering benefits to many of the technical areas of concern in zero-emission clean coal facilities. These benefits include high-efficiency electric and fuel-fired air compressors to reduce the significant operating and capital cost of the supporting Air Separation Unit (ASU) subsystem, both for cryogenic and Ionic Transport Membrane (ITM) technologies. (DOE award: $11 million; plus contractor cost-share, project duration: 60 months) Project Summary: Novel Concepts for the Compression of Large Volumes of Carbon Dioxide (CID: 42650) Participant: Southwest Research Institute In the effort to reduce the release of CO2 greenhouse gases to the atmosphere, sequestration of CO2 from IGCC and oxy-fuel power plants is being proposed. This approach, however, requires significant compression power to boost the pressure of CO2 to typical pipeline levels. The penalty can be as high as 8–12% on a typical IGCC plant. The goal of this project is to reduce this penalty through novel compression concepts and integration with existing IGCC processes. The primary objective of this project is to boost the pressure of CO2 to pipeline pressures with the minimal amount of energy required. First, fundamental thermodynamics will be studied to explore whether pressure increases in liquid or gaseous states would be preferred. Since the first phase of the project involves conceptual brainstorming, flexibility has been built into the project to permit investigation of several concepts. For gaseous compression, the project seeks to develop novel methods to compress CO2 while removing the heat of compression that is internal to the compressor. The high pressure ratio compression of CO2 results in significant heat of compression. Since less energy is required to boost the pressure of a cool gas, both upstream and inter-stage cooling is desirable. While isothermal compression has been utilized in some services, it has not been optimized for the IGCC environment. This project will determine the optimum compressor configuration and develop technology for internal heat removal. Furthermore, other process
10
Richard Dennis
streams within the IGCC environment will be utilized to provide a total system solution by fully integrating the air separation units, combined cycle, and the gas cleanup system. Other concepts that liquefy the CO2 and boost pressure through cryogenic pumping will be explored as well. Phase I will identify the concept that best meets the efficiency goals and integrates into the IGCC environment. Based on the selected concept, Phase II will design the optimum solution and perform prototype development testing. Phase III will apply a full-scale compression solution to an existing IGCC plant. This project is being co-funded by Dresser-Rand Company. (DOE award: $175,033; plus contractor cost-share, project duration: 12 months, Phase II and III were not awarded). Project Summary: Systems Analyses of Advanced Brayton Cycles for High Efficiency Zero-Emission Plants (CID: 42652) Participant: Advanced Power and Energy Program (APEP), University of California at Irvine The FutureGen plant concept is aimed at reducing the environmental impacts of fossil fuel usage while generating electric power and providing a clean fuel for transportation and for distributed power generation. Developing turbine technology to operate on coal-derived synthesis gas and hydrogen is critical to the development of advanced power generation technologies and the deployment of FutureGen plants. The FutureGen plant concept may also be deployed in natural gas based plants with respect to generating power with near-zero emissions, while utilizing these advanced Brayton cycle machines and securing fuel diversity. This APEP project therefore represents a key investment in implementing the FutureGen concept, and in helping to secure clean, efficient, affordable and fuel-flexible electric power generation for the U.S. As with the other turbine projects, APEP also will help make possible the continued use of our nation’s largest domestic fossil energy resource, coal. Numerous projections estimate that gas turbines will comprise a significant portion of the required generation capacity in the 21st century. Novel advanced gas turbine cycle modifications, intended to improve the basic Brayton cycle performance and reduce pollutant emissions, are currently under development or being investigated by gas turbine manufacturers and R&D organizations. Preliminary conceptual analyses of advanced cycles indicate that it may be possible to achieve an improved combination of efficiency, emissions, and specific power output, which in turn should reduce the power generation equipment cost on a $/kW basis. Thus, a need exists to evaluate advanced Brayton cycles and identify the best opportunities worthy of support by DOE for their development, and to assess their R&D needs and the most likely commercialization path. APEP will focus this study on defining advanced Brayton cycles and addressing the key technologies needed to enable development of such advanced turbines and turbine-based systems that will operate cleanly and efficiently when fueled with coal-derived synthesis gas, hydrogen fuels, and natural gas. System integration issues will be addressed that will allow the combination of high-performance technology modules and subsystems into safe, reliable, environmentally friendly, and economic power plants. Specifically, the project will develop concept(s) and present approach(es) that will increase the state-of-theart Brayton cycle (in a combined-cycle application) from today’s 58–60% efficiency (LHV on natural gas) to >65% equivalent efficiency. The proposed machine(s) will consider integration into advanced coal-based and natural gas-based zero-emission systems, with the ability to attain a 60% (HHV coal) efficiency and 75% (LHV natural gas) efficiency respectively (prior to carbon separation and capture). Options for zero CO2 emissions will be considered for both coal- and natural gas-based plants, and will show how the turbine design, operation, and overall system performance are affected. The integration of subsystem technologies such as advanced gasifiers, membrane technology for air and H2 separation, and fuel cells as they evolve, will be accounted for in the advanced Brayton cycle design(s), while performance of the resulting integrated advanced systems will be quantified. Start-up, shutdown, and off-design operating needs will be taken into account while configuring the advanced cycles. (DOE award: $603,012; Plus contractor cost-share, project duration: 24 months) Project Summary: Catalytic Combustion for Fuel Flexible Turbine (CID: 41891) Participant: Siemens Westinghouse Power Corp.
11
6.0.1 The DOE Turbine Program: Overall Program Description
Under the sponsorship of NETL, a team of organizations led by Siemens Westinghouse Power Corporation (SWPC) proposes a 3-year R&D program entitled “Catalytic Combustor for Fuel Flexible Turbines.” In this program, the team will develop and demonstrate a cost effective catalytic-based turbine combustor that will achieve the aggressive target of 2 parts per million NOx emissions at the turbine exhaust without the need for expensive back-end after treatment systems currently employed in gas turbine combined cycle generating plants. The catalytic combustor will be suitable both for retrofit into the installed base of operating turbines and also for deployment in the latest generation of advanced, high firing-temperature turbines, while achieving the low emissions objective at even the high firing temperatures of advanced turbines. The combustor will support fuel-flexible power generating facilities, with equal performance capabilities when operating on either conventional natural gas fuels or on synthetic fuels derived from coal. The program supports objectives of highly efficient, environmentally friendly power generating plants operating on our nation’s abundant resource of coal reserves. The program culminates in the demonstration of the combustor on syngas at the Power Systems Development Facility in Wilsonville, Alabama. SWPC has teamed with Solar Turbines, Penn State University, and Southern Company Services in the pursuit of the program objectives. (DOE award: $6,998,071; Plus contractor cost-share, project duration: 45 months) Summary: Heat Transfer in Advanced Hydrogen Fueled and Oxy-fuel Turbines Participant: Ames Laboratory/Iowa State University The purpose of this project is to analyze gas turbine thermal performance with a variety of new fuels, and optimize heat transfer within the turbine. Initially, the work will focus on cooling needs of turbines with new fuels, and then work toward a system-based understanding of turbine performance. Developing turbine technologies to operate on coal-derived synthesis gas (syngas) and hydrogen fuels is critical to the development of advanced power generation technologies such as integrated gasification combined cycle (IGCC) and the deployment of FutureGen type power plants that can lead to the capture and separation of carbon dioxide (CO2). The goal of this project is to develop an analysis tool that can be used to examine and explore heat transfer design and operation issues in turbine components to support the development of turbine technologies used in advanced coal-based power systems. The tool will consider heat transfer from the hot gases, thermal barrier coating systems that protect the superalloys, and cooling strategy as a function of: 1. Fuel used (natural gas, syngas with different ratio of CO/H2, H2) 2. Firing temperature and turbine inlet temperature that account for convective and radiation heat transfer to the combustor walls and combustion kinetics 3. Amount of diluents and the diluents used (e.g., N2, steam, or CO2 to control NOx formation) 4. Water steam content and different ratio of CO2/H2O in the working fluid 5. Mass flow rate 6. Cooling strategy with and without turbine blade cooling 7. Different thermo cycles (e.g., coal-based oxy-fuel Rankine cycles or advanced Brayton cycles). As a first step for this project, the primary goal is to conduct a thorough literature survey on what has been done and assess available modeling methods and codes. This includes: 1. Available studies on applications of gas turbine (originally designed for natural gas) fueled with syngas, and reports on existing gas turbine fueled with syngas in IGCC power plants. The focus is on thermal management, heat transfer, and failure mode. 2. Studies on H2 fuel combustion and its effect on turbine heat transfer. 3. Studies on syngas fuel combustion with CO2 sequestration and its effect on turbine heat transfer 4. Studies on different working fluid (e.g., CO2, CO2 plus water vapor, or water vapor only) 5. Assess available turbine design tools.
