Combustors for Micro-Gas

Ian A. Waitz
Associate Professor,
Massachusetts Institute of Tectinology,
Turbine Engines
Gas Turbine Laboratory,
Cambridge, MA 02139 The development of a hydrogen-air microcombustor is described. The combustor is
intended for use in a 1 rrmf inlet area, micro-gas turbine engine. While the size of
the device poses several difficulties, it also provides new and unique opportunities.
Gautam Gauba The combustion concept investigated is based upon introducing hydrogen and pre-
Consuitant, mixing it with air upstream of the combustor. The wide flammability limits of hydro-
Arttiur D. Littie inc., gen-air mixtures and the use of refractory ceramics enable combustion at lean condi-
Cambridge, MA 02140 tions, obviating the need for both a combustor dilution zone and combustor wall
cooling. The entire combustion process is carried out at temperatures below the
limitations set by material properties, resulting in a significant reduction of complexity
Yang-Sheng Tzeng when compared to larger-scale gas turbine combustors. A feasibility study with initial
Graduate Researcti Assistant, design analyses is presented, followed by experimental results from 0.13 cm' silicon
Massachusetts institute of Technoiogy, carbide and steel microcombustors. The combustors were operated for tens of hours,
Gas Turbine Laboratory, and produced the requisite heat release for a microengine application over a range
Cambridge, MA 02139 of fuel-air ratios, inlet temperatures, and pressures up to four atmo.^pheres. Issues
of flame stability, heat transfer, ignition and mixing are addressed. A discussion of
requirements for catalytic processes for hydrocarbon fuels is also presented.

1 Introduction and Groshenry (1995). In this section, we present only a short

review of the technology in order to establish the design goals
Micromachining of silicon and refractory ceramics is en-
and constraints for a microcombustor.
abling the development of a new class of miniature devices
including micromotors, microvalves, pressure transducers, mi- Microengine technology has been enabled by the advent of
croaccelerometers, and numerous others (Bryzek et al., 1994). micromachining processes for refractory, structural ceramics
Arguably one of the most challenging and innovative endeavors such as silicon nitride (Si3N4) and silicon carbide (SiC). These
being pursued in this area is the development of microscale materials have excellent mechanical, thermal, and chemical
turbomachines as described by Epstein et al. (1995). These properties for hostile environment applications, and indeed have
devices include micromotor compressors, micro-turbine genera- been the focus of several development efforts for larger-scale
tors, micro-gas turbines, microrefrigerators, and micro-rocket devices (Nakazawa et al., 1996; Measley and Smyth, 1996;
engines. Several of these heat engine applications require con- Tanaka et a l , 1996). Components manufactured from these
version of chemical energy to kinetic and thermal energy, and materials can survive uncooled operation at temperatures as
thus it is necessary to develop combustion strategies suitable high as 1750 K (Tanaka et al., 1996; The Carborundum Com-
for use in these miniature devices. pany, 1989). Therefore, use of these materials for a micro-gas
The objective of this paper is to elucidate many of the chal- turbine engine insures sufficient power per unit air flow to allow
lenges and opportunities associated with combustion in small for useful power output despite the fact that individual compo-
volumes, and to offer possible strategies for microcombustion nent performances are likely to be below those of larger ma-
.systems. Following a brief overview of micro-gas turbine tech- chines. Further, millimeter-size parts can be fabricated with 1 -
nology in Section 1.1, the specific challenges and opportunities 2 micron tolerances using micromachining processes (Juan et
for microcombustor development are discussed in Section 1.2. al., 1996), providing a dimension-to-tolerance ratio that is com-
Two viable microcombustion alternatives are presented in Sec- petitive with larger-scale turbomachinery components. Micro-
tion 2: the first based upon hydrogen-air combustion (Section machined parts however, are typically limited to extrusions of
2.2.1), and the second upon a catalytic hydrocarbon-air reaction two-dimensional shapes, thus the geometries are somewhat ru-
(Section 2.2.2). Results from experiments in a full-scale micro- dimentary when compared to their larger counterparts.
combustor are presented in Section 3 to provide a first proof- An example of a micro-gas turbine design is shown in Fig.
of-concept demonstration. The paper ends in Section 4 with a 1. The overall dimensions of the device are roughly 1 cm in
summary, conclusions and recommendations for future re- diameter and 3 mm in height. This design is the result of a
search. preliminary analysis for an engine geometry and layout that
are compatible with individual component characteristics and
1.1 Overview of Microengine Technology. The primary
constraints of micromachining technology. Thus the design em-
motivation for the work described in this paper is the develop-
bodies trade-offs between power output requirements, cycle pa-
ment of micro-gas turbine generators capable of producing 10-
rameters, material limitations, physical dimensions and manu-
100 watts of electrical power while occupying less than 1 cm-'
facturing processes. The device consists of a radial flow com-
and consuming approximately 7 grams of jet fuel per hour. If
such a device is successfully developed, it will possess an en- pressor and a radial flow turbine mounted on the same shaft and
ergy density 10 times that available from the best existing batter- separated by a combustor. An integral electrostatic induction
ies. Feasibility studies, preliminary designs, and performance generator is located on the top face of the compressor shroud.
estimates have been discussed in detail by Epstein et al. (1995) The combustor is the largest component of the engine.
The characteristics of the micro-gas turbine are summarized
in Table I; the device is a factor of 500 smaller than conven-
Contributed by the Fluids Engineering Division for publication in the JOURNAL tional gas turbine engines. Preliminary analyses by Epstein et
OF FLUID."; ENGINEERING , Manuscript received by the Fluids Engineering Division
October 14, 1996; revised manuscript received September II, 1997. Associate al. (1995) suggest that such a machine operating at a tip speed
Technical Editor: Wing-Fai Ng. of 500 m/s, with a compressor pressure ratio of 4.5; 1 and a

Journal of Fluids Engineering Copyright © 1998 by ASME MARCH 1998, Vol. 1 2 0 / 1 0 9

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 Compressor
Flame Diffuser Rotor Inlet Starter/ Fuel Fuel
Holders Vanes Blades I Generator Injectors Manifold
I
- ^^^
. «;_-—

W4\ rr
Combustion Exhaust Turbine Turbine
Chamber Nozzle Rotor Nozzle
Blades Vanes

