INTERNATIONAL JOURNAL OF ENERGY RESEARCH, VOL.

20,693-71 1 (1996)

CATALYTIC COMBUSTION FOR INDUSTRIAL GAS

TURBINES
DONALD ANSON

Process Engineering Dept, Batteiie, 505 King ALE,Columbus,OH 43201-2693, U . S A

MARIO DECORS0

Power Tech. Associates

AND

WILLIAM P. PARKS, JR

US.Department of Energy
This brief review provides a general account of work directed at the use of catalytic combustion in gas turbine
engines. A major potential advantage of using catalytic combustion is that the fuel can be burnt efficiently at
temperatures low enough ( < 15OOOC) to avoid significant oxidation of atmospheric nitrogen. This advantage was less
important when catalytic combustion was demonstrated in the 1970's than it is today and received relatively little
attention until the following decade.
After discussion of the principles involved in the design of a combustor that must meet the mixing, size,
performance and durability goals of a based gas turbine application, the review turns to accounts of experiments
conducted on a laboratory scale with simple configurations. These established basic operating parameters for
satisfactory combustion performance and led to larger scale work and to prototype design concepts for industrial gas
turbines in the late 70's and early 80's. Test results were encouraging but were not pursued definitively in the U.S.A.
Activity continued at several centres in Japan, with exploration of a number of different catalyst arrangements,
geometries, and control systems, again with encouraging results. At the same time, there has been renewed interests
in the U.S.A. and in Europe, spurred largely by the emphasis on reducing emissions of nitrogen oxides (NOx).
The paper concludes with suggestions for further development of catalytically stabilized combustion systems for
gas turbines. These systems must ensure adequate pre-catalyst temperature, with evenly premixed fuel and air, and
sufficient temperature rise across the catalyst to ensure effective completion of reaction in a homogeneous reaction
mode. The outstanding problems are largely concerned with questions of catalyst integrity and longevity in practical
configurations and realistic engine operating conditions.

KEY WORDS: catalytic; combustion; emission; gas turbine; nitrogen oxides

1. INTRODUCTION
Work by Westinghouse, with Engelhard Industries, in the 1970s showed the feasibility and desirability of
applying catalytic combustion to gas turbines. The use of catalysts facilitates efficient combustion of
hydrocarbons at relatively low temperatures without promoting oxidation of nitrogen. The reaction of
atmospheric nitrogen with oxygen is slow at temperatures below 1500"C, whereas the oxidation of
hydrogen, carbon monoxide and many hydrocarbon fuels in the presence of a catalyst proceeds freely at
temperatures as low as 200" C and can near completion in a few milliseconds at temperatures below
1Ooo"C.
Reconsideration of catalytic combustion is appropriate at this time because developments in ceramics
have made it possible to produce improved substrates for high temperature catalysts and to conduct

CCC 0363-907)3/96/080693- 19 Received 12 August 1993

0 1996 by John Wiley & Sons, Ltd. Revised 12 September 1993
694 D.ANSON, M. DECORSOAND W. P. PARKS

combustion in a near adiabatic (hot wall) environment. Structural ceramics are being considered for
other applications in gas turbines and these developments increase their acceptability as combustor
materials. Combustors fabricated from ceramics, with ceramic catalyst elements, could produce benefits
in lower maintenance costs for the combustor itself and for downstream components. Specifically, they
could
(llimprove combustion efficiency; (2)reduce emission of NO,, CO and hydrocarbons; (3)eliminate
impact damage to the turbine by carbon residues from the combustor; (4)reduce peak gas temperatures
in the hot gas path.The purpose of this review is to cover some selected literature relevant to the
application of catalytic combustion in gas turbines and to make a case for further development in this
field of work.

1.1 Catalyst Technology

Catalysts facilitate reactions, and thus increase reaction rates between substances, without themselves
undergoing permanent change. The way in which they work is imperfectly understood, but is not
important to this discussion.
Catalytically stabilized combustion has been thoroughly reviewed by, for example, Pfefferle and
Pfefferle (1986) and Prasad et al. (1984). Combustion catalysts typically are metals or metal oxides. In
most practical systems catalysis occurs when two reacting gaseous species contact the solid catalyst
surface simultaneously and the reaction is therefore heterogeneous. Many common metal oxides,
including those of iron and copper, are effective combustion catalysts, but platinum, palladium and other
noble metals are more often used because of their superior activity at low temperatures. The formula-
tions of commercial catalysts and methods of manufacture are essentially empirical, and are considered
to be highly confidential in the catalyst supply industry. In catalytically stabilized combustion the energy
released by heterogeneous catalytic reaction at the surface supports non-catalytic homogeneous reaction
in the free reactant stream.
Figure 1 represents qualitatively the relationship between reaction rate and the temperature at the
catalyst surface in a catalytically stabilized system. At low temperatures, the process is catalytic and is
kinetically controlled. At higher temperature, the process becomes self-accelerating, as represented by
the steep rise in the curve. The limiting rate of catalytic reaction is determined by mass transfer
considerations which control the transport of reacting species to, and products from, the catalytic surface.

Mass Transport
--
w
Kinetic Control
(NonCatalytic)
t

Kinetic Control Control

1-( ytiC
8
-d Limiting
Catalytic

!i
W
.L
0
d
B

Gas Tompewatwe
Figure 1. Reaction rate limitations in catalytically stabilized combustion.
 TURBINE CATALYTIC COMBUSTION 695

A further increase in the overall reaction rate occurs with increasing temperature, as homogeneous
reactions accelerate in the free stream.
Typically, catalysts are supported on a honeycomb structure to enhance contact with reacting gases
flowing through the honeycomb passages. Because the rate of diffusion of unreacted species towards the
catalyst is directly related to the transverse concentration gradient, the diffusion rate of reactants to the
catalyst surface declines in the direction of flow. As long as transport phenomena are limiting, the rate of
catalytic reaction per se will be independent of temperature but the overall reaction rate, including free
stream homogeneous reactions, will be sensitive to gas temperature. The final temperature is determined
by the mixture stoichiometry.
In catalytically stabilized combustion, homogeneous combustion occurs in the free gas stream as a
result of the transport of active species and thermal energy from the catalytic reaction zone, which acts as
a flame stabilizer. The rate of homogeneous reaction in a premixed system is dependent on the
stoichiometry of the mixture, the pressure and the temperature, which must be above the autoignition
temperature for the mixture.
The minimum temperature for non-catalytic combustion of methane to proceed at an acceptable rate
in a practical gas turbine combustor is about 1000° C. This is near or above the design turbine inlet
temperatures of some small industrial turbines and above the low load operating temperature of most gas
turbines.
Figure 2 shows qualitatively how the temperature and reaction rate change in the direction of gas flow
in a simple catalytically stabilized combustor. The catalyst section is terminated at a point where stable
non-catalytic reaction can proceed. At this point, the overall reaction rate may fall momentarily.
Inlet air temperatures to turbine combustors are typically above 250 C. At this temperature, a
O

number of materials are catalytically active, and as temperatures rise following ignition the choice of
catalysts become broader. Pfeffele, (1974, 197.51, for example, used copper chromate in an experimental
combustor simulating gas turbine conditions. The inlet mixture was preheated to start the system but
combustion was apparently self-sustaining with ambient temperature feed after ignition had been
established. Common constituents of ceramics such as alumina and silica are catalytically active at the
higher temperatures (> 750 C) developed in combustors. Work at Lawrence Berkeley (Schefer and
O