12
Richard Dennis
IGCC Turbine Issues • Turbine inlet temperature. So far, power generation gas turbines have been designed for the utilization of natural gas fuel. When they are used in IGCC applications, i.e., using syngas as fuel, the machines are derated in firing temperature to accommodate long term operational issues associated with excessive temperature of materials in the hot gas path. This temperature reduction is believed to be on the order of 200–300_°F for the current F-frame machines when compare to the same machine fired on natural gas. This temperature reduction has a directed affect on system efficiency. With the increase of turbine inlet temperatures, material degradation issues will be evaluated and improvements to thermal barrier coatings explored. • Heat flux increase. Besides the firing temperature, the heat flux conditions can also affect the material temperatures. For the combustion of the coal derived syngas, oxidants and diluents determine heat flux conditions at critical hot gas path locations. Heat flux can increase depending on the combustion by products and the diluent used to control NOx emissions. Higher heat flux conditions gives rise to higher material temperatures. • Mass flow increase. Furthermore, commercially available gas turbines have been developed for the use of natural gas, i.e., a fuel with high calorific value (LHV). With these turbines adapted to the use of syngas, a low LHV fuel, the gas turbine encounters two major changes: 1. For the same fuel heat input, the fuel mass flow is several times greater than for natural gas, due to the lower LHV. 2. Diffusion burners are used with syngas, and control of NOx is achieved by diluting the syngas with nitrogen, steam or carbon dioxide. These two factors increase substantially the overall mass flow through the turbine. The higher mass flow coupled with rotor torque limitations results in higher average temperature profiles at individual turbine stages. Due to this situation, last stage blades may experience temperatures higher than the original design specification. It is also believed that this higher mass flow and associated volume increase leads to higher local velocities and higher local heat transfer coefficients. The following factors will be considered in the heat transfer analysis tool development: (1) fuel used (e.g., natural gas, syngas (including different ratio of CO/H2), H2; (2) firing temperature; (3) amount of diluents and diluents used (e.g., N2, water vapor, or CO2 to control NOx formation); (4) amount of water and different ratios of CO2/water in the working fluid; (5) mass flow rate; (6) different cooling strategies with and without a thermal barrier coating; and (7) different thermo cycles (e.g., coal based oxy-fuel Rankine cycles and advanced Brayton cycles). Ames will conduct a thorough survey on what has been done so far, and available modeling methods and codes will be assessed. This study will include: (1) available studies on applications of gas turbine (originally designed for natural gas) fueled with syngas, and reports on existing gas turbine fueled with syngas in IGCC power plants, the focus is on thermal management, heat transfer and failure mode, (2) studies on H2 fuel combustion and its effect on turbine heat transfer, (3) studies on syngas fuel combustion with CO2 sequestration and its effect on turbine heat transfer, (4) studies on different working fluid, i.e., CO2, CO2+water vapor, or water steam only, and (5) an assessment of available turbine design methods, codes and tools. For this project, Ames will work closely with engineers and researchers at NETL and at the Oak Ridge National Laboratory to ensure that the work is relevant and compliments other related activities on-going at those laboratories. Project Summary: Material Issues in Coal-Derived Synthesis Gas/Hydrogen-Fired Turbines Participant: Oak Ridge National Laboratory Large gas turbines (i.e., about 250 MWe) firing on natural gas have been operating in combined-cycle systems since the mid 1990s, providing around 400 MWe of power output with efficiencies in excess of 55% in many cases, and with very low NOx and SOx emissions. Also, advanced concepts were evaluated to give efficiencies of up to 60%. With the rising cost of natural gas in recent years (the price generally tracks that of oil), attention has turned to the opportunities associated with the production of gas from coal (and from other feedstocks, including waste products). One consequence has been an interest in improving the technology of
13
6.0.1 The DOE Turbine Program: Overall Program Description
gas turbines burning low-Btu gases, with the aim of building on the advances made with the current generation of engines to improve generating efficiency with IGCC plant. In the United States, the FutureGen Program is specifically focused on producing electricity, hydrogen, or chemical feedstocks via the gasification of coal; in the power generation mode, gas turbines will be required to deliver efficiencies comparable to the machines resulting from the U.S. Advanced Turbine Systems (ATS) Program (1), and NOx levels < 5 ppm, while burning coal-derived syngas and / or hydrogen (2). Not unexpectedly, there are specific problems associated with the combustion of low-Btu gas, depending on its source, which will require additional development efforts, both in specific aspects of turbine design, and in materials performance, in order to provide cost-effective solutions. The objectives of this project are to provide materials guidelines for the reliable operation of gas turbines when fired with syngas and H2-enriched fuel gases, in terms of firing temperature and fuel impurity levels (water vapor content; sulfur; condensable species). The research effort in place aims to provide underpinning understanding needed in the consideration of materials issues associated with these new operating conditions. The intended outputs of this project are: • understanding of the factors limiting the firing temperatures of syngas turbines; • assessment of the potential for deposition, erosion, or corrosion (D-E-C) when firing syngas; and • evaluation of approaches for improved coatings to provide the basis for more robust hot gas path components. Current activities involve the development of a plan for addressing the overall materials and manufacturing needs of syngas-fired turbines. Input for this plan is being obtained from a detailed review of published literature concerning issues confronted in the combined cycle operation of gas turbines, with emphasis on design and operating changes necessitated to allow operation on fuels other than natural gas; cycle analyses that address the trade-off issues associated with optimizing the combined gas turbine and steam generation system; as well as practical experience (where this is available and accessible). Reports from international conferences, from demonstration programs in the U.S., Europe, and Japan, and from operating IGCC plants in particular has been sought and critically reviewed. A report summarizing the outstanding materials and manufacturing issues is being compiled, and the views of the GT manufacturers and materials suppliers and other specialists on suggested priorities are being sought and incorporated. This preliminary listing of materials needs and priorities will be tested at a workshop of turbine materials specialists, and the findings of this workshop will be incorporated into the final draft of the Materials Needs Report. This report is intended to summarize available information concerning the critical materials issues resulting from the desire to increase the efficiency of operation of gas turbines applied to power generation and, in particular, to achieve high efficiencies (and reduced emissions) with turbines fired by syngas and/or hydrogen derived from coal. The effort has involved a review of published information from the U.S., Europe, and Japan, including input from various current major programs (where available) which are mainly focused on the materials needs for advanced, natural gas-fired turbines, as well as an attempt to understand differences that arise from adaptation of these technologies to firing the coal derived fuels of interest. Since there is little published information concerning changes in design or materials needed because of specific influences of alternative fuels on the performance of gas turbines, contacts have been made with key organizations involved in pilot/demonstration IGCC projects to obtain reports and/or first-hand information, and visits have been made (or are planned) to the major U.S. gas turbine manufacturers. It was considered particularly important to initiate an interaction with the General Electric Company, because of its activities (including the recent acquisition of the Texaco/Chevron gasification technology) intended to position the company as a leader in supplying complete IGCC plants. Project Summary: Low-swirl Injectors for Hydrogen Gas Turbines in FutureGen Power Plants Project Participant: Lawrence Berkeley National Laboratory This goal of this research is to develop a robust ultra-low emission combustor for the gas turbines in FutureGen power plants that burn hydrogen derived from gasification of coal. The objective is to adapt lowswirl combustion (LSC) to these utility size turbines. LSC is a dry-low-NOx method conceived at LBNL. Under DOE-EERE, this technology has been commercialized for industrial heaters by Maxon Corp. of Muncie, Indiana. DOE Office of Electricity is supporting its development for natural gas and fuel-flexible
14
Richard Dennis
industrial turbines (5 – 7 MW) in partnership with Solar Turbines of San Diego, CA. The California Energy Commission is supporting a combined heat and power project that includes the development of LSC for Elliott Energy Systems’ 100 kW microturbine. This research leverages the knowledge and experience gained from these R&D activities. FutureGen power plants produce hydrogen which is separated from a concentrated CO2 stream that is then captured for subsequent sequestration. One of its key components is a cost-competitive all-hydrogen fueled turbine with ultra low NOx emission and high efficiency. To lower NOx, the current approach is to operate the H2 turbine at lower firing temperatures in combination with selective catalytic reductions (SCR). This approach sacrifices efficiency and impacts costs of electricity (via capital cost, efficiency and capacity output). Therefore, a cost-effective combustion technology that meets the FurtureGen emissions and efficiency targets is critical to achieving its ultimate goal of no more than a 10% increase in cost of electricity for mature FutureGen type plants that include CO2 capture and sequestration. Preliminary laboratory studies have shown that LSC has good promise to be an effective enabling technology for the H2 turbine to meet the FutureGen goals of 2 ppm NOx (@ 15% O2) at a firing temperature of 2500to 2600F. LSC is a sophisticated yet simple and very cost effective combustion technology that can operate with a variety of gaseous fuels including H2 fuel blends under a broad range of inlet and outlet conditions. As one of the components of a complex and fully-integrated FurtureGen power plant, the H2 turbines have to be reliable and sufficiently flexible and adaptable to meet the inlet and outlet requirements without compromising electricity output efficiency and emissions. With LSC, the H2 turbine will have greater flexibility in their operations than is achievable by current technology. Greater flexibility provides more options for developing a power plant scheme that offers an optimum balance between efficiency, reliability, emissions and costs through intelligent integration of technologies including gasification, separation technologies, combustion turbines, and steam turbines without the need to