Title: Author/Presenter:

Controlled Catalytic Combustion Module (CCM) William A. Whittenberger, PE, President, Catacel Corp. 5950 Eagle Creek Road, Leavittsburg, OH 44430 USA Ph: 330-898-3406, Fx: 413-638-3271, Email: [email protected] William B. Retallick, PhD, PE, Vice President, Catacel Corp. 1432 Johnnys Way, West Chester, PA 19382 USA Ph: 610-399-1371, Fx: 610-399-1478, Email: [email protected]

Co-Author:

Abstract: Catalytic combustion reactors for gas turbines present many challenges to the designer. The reactor must have enough activity to sustain the desired conversion. It must be small enough to meet packaging constraints. If possible, the reactor should light off at compressor discharge temperature, avoiding the need for a pre-burner. Heat from the reaction must be carefully managed to avoid damage to the catalytic and substrate materials. The reactor and its mounting system must be mechanically durable under a wide range of conditions. This paper reviews these issues, and proposes solutions that meet the design challenges. Keywords: catalytic combustion, catalytic reactor, catalysis, metal substrate catalyst, palladium catalyst, controlled combustion, NOx control. 1.0 Introduction Catalytic combustion has been successfully demonstrated as a means for significant NOx reduction on gas turbines. Good results have been demonstrated with a catalytic combustion system implemented on a Kawasaki M1A turbine [1]. Even with this success, broad-based implementation of catalytic combustion across range of turbine designs has been limited to date by cost and risk factors. One approach to reducing cost is the development of a high-performance reactor a reactor that is relatively small, and requires no preburner. To this end, Catacel Corp. has been developing a CCM, or controlled catalytic combustion module. The CCM promises to have performance that will enable widespread new and retrofit implementation of catalytic combustion on gas turbines. 2.0 Typical Performance Objectives Gas turbines operate at a wide range of conditions. Operating conditions are influenced by the size of the turbine. Considering that the CCM may be used for retrofit, one must understand the operating conditions of the installed base, which represents turbines that were designed 20 or 30 years ago. Table 1 shows key variables that influence reactor design, across a wide range of gas turbine models. Pressures, temperatures, and flows have been taken from manufacturers data [2,3,4]. Module dimensions are estimates of the largest reactor size that will comfortably package with the turbine. Space velocities are calculated as a function of the other values.

Turbine Pressure ratio Combustion temp, degC Compressor discharge, degC Flow per module, kg/sec Module diameter, mm Module length, mm Space velocity, 1/sec

Microturbine (estimate) 3.5 850 400 0.3 250 250 20

S-W W251AA 9.8 925 325 14 300 300 541

GE GE Alstom (S-W) MS7001EA MS7001FA GTX100 11.8 15.7 20.0 1370 1480 1550 350 410 440 24 45 3.2 350 400 150 350 400 150 584 734 990

Table 1 Turbine Design Variables Three important pieces of information are apparent from the table. First, the combustion temperatures vary widely, which means that reactor temperature management must be designed specific for each class of turbine. Second, the highest space velocity occurs on the GTX100 turbine. If a reactor has enough activity to perform in that application, it will probably work on the other applications. Third, the lowest compressor discharge temperature occurs on the older W251AA class turbines. This will create the biggest challenge for reactor lightoff. 3.0 Fundamental Design Concepts The CCM fundamental building block is a metal structure layered from flat and corrugated foils. Flat and corrugated foils are cut to a specific leaf length. Each flat and corrugated leaf is coated in specific areas with catalytic materials. Leaf assemblies are made by spot welding each flat to a corresponding corrugated leaf. The leaf assemblies are then arranged in a spiral pattern to form a generally round or annular reactor. The outer ends of each leaf assembly are attached to each other and to the container, thus forming a rigid structure. Figure 1 illustrates the various assembly steps.