Robben, 1980) suggests that at about 1000" C silica has the same catalytic activity as platinum in the

2!
a
CI

i
E
8 *

- E
a
-
Gas Flow

Gas Flow

-- --
I I I
k
Catalytic Zone ' Homogeneous A
Reaction Zone
Figure 2. Reaction rate and temperature changes in catalytically stabilized combustion (qualitative).
696 D. ANSON, M. DECORSOAND W. P. PARKS

oxidation of ethane and propane. This is important because at this temperature the noble metal catalysts
tend to deteriorate rapidly.
According to Prasad et al. (19841, of the noble metal catalysts only platinum and palladium appear to
have reasonable resistance to volatilization. Both are liable to sinter, and have been found to disperse on
the surface of an A,O, substrate. Palladium forms thermally stable products with alumina, whereas
platinum tends to form large crystallites, with a consequent reduction in area. The noble metals have a
high level of catalytic activity at lower temperatures (below 700" C) and are therefore of special value in
the ignition zone, where their thermal instability is of less concern.
Among base metals, the oxides of cobalt, nickel, chromium an copper are good combustion catalysts.
Combinations of such oxides, or monolithic ceramics which include them, offer promise for the higher
temperature zones of catalytic combustors. Thus a two state catalyst concept, using noble metals for
ignition and base metal oxides as the main catalyst, may provide the needed combination of low
temperature light-off and adequate catalyst life. The R & D Center of the Osaka Gas Co. Ltd. (1990) has
reported on the use of a monolithic catalyst incorporating lanthanum, strontium and manganese with
alumina. This is referred to as manganese hexaluminate. The total catalyst stack used by Osaka Gas
consisted of palladium on cordierite in the first stage and two grades of Mn hexaluminate in series.
The durability of the catalyst/substrate combination at the temperatures found in gas turbine
combustors, with desired life cycles of 10000 hours or more, is as important as the level of catalytic
activity. Some loss of activity in the zones of highest temperature may be acceptable, but activity must
remain high in the ignition zone. Combustion catalysts used for hydrocarbons are generally ineffective in
promoting the formation of nitrogen oxides, which gives catalytic combustion systems an inherent
advantage in controlling undesirable emissions.

2. DESIGN OF CATALYTICALLY STABILIZED COMBUSTORS

Figure 3 illlustrates the essential parts of a three-stage catalytically stabilized combustion device. These
stages are:
(1)a premixing stage, in which a homogeneous fuel/air mixture is achieved; (2)a heterogeneous
catalytic combustion stage in which ignition occurs and the mixture is heated to at least 1000"C; and
(3)a homogeneous non-catalytic stage in which combustion is completed at temperatures up to 1500" C.

To start the system, preheat is provide by a small diffusion type combustor upstream of the catalytic
surface. In a gas turbine, the rise in temperature across the compressor may be enough to ensure ignition
when the engine approaches full speed.
Because the output of a gas turbine is controlled by the temperature of the combustion products, the
concept allows for diluting the gases after combustion. With existing turbine technology, the final
homogeneous combustion stage probably could not be completed under low load conditions without this

UTALIST I
BED DlLUTlCU
A I R WORTS
(VrmIULE)

Figure 3. Catalytically stabilized combustor (schematic).

 TURBINE CATALYTIC COMBUSTION 697

provision. With it, homogeneous combustion will occur at 1000" C to 1500" C, while the turbine inlet
temperature will vary in a lower range, typically 850" C to 1350" C .

Some of the more important factors in the design of catalytically stabilized combustion systems are
discussed in the following sections, They are: fuel/air mixing;catalyst surface area and geometry;
homogeneous reaction space; dilution zone; catalyst deterioration; substrate integrity; and control.
Preheat requirements can be met most easily by using a small diffusion flame burner installed in the
compressor delivery to the combustor. The thermal input of the preheat burner needs to be about 10 to
15% of the design full load engine fuel demand. This burner must be fully self-stabilizing (using for
example bluff body or swirl type stabilizers) and must incorporate an igniter. In practice, mixture preheat
levels may be significantly above the catalytic ignition temperature in order to permit a reduction in
overall residence time and pressure drop in the catalyst.
The preheat burner must not be located in a position where it could act as a flame stabilizer for the
main air/fuel mixture under full load conditions. Such preignition could lead to serious overheating of
the catalyst and the premix zone hardware. A preheat burner used only for ignition does not pose any
special durability problems. If, however, it is operated as a continuous preheat burner it can introduce
both thermal and chemical inhomogeneities which will affect catalyst performance and endurance, and it
can be a major source of NO,.
In a catalytic combustion system, the existence of fuel-rich streaks would led to local overheating of
the catalyst and its support and to premature catalyst deterioration. Conversely, fuel-lean streaks may
lead to incomplete combustion in the homogeneous reaction zone. The fuel must be injected sufficiently
upstream of the catalyst to ensure the attainment of a uniform mixture in the space available. For liquid
fuels, prevaporizers may be required. The design of premix sections must be given careful attention to
avoid dangerous volumes of premixed air and fuel within the engine casing and obstructions which could
act as bluff body flame stabilizers and lead to ignition upstream of the catalyst. Flow velocity in the
premix zone must be high enough to avoid flashback from the catalyst.
It is anticipated that reaction rate in most of the catalytic combustion zone will be mass transfer
controlled. The unreacted fuel gas must be transported to the catalytic surface by a combination of
forced turbulent mixing and diffusion. Turbulence absorbs energy and causes pressure drop. The surface
profile of the catalyst and the width/length ratio of the gas flow passages need to be adjusted so that a
sufficient proportion of the gas mixture contacts the catalyst to result in combustion being around 80%
complete. The pressure drop will continue to increase with length of the catalyst passages while the
concentration of unreacted fuel molecules will decline exponentially in the direction of flow, and with it
the reaction rate. From both physical and chemical points of view, therefore, the passage geometry is of
prime importance and needs careful optimization. Catalyst geometry must be based on turbulent flow and
boundary layer theory as well as reaction kinetics. The essential criterion will be to achieve, in a given
overall space, the required fuel conversion (or temperature) with the lowest pressure drop and a near
homogeneous outlet condition. While honeycomb structures have been favored as catalyst supports, they
have the disadvantage of eliminating cross mixing and thus preserving any mixture inhomogeneities that
may exist. Improved designs may emerge in which transverse mixing will occur, resulting in more uniform
exit temperature profiles.
The existence of slightly fuel-rich or fuel-lean streaks in the mixture downstream of the catalyst will
not be critical unless it leads to a significant fraction of the gas (on the order of 10% or more) reaching
temperatures above 1500 C and maintaining those temperatures long enough to effect oxidation of
O

nitrogen, or unless they persist into the turbine. Therefore, if the fuel and air are mixed properly before
entering the catalyst zone, the homogeneous reaction zone can be a low turbulence, low pressure drop
region.
The dilution air introduced after combustion can be used to present either a uniform temperature
profile or a controlled temperature gradient to the turbine inlet plane. The dilution air will be at a
pressure equivalent to the combustor inlet pressure, which will be higher than the combustor exit
pressure by about 2% of the absolute pressure.
698 D. ANSON. M.DECORSOAND W. P. PARKS