Fig. 1 Schematic of a micro-gas turbine and electrical generator

turbine inlet temperature of 1600 K, may be capable of produc- engine. The estimates cited for the microcombustor were ob-
ing 10-20 watts for a canonical inlet area of 1 mm^. tained by making conservative assumptions about achievable
The overall goal is to develop a complete microengine system performance and size of a microengine operating with a total
that is contained within a volume on the order of 1 cm^. Before pressure ratio of 4.5:1. Compared to the larger-scale device, the
such a device can be realized however, many difficulties must size of the microcombustor has been reduced by a factor of 100
be overcome. Micro-gas turbines and other applications of this and the volume by over 6 orders of magnitude. The space heat-
technology are not simply scaled-down larger machines. The ing rate is 10 times larger than that of current technology aircraft
small size and constraints on fabrication and testing pose many gas turbine combustors.
challenges: the surface area-to-volume ratio is increased, vis- The differences between design parameters listed in Table
cous effects are more important, time-scales are shorter, and 2 for the micro and conventional applications are largely due
the range of three-dimensional shapes that can be fabricated is to the reduction in scale, but are also influenced by the rela-
limited. All of these effects either directly or indirectly impact tive size of the combustor with respect to the engine, the
the selection and development of appropriate combustion strate- cycle pressure ratio, and material temperature limitations.
gies for a microengine. These effects are discussed in greater detail in the following
sections.
1.2 Challenges and Opportunities for Microcombustor
Development. The functional requirements of a microcom- 1.2.1 Shorter Residence Time for Mixing and Combustion.
bustor are similar to those of a typical gas turbine combustor. The most significant and technically challenging aspect of a
Primarily, chemical energy must be converted into thermal and microcombustor is its limited residence time. If the dimensions
kinetic energy with high efficiency and low total pressure loss. of a conventional engine were to be reduced by a factor of 500,
In addition, there are requirements for introduction and mixing while maintaining the same mass flow rate per unit area, then
of fuel and oxidizer, reliable and smooth ignition, wide flam- the flow residence time within the combustor would be approxi-
mability limits so the flame stays lit over a range of operating mately 0.05-0.1 ms. This is on the same order as the character-
conditions, low pollutant emissions, and freedom from combus- istic chemical kinetic time scale for hydrocarbon-air reactions
tion instabilities. Likewise, the principal constraints on micro- (0.01-0.1 ms). Thus for the microengine, it will be necessary
combustor design mirror those of the larger-sized counterpart, to increase the relative size of the combustor compared to the
including the maintenance of low-stressed, cooled structures, engine.
minimal weight, and an overall shape and size that are compati- The required increase in relative combustor size can be ap-
ble with the rest of the engine layout. proximated using a simplified scaling for combustor residence
A comparison between design parameters of a conventional time given by Kerrebrock (1992)
gas turbine combustor and those of a microcombustor are shown
in Table 2. The parameters listed for the conventional combustor L-(A,M2)-7rll / y
are representative of a modern 30:1 pressure ratio gas turbine (1)
rhIAo

Nomenclature

A2 = compressor flow area h = convective heat transfer coeffi- T = flow temperature [K]
A, = combustor cross-sectional area cient [W/m^-K] r,,, = wall temperature [K]
.Ah = pre-exponential factors for re- JD = mass transport number Tn = compressor exit temperature [K]
action rates JH = heat transport number 77-4 = turbine inlet temperature [K]
A., = surface area [m^] k = thermal conductivity [W/m- K] V„ = reference velocity [m/s]
= specific heat at constant pres- L = length [m] Vol = volume [m-*]
sure [ J / k g ' K ] nia = air flow rate [kg/s] a = mass transfer coefficient [m/s]
dh = hydraulic diameter [m] OT/ = fuel flow rate [kg/s] T] = combustor efficiency
Da, = catalytic Damkohler number Nurf = Nusselt number based on diameter TTc = compressor pressure ratio
Da* = homogeneous Damkohler num- Q = volumetric rate of energy/heat y = ratio of specific heats
ber generation [W/m^] p„ = reference density [kg/m^]
E" = rate of energy/heat flux [W] Qf = fuel heating value [J/kg] Tres = combustor residence time
E = rate of energy/heat generation Re,, = Reynolds number based on diame- (p = equivalence ratio
[W] ter