Figure 1 Reactor Assembly

The foils used are preferably a fecralloy variant (nominal Fe-20Cr-5Al-Y), available from a number of suppliers. These alloys are relatively inexpensive, have good oxidation resistance to about 1000C, but have minimal strength at high temperature. Some reactor designs will require additional high temperature strength. In those cases, one would choose a high-performance alloy such as Haynes 214 or Haynes 230. These alloys are more expensive, have similar oxidation performance, and much higher strength at temperature. It is important to note that with either alloy system, temperature within the reactor must be limited to 1000C or less to limit oxidation of the foil. This means that for newer turbines with high firing temperatures, only a part of the combustion can take place inside the reactor. An appropriate burn-out zone must be located downstream of the reactor to complete the combustion. In those cases where only part of the combustion must take place in the reactor, a one-sided coating strategy is employed. Retallick [5] and Dalla Betta [6] pioneered this strategy in the late 1980s. One side of each foil leaf is coated with catalyst, and the other side is left uncoated. The layers are arranged so that coated sides face each other to form catalyzed channels, and uncoated sides face each other to form uncoated channels. Combustion goes to completion in the coated channels. However, no combustion takes place in the adjacent uncoated channels. The net effect is that only half of the fuel is burned (assuming the channels are of equal size), and the bulk temperature rise is limited to half of the adiabatic rise. Figure 2 illustrates the process.

Figure 2 One Sided Coating Note that the bulk temperature must remain below the auto-ignition temperature of the fuel, or the strategy fails due to combustion taking place in the uncoated channels. Since the auto-ignition limit for natural gas is on the order of 900C (at 1 atm), one must limit the reactor temperature to approximately that level. Fortunately, this level is very compatible with the oxidation performance of the foils, and thermal durability performance of many known catalyst systems.

4.0 Temperature Management In the case of the microturbine and W251AA applications in Table 1 where the combustion temperature is 850-925C, the one-sided coating strategy need not be employed. Here, all the combustion can take place in the reactor without damage, and the reactor can be completely coated with catalyst. In the case of the other three applications, a one-sided coating strategy would result in reactor temperatures of 860, 945, and 995C respectively. The 945 and 995C cases could be subject to auto-ignition and should be carefully analyzed to make sure that auto-ignition cannot occur. A class of turbines exists in the field with combustion temperatures in the 11001200C range. Clearly, these temperatures are too high to permit full combustion in the reactor. The use of one-sided coating on these systems results in reactor temperatures on the order of 750C, which is quite low. To get maximum NOx control, one would like to see reactor temperatures in the 900C range. In principle, this temperature situation can be managed by stacking two onesided coating reactors, while providing a small mixing area between the reactors. The first reactor burns half of the fuel, resulting in a temperature in the mixing area of approximately 750C. The burned and unburned gases mix before entering the second reactor. The second reactor burns half of the remaining fuel (75% total), resulting in a desirable outlet temperature of about 950C. In practice, the CCM can use an internal mixing area to accomplish this same effect. Two half-width corrugated leaves are welded to a common flat leaf to form the leaf assembly. Corrugations on the corrugated leaves are aligned 180 degrees out of phase from one another, so that gas exiting each channel of the first section splits between a coated and uncoated channel in the second section. This construction method allows the overall reactor to retain the structural integrity provided by the wide flat leaves. Figure 3 illustrates the technique.

Figure 3 Offset Corrugations for Mixing Note that the mixing technique may be used more than once within the reactor to achieve a variety of temperatures within or at the exit of the reactor. In certain circumstances, this technique is employed as a way to keep catalytic materials within their appropriate operating windows.

Figure 4 shows typical performance curves of two families of palladium catalysts that are employed in the CCM. Palladium catalysts always exhibit a reversible transition from palladium oxide to palladium metal. The manifestation of the transition is highly dependent on the catalyst composition, and is a function of temperature and pressure.

Figure 4 Catalyst Performance In this figure, a perfect catalyst would approach the adiabatic temperature line. Catalyst A lights off slowly, then behaves near perfect until the temperature reaches approximately 850C, where the palladium begins the transition. Here the transition manifests as a flattening of the curve. Catalyst B lights off quickly, and behaves near perfect until about 750C, where activity drops sharply when the palladium transitions from oxide to metal. This transformed catalyst then gains activity until it is near perfect again above 900C. Clearly neither of these catalysts are suitable individually for the reactor. Yet, when used together and made to operate in appropriate temperature windows, they can be very effective. Catalyst B works well for lightoff, and also in the high temperature areas. Catalyst A works well in the mid-range where Catalyst B loses its effectiveness. Note that Catalyst A can be used as a temperature regulator. No matter how much fuel you add, the temperature will not rise above the flat area of the curve. Catalysts are normally applied to the foil leaves in zones that correspond to the expected temperature profiles. The use of internal mixing helps to better define those temperature zones, to make sure that each catalyst system remains in its appropriate temperature window, regardless of turbine operating condition. 5.0 Achieving Lightoff Catalyst lightoff is an important system cost driver. If a pre-burner can be avoided, the combustion system can be greatly simplified. An unmodified CCM using conventional catalytic materials will light off on natural gas at about 375-400C. Ongoing