invoke SCR for exhaust gas cleanup. The feasibility of burning H2 in a LSI has been demonstrated in a recent laboratory study of the fuel effects on LSI flow fields and flame characteristics [3]. The fuels tested in this study consist of seven diluted and undiluted hydrocarbon mixtures, pure H2 and a fuel mixture consisting of 50% H2 and 50% CO2. The lean blowoff limits for the two H2 fuels are found to be less than 0.2 and are close to the theoretical flammability limit. These results demonstrate the LSI’s capability to support ultra-lean premixed turbulent flames with H2 fuel mixtures. Within the velocity range afforded by our experimental setup (3 < U < 9 m/s) intermittent attachment of the H2 flame to the LSI rim occur at greater than 0.3 showing that the high diffusivity of H2 can lead to phenomena that are unique to H2 firing. However, the significant conclusion of this study is that the NOx emissions from the hydrocarbon fuels depends primarily on the adiabatic flame temperature set by the fuel air equivalence ratio. For the very lean H2 flames (< 0.3), the NOx emissions are below the detectable limit of our instrumentation (0-5 ppm). This study demonstrates that the LSI concept is amenable to a very wide range of gaseous fuel mixtures. The reason is due to the self-similarity of the LSI flow field and its linear coupling with the turbulent flame speed. By invoking an analytical equation for the flame position that involves the self-similarity parameters, the turbulent flame speed and turbulence intensity, we obtain a theoretical proof on why the LSI enables the lifted flame to remain stationary throughout a very wide range of velocities from 5 to 80 m/s. The analytic model based on this equation [4] also shows that the higher H2 flame speeds are not expected to cause a significant change in the overall behaviors of the flame and the LSI flow field. It indicates that the first order effect of switching from hydrocarbon to H2 is associated with a change in the correlation of its turbulent flame speed with the turbulence intensity. The change can be accommodated by adjusting the swirl number of the LSI. The second order effects will be associated with heat release and higher H2 diffusivity. Therefore, the knowledge and insights gain from this study and the analytical model grounded on fluid mechanics and turbulent combustion theories will be useful for guiding the developmental effort to optimize the LSI for FutureGen turbines. FY06 Define H2 LSI specifications and develop a skeletal R&D plan As in our prior technology developments, a close partnership with gas turbine OEM is critical to the success of this research. In FY06, the initial step is to discuss with their combustion engineers to obtain a preliminary set of specifications for the configuration as well as the operational and performance criteria of their H2 turbines being developed for FutureGen. These include information on the geometric arrangement, the
15
6.0.1 The DOE Turbine Program: Overall Program Description
physical size, the form factor and the number of injectors and the size and shape and size of the combustion chamber, the anticipated fuel compositions (i.e. H2 and the anticipated concentrations of N2 as a diluent) the firing temperatures, firing pressures, firing rates for the injectors, and exit temperature etc. The outcome will be an assessment of the compatibility of LSI with their hardware and a skeletal research and development plan that highlights the critical issues as well as the pathways, and the roles and responsibilities toward resolving them. In parallel to the discussion with OEMs, laboratory studies will be conducted to obtain the basic information needed to optimize the LSI for H2. These experiments will include the measurement and correlation of the turbulent flame speed for diluted H2 fuels as well as simulated syngases. Our air flow supply systems will be upgraded to enable experiments at velocities up to 20 m/s. Operating the LSI at higher velocities will allow us to better understand the second order effects such as intermittent flame attachments and heat release. FY07 Design and fabricate H2 LSI prototype For H2 turbines, the challenges in achieving ultra-low emissions while balancing the tradeoffs between efficiency, complexity, reliability and costs are similar to those for natural gas turbines. The design of the H2 fueled LSI, however, needs to address additional specific issues concerning with the high flame temperatures, faster flame speeds, auto-ignition risk, shorter premixing, preferential diffusion, and inherent H2 flame instability. Our approach in FY07 is to follow the development pathway for natural gas turbines by applying our understanding of the LSC principle to optimize the LSI for burning lean H2 flames in a configuration that is compatible with the proposed H2 combustor. In parallel, laboratory and analytical studies will continue and focus on assessing the potential impact of the H2 fuel specific issues and seeking effective solutions. We anticipate that the FutureGen H2 turbine will utilize diluents and staged premixing schemes (e.g. premixing of H2 with N2 before injecting in air) to control auto-ignition and flashback. Mixing of an inert gas such as N2 readily available from the coal gasification process into the H2 stream is an effective means to reduce flame speed, lower flame temperature and increase ignition delays. For example, our estimation shows that H2 with 50% N2 dilution at = 0.4 produces a flame temperature of about 2600F. A properly designed LSI should be able to burn this mixture or similar mixtures of various combinations of N2 concentrations and equivalence ratios. As to mitigating the hazards associated with auto ignition, the established approach is to reduce the residence time by injecting the fuel as close as possible to the burner tip. The LSI is conducive to this treatment because it can tolerate some variations in mixture homogeneity without sacrificing flame stability and emissions. Additionally, there are other means that can further reduce the risks of auto-ignition and flashback. One simple method worthy of consideration is by blending of N2 into the H2 stream prior to injection into the air stream. Obviously, the optimum solution will depend on which combination of fuel blend, fuel treatment and injection scheme would be the best to meet the specifications and requirements of the OEM. The outcome of these studies will be applied to develop and fabricate a full-scale or pilot-scale prototype H2 LSI. This prototype will include mixers and fuel injectors that can mitigate the H2 fuel related issues. Project Summary: Fuel Flexible Combustion System for Co-Production Plant Operations Project Participant: GE High-efficiency, low-emissions co-production plants that produce electric power, transportation fuels, and/or chemicals from fossil fuel feed stocks require a new class of fuel flexible combustors. In this 36-month program, a validated combustor approach will be developed which will enable single-digit NOx operation of cogeneration plants with low-Btu off gas and high-hydrogen fuels, with the flexibility of process-independent backup with both natural gas and liquid fuels. This combustion technology will overcome the limitations of current syngas gas turbine combustion systems, which are designed on a site-by-site basis, and enable improved plant designs. In this capacity, a fuel-flexible combustor will enhance the efficiency and productivity of IGCC based coproduction plants. One of the major challenges for coproduction plants is handling a fuel stream with a time varying heating value and hydrogen content. In current Integrated Gasification Combined-Cycle (IGCC) practice, the combustor is tailored to the fuel properties at each site. In addition, there are emerging needs for highhydrogen fuels, which currently require diluent injection to meet emissions and safety constraints.
16
The approach in this program is to unify and improve these existing designs and introduce the latest technology, where appropriate. A hybrid combustor, successfully incorporating the low-NOx performance of our most advanced premixed combustion systems with enhanced versions of the Integrated Gasification Combined-Cycle (IGCC) nozzles currently in production, will lead to a fuel-flexible combustor design capable of meeting fuel flexible IGCC performance requirements. The success and the resultant quality of the fuel-flexible combustion system is enhanced by the Design for Six Sigma (DFSS) quality process, which is a statistically based methodology focused on flowing performance specifications and tolerances from the high level of customer or power plant objectives down to the low level of component parts. The current process capability of each component flows back up to understand the influence of its variability on system performance. Using this methodology with a conceptual plant configuration will ensure that the combustion system is robust and flexible enough for highly efficient operation. The program focuses on plant optimization, low emission combustor design, and development of tools for syngas flame modeling. A study of market fuel variations and gas turbine combustor operating conditions will be studied to determine optimal plant efficiency. The fuel space definition will be used with a combined cycle plant model to determine combustor inlet and required firing conditions. The combustor design study will evaluate several design options in the quest to define a design space that will meet the operating requirements. The flame modeling tools are based upon fundamental data characterizing the syngas flames. Data for H2 flames and H2/CO mixtures has been obtained at atmospheric pressures. Project Summary: System Study for Improved Gas Turbine Performance for Coal IGCC Application Project Participant: GE This 15-month study will identify vital gas turbine parameters and quantify their role in meeting the overall DOE Integrated Gasification Combined-Cycle (IGCC) plant goals of 50% net HHV efficiency, $1,000/kW capital cost, and low emissions. The proposed project will analyze and evaluate gas turbine conceptual cycle designs, and quantify their influence on IGCC plant level performance. The study will provide DOE with information as it develops strategies for identifying future technologies needed to advance IGCC gas turbine performance. A baseline conceptual IGCC system design will be established utilizing current General Electric (GE) F-class gas turbine technology, based on a U.S. IGCC site such as the Tampa Electric Polk IGCC Project or the Wabash River Coal Gasification Repowering Project. Confirmation of plant level performance goals would help lead to the selection of gas turbine cycle concepts to be further investigated. An overall IGCC system performance model will be constructed utilizing GE in-house proprietary software for the gas turbine and steam turbine, and commercially available software for the balance of the systems. The model will be exercised through parametric analysis to quantify gas turbine performance impact at IGCC plant system level. Results from the system analysis will be used to identify gas turbine technology improvements for development consideration in future program phases. The proposed program will be performed through the following five major tasks utilizing GE’s Design for Six Sigma methodology: • Overall System Requirements Identification • Requirements Prioritization & Flow-Down to Gas Turbine Subsystem Level • IGCC Conceptual System Analysis • Gas Turbine Cycle Options vs. Requirements Evaluation • Recommendations for Gas Turbine Technical Improvements
In conclusion, the goals and project summaries outlined above represent the approach for the Advanced Turbine program in the 2010, 2012, and 2015 time frames, and how these goals will be realized by way of each project.