110 / Vol. 120, MARCH 1998 Transactions of the ASIVIE

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 Table 1 Performance estimates for a micro-gas turbine For a combustor, the ratio of surface heat transfer losses to
engine total heat released in the combustion process can be written as
Single spool AMT- TJ
Type turbojet/turbogenerator El (2)
E Vol-Q
fuel
recuperator No For given fuel, equivalence ratio, inlet temperature, and pres-
engine weight sure, the energy per unit volume released in the combustion
inlet area '1^ mma process, Q, is constant. Since the dependence of Nusselt number
air flow rate 0.2 g/s on Reynolds number differs between the flow regimes of a
pressure ratio 4.5:1 conventional combustor and a microcombustor, the turbulent
turbine inlet temperature 1600 K
power output (electrical) 10-20 W flow case, which is more sensitive to Reynolds number
thrust (as turbojet) 0.1-0.2 N
fuel consumption 7 g/hr Nurf ^ VReJ (3)
specific fuel consumption 0.45 kg/kW-hr
thrust/wt 20:1 is adopted for both flow conditions. As a result, the convective
energy/air flow 84J/g heat transfer coefficient
k • NUJ
(4)
Note that either the length or the area of the microcombustor
is inversely proportional to the fifth root of the hydraulic diame-
must be enlarged by a factor of 4 solely to account for the
ter. Finally, if the temperature difference between the wall and
reduced pressure ratio of the cycle (4.5:1 compared to 30:1
the flow is assumed to be roughly equal for the two combustors,
for the conventional application), for fixed mass flow per then the ratio of heat lost to that generated would scale with
unit area. It will require a further ten-fold increase in either the hydraulic diameter as follows:
microcombustor length or area to raise the combustor resi-
dence time by a factor of 10. This will result in a residence 1
time of 0.5 to 1 ms which is several times larger than the
El (5)
E
chemical time scale. Therefore, if a full-sized engine is
scaled down by a factor of 500, the volume of the microcom- The hydraulic diameter of the microcombustor is on the order
bustion chamber must grow relative to the engine by a factor of 2 mm, several hundred times smaller than that of a conven-
of approximately 40 to provide sufficient residence time for tional gas turbine combustor. Thus the ratio of heat lost to heat
complete reaction. Even after this factor of 40 increase, the generated will be about 2 orders of magnitude greater than that
entire microengine can still be contained in less than 1 cm' of typical combustors. This is likely to influence the perfor-
volume. mance of a microcombustor in two ways: 1) typical large-scale
Note, however, that the residence time in conventional gas combustor efficiencies of greater than 99.9 percent may not be
turbine combustors is not driven primarily by the reaction feasible due to the significant surface heat transfer losses, and
rate. Rather it is governed by requirements for fuel-air mix- 2) flammability limits are likely to be affected because of flame
quenching.
ing as well as dilution with air to reduce the flow temperature
and meet efiiciency and emissions requirements. Indeed, of 1.2.3 Use of Refractory Structural Ceramics. A notewor-
the typical 5 - 8 ms combustor residence time in a conven- thy item listed in Table 2 is the maximum material temperature
tional gas turbine, approximately 3 - 5 ms is devoted to fuel allowable for each application: 1200 K for uncooled structures
vaporization and mixing, and about 2 - 3 ms to mixing of in current gas turbines, and approximately 1600 K for the struc-
dilution air (Dodds and Bahr, 1990; Lefebvre, 1983). Thus tural refractory ceramics targeted for use in micro-gas turbine
if conventional combustion strategies are adopted for the engines. The increase in operating temperature projected for the
microengine application, augmenting the fuel-air mixing rate microengine structure alleviates some of the requirements for
will be necessary. combustor wall cooling.
Fracture failure has generally hindered the use of refractory
1.2.2 Heat Losses Due to the High Surface-to-Volume Ra- ceramics for large-scale applications. Parts with larger volumes
tio. Energy loss due to heat transfer at the walls of the combus- statistically entail flaws that are greater in quantity, larger in
tion chamber is typically neglected in the design of conventional size, and deeper in location (Scott, 1979). Failure of ceramic
gas turbine engines. However, for the combustor designs pre-
sented in Table 2, the surface area-to-volume ratio, which is
proportional to the inverse of the hydraulic diameter, increases Table 2 Comparison between conventional and microcom-
from 3-5 m^' for the large-,scale combustor to 500 m ' for bustors
the microcombustor, and thus heat transfer losses may pose a Design requirement Conventional
significant problem. (sea level takeoff) combustor Microcombustor
The effects of surface heat loss on combustion have been
investigated in several studies. In premixed gases, flame extinc- length 0.3 m 0.003 m
volume 6 X IQ-" m' 4 X 10"* m'
tion occurs when the amount of heat liberated by combustion, cross-sectional area 0.2 m' 4 X 10"' m^
minus the heat transferred from the gas, no longer exceeds the inlet total pressure 30 atm 4.5 atm
amount needed to ignite the mixture (Dodds and Bahr, 1990; intlet total temperature 800 K 500 K
Ballal and Lefebvre, 1979). Studies of premixed combustion mass flow 55 kg/s 20 X 10^" kg/s
in flame tubes (Zeldovich et al., 1990; Zamaschikov, 1995) average flow speed 40-60 m/s 6 m/s
residence time 5-8 ms 0.5 ms
have shown that if the inner diameter of a flame tube is less efficiency >99.5% >99.5%
than some critical size, heat transfer from the flame front to the combustor pressure ratio >0.95 >0.95
tube wall quenches the reaction. Below this critical diameter, a exit temperature 1800 K 1500 K
combustion wave can only be stabilized through external heat- allowable wall temperature 1200 K 1600 K
ing of the tube wall. space heating rate (kW/m'/atm) 3.8 X 10" 3.3 X 10-'