Catalytic surfaces are subject to deterioration by a number of processes, which include the following:
(1) poisoning by specific substances present as impurities in the reactants, for example, sulphur;
(2)modification of surface structure, for example by sintering or glazing, which reduces surface area;
(3)modification of surface chemistry by reaction with the substrate; (4)fouling by deposits; (5)loss of
active species by volatilization; and (6)loss of catalyst by spalling or flaking from the substrate.
For natural gas fired combustion systems, poisoning is not expected to be important, and at the
temperatures existing in combustors many poisons will in any case be volatile or unstable. Poisoning is
mostly a concern in the ignition zone where temperatures are moderate.
Modification of the surface structure will be important if diffusion of fuel molecules into the catalyst
pore structure is critical. Pfefferle et al. (1975) suggest that this is not the case, at least for most catalysts,
and that turbulent diffusion to the surface is more likely to be limiting.
Fouling of the catalyst is also unlikely with natural gas fuel. Thus the most likely causes of
deterioration will be catalyst volatilization, chemical modification or physical breakup of the catalyst
coating, assuming that it is applied to a ceramic substrate. All of these are potential problems at
temperatures between 1000 and 1500" C, which may be encountered. It is clearly desirable to design for
combustion to be completed in the non-catalytic homogeneous reaction mode, in which case the catalyst
temperature can be kept below about 1200" C . Even so, the catalyst coating must be fully compatible with
the substrate in terms of its chemical affinity and physical properties such as thermal expansion. Thin
coatings are less likely to spall, but more likely to fail by volatilization or chemical damage.
Catalyst deterioration depends on the design of the catalyst matrix (which can be manipulated to limit
the extent of reaction and avoid overheating), on the selection of mutually compatible catalyst/substrate
pairs and on the method of catalyst application.
Substrates for catalysts are often made in the form of honeycomb structures in which the passages are
usually of square or hexagonal cross section. These honeycombs may be extruded from ceramic pastes or
glasses using any one of several starting materials. Cordierite (2Mg0.2Al20,-5Si0,) is frequently used
because of its very low coefficient of thermal expansion (1-2 x lo-"/ " C) and resistance to thermal
shock.
Catalyst support substrates are subject to severe temperature transients during light-off and on
shutdown, and also to temperature gradients in the direction of flow during steady state operation.
Thermal stresses resulting from these conditions are the major cause of physical failure. The integrity of
the catalyst support structure is a major concern in the design of catalytic combustion systems for gas
turbine applications. Any breakup of this structure is liable to cause impact damage to the turbine vanes
and blades. Desirable materials properties are low thermal expansion coefficient and elastic modulus, and
high thermal conductivity and strength. For combustion catalyst supports, the operating temperatures (up
to 1200" C) are such that the properties of cordierite are probaly inadequate or at best marginal, taking
into account the possibilities of locally higher temperatures and the fact that the substrate properties may
be modified by the catalyst coating or by ceramic wash coats that are sometimes applied to increase
surface roughness and increase area before the catalyst is deposited. Other possible support materials are
alumina and mullite. These maintain their strengths at higher temperatures, but have higher rates of
expansion, as shown in Table 1.

Table 1. Properties of potential catalyst support materials

Elastic Coeff., Thermal Max Fracture

Strength Mod. T . Exr Cond. Service Toughness
Material Composition (MORXMPa) (CPa) ( X 10 ) (W/mK.) Temp (" C) (MPam''')

Alumina Al,O, 300-1000 380 7.2-8.5 27.2 2000 3.2

Mullite 3 A , o , 2Si0, 180-500 145 5.7 5.2 1800 3.0
-
Cordierite 2 M g 0 2 N 2 O 15Si02 206-334 117 0.6-2.4 1470 2.5
 TURBINE CATALYTIC COMBUSTION 699

As well as the properties listed in Table 1, the chemical compatibility of the substrate with the catalyst
and the combustion gases must be considered. Oxide ceramics are more likely than non oxides to resist
combustion gases, but they may react with wash coats or oxide catalysts.
It is possible to reduce the vulnerability of catalyst support structures to thermal stress by modifying
the geometry. Stress analysis methods developed for thermal regenerators (Garrett 1987) are applicable
to this problem. Designs other than those based on honeycomb concepts may emerge. Toughened
(composite) ceramics may also be applicable. Of these, silicon-carbide-reinforced glass ceramics offer
good physical properties but probably lack the necessary high temperature capability, and they are more
difficult to form than the monolithic ceramics. Some of the problems described may be avoided by using
catalytically active structural materials in monolithic form rather than by coating catalysts onto an inert
matrix, at least in higher temperature zones.
For industrial gas turbines, there appears to be no major problem in designing a viable control system
for catalytically stabilized combustors and in many ways it could be simpler than systems currently used
for diffusion type combustors because of the inherently high stability of catalytic combustion. The
combustion controls must achieve the following objectives:
(1) keep temperatures within the catalyst in a range that will minimize catalyst deterioration; (2)
maintain catalyst exit temperatures at a level that will ensure efficient combustion without NO,
formation; and (3) modulate the combustion process to match the engine load demand.

2.1 Summay of Design Concerns

The major concerns in the design of catalytically stabilized combustors are: achievement of preheat;
space requirement for mixing, catalytic and homogeneous reaction sections; optimization of catalyst
surface area and geometry with low pressure drop; selection of substrate material/catalyst pair; substrate
integrity; and control to maintain temperatures in acceptable ranges.
Of these, the substrate integrity is very critical because it affects the reliability and maintainability of
the engine. The selection of compatible and effective catalyst/substrate pairs and procedures for
manufacturing durable catalytic matrices are closely related. The other concerns are less critical to the
technical feasibility of the development but are important to the overall cost, performance and durability.
They are also affected by the overall engine design parameters.

3. EXPERIMENTS IN CATALYTIC COMBUSTORS FOR GAS TURBINES

Workers at NASA Lewis Research Center (Anderson et al., (1975) and at Wright-Patterson Air Force
Base (WPAFB) (Blazowski and Walsh, 1975) published papers describing the performance of catalytic
combustion systems aimed at applications in gas turbines. Blazowski and Walsh (1975) at WPAFB
pointed out the potential advantages of lean mixture catalytic combustion in terms of low NO, formation,
combustion stability, and improved temperature profiles. They also recognized the problems discussed
earlier. In their paper there is considerable discussion of catalysts, some of the more important points
being:
(llnoble metals (Pt Pd, etc) are active in the oxidation of CH, above 150" C; (2)cobalt, chromium,
copper and nickel oxides are active in the range 250 to 600" C, and are not much inferior to platinum
above 400" C; (3)mixed oxide catalysts (spinels, chromates, etc) are often found to be superior to single
oxides.
Data for metal oxides are given in Table 2 for the oxidation of methane, ethylene and propane.
For their experimental program, Blazowski and Walsh used proprietary catalysts supplied by Engel-
hard. The catalyst pack was 19cm in diameter and 18cm in length, and was followed by a 42cm length of
plain pipe with a sampling section at its end. The apparatus was capable of operation to 10-5atm pressure
and 480" C inlet temperature. Fuel was JP4 (liquid) with catalyst bed inlet velocities up to 20-3ms-'. The
arrangement was purely experimental and not configured to fit into an engine.
700 D. ANSON, M.DECORSOAND W. P. PARKS

Table 2. Catalysts for oxidation of selected hydrocarbons [Source: Blazowski and Walsh
(197511