materials research promises to push this temperature lower, but there will always be risk on the older turbines with compressor discharge temperatures in the 325-350C range. Two strategies may be employed in the CCM to promote better lightoff. Both involve getting additional catalytic materials near the leading edge of the reactor. First, a narrow band of catalyst may be applied in the normally uncoated channels on the leading edge. This allows a small amount of combustion to take place in the uncoated channels, generating additional heat and activity on the leading edge, and allowing system lightoff to occur at a somewhat lower temperature. Care must be taken to make this band just wide enough to help with the light off, while not significantly impacting the uncoated behavior. Second, narrow strips of corrugated and coated foil (coated both sides) may be embedded in the leading face of the coated channels. Figure 5 illustrates the concept. The embedded strips have a corrugation height approximately half the height of the primary corrugations, and a pitch that is identical. This allows the strips to embed. The strips create additional, smaller catalyzed channels near the leading edge. The strips are somewhat insulated from the cooling effect of the uncoated channel, and will light off much faster than the main strips.

Figure 5 Embedded Strips in Lightoff Zone 6.0 Mounting Issues The CCM must be correctly mounted and sealed in the turbine combustion system so that thermal expansion, heat transfer, and bypassing are all properly accounted for. The foil leaves are welded securely into a relatively thin metal shell. If unrestrained, the shell will track the thermal expansion of the foil and maintain structural integrity with and a seal to the foil. Ideally, the shell should be sealed to, thermally insulated from, and allowed to float within the primary combustor structure. However, these objectives are in conflict with one another. The arrangement shown in Figure 6 provides a good compromise. A ring is welded to the CCM metal shell near the inlet, where the CCM is coolest, the material is strongest, and thermal expansion is minimal. This ring is trapped by a pocketed surface in the primary combustor structure, and held and sealed in place by an annular cap. Near the downstream end of the CCM, an air gap is provided which insulates the CCM from the combustor structure, and allows room for the hot end to expand.

Figure 6 CCM Mounting Arrangement 7.0 Performance Evaluation Much of the fundamental catalyst validation for the CCM has been done on Catacels 1 and 5-atmosphere combustion rigs. Two catalyst combinations have operated for over 500 hours at 1 atm., with realistic temperatures and space velocities, with no evidence of deterioration after the initial burn-in. Figure 7 shows the performance of one such combination over a period of 500 hours. During the first third of the chart (the burn-in period), some deterioration is evident, as the relationship between the temperatures change somewhat. During the remainder of the chart, performance is very stable.
41-77 Performance
1000

900

Temperature, degC

800

700

600

500

400
1

500 hours

Figure 7 Performance over 500 hours Alstom Power has also evaluated catalysts on their high-pressure test rig as part of their Hybrid Burner project, with very encouraging results. This rig has also been used to quantify performance and validate some of the concepts presented here, such as one-sided coating, light-off bands, and internal mixing.

8.0 Summary A variety of concepts have been presented that enable a high-performance catalytic reactor for gas turbines. Preliminary work on this controlled catalytic combustion module (CCM) shows promise for achieving the desired high performance at reasonable cost. A successful completion of this development effort should lead to widespread implementation of catalytic combustion for gas turbines. Acknowledgements: The author wishes to thank his co-author and business partner, Dr. William B. Retallick for his help and effort on the CCM project. He also wishes to thank Richard Carroni and Tim Griffin at Alstom Power for their cooperation and wise counsel on the Hybrid Burner Project. References: 1. Dr. Robert Peltier, Gas turbine combustors drive emissions toward nil, POWER, March 2003. 2. Catalytic Reactor Functional Specification for Application to the GEPS MS7001EA Gas Turbine Catalytic Combustion System, draft circa December 1998. 3. Data from General Electric website. 4. Data from Alstom Power website. 5. William B. Retallick and William R. Alcorn, U.S. Patent # 5,202,303 6. Ralph A. Dalla Betta et. al., U.S. Patent # 5,250,489