Journal of Fluids Engineering MARCH 1998, Vol. 1 2 0 / 1 1 1

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 parts typically occurs through propagation of existing material Table 3 Comparison of selected microengine fuels (winter,
imperfections, but this is less of a concern for microengines 1990)
because of the small component size. Quality control and re-
Hydrogen/ Hydrocarbon/
finement during manufacturing can be more readily applied to Fuel property* air mixture air mixture
reduce inherent flaws, so the structural capabilities of these high
temperature refractory ceramics can be more fully realized. nominal composition H2 CH,.8
fuel specific heating value (kj/g) 120 42.8
diffusion velcoity (cm/s) 2 0.2
2 Combustors for Microengines flammability limits (% by volume) 4-75 0.6-4
vaporization rate w/o burning (cm/min) 2.5-5 0.05-0.5
Before discussing the proposed combustion strategies for mi- minimum ignition energy (mj) 0.02 0.25
croengines, it is instructive to review the physical basis for the autoignition temperature (K) 858 500
combustion scheme that is currently employed in most gas tur- characteristic reaction time at 5 atm (s) 1 X 10"- 1 X 10'
bine applications. flame propagation velocity (cm/s) 300 20
stoichiometric adiabatic flame
2.1 Combustion in Large-Scale Engines. The enthalpy temperature (K) 2318 2200
rise requirement of a gas turbine combustor is set by the cycle fraction of thermal energy radiated (%) 17-25 30-42
pressure ratio, properties of the working fluid and the fuel, and
material limitations. A balance of these influences is described *(in air at STP, unless otherwise specified).
by the equation
faster diffusion velocity, shorter reaction time, a significantly
TVs) (6) higher flame speed, wider burning limits, lower ignition energy,
and radiates less heat to its surroundings. Most importantly, the
an expression of the first law of thermodynamics. The required broad flammability Umits remove requirements for a relatively
fuel-air ratio is a function of the combustor inlet temperature rich primary burning zone followed by a dilution zone as is often
(Trj), which is set by the compressor pressure ratio, the com- necessary for hydrocarbon fuels. For a hydrogen-air reaction to
bustor exit temperature (7^4), which is typically set by maxi- provide a turbine inlet temperature of 1600 K, the required
mum material temperature limits downstream of the combustor, equivalence ratio of 0.34 (assuming no surface heat loss) falls
and properties of the fuel and air (g/, and Cp, respectively). well above the lean flammability limit of <^ = 0.1.
For gas turbine applications, the fuel-air ratio mandated by these Although the diffusion speed of hydrogen in air is an order
constraints typically falls between 0.015 and 0.045 over the of magnitude greater than that of a hydrocarbon fuel, providing
operating range of the engine. This proportion may also be adequate fuel-air mixing is still a critical requirement. Requisite
expressed as an equivalence ratio mixing can be achieved if the hydrogen gas is introduced well
upstream of the combustor, possibly even upstream of the com-
rhflrha pressor. Similar lean, premixed, prevaporized combustors have
(/.= (7)
{rhflrha^ stoichiometric been developed for hydrocarbon fuels for larger-scale gas tur-
bines to meet low-emissions requirements. Two significant dif-
which would then range between 0.21 and 0.63. ficulties have been identified in studies of these combustors:
The combustion strategies employed in current combustors lean blowout, and flashback or autoignition at high inlet temper-
are set to a large extent by the fact that most hydrocarbon-air atures (Grieb and Simon, 1990). For the current application,
mixtures will not burn at equivalence ratios less than approxi- lean blowout is not a concern because of the wide flammability
mately 0.5. The restrictive flammability range mandates a two- limits for hydrogen-air mixtures. Further, at the low operating
zone combustion scheme: a primary zone with relatively high pressure of 4.5 atm projected for a micro-gas turbine engine,
equivalence ratio, which also implies high temperatures beyond the combustor inlet temperature is around 500 K, which is below
material limits, and a secondary or dilution zone where addi- the autoignition temperature for a hydrogen-air mixture. There-
tional air is supplied to reduce the mixture temperature to fall fore, introduction and premixing of hydrogen well upstream of
within material limits. The dilution zone is also the region where the combustor is feasible in a microengine to enable nearly
pollution processes are of critical concern. These factors lead complete mixing of fuel and air.
to a residence time requirement of between 5 and 8 ms which
Adoption of a lean, premixed hydrogen system also results
is relatively independent of the physical size of the engine.
in a significant reduction in complexity because of the higher
2.2 Combustion Strategies for Microengines. Turning operating temperatures allowed for the microengine's materials.
to a discussion of the microcombustor applications, the strate- Transpiration and film-cooling are used in conventional gas
gies presented are based upon three general concepts: 1) in- turbine combustors to maintain the combustor liner below mate-
creasing the size of the combustor relative to the engine to rial temperature limits. However, as discussed in Section 1.2.3,
increase residence time, 2) premixing, and 3) lean burning. the ceramic materials intended for use in microturbomachinery
Removal of fuel-air mixing from the combustion chamber is have a distinct thermal advantage over the materials currently
driven by the severe residence time requirements for the micro used in gas turbines, and are expected to withstand local wall
application, and the realization that a large part of the residence temperatures in the vicinity of 1600 K. Thus a lean burning
time in current combustors is devoted to mixing. However, if hydrogen-air system not only obviates the need for a dilution
the reactants are mixed upstream of the combustor, then the zone, but also removes any requirement for combustor wall
stability benefits of a near stoichiometric primary zone are lost. cooling. A turbine inlet temperature of 1500 K permits the entire
Two alternatives for offsetting this difficulty and achieving sta- combustion process to be carried out without exceeding material
ble burning at low equivalence ratios are the use of hydrogen limits.
fuel, and the use of hydrocarbons with the assistance of surface While storage requirements currently prohibit the use of hy-
catalysis. Concepts based upon these ideas are discussed in drogen in commercial aircraft, several studies (Winter, 1990)
Sections 2.2.1 and 2.2.2, respectively. support its use as a fuel for future transport aircraft. For micro-
gas turbines, the use of hydrogen fuel is not necessarily viewed
2.2.7 Lean Burning Hydrogen-Air System. Hydrogen is as an endpoint, but rather as a first step to enable the realization
an ideal fuel in many respects. Table 3 compares selected fuel of a workable microscale device. This hydrogen-air system is
properties of hydrogen with a typical hydrocarbon. Hydrogen the subject of the experimental development efforts described
has a greater heatirig value, more rapid rate of vaporization. in Section 3.

112 / Vol. 120, MARCH 1998 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 2.2.2 Catalytic Hydrocarbon-Air Combustion. To achieve the value of 1600 K, the minimum adiabatic flame temperature
stable, lean combustion with a hydrocarbon fuel, catalytic pro- typicaUy associated with mixtures at the lean flammabiUty limit.
cesses must be employed. During the last 30 years, catalysts Cerkanowicz, Cole, and Stevens (1977) present several di-
have been investigated both for augmenting primary heat releas- mensionless groups that are useful for scaling the performance
ing reactions in gas turbines and for post-combustion treatment of catalytic combustors. The reaction in Region I is primarily
of pollutant emissions in aircraft, automobiles, and power gener- dependent on the catalytic Damkohler number, Da^., the ratio
ation applications. The focus here will be on situations where of residence time to the characteristic heterogeneous reaction
catalytic reactions are used to promote the primary combustion time;
reactions.