Active
Hydrocarbon catalysts Temperature (" C)

Methane (CH,) C030, > NiO, CuO, Cr203 300

Ethylene (C,H,) c0304, Cr203 250-450
Propane (C,H,) Cr203> NiO, V20, 450-600

Fuel pre-ignition occurred at an inlet temperature of 454" C, resulting in some damage and all later
testing was at 400" C or below. Flashback was also experienced at velocities below 12.5 m s-' during
startup. For inlet velocities below about 14m s-', inlet temperatures of 377 " C and fuel/air mass ratios of
0-0212, corresponding to a calculated adiabatic exhaust temperature of 1093 C, CO and hydrocarbon
O

emissions accounted for only 0.15% of fuel by weight. NO, was not measured. Most tests appear to have
been run at 7atm pressure, but pressure had little effect in the range 4-105 atm. Efficiencies were
strongly reduced by lower mixture inlet temperatures, higher velocities and lower mixture strengths,
suggesting that the attainment of high overall efficiency was kinetically dependent. Homogeneous
reaction in the plain tube following the catalyst section would be strongly temperature sensitive under the
experimental conditions. In the range of satisfactory operation, the specific heat release rate was 3.6 x
lo6 W rn-, atm-' based on catalyst volume, which approaches the heat release rates found in gas turbines
with conventional turbulent mixing combustion.
The work at NASA Lewis R.C. by Anderson et ~1.0975) was also carried out on a cylindrical research
combustor similar in many respects to that used at WPAFB, but rather smaller in diameter (lOcm), with
a 12-1cm dia X 17cm long catalyst section, again of proprietary composition supplied by Engelhard. The
fuel was propane, preheated to 97" C at 10atm and was premixed with the combustion air by a swirler
ahead of a 2m long premix section. The performance was also broadly similar to that observed at
WPAFB, efficient combustion occurring above a threshold condition related to equivalence ratio, preheat
temperature, and velocity. Figure 4 is characteristic of the performance profiles obtained. At lower
preheat temperatures, the critical equivalence ratio for a given velocity is higher. Gas temperatures were
measured 10cm downstream from the catalyst, so that if homogeneous reactions contributed to the
overall efficiency they must have been rapid. The time available for homogeneous reaction would be less
than 5ms for the range of experimental conditions studied. This is about half the time allowed for
catalytic reaction.
Ignition in the NASA Lewis reactor required inlet temperatures above 327" C (427" C was usual)
although the temperature could be reduced to 327" C after ignition stabilized.
Flashback occurred above 427" C, with flames stabilizing on the thermocouples in the upstream
section. Total NO, emissions were below 0.5 ppm for all tests, maximum temperatures being about
1280" C. The NASA workers observed significant deterioration in catalyst activity over a total of about 20
hours operation, but the cause was not determined definitely.
Also in that year, Pfefferle et al. (1975) at Engelhard Minerals and Chemicals Corp reported on
smaller scale experiments in which five different fuels were used. They described the proprietary catalysts
as 'containing a precious metal or metals' over a ceramic honeycomb with a surface-extending coating. In
these experiments, preheat temperatures were always below 600 C. Operating conditions and emission
O

levels for the various fuels with one of the catalyst formulations are listed in Table 3. The inlet
temperature for methane, which is considered to be the most difficult hydrocarbon to oxidize catalyti-
cally, was 470" c , whereas the low Btu gas, which contained 15% H,, required only 220" C preheat. This
is significant because this gas composition was fairly typical of synthetic gas from coal, which is likely to
become an important gas turbine fuel. Gas actually derived from coal may also contain potential catalyst
poisons such as sulphur.
 TURBINE CATALYTIC COMBUSTION 70 1

(r)lwLLI 1 O I L M t l l l L W O E ~~ t u r . 0 a c n
Figure 4. Effect of equivalence ratio and velocity on catalytically stabilized combustion of propane [source:Anderson, et al. 19751.

The Engelhard work produced results very similar to those of the Lewis and WPAFB teams. In
general, mixtures producing a final flame temperature of 1200" C or higher were needed to achieve
complete combustion. Figure 5 illustrates the consistency between the results of Engelhard and those of
NASA Lewis, with a scale factor of over 20 between the cross sections of the combustion ducts. This
feature validates the use of reduced scale experimentation in catalytic combustion work.

Table 3. Emissions data for various fuels [Source: Pfefferle et al. (197511

Operating conditions Emissions (ppmv)

Fuel/air
Inlet ratio Space vel. UHC
Fuel temp. (" C) (W/W) (S-l) (asCJ CO NO,

Methane 470 0025 83 0 10 <1

Propane 370 0022 28 3 10 <1
#2 Diesel 330 0.026 51 2 10 7
JP-4 410 0022 49 0 15 <1
Synthetic
low Btu-gas 220 0313 83 2 25 <1

Unburnt hydrocarbons.
Fixed operating parameters:
catalyst = Engelhard CATCOM DXA-11; catalyst length = 0076 rn; pressure = 1 atm.
702 D. ANSON,M.DECORSOAND W. P. PARKS

ao! 1
01 01 M oa II oao Q~I Ma oru on
CQUlVALtlCL RATIO

Figure 5. Efficiencies of catalytically stabilized combustors as a function of equivalence ratio [source: Pfefferle, (1975)l.

In Table 4, results of the Engelhard experiments are compared with contemporary (1975) and
projected gas turbine performances, illustrating the anticipated improvements in emissions performance.
Note that the fuel/air ratios listed are below or near the lower limits for stable uncatalysed combustion.
It is also interesting that at the time of writing (1975) the authors did not foresee the 1990s trend towards
premixed non-catalytic combustion, which has succeeded in reducing NO, emissions to 10-25 pprn when
firing natural gas, despite early experience of flame stability problems.
In a paper to the 14th A I M Aerospace Science Meeting, Rosfjord (1976) discussed the performance
and design implications of the results from the studies described above, but added no new data. His
conclusions were favourable, predicting increased specific thrust (in aircraft jet engines), lower fuel
consumption, enhanced combustion stability, and reduced exhaust emissions.
Westinghouse had co-operated with Engelhard since about 1973, and carried out a number of tests
starting in that year, and reported in 1976 (decorso et al. 1976). These tests were run in equipment of
size similar to the NASA and Wright-Patterson rigs (152cmdia), using #2 distillate oil and simulated low

Table 4. Comparison of CATATHERMAL combustion and power generation gas turbine

requirements [Source: Pfefferle et al. (1975)]

Gas turbine CATCOM

Catalyst
Parameter Present Projected Performance

Pressure ratio 8-10 12-14 10 atm

Combustor inlet temp. 280-380" C 380-480" C 320-600" C
Turbine inlet temp. 1130-1300" C 1330-1530" C 1090-1430" C
Combustor pressure drop 3-4% 3-4% 3-6%
Fuels Gaseous Gaseous Gaseous
Distillate Distillate Distillate
Treated bottoms Treated bottoms Not demonstrated
Emissions-CO 90-100ppm 30PPm
Emissions-NO, (gas fuel) 55-70 ppm < 2PPm
Emissions-NO, (liquid fuel) 75- 100ppm 8 PPm
 TURBINE CATALYTIC COMBUSTION 703