17
6.0.2 6.0.2 Combustion Strategies for Syngas and HighHydrogen Fuel The technical challenges surrounding syngas and hydrogen fuel combustion have been outlined in section 3.1. Given the issues presented there, various options can be considered for combustor design and operation. First, it is critical to define the type of combustion system that will be used. There are two broad categories: diffusion flame combustors, and premixed combustors. These are described below, but before discussing the combustion strategies, it is useful to review how NOx pollutants are formed. NOx formation There are several routes to form NOx pollutants and these may be broadly catalogued as thermally-generated, flame-generated, or fuelbound NOx. Different authors use different names to catalogue these mechanisms and there is still continuing research to understand the most prominent mechanisms at ultra-low NOx conditions. For example, in hydrogen fueled systems, the prominence of H radicals may contribute to NOx in a manner that is different than in systems fueled by natural gas [ 1]. Thermal NOx is formed by oxidation of nitrogen in air and requires sufficient temperature and time to produce NOx. A rule of thumb is that below approximately 1700K, the residence time in typical gas turbine combustors is not long enough to produce significant thermal NOx. Where temperatures higher than 1700K cannot be avoided, it is necessary to limit residence time to control NOx formation, which favors very short combustor designs. Thermal NOx production also increases with the square root of operating pressure, making it more difficult to reduce in higher-pressure aeroderivative gas turbines.
As the name implies, flame-generated NOx occurs in the flame front, created on the short time scale associated with primary combustion reactions. There are a variety of chemical mechanisms involved, all linked to intermediate combustion species that exist only in the reaction zone of the flame. It is important to understand that in practical combustors, the reaction zone is just a small portion of the total combustor volume –most of the combustor volume is sized to complete the relatively slow approach to equilibrium products (notably CO to CO2 oxidation). Thus, residence time in the whole combustor does not affect the flame-generated NOx produced – a significantly different behavior compared to thermal NOx. A convincing demonstration of this point was presented by Leonard and Stegmaier 2 who studied multiple flame holders, operating conditions, and residence times from 2 to 100 milliseconds, demonstrating that the flame temperature alone (not residence time) determined the NOx production for emissions under 10 ppmv. Fig. 1, adopted from [2], is useful to estimate the flame NOx produced at a given flame temperature, accounting for ideal, and “poor” premixing (not carefully defined in [2]). Note that the effect of poor premixing raises the NOx levels by as much as a factor of three. These data were recorded in turbulent flames, where combustion products are mixed with the fresh reactants right at the flame. It has been suggested that other combustion configurations, without significant stirring between the flame front and products, may reduce the flame generated NOx [ 3]. This may be the basis for NOx reductions reported using low-swirl combustion in section 3.2.1.4.3 Finally, fuel-bound NOx is produced by nitrogen species in the fuel reacting with air during combustion. For coal syngas, the most prominent fuel nitrogen species is ammonia, generated during gasification from nitrogen compounds in coal. The ammonia should ideally be removed from the fuel before entering the combustor, or it will be converted to NOx by most combustion strategies. Where this is not possible, rich-lean strategies have the most potential to reduce NOx pollutants. In this approach, combustion is first carried out under fuel-rich conditions, followed by completing combustion under fuel lean conditions. In fuel rich conditions, with sufficient residence times, the ammonia can be reduced to nitrogen and water, rather than atmospheric oxygen. A number of studies have been conducted to evaluate rich-lean combustion as an approach to reducing fuel bound NOx. These studies have shown as much as 95% of the fuel ammonia can be reduced to nitrogen and water using rich-lean combustion, with the remaining 5% converting to NOx. [ 4, 5, 6, 7, 8]. Untreated syngas ammonia concentrations can exceed 1000ppm, where even 5% conversion would lead to 50ppm NOx, which are well above desired emissions levels. Thus, it is desirable to remove fuel ammonia during gas cleanup, rather than rely on combustion techniques to reduce it to water and nitrogen.
NOx (ppmvd) @15% O2
George Richards
10
Poor Premixing
6 4 2 Ideal Premixing 1 1650
1700
1750
1800
1850
1900
1950
Flame Temperature (K)
Fig. 3.1.1 NOx emissions, adopted from Leonard and Stegmaier [2]. Diffusion flame combustor In this style of combustion, fuel and air are introduced in separate passages, and the flame is stabilized where the fuel and air streams mix. Combustion reactions are typically so fast that fuel and oxidant consumption is limited by transport to the reaction zone (i.e., diffusion), and the reaction proceeds locally at nearly stoichiometric conditions. The Lewis number (Le) describes the ratio of thermal transport to species transport from this reaction zone. Where Le = 1, the temperature in the reaction zone will equal the adiabatic flame temperature because thermal energy diffuses away as fast as the reactants are supplied. The fuel species in hydrocarbon combustion typically have fuel Lewis numbers ( α mix / Dij fuel ) in the range of 0.9 to 1.2, meaning that diffusion flame combustors will have flame temperatures near the adiabatic flame temperature. These temperatures are high enough to oxidize nitrogen in air, producing appreciable NOx pollutants. Hydrogen itself has a fuel Lewis number as low as 0.4, making it even more difficult to reduce NOx because the peak laminar flame temperatures are higher than adiabatic due to differential diffusion effects. The effect of fuel Lewis number on flame temperature has been observed experimentally as well as with direct numerical simulations (DNS). [ 9, 10] Because of their high flame temperatures, diffusion flame combustors require some method to achieve low-NOx performance. An obvious technique is to dilute the fuel, lowering the adiabatic flame temperature. A common diluent is steam, which can both lower the flame temperature, and reduce the production of non-thermal NOx. The hydroxyl radical OH is increased by the presence of additional water, and these radicals favorably scavenge HCN fragments which might otherwise produce NOx. Steam dilution is already used on IGCC applications, but it is not completely desirable. The extra energy that is needed to make steam from water is not recovered in the turbine expansion, penalizing cycle efficiency, (but raising power output from the added mass flow). The additional steam in the exhaust produces a modest increase in the turbine nozzle heat transfer, raising metal temperatures. The protective thermal oxide layers in turbine material sets can be affected by increased moisture levels. Finally, steam consumption by stationary turbines should be minimized to conserve water resource. For these reasons, any further development of diffusion flame combustors for IGCC applications
6.0.2 NETL Internal Combustion and Turbine Research
would ideally use nitrogen from the air separation plant, rather than steam. The amount of nitrogen available for flame dilution is established by the engine cycle and the ASU, and it can be shown that for example, hydrogen could be diluted up to about 50% with nitrogen in a typical IGCC configuration. Unfortunately, this level of dilution produces an adiabatic flame temperature around 2025 K, which is still too high for ultra-low NOx performance. Given the dilution limit on adiabatic flame temperature, it is important to consider other methods to reduce the diffusion flame temperature. As noted above, the diffusion flame temperature is set by the ratio of thermal diffusion away from the reaction zone to heat generated by reactants. If the reaction zone is “strained” by fluid shear, it is possible to change the balance between diffusion and reaction in the reaction zone, changing the flame temperature. Strongly sheared flows can locally extinguish the flame, providing opportunity for fuel air mixing before combustion is initiated elsewhere. This raises the possibility that strong shearing could be used to make a diffusion flame combustor behave more like a premixed combustor. The required levels of shearing (known as “stretch” or “strain”) have not been fully characterized. These concepts are discussed next. Lean Direct Injection Lean Direct Injection (LDI) combustion was developed as a low NOx alternative to Lean Prevaporized Premixed (LPP) combustion for aircraft gas turbines, where the inherent flashback and dynamic instability concerns of LPP combustion are considered too great of a risk for flight application. In LDI combustors, liquid fuel is directly injected into the combustion chamber, where it is mixed with air in the shortest possible distance. The intent is to provide an essentially lean premixed fuel/air mixture that burns in a low-NOx flame, similar to LPP combustors, which are discussed in the Premixed Combustion section below [ 11, 12]. NOx performance is compromised in an LDI combustor if the fuel and air are not perfectly mixed before combustion, creating regions with higher fuel content that burn hotter and generate more NOx. Similarly, the mixture may burn upstream of the premixed zone in a diffusion flame, with combustion occurring at stoichiometric conditions that result in higher temperatures and NOx production. Nevertheless, flashback and auto-ignition concerns are nearly eliminated in LDI combustors, and they can operate over a wide turndown range with a high degree of static and dynamic stability using a wide range of fuels. The desire to burn high-hydrogen fuels in gas turbines used for power applications raises similar concerns of flashback and instability when operating in the Lean Premixed mode of combustion, so LDI combustors seem to be a natural fit for burning these fuels in a low NOx gas turbine system. To demonstrate the potential of LDI combustors, researchers at NASA Glenn have recently studied various low NOx LDI concepts for pure hydrogen combustion in aircraft gas turbine combustors.[ 13] Five separate injector concepts from different manufacturers were tested at aircraft gas turbine conditions (4.8 – 13.6 atm., Tin = 600 – 1000 °F). At low combustor exit temperatures, it was possible to achieve very small NOx levels (~1 ppmv, wet, uncorrected). NOx
George Richards