Overview of Catalytic Combustion. The principal advantage

Da. = ( ^ ] i ^ ^ ^ : ^ (8)
Vn) \ d.
of using catalysts for primary combustion reactions is their
ability to sustain reactions for hydrocarbon-air mixtures that are
well below the lean flammabiUty limit for homogeneous gas- Because of the dependence of the reaction rate on surface area,
eous combustion. Indeed, combustion of JP-4 at equivalence Da. a 1 id^ for fixed residence time. The transition from Region
ratios as low as 0.1 has been reported (Rosfjord, 1976). This I to Region II, and the progression of the reaction in Region II
is attractive for large-scale applications because the extended are expected to scale with the mass transport number which is
lean flammabiUty limit allows combustion to proceed at rela- the ratio of the characteristic residence time of the gas in the
tively low temperatures providing the potential for a factor of reactor to the mass transport time
100 decrease in emissions of nitrogen oxides.
Catalytic combustors typically employ proprietary platinum- (9)
and palladium-based catalysts on monolithic, parallel channel,
substrates. Silicon carbide is a leading substrate material. The
residence times in the catalyst bed are typically 5 to 30 ms for and the heat transfer number
tubes with hydraulic diameters on the order of I to 4 mm. For
combustor inlet temperatures ranging from 650 K to 900 K,
combustion efficiencies of greater than 99.5 percent have been JH = (10)
poCp di,
reported over a range of fuel-air equivalence ratios that result
in adiabatic flame temperatures between 1100 K and 1800 K
(e.g., Wampler et al., 1976; Mori et al, 1987). which is the ratio of the residence time of the gas to the charac-
The important technology limiters for catalytic gas turbine teristic heat transfer time. Both the heat and mass transfer scale
combustors are two-fold. First, the minimum operating tempera- inversely with the hydraulic diameter, and thus both are favored
ture of approximately 600 K rules out use of catalysts for com- as the size of the tube is decreased, as is the overaU progression
bustor inlet conditions corresponding to idle and low power to the homogeneous reaction region. Finally in Region III the
cruise operation. Second, the maximum temperature limit of process scales with the homogeneous Damkohler number. Da;,
around 1800 K, set by catalyst and substrate thermal degrada- the ratio of gas residence time to the homogeneous reaction
tion, does not allow use for hydrocarbon mixtures with equiva- time.
lence ratios greater than around 0.6. In practical devices these
limits are offset by using preburners, pilot flames, fuel and air A,[iusir[02]"e^
Da. = I A (11)
staging, and a variety of other hybrid catalytic/homogeneous [fuelL
combustion concepts.
The reacting flow processes within the catalytic bed are com- For fixed residence time. Da. is independent of the scale of the
plicated, and a brief review of these is necessary to understand device.
how the processes scale with combustor size. The following
discussion is based on descriptions given by Groppo et al. Application to Microengines. Based on the preceding dis-
(1993), Rosfjord (1976), and Pfefferle (1978). Consider pre- cussion of performance and scaUng of catalytic combustion
heated, premixed air and fuel as it passes through a long channel systems, the following conclusions can be drawn with respect
with catalytic surfaces. Reactants are absorbed by the catalyst, to a microengine application. The size of the combustor relative
surface reactions occur, and then the products are desorbed and to the engine can be increased to obtain a factor of 10 increase
along with heat are transported into the flow by diffusion. Three in residence time, as was proposed for the hydrogen system.
different regions exist in the channel. Region I: Near the inlet However, the resultant residence time, at 0.5 ms, will still be
of the channel, the fuel conversion is controlled by the kinetics an order of magnitude less than that in any successfully demon-
of the heterogeneous reaction at the surface of the catalyst. strated catalytic combustion systems. Second, the combustor
The bulk gas temperatures are typically too low to support inlet temperature of 460 to 500 K in the microengine appUpation
homogeneous combustion. This region of the flow usually ex- is below the 600 K limit required for catalytic ignition. We will
tends for a distance of less than five tube diameters. Region discuss each of these challenges in turn, beginning with the low
II: As the near-wall gas and substrate temperatures rise, the residence time.
heterogeneous reaction rates increase. The reaction rates in- The reduced residence time is largely offset by the favorable
crease to an extent that the fuel conversion is no longer con- effects associated with the high surface area-to-volume ratio of
trolled by the surface reaction rate, but rather by the rate at the microcombustor, including order of magnitude increases
which new gas phase reactants are transported to the surface. in both the heterogeneous Damkohler number and the non-
Pfefferle (1978) notes that, for any reasonably active catalyst dimensional heat and mass transfer rates that are the primary
at temperatures above 800 K, the catalytic reaction at the wall controlling parameters for the rate of reaction in the system.
will be limited solely by the rate of mass transfer. This diffusion This suggests that catalytic combustion of hydrocarbons can be
limited region extends for 10-100 tube diameters in typical completed for flow residence times on the order of 1 ms if the
catalytic combustors. Region III; Further downstream, as heat hydraulic diameter of the flow path is reduced by an order of
continues to diffuse into the bulk gas flow, thermal ignition of magnitude from larger-scale applications. For the microengine
homogeneous or gas phase reactions occurs. These homoge- application then, the catalytic flow passages should be approxi-
neous reactions are initiated at bulk gas temperatures in the mately 100 microns in diameter. Manginell et al. (1996) have
range of 1200-1300 K (Pfefferle, 1978) which is far below used chemical vapor deposition technology to apply platinum

Journal of Fluids Engineering MARCH 1998, Vol. 120 / 113

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 catalysts to micromachined polysilicon, providing proof that
miniature catalytic elements can be microfabricated.
The second problem mentioned above was that the combustor
inlet temperature was too low to support catalytic ignition. How-
ever, the experimental results presented in Section 3.4 suggest
that, for the current configuration, enough heat transfer may be
occurring between the combustor and the compressor exit flow
passage to overcome this difficulty. If not, a recuperator can be Fig. 3 Schematic of the model microcombustor
employed. A recuperator is a heat exchanger used to transfer
waste heat in the turbine exhaust to the compressor exit flow
prior to its entrance into the combustor. Recuperated cycles can manufactured, a 0.4 mm thick steel piece with 24 combustor
provide significantly higher thermal efficiency for low pressure inlet holes of 0.4 mm diameter around a circle of 5 mm radius,
ratio engines because of the relatively large difference in tem- and a 0.5 mm thick silicon carbide plate with 40 inlet holes of
perature between the combustor exit flow and the turbine ex- 0.38 mm diameter also around a circle of 5 mm radius. The
haust. For the microengine with a 4.5:1 pressure ratio, the com- large backward facing area of the inlet plates was designed to
pressor exit flow will be approximately 460 K and the turbine set up recirculation zones to provide flameholding.
exhaust approximately 1200 K, so heat exchangers with effi- The apparatus for holding the microcombustor is shown in
ciencies as low as 20% will allow catalytic combustor operation. Fig. 4. The microcombustor plates are held together in compres-
A typical heat exchanger efficiency reported in the literature sion by a spring. This modular design permits replacement of
for 50 kW automotive gas turbine applications is about 90% plates to simplify the task of conducting parametric studies for
(AlliedSignal Aerospace Company, 1988; Allison Gas Turbine various geometric configurations. The residence time for the
Division, 1988), suggesting that sufficient heat exchanger effi- microcombustor flow can be adjusted by adding or removing
ciency can be attained to enable catalytic combustion of hydro- plates, thereby changing the flowpath length and combustor
carbons. volume. The downstream side of the combustor is supported
The envisioned configuration for a recuperated, catalytic hy- by a clear quartz tube to allow limited visual access to the
drocarbon combustion system is shown in Fig. 2. The compres- combustor exit flow field. The entire combustor rig is housed
sor exit flow exchanges heat with the engine exhaust before in a water-cooled pressure chamber to enable testing at elevated
passing into a catalytic combustor. The catalytic combustor is pressures. Visual access is provided by a quartz window.
in the form of a radial labyrinth. An additional benefit of such
3.2 Diagnostics and Ignition. Due to the small size and
a system is that a large fraction of the heat lost from the walls
restrictive geometry of the microcombustor, diagnostics for the
of the combustion chamber can be retained in the cycle thus
combusting flow were difficult to apply. Two diagnostic tech-
improving the overall efficiency of the chemical to thermal
niques were used: 0.25 mm type K thermocouples and an ioniza-
energy conversion process. The primary drawback of this sys-
tion probe. Temperature measurements included inlet and exit
tem is its geometrical complexity, which may pose a significant
flow temperature, combustion chamber temperature, and com-
challenge to existing microfabrication technology.
bustor wall temperature. The ionization probe was used for
evaluating whether a flame existed within the microcombustor.
3 Hydrogen-Air Microcombustor Experiments and A schematic of the steel combustor illustrating the location of
Simulations some of the thermocouples, the ion probe, as well as an igniter,
is presented in Fig. 5. For the silicon carbide microcombustor,
A microcombustor test rig was developed to study combus-
only a single thermocouple and an ignitor were used. Both the
tion phenomena for microengine applications. The apparatus
thermocouple and the ignitor were introduced axially into the
and diagnostics are described in Sections 3.1 and 3.2, respec-
exit of the combustion chamber. Additional details of these
tively. Numerical simulations of the flowfield are presented in
measurement techniques and the ignition procedure are dis-
Section 3.3, followed by initial experimental results for lean,
cussed below.
premixed hydrogen-air combustion in Section 3.4.
Generally, the temperature indicated by a thermocouple probe
3.1 Apparatus. A schematic of the model microcombus- is not the same as the gas temperature because of the energy
tor is presented in Figure 3. As depicted in the figure, the model exchange between the thermocouple probe and the surrounding
consists of a series of plates stacked on top of each other. The environments. In steady-state, the temperature attained by the
hydrogen-air mixture is introduced into the dump combustor probe is a net result of convective energy transfer from the
through a ring of holes in the combustor inlet plate. After com- surrounding gas, radiative energy flux to the cooler walls, and
bustion, the flow exits axially. This differs from the flow in the
microengine shown in Fig. 1 where the combustion products
exit radially into the turbine. Two model microcombustors were
constructed, one of silicon carbide, which is the proposed mate-
rial for the micro-gas turbine, and the other of steel, which
allowed easier addition of diagnostics since it was less trouble-
some to machine. The volume of both combustors was approxi-
mately 0.13 cm'. Likewise, two combustor inlet plates were