Btu coal derived gas. Because of the minimum combustion temperature required for efficient catalytically
supported combustion, tests were run at temperatures above those used at that time for stationary gas
turbines. Despite this, much lower NO, emission levels were observed. Table 5 summarizes the results of
the Westinghouse 1973-1974 tests. The premix non-catalytic combustor used in the second test series
gave emission figures close to those of the catalytic combustors, but operated at significantly lower
temperature.
The simulated coal gas mixture, which contained 15.4% H,, 11.6% CO and 5.1% CH,, required only
7.6 cm catalyst length compared with 17.8cm used for the other fuels. Pressure drop was proportionally
lower. The transition from mass transfer limited catalytic combustion to catalytically supported homoge-
neous combustion occurred at only 700 C, compared to 870 C with #2 oil.
O O

Most of the Westinghouse tests were carried out with #2 oil. At reference velocities above 20 m s-l and
mixture inlet temperatures below about 340 C, flat temperature profiles were obtained at levels close to
O

the adiabatic value. At higher velocities, lower temperatures were observed towards the centre of the
duct indicating incomplete reaction-possibly caused by non-uniform inlet temperatures. The problem
was corrected by increasing the fuel/air mixture approach temperature to over 400 C, which would have
O

accelerated light-off and temperature rise in the catalytic honeycomb. With this high approach tempera-
ture, some reaction appears to have occurred ahead of the catalyst, since the fuel/air mixture
temperatures were above those of the air supply. In all cases at velocities above 20ms-', there was some
homogeneous reaction downstream of the honeycomb, as evidenced by an increase in average gas
temperature.
Westinghouse observed no deterioration in catalyst performance over 100 hours of testing, and no
effect on emission levels of pressure in the range 3-6atm.
Investigations into the practical problems of fuel/air mixing and ignition were reported by Anderson
et af. (1980) at Acurex. As might be expected, premixing of gaseous fuel was considerably simpler than the
preparation of a uniform mixturg starting with a liquid fuel. Several possible methods were explored, and
it was demonstrated that the desired results could be achieved without pre-vaporization. Conversely,
ignition was easier with liquid fuel when using an upstream jet of air/gas mixture to form a stagnation
zone. With gaseous fuel in the main stream, ignition was explosive with this device. Earlier experience at
NASA, Wright-Patterson and Westinghouse showed that air preheat before fuel gas admixture was a
satisfactory (and safer) procedure.
Under contract to EPRI, Hung et al. (1978) at Westinghouse carried out a preliminary design study for
a catalytic combustor to replace a conventional combustor in an industrial turbine. Even in 1978, the
functional parameters could be defined from available data. It was estimated that the catalytic combustor
would need to be of larger cross section than the one it replaced in order to meet the needs of adequate
contact time without excessive pressure drop. Pressure drop data from this paper are shown in Figure 6.
Pressure drop was found to be proportional to reference velocity said by the authors to be consistent with

Table 5. Summary of results of Westinghouse catalytic combustion tests 1973-1974 [Source:

DeCorso et al. (197611

Combustor Ref vel. Outlet Emissions (ppm)

Type Fuel (ms-') Temp("C) NO, CO HC,

Conventional #2 oil 40 980 6 18 8

Conventional LBtu gas 28 950 9 160 80
Catalytic #2 oil 22 1139 <1 10 -
Catalytic LBtu gas 32 1156 <1 40 14
Catalytic #2 oil 22.6 1189 5 30 1
Catalytic LBtu gas 34 1216 1 26 1
Premix non-cat #2 oil 21.3 1066 6 20 18
704 D. ANSON. M.DECORSOAND W. P. PARKS

earlier NASA experience. The matrix design adopted for this study was a honeycomb structure of
cordierite. Hung et al. recognized the need for detailed stress analysis of the catalyst support and
substrate and this need can be met much more adequately today.
The combustor was configured to fit into a standard commercial gas turbine and to use liquid fuel.
Most of the combustor length was taken up by the preheat and premix sections. The catalyst section was
only 17.8cm long and 38cm in diameter. Much of the study was concerned with structural design of the
metal combustor body and its cooling problems. Today's concept could use uncooled ceramic elements,
with different problems (associated with support of the body and interfacing with the metal turbine
casing). This early study demonstrated the feasibility of accommodating a practical catalytic combustor in
an existing industrial engine envelope.
In 1982, two publications by workers at EPRI, Westinghouse and Engelhard (Angello et al.., 1982;
Scheihing et al., 1982) described work related to the application of catalytically stabilized combustion to
the W501D oil fired industrial gas turbine, which was rated at 92MW, and used 16 can-type combustors.
The first paper (Angello et al.) describes an EPRI supported design exercise that explored, amongst other
variables, ways in which to satisfy the simultaneous requirements for a minimum catalyst inlet tempera-
ture of 427 " C a maximum outlet temperature of 927 " C, and a capability for continuous operation over
the entire load range, which required combustor outlet temperatures up to 1330°C. The approach
adopted depended on control of the compressor air flow by means of variable inlet guide vanes (IGV's)
together with the use of a primary combustor of conventional aerodynamically stabilized type. The
primary combustor was designed to permit operation of the turbine up to 55% capacity with the IGV's
fully open. In this range the catalyst was inoperative as it saw only the products of combustion. At 55%
load the primary fuel was cut back, the catalytic fuel system was brought in and the IGV's were closed.
Transition to nearly full catalytic combustion occurred at essentially constant load, the primary combus-
tor providing only enough heat to maintain the required catalyst inlet temperature (530" C).A prototype
combustor, built according to this concept was subjected to preliminary testing and demonstrated
feasibility, low NO, capability and acceptable pressure drop. This prototype used a slightly less than full
scale catalyst element (35.6 cm dia versus 40.6 cm) because of manufacturing limitations.
As well as the temperature limitations mentioned earlier, the Westinghouse/Engelhard design team
addressed concerns about fuel/air mixing, evaporation and distribution, and the possibility of pre-igni-
tion of the main fuel introduced just before the catalyst section. Substantial effort was also put into
mechanical design, including catalyst support and casing cooling. These and other design details received

-
--
FUEL NO. 2 OIL
INLET TEMPERATURE W O O F
EXIT TEMPERATURE t 2 0 0 0 F 112(y°CI
0 3 A W
45ATM

0 10 20 ~ 4 0 0 5 Q W Mls)
REFERENCE VELOCITY
Figure 6. Pressure loss as a function of velocity for a catalytic element [source:Hung, el al. 19781.
 TURBINE CATALYTIC COMBUSTION 705

further attention described in the second paper (Scheihing et al. 1982), which was based on parallel
efforts supported by Westinghouse and Engelhard independently of EPRI.
A major part of the initial activity was directed at achieving a uniform air/fuel distribution and
complete evaporation of the fuel spray ahead of the catalyst. Using the preferred geometrical arrange-
ment, a number of experimental runs were made, but without the use of the primary combustor described
above. Thus the experiments did not demonstrate the concept of using IGV control to bring the
combustor up to load while controlling catalyst temperature. Catalyst inlet temperature was controlled at
370-450 C by external means. Combustion profiles inferred from temperature measurements at the
O