emissions were primarily controlled by lowering equivalence ratios to limit combustion temperatures, and no hydrogen dilution cases were considered. One of the tested injectors at NASA Glenn was similar to those used in current IGCC gas turbines that burn syngas, where fuel is injected axially into a swirling airflow. Although this injector was very robust, it produced substantially higher NOx than the other tested injectors. Some of the other tested injectors were similar to those studied recently at GE Energy, where multiple fuel jets were injected at an angle into a central air jet [ 14]. Their results show that more fuel injection ports per air jet reduce NOx emissions due to higher fuel jet momentum and mixing. Increasing the number and decreasing the size of the air jets is shown to reduce NOx by reducing the length of the combustion zone, although this comes at the expense of increased combustor pressure drop. Similar injector configurations studied at NASA Glenn had better NOx emissions, due in part to the shortened combustion zone. However, in some cases, this also led to overheating problems that led to failure of the injector, since the combustion zone was located much closer to the fuel and air injectors. Pressure drops in the NASA Glenn injectors were sometimes very large (4-25%). Redesign and optimization for power gas turbines could reduce these pressure drops. In addition, large pressure drops may have been required to reduce the flashback or flameholding potential in those injector designs that operated more in a premixed combustion mode than a diffusion combustion mode. As the injectors were tested on pure hydrogen, dilution with nitrogen will reduce flame speeds and may decrease the necessity for large injector pressure drops and high air velocities to avoid these issues. Highly-strained diffusion flame combustors Though not discussed explicitly in the above studies, successful LDI diffusion flame combustors use jets of fuel and air that introduce high strain rates in the combustion zone. In a pure diffusion flame, strain rate can be quantified by measuring or calculating the velocity gradients in the mixing flow field. In regions of high strain and 3000
0.12 0.1
2000
0.08
1500
0.06 wNO
1000
0.04 10 XO
500 0 1
10 100 1000 Strain Rate (1/s)
0.02
3
2500
w NO (kg/m s) and X O
Tmax (K)
Tmax
0 10000
Figure 3.2.2: Strain rate effects, adapted from Sanders et. al. [15]. wNO = NO formation rate, XO = O-atom mole fraction, Tmax = peak temperature
6.0.2 NETL Internal Combustion and Turbine Research
fluid shear, mixing rates and bulk transport rates are faster than chemical reaction rates, thus local reactions are not allowed to go to completion before the flow carries the combustion radicals away from the reaction zone. The net result of this process is a reduction in peak flame temperature of a highly strained flame, which in turn reduces thermal NOx production. It should be pointed out, however, that thermal NOx is not only a function of temperature, but also of flame residence time and O-atom concentration in the reaction zone. Increasing the flame strain also tends to reduce the residence time in the flame, but it also can increase the O-atom concentration in the flame by an order of magnitude. This effect is shown in Figure 3.2.2, where intermediate strain rates tend to increase the production rate of NO due to the increased O-atom concentration, while at high strain rates, the reduction in flame temperature overcomes the influence of the Oatom concentration, and NO production rates are reduced [ 15]. This shows increased strain rates as a possible path to reducing or effectively eliminating thermal NOx in a diluted diffusion flame, where dilution of the fuel alone does not reduce flame temperatures enough to satisfy ultra-low NOx emission goals. Increased strain rates are typically attained by increasing the fuel and/or air jet velocities to increase fluid shear, though at the expense of increased combustor pressure drop. In addition, the static stability of the flame is a strong function of these jet velocities, where too high of a jet velocity could cause the flame to blowout. Thus, flame stability concerns place limits on allowable levels of flame strain, particularly for diluted high-hydrogen content fuels, since flameholding ability is closely linked to the flame speed of the fuel/air mixture, which decreases as more diluent is added to the fuel stream. From this perspective, impinging fuel and air jet injector configurations [13,14] hold an advantage over co-axial jet configurations, as forced mixing of the fuel and air should improve the flameholding abilities of these diffusion flames. Much more study could be done in this area to determine injector configurations that maximize flame strain while minimizing stability and combustor pressure drop concerns. In addition, the effect of strain rate on NOx emission from diffusion flames has only been partially quantified for simple diffusion flames, and there are no such studies in practical LDI-type diffusion flame combustors using hydrogen, syngas, and/or fuel diluents. Other areas requiring further study include the effects of increased flame strain on combustion efficiency and on in-flame NOx production mechanisms. Premixed Combustion As the name implies, premixed combustion is accomplished by mixing the fuel and air upstream of the flame. The fuel-air ratio normalized by the stoichiometric value is known as the equivalence ratio φ, and in many practical premixed turbine combustors, has a value of slightly more than 0.5. Thus, there is approximately ½ the fuel needed to burn all the air, or conversely twice as much air as needed to burn all the fuel. The excess air serves to dilute the combustion and keep the flame temperatures low enough to avoid thermal NOx formation. While the concept of premixed combustion is simple and effective at reducing NOx, it also has drawbacks. The combustor must operate in a very narrow range of equivalence ratio to avoid blowout at (typically) φ < 0.5, and increasing NOx formation for φ somewhat greater than 0.6. The combustor controls must include some form of staging, since the range of desired exit temperatures usually cannot be achieved with such a small range of φ . For example, if four fuel injectors are used in a
George Richards
combustor, it is possible to reduce the heat input 50% keeping two injectors operating, but turning two off. The difficulty with this approach is that the air flow from inactive injectors can quench the boundary of the flame from operating injectors, raising CO emissions, but this can be addressed with good aerodynamic design. Staging in this manner is used on commercial engines [ 16, 17]. Beyond simply de-activating injectors, staging is also accomplished by operating some injectors at slightly richer equivalence ratios, to improve flame stability. This can also be accomplished using “pilots” on individual injectors. The pilot flame is typically supplied with some air for partial premixing, and the pilot fuel circuit is controlled to achieve stable combustion at the lowest possible NOx emissions, as described in the following section. Tuning and Combustor Control Balancing the fuel delivery among various fuel circuits to meet operating requirements is known as “tuning” and has become a critical part of both commissioning and operating low-emission gas turbines. Various strategies have been used, or are being developed so that the emissions targets can be met with stable combustion. Because combustion stability is affected by inlet temperature and fuel composition, tuning may need to be adjusted to accommodate ambient environment temperatures and even fuel composition [ 18]. In addition to controlling the fuel split, for some turbines, tuning may include adjustment of compressor inlet guide vanes or bleeding compressor flow [18]. This allows an adjustment of the combustor air flow at fixed compressor speed, providing another tuning option even on single-speed (synchronous) gas turbines. It is important to understand that turbines must be able to contend with requirements for load rejection while low-emission combustors operate near the blowout condition. Without careful development, cutting the fuel during load rejection can lead to flame blowout, requiring (sometimes) unacceptable time to re-light and establish power, or making the engine unable to meet grid requirements. An interesting account of the development of combustor and control system required to meet stringent rejection requirements is given in the references [ 19, 20]. On some engines, fast acting valves are used to enhance lean-blowout performance [ 21] and allow operation right near the limits of stable combustion. A more advanced concept is to modulate the fuel to counteract combustion oscillations, usually called active combustion control. Active control has been studied in many research projects [ 22, 23, 24], but has only been deployed on one test engine [ 25] and on one commercial engine installation [ 26] to date. An important aspect of combustion tuning and control is diagnosing conditions in the combustor so that the control system can respond to maintain stable, low-emission operation. For example, it is possible to improve engine operation by monitoring combustion performance from available engine sensors [27]. A number of recent papers have shown the potential of using flame optical signals, acoustic signals, or flame ionization to monitor and control the combustion process [ 28, 29, 30].
Oxy-Fuel Combustion As noted in section 1.3.1, advanced engine cycles using oxy-fuel combustion have been proposed as a means of capturing CO2 from engine operation. These oxy-fuel cycles require a different approach to combustor design because the combustion is ideally operated at stoichiometric conditions – having just enough oxygen to completely oxidize the fuel. Oxygen is produced from air separation, such that any excess oxygen is produced with an accompanying penalty to the overall cycle efficiency. In addition, after the water is condensed from the exhaust, any excess oxygen should be eliminated from the compressed CO2 to avoid corrosion in handling the CO2 gas. For these reasons, the combustor design must achieve very high combustion efficiency at conditions with little excess oxygen. This requirement places a premium on achieving high levels of mixing uniformity in the combustor, because even modestly unmixed fuel stream will be starved for oxygen. It should be noted that boiler designs also ideally operate near stoichiometric, but typically use 1-3% excess oxygen, and have relatively long residence times to complete fuel oxidation. For the oxy-fuel turbine, the excess oxygen would ideally be lower, with much shorter residence times (~30ms) to avoid excessively large pressurized combustion chambers. Oxy-fuel combustion for power cycles has been studied in a number of papers [ 31, 32]. The easiest combustion strategy is to employ a diffusion flame combustor. The stability and simple operation of diffusion flame systems make them appealing for oxy-fuel systems. There is no need to control NOx, since the products are sequestered, and there is otherwise little nitrogen in the combustor. Even without sequestration, the peak flame temperature in diffusion flames can be controlled by the level of diluent added, thereby avoiding NOx formation. Nevertheless, a potential advantage of premixed combustion is that premixing the fuel and oxidant can reduce the unmixed streams of fuel and oxygen that are created in diffusion flame systems where relatively small fuel jets must penetrate and mix in the large combustion volume. There is relatively little fundamental data on premixed oxy-fuel flames diluted by water or CO2 [ 33, 34] such that proposed designs must include some margin with respect to fundamental issues like flame speed. 1
Konnov, A.A., Colson, G., De Ruyck, J. (2000). The new Route to Forming NO via NNH, Combustion and Flame, Vol. 121, pp. 548-550. 2 Leonard, G., Stegmaier, J. (1994). Development of an Aeroderivative Gas Turbine Dry Low Emissions Combustion System, ASME J. Eng. For Gas Turbines and Power, Vol. 116, pp. 542 – 546. 3
Sattelmayer, T., Polifke, W., Winkler, D., Dobbeling, K., (1998). NOx-Abatement Potential of LeanPremixed Gas Turbine Combustors, ASME Journal of Engineering for Gas Turbines and Power, Vol. 120, pp. 48- 59. 4 Fietelberg, A. S., Lacey, M. A., (1997). The GE Rich-Quench-Lean Gas Turbine Combustor ASME 97GT-127. 5
Hasegawa,T., Sato, M., Ninomiya, T. (1997). “Effect of Pressure On Emission Characteristics In LBGFueled 1500C-Class Gas Turbine, ASME 97-GT-277.