H.at Exchanger Exhaust^

Fig. 2 Schematic of a recuperated microengine with a catalytic combus-

tor for burning hydrocarbon fuels Fig. 4 Schematic of the microcombustor testing facility

114 / Vol. 120, MARCH 1998 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 —r:; ===-:i:=.T. - , - - - • • - • • -

BOOK - - » f * f | P ' \ 1
^«- Lj | L | - BOOK

..' ••'\', ' ' .J '

\^l
,1 ll''* -!.• [•
10 (m/s60) t
lli iiP' J-1
1
Fig. 6 Results of a numerical simulation showing temperature and ve-
locity fields for ttie microcombustor for if) = 0.4, Ti„ = 450 K, P,„ = 1 atm,
and m = 0.045 g/s
Fig. 5 Microcombustor diagnostics and ignitor layout

mixture was specified an inlet temperature of 450 K at atmo-

conduction along the wires away from the thermocouple junc- spheric pressure, to provide a comparison with the microcom-
tion. Corrections for these effects must be made to obtain accu- bustor experiments.
rate temperature measurements (see, for example, Bradley and The velocity and temperature fields are shown in Figure 6.
Matthews, 1968). For most applications, the length of thermo- Flameholding is seen to occur in the immediate vicinity down-
couple wire exposed to the flow is long enough, and the gradi- stream of the combustor inlet plate hole. The maximum flow-
ents along the wire are small enough, so that conduction losses field temperature is 1650 K. The location of the maximum
are small. However, for the microcombustor, the length of ther- temperatures indicates that there is sufficient time for reactions
mocouple wire that protrudes into the flow is limited to only a to be completed well within the current combustor volume of
few millimeters by the scale of the device. Further, the large 0.13 cm'. This conclusion is supported by the results obtained
temperature gradients within the combustion chamber imply in the experiments, which are discussed in the following section.
unusually high uncertainties in flow temperature measurements
due to conduction losses. These uncertainties were aggravated 3.4 Preliminary Experimental Results. Tests of both
by lack of knowledge of the interior temperature field and addi- the steel and silicon carbide combustors were completed for the
tional inaccuracies introduced by the combustor wall measure- two combustor inlet plates described in Section 3.1. Equivalence
ments. Finally, since radiation corrections require the use of ratios between 0.4 and 1.0 were tested with mass flow rates
surrounding surface temperatures, they are also impacted by from 0.045 g/s to 0.2 g/s for combustor pressures of 1 to 4.5
inaccuracies in the combustor wall temperature measurements. atmospheres respectively, such that the combustor residence
Because of these accumulated errors, the estimated uncertainty time was maintained at approximately 0.5 ms. These conditions
of the temperature measurements is about ± 100 K. were chosen to correspond to typical microengine operating
An ionization probe was used to provide a secondary indica- parameters. All of the tests were carried out by varying the
tion of the presence of a flame within the combustion chamber. equivalence ratio from </> = 1.0 down to <^ = 0.4 in increments
The underlying concepts of ion probes are discussed in several of 0.1, and then back up to </> = 1.0 in increments of 0.1. Stable
references (Smy, 1976; Calcote, 1963; Travers and Williams, combustion was observed for both the silicon carbide combustor
1965; Zsak, 1993). The probe consisted of two platinum elec- and the steel combustor over the full range of equivalence ratios
trodes separated by an air gap. The two electrodes were main- tested.
tained at a fixed electrical potential difference. When ionized Internal temperature measurements obtained with the steel
radicals associated with the combustion zone were present, a microcombustor are shown in Fig. 7 over a range of hydrogen-
current was induced across the air gap, and this was manifest air equivalence ratios at atmospheric pressure. The flow rate for
as a change in voltage across a resistance placed in series with the case shown was 0.045 g/s. The measurements have been
the probe. corrected for radiation and conduction errors, however the un-
Ignition of the microcombustor was achieved by resistance certainty in these measurements is expected to be ±100 K due
heating of a 0.2 mm platinum wire. The ignitor consumed ap- to inadequate knowledge of thermal gradients within the com-
proximately 50 watts of power. Ignition usually occurred in a bustion chamber as described in Section 3.2. The magnitude of
matter of seconds, but was sensitive to wire location. Ignition each of the corrections is shown on the plot along with the raw
occun-ed more readily if the ignitor was positioned toward the temperature measurement. Also included in the figure are the
center of the combustion chamber; however, the wires would estimated adiabatic flame temperatures over the range of equiva-
only survive one test in this position. Moving the ignitor closer
to the combustor wall substantially increased its lifetime. As
with other experiments reported in the literature (Lewis and ESD measurement
von Elbe, 1987) catalysis on the heated platinum may have • 1 conduction
IB radiation
played a role in the ignition behavior. This will be investigated
in future tests.