catalyst outlet plane and at two locations 12.7cm and 22.9cm downstream were consistent with the
mixing data. Efficiencies were in the range 80-90% at the catalyst outlet plane, rising to 85-93% at the
downstream (22.9cm) location. This is at first sight not a good result, but the maximum temperatures
were only about 1210"C, and probably as low as 1030-1130" C at the catalyst outlet plane. The practical
limitation was catalyst and catalyst support life, which mandated conservative operating conditions. The
substrate was cordierite, with a noble metal catalyst, possibly using mixed metals. NO, levels were as
expected (1-3 ppm), but CO and HC, were high, consistent with the low combustion efficiencies.
These tests confirmed the general feasibility of catalytic combustion within the physical constraints of
a practical engine but did not go beyond the first stage of development. Combustion efficiency would
undoubtedly be better with a hot wall (ceramic) combustor.
Since these earlier experiments, pressure to reduce NO, emissions has revived interest in catalytic
combustion especially in Japan and the U.S.A. Two papers from Japan presented at 1987 ASME Gas
Turbine Conference exemplify this trend. Authors from Toshiba and Tokyo Electric Power (TEPCO)
(Furuya et al., 1987) described a hybrid system in which fuel was introduced in three stages. In stage 1,
fuel was burnt non-catalytically to provide preheat ( * 450 O C after dilution) followed by the introduction
of Stage 2 fuel in the main premixing zone. Only Stage 2 combustion was premixed. The third stage fuel
was introduced after the catalyst. Air was introduced with the first two stages of fuel admission, and after
-
the Stage 3 fuel for dilution to modulate the relatively high gas temperatures ( 1500" C) reached in the
Stage 3 homogeneous reaction zone. This also reduced gas velocities in the catalyst bed. Figure 7 shows
the arrangement in schematic form. Bypass (dilution) air varied from nearly 80% at no load to about 30%
at full load. The system was expected to produce virtually constant temperature profiles through the
premix and combustion zones over the load range. The authors of this paper describe the results of an

Compressed Air
I

Preheating Homogeneous
Combustion Reaction Zone
Zone
Mixing
Zone

F1, Fz. F3 - Fuel Feed P o i n t s

Figure 7. Toshiba hybrid catalytic combustor (schematic).

706 D. ANSON, M. DECORSO AND W. P. PARKS

extensive computer investigation of the effects of catalyst properties (activation energy and pre-exponen-
tial factor) on the catalyst temperature profiles for different preheat levels. They point out that a low
activation energy and moderate pre-exponential factor are desirable to permit the control of catalyst
temperatures at levels assuring long catalyst life. The catalysts actually used in experiments were
palladium and platinum on alumina wash coat over a cordierite honeycomb.
Experiments were carried out on two facilities, with 3.0cm and 10-0cm diameter catalyst beds,
respectively. The first used five catalyst packs, each 3.0cm long, in series, and the lOcm combustor used
four packs, also 3.0 cm long. Experiments, using LNG and methane as fuels, demonstrated satisfactory

-
performance, generally in line with predictions. However, at the design catalyst outlet temperature of
800" C tertiary methane fuel did not autoignite and a separate ignition source was required. Pressure
drop was below 4% at inlet gas velocities of 40ms-' or below, and overall combustion efficiency was
-
above 99.9%. NO, was 4ppm for the 3.0cm combustor, independent of pressure to 7atm but was
higher for the 10.0cm combustor, data for which are shown in Figure 8. These data are for ambient
pressure operation. Note that only the larger combustor included a pre-burner (preheat burner) which
accounts for the difference in NO, levels at 15-20ms-' velocity. The large increase in NO, at lower
velocities was attributed to local high combustion temperatures in the final combustion stage and was
virtually all attributable to the non-catalytic stages.
Hoshino et al. (Kawasaki) (1987) described a somewhat different approach to catalytic combustor
design for a 228 kW gas turbine. The experimental combustor used essentially the same design concept as
that shown in Figure 3 of this report. The engine was started using only the output of the preheat burner.
Once the catalyst was warmed up fuel was admitted to the main premixing section and the preheat fuel
was cut back. This transition was accomplished manually, as was dilution air control. A step change of
130kW in load was accommodated without use of variable geometry, but resulted in substantial
temperature changes. Exit gas temperature from the catalyst was controlled at 1150" C. Combustion
efficiency was poor at first, but was improved to about 98% by introducing spacers between the three
catalyst sections. It may be that the cross mixing which occurred between sections was beneficial. The

-
maximum reported load was about 150 kW, at which emissions were CO * 450ppm, HC 150ppm and
NO, 6 ppm. These performance levels were not considered adequate, and further development was
-
planned. Catalyst damage was experienced after only 5 hours operation, and was caused by overheating
near the periphery because of partial flow obstruction. The catalyst, supplied by NSKK, used platinum
group metals on an alumina wash coat on cordierite with 31 channels per square centimetre.
Information on the next phase of this work was presented in 1989 (Kitajima and Kajita, 1989). The
goals were 170kW output with > 98% combustion efficiency and < 15ppm No,. The combustion

I500
x)O
n FUEL METHANE
COMBUSTION TEMP. -
0
5
- 1400
s W
I-
z

-1300 3
6AS PHASE COMBUSTION ZONE
P T ! U T l O N
3
0-
PRCBWNCR COIITRHVTION

GAS VELOCITY (ot 500 C) (m/s)

Figure 8. Performance of toshiba catalytic combustor [source:Furuya, et al. 19871.
 TURBINE CATALYTICCOMBUSTION 707

TIaxit) - 1175 - 1375 1:

Combustor
0 TmoA

0 TmeB

Ul I
1
510 360 400 468 600 661 I00
Catelvet Inlet Tonwreturo

Figure 9. Effect of catalyst inlet temperature on combustion efficiency [source:Kitajima and Kijita er ul. (1989)l.

loading on the catalyst was reduced by a factor of 0.65 using two alternative approaches. The first
approach was to increase the diameter from 18.4 cm to 22 cm and the total length from 15.0 to 16.0 cm. In
the second approach, the cell density was increased from 31 to 62 per sq. cm and the length was again
16.0cm. The catalyst section incorporated two 2.0cm spacers with 3 active catalyst honeycombs. The
combustor layouts were also modified to increase penetration of the air jets in the premix zone. The 62
cells per sq. cm catalyst was installed in a combustor (designated Type B) which incorporated a number
of mechanical changes, including a redesigned variable geometry dilution air section and an enlarged
primary zone swirler. This Type B combustor produced significantly less NO,, attributed to the leaner
primary mixture. Nearly all NO, was apparently produced in the primary section. CO emissions were also
significantly less at lower loads but HC emissions tended to be higher over most of the load range.
The higher catalyst inlet temperatures ( > 450 C ) resulted in significant increases in NO, emissions
O

(probably from the larger preheat burner input) but also in higher combustion efficiency (Figure 9). The
Type A combustor achieved 99% combustion efficiency with 1050 C catalyst outlet temperature, the
O

Type B combustor 98%, with a less uniform temperature profile. Because of its relatively high pressure
drop, the engine with Type B combustor was unable to achieve power outputs above about 130 kW, while
with Type A combustor it reached 184 kW with significantly lower CO and HC emissions over much of
the load range.
The most important observation from the Kawasaki engine tests concerns catalyst life. Figure 10 shows
how the combustion efficiency varied between new and used catalyst after only 3 hours of operation.
Physical damage to the catalyst outer regions was also noted. Catalyst durability was identified as the
major obstacle to commercial engine development. As might be expected, efficiency is more sensitive to
catalyst deterioration at lower inlet temperatures.
Workers at Toshiba and TEPCO (Kawakami et al., 19891, also reporting on follow-on activities to their
1987 report (Furuya et al. 1987) were clearly concerned about the physical durability of ceramic
honeycomb catalyst substrates. They used both analytical approaches and direct measurements on
honeycomb structures to arrive at the conclusion that adequate strengths could be achieved for survival
of cordierite honeycomb at 1000°C. As might be expected, peak stresses from both mechanical and
thermal effects occurred at corners. The authors of this work are careful to point out that in their hybrid
catalytic combustor the catalyst is not exposed to temperatures above about 1000 C because combustion O

is completed in a non-catalytic stage.