6
97-GT-38 Constant, D. R., Bevan, D. M, Cannon, M. F., Kelsall, G. J. (1997). Development of an LCV Fuel Gas Combustor for an Industrial Gas Turbine ASME 97-GT-38.
7
Folsom, B.A., C.W. Courtney, Heap, M. P. (1980). “The Effects of LBG Composition and Combustor Characteristics on Fuel NOx Formation,” ASME J. Eng. Power, V102, pp459-467.
8
Domeracki, W.F., Dowdy, T. E., Bachovchin, D. M. (1997). Topping Combustor Status for SecondGeneration Pressurized Fluidized Bed Cycle Applications, ASME J. Eng. Gas Turbines and Power, Vol. 119, pp. 27 – 33.
9
Takagi, T., Xu, Z. and Komiyama, M., Preferential Dissusion Effects on the Temperature in Usual and Inverse Diffusion Flames, Comb. and Flame 106: 252-260 (1996). 10 Gabriel, R. Navedo, J. E. and Chen R.,, Effects of Fuel Lewis Number on Nitric Oxide Emission of Diluted H2 Turbulent Jet Diffusion Flames, Comb. and Flame 121:525-534 (2000). 11 Tacina, R., Wey, C., Liang, P., and Mansour, A., “A Low NOx Lean-Direct Injection, Multipoint Integrated Module Combustor Concept for Advanced Aircraft Gas Turbines,” Clean Air Conference, Porto, Portugal, NASA/TM-2002-2111347.
12
Tacina, R. R., Wey, C., Choi, K. J., “Flame Tube NOx Emissions Using a Lean-Direct-Wall-Injection Combustor Concept,” 37th Joint Propulsion Conference and Exhibit, Salt Lake City, Utah, July 8-11, 2001, AIAA-2001-3271. 13
Marek, C. J., Smith, T. D., and Kundu, K., "Low Emission Hydrogen Combustors for Gas Turbines Using Lean Direct Injection," 41st Joint Propulsion Conference and Exhibit, Tuscon, Arizona, AIAA-20053776, July 10-13, 2005. 14
GE Energy, “Premixer Design for High Hydrogen Fuels – Final Report,” DOE Cooperative Agreement No. DE-FC26-03NT41893, November, 2005. 15 Sanders, J. P. H., Chen, J.-Y., and Gokalp, I., “Flamelet-Based Modeling of NO Formation in Turbulent Hydrogen Jet Diffusion Flames,” Combustion And Flame, Vol. 111, pp. 1-15, 1997. 16
Joshi, N. D., Mongia, H. C., Leonard, G., Stegmaier, J. W., Vickers, E. C. (1998). Dry Low Emissions Combustor Development, ASME 98-GT-310. 17 Lefebvre, A.H. (1998). Gas Turbine Combustion, 2nd ed, pp. 349, Taylor and Francis. 18 Sewell, J. B., Sobieski, P. A., (2005). Monitoring of Combustion Instabilities: Calpine’s Experience, in Combustion Instabilities in Gas Turbine Engines, Lieuwen, T. C. , Yang, V. [eds.], American Institute of Astronautics and Aeronautics, pp. 147 – 162. 19 Myers, G., Tegel, D., Feigl, M., Setzer, F., Bechtel, W., Fitts, D., Couture, B., Tuthill, R. (2003). Dry, Low-Emissions For the ‘H’ Heavy Duty Industrial Gas Turbines: Full-Scale Combustion System Rig Test Results, ASME GT2003-38193. 20 Feigl, M., Setzer, F., Feigl-Varela, R., Myers, G., Sweet, B. (2005). Field Test Validation of the DLN2.5H Combustion System on the 9H Gas Turbine at the Baglan Bay Power Station, ASME GT200568843. 21 Mongia, H.C., Held, T. J., Hsiao, G. C., Pandalai, R.P. (2003). Challenges and Progress in Controlling Dynamics in Gas Turbine Combustors. AIAA Journal of Propulsion and Power, Vol. 19, No. 5, pp. 822829. 22 Cohen, J. H., Rey, N.M., Jacobson, C. A., Anderson, T.J. (1999). Active Control of Combustion Instabilities in a Liquid-Fueled Low-NOx Combustor. ASME Journal of Engineering for Gas Turbines and Power, Vol. 121, No. 2, pp. 281 - 284. 23 Sattinger, S.S, Neumeier, Y., Nabi, A., Zinn, B. T., Amos, D. J., Darling, D. D. (1998). Subscale Demonstration of the Active Feedback Control of Gas Turbine Combustion Instabilities, ASME Paper 98GT- 258. 24 Jones, C. M., Lee, J. G., Santavicca, D. A. (1999). “Closed-loop Active Control of Combustion Instabilities Using Subharmonic Secondary Fuel Injection, Journal of Propulsion and Power, Vol. 15, No. 2, pp. 1-7.
25
Richards, G. A., Thornton, J. D., Robey, E. H., Arellano, L (2004). Open-Loop Active Control Of Combustion Dynamics On A Gas Turbine Engine, ASME IMECE2004-59702 26 Seume, J. R., Vortmeyer, N., Krause, W., Hermann, J., Hantschk, C.-C., Zangl, P., Gleis, S., Vortmeyer, D., and Orthmann, A., (1998). Application of Active Combustion Instability Control to a Heavy Duty Gas Turbine. ASME Journal of Engineering for Gas Turbines and Power, Vol. 120, No. 4, pp. 721 -726. 27 Angello, L. C., Castaldini, C. (2004).Combustion Instability Tuning Guidelines: Understanding and Mitigating Dynamic Instabilities in Modern Gas Turbine Combustors, ASME GT2004-54081. 28 Muruganandam, T., Seitzman, J.M. (2003). Optical Sensing of Lean Blowout Precursors in a Premixed Swirl Stabilized Dump Combustor. ASME GT 2003-38104. 29 Lieuwen, T. (2004). Online Combustor Stability Assessement using Dynamic Pressure Data, ASME GT2004-53149 30 Benson, K., Thornton, J. D., Straub, D. L., Huckaby, E. D., Richards, G. A. (2005). Flame Ionization Sensor Integrated Into a Gas Turbine Fuel Nozzle, ASME Journal of Engineering For Gas Turbines and Power, Vol. 127 pp. 42 - 48 31 Chorpening, B. Richards, G. A., Casleton, K. H., Woike, M., Willis, B., Hoffman, L., (2005). Demonstration of a Reheat Combustor for Power Production with CO2 Sequestration. ASME Journal of Engineering For Gas Turbines and Power, Vol 127, pp. 740 – 747. 32 Richards, G. A., Casleton, K. H., Chorpening, B. T., (2005). –CO2 and H2O Diluted Oxy-Fuel Combustion for Zero-Emission Power, Proc. IMecheE, Vol 219, Part A, J. Power and Energy, pp. 121 – 126. 33 Lewis, B., von Elbe, G., (1987). Combustion, Flames, and Explosions of Gases, 3rd ed. , ppAcademic Press. 34 Koroll, G. W., Mulpuru, S. R., (1986). The Effect of Dilution with Steam and the Burning Velocity and Structure of Premixed Hydorgen Flames, The Twenty First Symposium (international) On Combustion, The Combustion Institute, pp. 1811-1819.