3.3 CFD Simulations of tlie Microcombustor. Before

presenting experimental results, we illustrate the general fea-
tures of the flowfield in the microcombustor using results from
two-dimensional reacting flow simulations. These simulations
were carried out using ALLSPD (Chen et al., 1995), a Navier-
Stokes solver incorporating finite-rate chemical kinetics, The
code has been validated by personnel at NASA Lewis Research
Center (ALLSPD, 1995) for laminar reacting flows like those
0.5 0.6
expected in the microcombustor. A hydrogen-air chemistry
Equivalence Ratio
model with 9 species and 18 reaction mechanisms was used in
the current study. The simulations were conducted for a 0 = Fig. 7 Exit gas temperatures measured in the hydrogen-air microcom-
0.4 hydrogen-air mixture flowing at a rate of 0.045 g/s. The bustor

Journal of Fluids Engineering MARCH 1998, Vol. 120 / 115

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 4 Summary and Conclusions
Microengines and other miniature thermal devices pose
unique challenges and opportunities for combustion in small
volumes. The principal difficulties are associated with the lim-
ited residence times, and heat transfer losses due to the high
surface area-to-volume ratio. Two viable combustion alterna-
tives for microengines were presented in the paper, one that
takes advantage of the wide flammability limits of hydrogen-
air mixtures, and a second that involves surface catalysis of a
95% Confidence O 1 atm
-n-2atm
hydrocarbon fuel. Both options require an order of magnitude
Interval
. ^ 3 atm increase in the size of the combustor relative to the engine, and
K>-4 atm premixing the fuel and the oxidizer to decrease residence time
requirements within the combustor. Both combustion strategies
0.4 0.5 0.6 0.7 also represent significant reductions in complexity from typical
Equivalence Ratio, ^ gas turbine systems by the removal of requirements for a pri-
mary combustion zone where temperatures exceed the material
Fig. 8 Combustor efficiency for tlie microcombustor over a range of
pressures and equivalence ratios
limits.
Preliminary test results from 0.13 cm^ silicon carbide and
steel microcombustors were presented for inlet pressures be-
tween 1 and 4 atm. Stable and complete combustion of hydrogen
lence ratios based on the measured inlet air temperatures. Lack and air was achieved over the full range of equivalence ratios
of any visible indication of a flame at the combustor exit (even required for a microengine application. Combustor efficiencies
when the hydrogen was seeded with five percent methane to between 0.7 and 0.9 were obtained. The low combustor effi-
increase the chemiluminescence) suggests that the combustion ciencies were largely a result of heat transfer losses from the
efficiency was near one. However, the combustor efficiency walls of the combustor. The experiments provide a first proof-
was reduced due to heat transfer losses as is suggested by the of-concept demonstration for the hydrogen-air system, and sug-
difference between the corrected temperatures and the adiabatic gest that combustion in small volumes for microengine applica-
flame temperatures. This difference increased with increasing tions is feasible.
equivalence ratio as is expected. The work presented in this paper also highlights the need for
The overall combustor efficiencies associated with these tem- significant additional research to more carefully understand flow
perature differences fall between 0.7 and 0.9 and are shown behavior within the microcombustor. This research would be
plotted in Fig. 8 for the range of equivalence ratios and pressures aided by the development of miniature diagnostic techniques for
tested. Results are not presented for many of the higher pressure these flows. Experimental and computational work is continuing
and equivalence ratio conditions because the associated high with the focus on issues of hydrogen injection and mixing,
temperatures generally resulted in failure of the fine wire ther- ignition, catalytic combustion of hydrocarbon fuels, and deter-
mocouple probes. The combustor efficiency variation with pres- mination of oxidation rates for silicon carbide in the post-com-
sure is consistent with the variation in heat transfer coefficients bustion gas environment. A parallel experimental effort is also
with pressure and flow rate. In sum, despite the low combustor ongoing using a miniature flame tube apparatus. The flame tube
efficiency, both the steel and silicon carbide microcombustors will be used to provide measurements to help validate reacting
were operated for tens of hours and throughout produced the flow simulations and finite element heat transfer models.
requisite heat release necessary to support the various heat en-
gine applications that are currently envisioned.
Other observations from these preliminary tests include:
Acknowledgments
We first recognize and thank Professor Alan H. Epstein, both
1) The ionization probe indicated the presence of a flame for conceiving and pursuing the concept of microengines, and
within the combustor over the full range of flow rates for being a constant source of encouragement and ingenuity. In
tested for equivalence ratios between 0.4 and 1.0. The addition, we thank the members of the MIT microengines team:
observed voltage was found to increase with increasing Kuo-Shen Chen, Eric Esteve, Luc Frechette, Chuang-Chia Lin,
equivalence ratio. Amitav Mehra, Steve Nagle, Ed Piekos, G. Ananthasuresh, Ar-
turo Ayon, Fredric Ehrich, Stuart Jacobson, Choon Tan, Ken-
2) At constant flow rate, the choice of the combustor inlet
neth Breuer, Mark Drela, Jeffrey Lang, Martin Schmidt, Ste-
plate did not significantly affect the temperatures mea-
phen Senturia, and Mark Spearing, for all they have taught us.
sured, probably indicating similar flameholding charac-
The skill and assistance of Viktor Dubrowski, Mariano Hellwig,
teristics, and thus similar temperature fields for the two
and James Letendre made the apparatus used in this study possi-
geometries. ble. We are grateful to Diana Park for superb editing and graph-
3) The microcombustor plates became sufficiently hot to ics, and to Holly Anderson for shouldering many of the adminis-
glow red at equivalence ratios greater than 0.5, and the trative burdens.
heat transfer to the inlet gases was significant (on the
This work was largely supported by the Army Research Of-
order of 50 K to 100 K temperature rise). Typically,
fice through Grant DAAH04-95-1-0093. We greatiy appreciate
combustor wall temperatures were about 750 K.
the interest and encouragement of the technical monitor. Dr.
4) The thermocouples located at two different radii within Richard Paur. We also thank the MIT Lincoln Laboratory for
the combustor sensed temperatures that differed by as providing support for the initial feasibility studies.
much as 180 K due to large radial temperature gradients
in the combustor. Lack of further knowledge of these
gradients is the source for the large errors in conduction References
corrections quoted for the thermocouples. AUiedSignal Aerospace Company, 1998, "Advanced Gas Turbine AGT Tech-
5) During some low equivalence ratio conditions, hystere- nology Development Project," Final Report to NASA Lewis Research Center,
DOE/NASA 0167-12, Mar.
sis in the thermocouple measurements of 200 K was AUison Gas Turbine Division, 1985, "Advanced Gas Turbine AGT Technology
observed suggesting that flameholding did not always Project," Final Report to NASA Lewis Research Center, DOE/NASA 0168-11,
occur at the same location within the microcombustor. Aug.