An approach to the problem of catalyst durability is suggested by a report from the R&D Center of
the Osaka Gas Co. Ltd. (OGC, 1990). From the illustration in that report the work would appear to refer
to the same small Kawasaki engine which was the subject of the work of Hoshino et al. (1987) and
Kitajima and Kajita (1989). OGC refers to the invention by Professor Arai of Kyushu University of a
monolithic ceramic catalyst in which manganese is substituted into a complex hexaluminate structure
708 D. ANSON. M. DECORSOAND W. P. PARKS

7B
808 8M B88 WW 1888 1BW 111B 1164
Cmtalmt Exit T o m r r t u r o I:
b I
iiw iiu 1 2 e ~ 12ia im iaw MH 146~
Cmtmtmt bit TomPrmturm X

Figure 10. Effect of the catalyst exit temperature on the combustion efficiency [source: Kitajima, et al., and Kijita (1989)l.

(La(, -,,Sr(,,MnAll10(9-,,). The honeycomb was constructed by gel forming and sintering at
1200-1300" C.
A four-stage catalyst bed was tested in a small scale test rig. The first stage used palladium on
cordierite, followed by three stages of manganese substituted hexaluminate. Temperature profiles show
an essentially linear temperature rise from the inlet (500" C) to about 1175" C after three stages, with a
slight drop to 1100" C after the final stage, for an adiabatic temperature of 1260" C. Profiles for leaner
mixtures (TAD= 1160 and 1090" C) were similar. Thermal losses probably account for most of the
difference between observed peak temperatures and adiabatic temperatures. Test results for the
intermediate condition (TAD= 1160" C) demonstrated 99.6 to 99.9% combustion efficiency at pressures
up to 4atm with NO, equivalent to -
10ppm at 15% 0,. Approximately 28% of the combustion
occurred in a non-catalytic pre-combustion zone, which may have accounted for most of the NO,. The air
velocity was 10ms-' at 300" C inlet condition. At the time of the report, this arrangement had been
tested for 1500hours at 1260" C and showed no deterioration in combustion efficiency. After 50 cycles,
thermal cycling tests between ambient temperature and
periphery of test pieces.
-
1100" C produced cracking near the

Pressure drop in this catalyst stack was not reported. However, the authors quoted the same
performance targets as Kawasaki, which included a A P value of < 5%. In addition, the targets for
catalyst life were 8000 hours and 400 thermal cycles. The intention was to proceed to full engine tests.
In 1993, workers at the Central Reserch Institute of the Electric Power Industry (CRIEPI) in Japan,
with engineers from Kansai Electric Power Co. (KEPCO), reported on test results from two different
catalytic combustors designed for a 20 MW turbine (Ozawa et al. 1993). In both designs, about 50% of the
throughput was heated to 1000" C in the catalyst section, stabilizing the combustion of premixed fuel and
air in a homogeneous reactor with a final temperature of 1300" C . In the first design, the homogeneous
reaction became unstable below 1250" C, but in the second design, it was stable down to 1150" C
although efficiency dropped steeply below 1200 C. By maintaining combustion temperature at 1300" C
O

and using bypass air to modulate the final gas temperature between 970 and 1300" C it was possible to
achieve stable combustion with an efficiency over 99.9% and NO, below 10ppm (6 ppm when normalized
to 16% 0 2 )Most
. of the NO, was generated in the preheat burner.
Catalytica in the U.S.A. has used a different approach to limiting catalyst temperature and avoiding
deterioration-palladium catalyst on a metal substrate. At about 800 " C, palladium forms a non-catalytic
ozide that limits temperature rise. In an editorial (Barker, 1994) Turbomachinery International reported
 TURBINE CATALYTIC COMBUSTION 709

that Catalytica was working with a number of turbine manufacturers. Stambler (19931, in another
editorial, also reported on Catalytica's cooperative efforts with Allison, G. E. and the Tokyo Electric
Power Co. (TEPCO). Some results obtained by Catalytica working with GE and TEPCO have been
reported by Della Betta et al. (1994). In pressurized test rigs with up to 12atm capability, multi-stage
catalysts on metal substrates were used to support combustion up to about 1000" C. Temperatures up to
1475" C were achieved after the final post catalytic combustion stage, with NO, of 1.2 ppm (normalized to
15% 0,)when no preheat burner was used. At the time of reporting, a catalyst durability of 4500 hours
had been demonstrated and life testing was planned. A major part of GE's development effort has gone
into achieving a uniform fuel/air mixture at the catalyst inlet (Beebe et al., 1987).
Solar Turbines, Inc. with the Engelhard Corp. also has reported on catalytic combustion work (Cowell
and Larkin, 1994). This work does not appear to be so advanced as that at Catalytica but it has been
carried out at pressures up to 9 atm on complete prototype systems suitable for incorporation in engines.
These systems incorporated non-catalytic preheat and part load burners. At an exhaust temperature of
1399" C, emission levels recorded (corrected to 15%0,) were NO,8ppm; CO 22ppm; and UHC 4ppm.
An interesting but unconfirmed observation was that NO, generated in the preheat burner appeared to
be slightly reduced in the catalyst. The Engelhard catalyst composition is proprietary. The substrate was
ceramic. Also, the post-catalyst combustion section was ceramic lined, and this was claimed to improve
combustion efficiency.
Some European work on catalytic combustion has been reported by Vortmeyer et al. (1994), who used
a single tubular reactor with short (20-60mm) catalytic sections and long (up to 2m) thermal reactor
sections. They achieved targeted efficiencies ( > 99.9%) with propane, methanol and diesel fuel. NO, was
found only with diesel fuel that contained combined nitrogen. Pressure drop was almost proportional to
velocity and catalyst length.