6.0.3
University Turbine Systems Research Program
6.0.3-1 Introduction The University Turbine Systems research (UTSR) Program began in 1992, as part of the U.S. Department of Energy’s major development program in gas turbines1. Between 1992 and 2001 the program was funded with about $400 million of DOE money and a similar amount of contractor money. Part of the program was a university research effort. Major emphasis areas were university research projects, internships, and technology transfer. The university program was coordinated by the South Carolina Institute for Energy Studies (SCIES), a part of Clemson University, under the overall direction of DOE’s National Energy Technology Laboratory. Total cost of the university research program from 1992 until 2001 was $35.5 million, of which $34.2 million came from DOE and $1.2 million came from industry. The UTSR activity, which began in 2002, is budgeted for $15 million in DOE funds and $750,000 from industry over 5 years. More details on the UTSR Program can be found in note 12. The DOE Program has continued since 2002 and is now called the Turbine Program. The university activity is continuing as the University Turbine Systems Research (UTSR) Program. In 2003, the UTSR program shifted from an emphasis on natural gas fuel to research that supports a future power industry needing turbines fueled by syngas and hydrogen (SGH). Under the university research program 108 universities in 40 states did/do research in the fields of aerodynamics and heat transfer, combustion, and materials. Besides the universities, the UTSR consortium includes leading gas turbine original equipment manufacturers (OEM’s) and users, and gas turbine component manufacturers. These companies comprise the Industry Review Board (IRB), who recommend and track research projects that are funded, and are the host sites for graduating seniors and graduate students from UTSR universities placed for summer assignments (Fellows). Current IRB Member Companies are BP, Capstone Turbine Corporation, Cinergy Energy Services, Clean Energy Systems, EPRI, ExxonMobil, General Electric Company, Ingersoll Rand Energy Systems, Parker Hannifin, Pratt & Whitney/UTRC, Precision Combustion Inc., RAMGEN, Rolls-Royce, Siemens Westinghouse Power Corporation, Solar Turbines, Inc., Southern Company Services and Woodward FST. Voting member companies are gas turbine manufacturers: GE, Pratt & Whitney, Rolls-Royce, Siemens Westinghouse and Solar Turbines. The others are Associate Member companies. 1.
The overall program in 1992-2001 was called the Advanced Turbine Systems (ATS) Program, and the university research portion of ATS was called the Advanced Gas Turbine Systems Research (AGTSR) Program.
An Academic Advisory Board (AAB) was formed in 2004 to provide a mechanism for obtaining input from the academic community to the UTSR program and to develop short courses on gas turbine technologies. Their first product was a short course in August 2004 on the impact of synfuels on gas turbines. Current AAB members represent Virginia Tech, Georgia Tech, U. of California Irvine, U. of Central Florida, Penn State, U. of Connecticut, Brigham Young and U. of Wisconsin. The program is geographically broad – based, see figure 1.66
43766
Fig. 1. The UTSR Process
The point of the research program is to address the goals of the DOE Turbine Program while producing results that are useful to the gas turbine industry. In order to achieve that outcome the program is designed with significant input from industry, through the IRB. The IRB and DOE decide which research topics are most relevant to their needs, in the areas of a) combustion, b) materials – focusing on thermal barrier coatings (TBC’s) – and c) aerodynamics / heat transfer – focused on the turbine section of the gas turbine. These topics define the requested research in the Request for Proposals released yearly to the UTSR universities. Then, after the proposals come in from the universities, the IRB ranks them within each of the three subject areas and recommends a short list to DOE for funding in the annual IRB meeting. The short list is created in rank order starting with the most important proposal and ending where the expected funding runs out, plus typically two backup proposals in case there is extra funding or enough cost savings can be achieved on the selected ones to fund more projects. The typical project lasts for three years, averaging about $150K per year. So each year, the available funding needs to cover projects that were started one and two years ago in addition to the first year costs of new projects. There is considerable communication between industry, the universities, DOE and SCIES, designed to keep the program relevant to the needs of industry, DOE’s priorities, and the capabilities and ideas from the universities. The overall process of the UTSR Program is illustrated in figure 2.
Fig. 2. Research Program
438
William H. Day, Richard A. Wenglarz and Lawrence P. Golan
Since its inception in 1992, the UTSR research program has launched 103 projects, including 77 completed, 21 underway and 5 announced but not yet started. By visiting the website mentioned in note 1 the reader can select descriptions of each project and search by Principal Investigator or University3. Considerable results of use to industry have been achieved from these projects. A sampling of some of the most significant are listed below in the three technical areas of concentration. MATERIALS (THERMAL BARRIER COATINGS): • Laser fluorescence (LF) was determined to be the most promising technique for non-destructive evaluation (NDE) of TBCs and a UTSR projected started the commercial development of a new low cost and portable NDE instrument. • Processing approaches have been identified that can increase TBC lifetimes by a factor of four and more. • A new Small Particle Plasma Spray (SPPS) process was shown to produce a factor of two lower internal oxidation rate of the bond coat and TBC coatings that experience lower fatigue damage. • Two superior alloys and one coating were identified for operation at surface temperatures above 700 C (1290 F which did not experience significant degradation associated with water vapor effects, significant to operation with syngas fuel. COMBUSTION: • Active control approach to overcome instabilities in low emission turbine combustors. A factor of four reduction in combustor pressure oscillations was demonstrated, and several gas turbine companies have started projects to evaluate application of the approach to their combustors. • Method to determine the stability margin of combustors before experiencing problems in the field. • Computer code for NOx and CO emissions prediction design of low emission turbine combustors. The code has shown a factor of forty reduction in computation times. • Devices using infrared light for measuring fuel-air mixedness in combustors. Less than one-third of the cost of laser devices for measuring mixedness and are more compact and rugged. AERODYNAMICS / HEAT TRANSFER: • Experiments showing a scientific foundation for use of a fine water mist in steam for cooling high temperature turbine components. The addition of 1% water mist can enhance cooling by 50 to 100%, and in best cases, as much as 700%. • Internal surface features within channels improve turbine blade cooling. Dimples on the interior of cooling channels can improve cooling effectiveness by as much as a factor of two, significant for operation with syngas fuels. • LES (Large Eddy Simulation) computational approaches improve predictions of heat transfer and design of turbine blade cooling. This enables less coolant air to improve turbine performance. Illustrations of a few of the completed projects are shown in figures 3 – 5.
439
6.0.3 University Turbine Systems Research Program
Fig. 3. Improvement in Life of Thermal Barrier Coatings
Fig. 4. Reduction in Combustor Pressure Oscillations
Fig. 5. Improved Turbine Cooling Design
440
William H. Day, Richard A. Wenglarz and Lawrence P. Golan 6.0.3-2 The Challenge of Synfuels The use of coal-derived synfuels poses particular issues to the designers of the hot parts of the gas turbine. Table 1 summarizes the major turbine research challenges and suggested means to overcome them. Table 1
441
Issue Syn-H2 Gas Introduces
Description
Challenge/Resolution
Heating Value Variability
Syngas to natural gas variability is as much as 20:1.
Current combustors cannot manage this variability and research is needed to aid design of combustors that will.
Gas Constituent Variability
Constituent proportions vary significantly; for example H2 can vary 8.6% to 44%. Amount of H2O in the combustion process will be significantly greater than for conventional fuels.
Flash back is an issue for H2 bearing fuels and research on flashback will also aid design of natural gas turbines. Increased mass flow has required down rating of current designs. Design to prevent combustion instabilities is complicated by fuel variability
Blade coatings
Mixed gases (H2O, salt, etc.) attack current airfoil base materials.
Research for more robust barrier blade coatings is justified in order to increase operating temperature and life.
Heat Transfer
Increased heat transfer to components results for syngas-hydrogen fuels.
Research to gain improved understanding of heat transfer mechanisms for syngas is justified.
Deposits
Syngas deposits will cause special problems not now fully understood.
Research needed to gain improved understandings of consequences such as blade wakes, reduced flow passageways, and clogging of film cooling holes.
Erosion-Corrosion
Erosion-corrosion issues not fully understood.
Research is needed to understand rates and consequence of blade passage wear, especially in blade tip region. TBC degradation and spalling need to be understand in order to allow operation at higher more optimal temperatures. Contribution of syngas contaminants need improved understanding.
Combustion
Flashback is an issue and increased residence time is needed for CO burnup.
Flashback due to H2 in the fuel, leanness limits to support low NOx and residence time needs for CO burnout, must be researched separately and in an integrated sense.
Prediction of Problems
Syngas-hydrogen fuels are likely to challenge predictive understanding of turbine robustness.
Research is very important to aid prediction of slow failures that would result in environmental issues, accelerated blade degradation, hindrance of engine cycling.
6.0.3 University Turbine Systems Research Program 6.0.3 Gas Turbine Industrial Fellowship As part of the UTSR Program, the Gas Turbine Industrial Fellowship Program offers students valuable work experience and the opportunity to practice the “art” of engineering in an industrial setting. Discipline areas, as applied to land-based gas turbine power generation systems, include mechanical design and manufacturing, heat transfer, aerodynamics, combustion, thermodynamic analyses, materials and coatings, and testing and evaluation. Emphasis is placed on gas turbine component design and manufacturing techniques, using state-of-the-art experimental and computational facilities. UTSR professors and industry engineering staff serve as mentors and advisors for the fellows. Students are exposed to gas turbine design techniques, analysis and system optimization methods, design limitations and practical problems encountered in the industry. Fellows participate in a 10-12 week work experience at turbine industry sites (manufacturers and end users). In order to participate in the program, students must be in good standing in an appropriate degree program at an accredited U.S. college or university that is a UTSR Performing Member. The program targets B.S. (graduating seniors), M.S. and Ph.D. graduate students. Applicants must be U.S. citizens or permanent resident aliens. The applicant’s selection is based on academic record, aptitude and gas turbine engineering interest, as well as the recommendation of the applicant’s advisor and engineering instructors. Fellows are paid a stipend of $700, $800, or $900 per week, depending on academic status of BS, MS or Ph.D. respectively, plus a one-time payment of $1200 to help cover travel and relocation expenses. Further details and background can be found in note 44. Process of the Fellowship Program Selection of Fellows: Annually, an announcement of the program is released, and applications are received and reviewed by industry. The applicants rank the companies they most want to work for, and the IRB companies rank the applicants they are most interested in, typically their top 4 or 5 since they get only 1 or 2 Fellows; one per Associate Member and two per Voting Member. The applicants respond, accepting or rejecting the offers. Typically the student acceptance rate is about 90%. For those cases where the applicant rejects the offer, SCIES, after conferring with the IRB company where that student was offered placement, makes offers to backup candidates as needed until all the available slots are filled. Reports and Presentations: At the conclusion of their projects, the Fellows are required to make a report on their projects, including objectives, procedure and results. Also they develop a PowerPoint description of the projects. One Fellow from each IRB company participates in a poster session at the annual Peer Review Workshop. The PowerPoint presentations are posted on the UTSR website. The presentations are cleared with the IRB companies for placement in the public domain.