116 / Vol. 120, MARCH 1998 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms

 ALLSPD Demonstration Workshop, Personal Communication, 1995. Microfilament Gas Sensors," Solid State Sensor and Actuator Workshop, Hilton
Ballal, D. R,, and Lefebvre, A. H., 1979, "Weak Extinction Limits of Turbulent Head, South Carolina, June.
Flowing Mixtures," Journal of Engineering for Power, Vol. 101, No. 3. Measley, M. L., and Smyth, J. R., 1996, "Ceramic Gas Turbine Technology
Bradley, D., and Matthews, K. J., 1968, "Measurement of High Gas Tempera- Development," ASME paper 96-GT-367.
tures with Fine Wire Thermocouples," Journal of Mechanical Engineering Sci- Mori, K., Kitajima, J., Kajita, S., and Ichihara, S., 1987, "Development of a
ence, Vol. 10, No. 4. Catalytic Combustor for Small Gas Turbines," ASME Paper 87-GT-62, May-
Bryzek, J., Peterson, K., and McCuUey, W., 1994, "Micromachines on the June.
March," IEEE Spectrum, Vol. 31, No. 5, May. Nakazawa, N., Ogita, H., Takahashi, M., Yoshizawa, T., and Mori, Y., 1996,
Calcote, H. F., 1963, "Ion and Electron Profiles in Flames," Ninth Symposium "Radial Turbine Development for the 100 kW Automotive Ceramic Gas Tur-
(International) on Combustion, The Combustion Institute. bine," ASME paper 96-GT-366.
Cerkanowicz, A. E., Cole, R. B., and Stevens, J. G., 1977, "Catalytic Combus- Pfefferle, W. C , 1978, "The Catalytic Combustor: An Approach to Cleaner
tion Modeling; Comparisons with Experimental Data," ASME Paper 77-GT-85, Combustion," AIAA J. Energy, Vol. 2, May-June.
Mar. Rosf jord, T. J., 1976, "Catalytic Combustors for Gas Turbine Engines," AIAA
Chen, K-H., Duncan, B., Fricker, D., Lee, J., and Quealy, A., 1995, ALLSPD- Paper 76-46, Jan.
3D, Version 1.0, Internal Fluid Mechanics Division, NASA Lewis Research Cen- Scott, W. D„ 1979, "Statistics of the Strength of Brittle Materials," Design
ter, Nov. with Brittle Materials, Mueller, J. I., Kobayashi, A. S., Scott, W. D. (Editors),
Dodds, W. J., and Bahr, D. W., 1990, "Combustion System Design," Design University of Washington.
of Modern Gas Turbine Combustors, Mellor, A.M. (Editor), Academic Press Smy, P. R., 1976, "The Use of Langmuir Probes in the Study of High Pressure
Limited. Plasmas," Advances in Physics, Vol. 25, No. 5.
Epstein, A. H., Oroshenry, C , Haldeman, C. W., Schmidt, M. A., Senturia, Tanaka, K., Tsuruzono, S., Kubo, T., and Yoshida, M., 1996, "Development
S. D., Tan, C. S., Waitz, 1. A., and Wong, J., 1995, "Microjet Engines," Final
and Fabrication of Ceramic Gas Turbine Components," ASME paper 96-GT-
Technical Report to MIT Lincoln Laboratory, Massachusetts Institute of Technol-
446.
ogy-
The Carborundum Company, 1989, "Physical Properties of Hexoloy® SA Sili-
Grieb, H., and Simon, B., 1990, "Pollutant Emissions of Existing and Future
con Carbide—Technical Data," The Carborundum Company, November 10.
Engines for Commercial Aircrafts," Air Traffic and the Environment—Back-
Travers, B., and Williams, H., 1965, "The Use of Electrical Probes in Flame
ground, Tendencies and Potential Global Effects, Proc. of the DLR International
Colloquium, Bonn, Germany, U, Schumann (Ed.), Springer-Verlag. Plasmas," Tenth Symposium (International) on Combustion, The Combustion
InslitLUe.
Groppo, G., Tronconi, E., and Forzatti, P., 1993, "Modelling of Catalytic
Combustors for Gas Turbine Applications," Catalysis Today, Vol. 17, Elsevier Wampler, F. B., Clark, D. W., and Gaines, F. A., 1976, "Catalytic Combustion
Science Publishers B. V., Amsterdam, pp. 237-250. of CaHs on Pt Coated Monolith," Combustion Science and Technology, Vol. 14,
Groshenry, C , 1995, "Preliminary Study of a Micro-Gas Turbine Engine," pp. 25-31.
S.M. thesis, Massachusetts Institute of Technology, May. Winter, C.-J., 1990, "Hydrogen Technologies for Future Aircraft," Air Traffic
Juan, W-H., Pang, S. W., 1996, "Released Si Microstructures Fabricated by and the Environment—Background, Tendencies, and Potential Global Atmo-
Deep Etching and Shallow Diffusion," Journal of Microelectromechanical Sys- spheric Effects, Proc. of the DLR International Colloquium, Bonn, Germany, U.
tems, Vol. 5, No. 1, Mar. Schumann (Editor), Springer-Verlag.
Kerrebrock, J. L., 1992, Aircraft Engines and Gas Turbines, 2nd ed., MIT Zamaschikov, V. V., 1995, "Combustion of Gases in Thin-Walled Small-Diam-
Press. eter Tubes," Combustion, Explosion, and Shock Waves, Vol. 31, No. 1, Jan.-Feb.
Lefebvre, A. H., 1983, Gas Turbine Combustion, Hemisphere Publishing Cor- Zeldovich, Y. B., Barenblatt, G. 1., Librovich, V. B., et al., 1990, Mathematical
poration. Theory of Combustion and Explosion.
Manginell, R. P., Smith, J. H., Ricco, A. J., Moreno, D. J., Hughes, R. C , Zsak, T. W., 1993, "An Investigation of the Reacting Vortex Structures Associ-
Huber, R. J., and Senturia, S. D., 1996, "Selective, Pulsed CVD of Platinum on ated with Pulse Combustion," Ph.D. thesis, California Institute of Technology.

Journal of Fluids Engineering MARCH 1998, Vol. 120 / 117

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/26/2015 Terms of Use: http://asme.org/terms