4. CONCLUSIONS
Together, the experiences summarized above suggest the following lines for development:

(1) Catalyst inlet temperature high enough to ensure rapid ignition is desirable so that the catalyst
length and pressure drop can be reduced while keeping catalyst exit temperature high enough to ensure
fast homogeneous reaction in the final combustion stage. Actual required temperatures depend on fuel
composition. (2) High catalyst exit temperatures compatible with homogeneous reaction conditions may
require the development of new catalysts for natural gas with demonstrated endurance to 10 000 hours or
more at 1100-1200" C. (3) High preheat must be achieved without a high NO, penalty in the preburner,
and without flashback or premature reaction in the premixing zone. (4)Premixing of fuel and air is
critical.These requirements apply throughout the range of physical layouts and control strategies that
have been used.
The failure of catalytic combustion to reach a definitive point of development prior to 1985 was
perhaps in part due to a lack of sufficient pressure to reduce NO, emission levels. The possibilities of
achieving low NO, had by then been demonstrated by work at Engelhard, Westinghouse and NASA
Lewis, and the problems that had been identified appeared to be surmountable with reasonable effort.
These problems still exist today, but there is a greater need to solve them, and there have been changes
in gas turbine technology and the turbine market that should help to justify the R&D process.
With today's emphasis on pollution control, NO, reduction by the use of catalytic combustion is a
logical extension of non-catalytic premixed combustion technology, with a prospect of better performance
and a wide load range. It would eliminate the equipment and performance penalties associated with
water injection and post combustion catalytic reduction strategies. The introduction of ceramic combus-
tor components in some smaller industrial gas turbines, with near adiabatic combustion, can facilitate
low-temperature catalytically stabilized systems. As turbine inlet temperatures are increased, so combus-
tor outlet temperatures match more closely the preferred levels for the final homogeneous reaction stage
710 D. ANSON, M. DECORSO AND W. P. PARKS

of such systems (1300-1500” C) with less need for attemperation by dilution. We may also see wider use
of recuperation, which will provide combustor inlet temperatures above catalytic ignition levels.
These circumstances should encourage the development of improved catalyst support systems and also
of monolithic catalysts such as the hexaluminates.
The outstanding problems are still catalyst life, catalyst substrate integrity and pressure drop. Given
the lack of severe load cycling in industrial gas turbines and the relative freedom to decrease velocities in
the catalyst zone, these do not appear unduly difficult to overcome. Both catalyst life and substrate
integrity can be explored experimentally on a small scale, and therefore economically. Pressure drop is a
design related parameter and the data exist that can provide an adequate basis for the design of
prototype catalytic combustors. Aerodynamic testing without combustion can also be used as an
economical development approach for fuel/air mixing studies.
As well as addressing the individual problems listed above, engineers need to devote attention to the
design of the entire combustion system and its control in order to meet the needs for ignition and
combustion efficiency over the whole load range without exposing the catalyst section to excessive
temperatures. The critical temperature limitations are by now reasonably well defined for all parts of the
system and, with the developments suggested, they should become less severe.

ACKNOWLEDGEMENT

This material was originally assembled as an informal topical report to the U.S. Department of Energy.
We thank the DOE for approval to publish it.

REFERENCES
Anderson, S. J., Krill, W. V. and Kesselring, J. P. (19801, ‘Fuel injector, ignition and temperature measurement techniques for
catalytic combustors’, Proc. 4th Workshop on Catalytic Combustion. EPA Report 600/9-80-035.
Anderson, DN., Tacina, R.R. and Mroz, T.S. (1975) ‘Performance of a catalytic reactor at simulated gas turbine combustor
operating conditions’, NASA, TM X-71747.
Angello, L. C., Pillsbury, P. W. and Toof, J. L. (1982). Catalytic burners for stationary combustion turbines-a preliminary design
analysis’, ASME.Paper 82-GT-58.
Barker, T. (1994). ‘Catalytic combustion promises lower NO,’, Turbomachinev International, March/April (1993).
Beebe, K., Ohkoshi, A,, Radak, L. and Weir, A. Jr. (1987). ‘Design and test of catalytic combustor fuel/air preparation system’,
Paper#51 to the 1987 Tokyo International Gas Turbine Congress, Tokyo, 26-31 October 1987.
Blazowski, W. S., and Walsh, D. E. (1975). ‘Catalytic combustion: an important consideration for future applications’, Combustion
Science and Technology, 10, 253-244.
Cowell, L. H . and Larkin, M.P. (1994). Development of a catalytic combustion for industrial gas turbines’, ASME.Paper 94-GT-254.
DeCorso, S. M., Mumford, S., Carruba, R. V. and Heck, R. (1976). ‘Catalysts for gas turbine combustors-experimental test results’,
ASME.Paper 76-GT-4.
Della Betta, R. A., Schlatter, J. C., Nickolas, S. G., Yee, D. K. and Shoji, T. (1994). ‘New catalytic combustion technology for very low
emissions gas turbines’ ASME Paper 94-GT-260.
Furuya, T., Yamanaka, S., Hayata, T., Koezuka, J., Yoshine, T., and Ohkoski, A., (1987). ‘Hybrid catalytic combustion for stationary
gas turbine-concept and small scale test results’, ASME.Paper 87-GT-99.
Garrett, A. P. (1987). Advanced Gas Turbine (ATD) Technology Development Project, Final Report DOE/NASA/0167-12,
Garrett A.P. Div., December 1987.
Hoshino, A,, Kajita, S., Hagiwara, Y., Fujimot, K. and Kitajima, J. (1987). ‘Preliminary tests of catalytic combustion in a small gas
turbine. ASME.Paper 87-GT-10.
Hung, W.S.Y., Dickson, W.H. and DeCorso, S. M. (1978). ‘Preliminary design analysis of a catalytic ceramic structure in a turbine
combustor’, Proc. 4th Workshop on Catalytic Combustion. EPA Report 600/9-80-035.
Kawakami, T., Furuya,T., Sasaki, Y., Yoshine, T., Furuse, Y., and Hoshino, M., (1989). ‘Feasibility study on honeycomb ceramics for
catalytic combustor’, ASME.Paper 89-GT-41.
Kitajima, J. and Kajita, S. (1989). ‘Catalytic combustor for small gas turbines-combustor development’, ASME.Paper 89-GT-265.
OGC (1990). ‘A development of high temperature catalytic combustion technology for gas turbine’, Osaka Gas Co. R & D Center
release. 11/7/1190.
Ozawa, Y., Hirano, J., Sato. M., Saiga, M. and Watanabe, S. (1993). ‘Test result of low NO, catalytic combustor for gas turbine’.
ASME.Paper 93-GT-344.
Pfefferle, W C. (1974). ‘Catalytically supported thermal combustion’, Belg. Patent 814,752, 1974 and U.S. Pat. 3,928,961, December
1975.
Pfefferle, W. C., Canuba, R. V., Heck, R. M., and Roberts, C. W. (1975). ‘CATATHERMALTMcombustion: a new process for low
emissions fuel conversion. ASME,75-WA/Ful.
 TURBINE CATALYTIC COMBUSTION 711
Pfefferle, W. E. and Pfefferle, L. D. (1986). ‘Catalytically stabilized combustion’, Progress in Energy and Combustion Science, 12,
25-41.
Prasad, R., Kennedy, L. A. and Ruckenstein, E. (1984). ‘Catalytic combustion’, Card. Reu.-Sci. Eng, 26(1), 1-58.
Rosfjord, T. J. (1976). ‘Catalytic combustors for gas turbine engines’, 14th A M Aerospace Sciences Meeting, Washington, January
1976, Paper No. 76-46.
Schefer R. W. and Robben, F. (1980). ‘Catalyzed combustion of lean fuel/air mixtures’, Proc. 4th Workshop on Catalytic Combustion,
EPA, Report. 600/9-80-035, pp. 266-284.
Scheihing, P.E., Pillsbury, P.W., Piaia, T.A. and Kitaplioglu, 0. (1982). ‘Full scale catalytic burner testing with No. 2 oil for
combustion turbine application’, 82-JPGC-GT-20.
Stambler, I. (1993). ‘Cool catalytic combustor design limits NO,, to less than 0.5 ppm’, Gas Tut6ine World, May-June.‘
Vortmeyer, N., Valk, M. and Kappler, G. (1994). ‘Catalytic combustor for high-temperature gas turbines’, ASME. Paper 94-GT-211.