Fig. 6 Sample project description from the 2004 Fellowship Assignments
442
William H. Day, Richard A. Wenglarz and Lawrence P. Golan Title: CFD Modeling of Rotor-Stator Cavity Purge Flow Author: Jonathan McGlumphy Virginia Polytechnic Institute & State University Industrial Mentor: Dr. Philip Andrew, Mgr., Turbine Aero Engineering Industrial Site: General Electric Power Systems, Greenville, SC facility Objective: Apply GE in-house CFD solver to simple rotor-stator cavity geometries to lay foundation for future work Achievements:
Reviewed literature to find simple cavity geometry with available experimental data Performed initial computations using GE in-house code Developed technique to allow density-based code to solve incompressible flow field Results compared favorably with published data and accompanying computations Confirmed that code can capture fundamental physics without modification to the code
Future Work
Code will be used to model flow fields of more complex geometries. Method could be used in the future to guide experimental set-ups
2005 Fellows The placement of the Fellows for 2005 is shown in Figure 4; a total of 20 applicants from 12 universities were placed in 13 companies. Fellow
University
Company Placement
Benjamin Asefa
University of Michigan
Parker Hannifin
Christopher Bolszo
University of California, Irvine
General Electric Company
Nicholas Cardwell
Virginia Polytechnic Institute
Pratt & Whitney/UTRC
Jeffrey Carullo
Virginia Polytechnic Institute
Solar Turbines, Inc
Jared Crosby
Brigham Young University
Pratt & Whitney/UTRC
Jason Habeger
Michigan State University
Siemens Westinghouse
Michael Hind
University of Wyoming
Ingersoll Rand Energy Systems
Jerrod Isaak
University of Wyoming
Clean Energy Systems
Richard Klop
Michigan State University
Rolls-Royce Corporation
Jonathan McGlumphy
Virginia Polytechnic Institute
Rolls-Royce Corporation
Adam Norberg
Virginia Polytechnic Institute
Ramgen
Clark Paterson
Colorado State University
Woodward FST
Travis Patterson
University of Central Florida
Siemens Westinghouse
Emmanuel Perez
University of Central Florida
General Electric Company
Stanton Peterson
University of Wyoming
BP
Patrick Sheppard
Vanderbilt University
Solar Turbines, Inc
Andrew Skoglund
University of Minnesota
Siemens Westinghouse
Ruwan Somawardhana
University of Texas, Austin
Precision Combustion
Paul Teini
University of Wyoming
Pratt & Whitney/UTRC
Scott Thawley
Virginia Polytechnic Institute
Capstone Turbine Corporation
In 2004 we contacted former Fellows concerning their permanent positions and heard back from 63 of them: To the gas turbine industry : To industry, non-gas turbine: To academia:
29 8 3
Subtotal who have started full-time employment: 40 Still in graduate school at last report: 23 So, of those who have started full-time employment, 73% went to the gas turbine industry, and an additional 7% went to academia. If we include those in academia, 80% of them are benefiting the gas turbine industry.
443
6.0.3 University Turbine Systems Research Program 6.0.3-4 Notes ___________________________ 1. SCIES website: http://www.clemson.edu/scies/; click on “University Turbine Systems Research” for information on the UTSR Program. 2. Ibid. 3. Ibid. 4. SCIES website: http://www.clemson.edu/scies/ then click on “Gas Turbine Fellowship”; W. Day, “University Turbine Systems Research Program: An Innovative Approach to Graduate Education” ASME Paper number MECE2005-82534 (to
be presented at the ASME International Mechanical Engineering Congress & Exposition, Orlando, FL, November 2005).6.0.3
444
BIOGRAPHY
6.0.3 University Turbine Systems Research Program
William H. Day Outreach Manager South Carolina Institute for Energy Studies (SCIES) Clemson University phone: (860) 404-0759 email: [email protected] www.clemson.edu/scies Dr. Day is a graduate of Cornell University BME 1960, Polytechnic University MSME 1966, and Polytechnic University Ph.D. ME 1970. Dr. Day started as an engineer with General Electric in 1960. During his career with GE’s Gas Turbine Division he was responsible for establishing and managing the High Temperature Turbine Technology Program, and managed the business with EPRI and the US Department of Energy. In 1979 Dr. Day joined the United Technologies Corporation, where he managed a joint gas turbine design study between Pratt & Whitney and Siemens. He was also responsible for negotiating the contract with the People’s Republic of China that launched the FT8 gas turbine and for directing the development of the FT8. He was responsible for negotiating the contract with Siemens that led to development of the V84.3A gas turbine and was in charge of that development effort at UTC. From 1995 until his retirement from United Technologies in 2002, Dr. Day was Manager of Advanced Engine Programs for Pratt & Whitney’s industrial gas turbine business. In 1995, Dr. Day led the founding of the Gas Turbine Association, the trade association for the gas turbine industry. He served as Chairman of the Board of GTA from its founding in 1995 until his retirement from United Technologies in 2002. In 2001 Dr. Day was named a Fellow of Pratt & Whitney in recognition for his expertise in industrial gas turbines. In 2002, Dr. Day joined the South Carolina Institute for Energy Studies (SCIES) as Outreach Manager. SCIES coordinates the university research programs for the U.S. Department of Energy’s University Turbine Systems Research (UTSR) Program. Responsibilities and accomplishments include a substantial expansion of the Gas Turbine Industrial Fellowship Program and of the Industry Review Board, publication of numerous articles in gas turbine trade journals and ASME papers on the UTSR Program, while working closely with the Department of Energy. Dr. Day has four patents and has published over 60 technical papers.
Richard A. Wenglarz South Carolina Institute for Energy Studies (SCIES) Clemson University phone: (864) 656-0142 email: [email protected] www.clemson.edu/scies Richard Wenglarz received B.S. and M.S degrees from the University of Illinois, and Ph.D. degree from Stanford University, all in Engineering Mechanics. He has held positions at the University of Newcastle Upon Tyne, Bellcomm, Bell Laboratories, Westinghouse R&D Center, Rolls Royce/Allison Division of General Motors, and South Carolina Institute for Energy Studies (SCIES) at Clemson University. His early experience involved dynamics and control for gyroscopic systems and manned space stations. Later experience concerned developing and applying analytical and experimental methods to evaluate deposition, erosion, and corrosion (DEC) in advanced energy systems (e.g., gas turbines and fuel cells) operating with alternate fuels. Currently, Dr. Wenglarz is Manager of Research at SCIES for the DOE sponsored University Turbine Systems Research (UTSR) program supporting university gas turbine research nationwide. Dr. Wenglarz has over 80 publications and presentations including invited presentations at the Von Karman Institute for Fluid Dynamics, Yale University, UK Central Electricity Research Laboratories, Cambridge University, and the Kentucky Energy Cabinet Laboratories.
Lawrence P. Golan South Carolina Institute for Energy Studies (SCIES) Clemson University phone: (864) 656-2267 email: [email protected] www.clemson.edu/scies
Lawrence P. Golan is presently Special Assistant to the Vice President for Research at Clemson University (2003 – present). In addition, Dr. Golan is Director of the South Carolina Institute of Energy Studies - SCIES. (1986-2003) He established the Energy Systems Laboratory, a unit blending energy facilities, academics and research. Dr. Golan maintains a working relationship with the Gas Turbine Association, the Alliance to Save Energy and the Southeast Energy Efficiency Alliance. He received his Ph.D. from Lehigh University and his B.S. and M.S. from West Virginia University. Dr. Golan’s professional activities include: Chairman 1996 National Heat Transfer Conference hosted by AIChE, ASME, AIAA and ANS; 1990, 1994 and 1999 Chair of the AIChE Heat Transfer Division; 1992 National Heat Transfer Conference Best Paper co-Chair; 1992, 1994 and 1996 Chair Kern and Jakob Award Committee; member National Heat Transfer Conference Coordination Committee; AICHE Chair 2004 Summer Heat Transfer and Fluids Engineering Conference, advisor to the State of Illinois Coal Combustion Program; and member of the Academic Advisory Committee West Virginia University Mechanical Engineering Department. Dr. Golan has authored 34 articles, offered 13 short courses and organized numerous energy conferences and workshops. Prior to his present position at Clemson University, Dr. Golan was employed by Exxon Research and Engineering Company (presently ExxonMobil) for nineteen years. Since 1992 Dr. Golan has directed a nationwide activity of 110 universities and 10 U.S. corporations for the U.S. DOE that is providing technology necessary for developing the next generation of advanced land-based power generation systems.