United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600.9 80-035
August 1980
Research and Development
Proceedings Fourth
Workshop on
Catalytic Combustion
(Cincinnati, OH,
May 1980)
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA 600/9-80-035
August 1980
PROCEEDINGS: FOURTH WORKSHOP
ON CATALYTIC COMBUSTION
(CINCINNATI, OH, MAY 1980)
John P. Kesselring, Compiler
Acurex Corporation
Energy and Environmental Division
485 Clyde Avenue
Mountain View, California 94042
Contract No. 68-02-3122
Project Officer:
G. Blair Martin
Office of Energy, Minerals, and Industry
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory - Research Triangle Park, N.C., U.S. Environmental Protection
Agency, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
ii
-------
PREFACE
This proceedings document covers the presentations of the Fourth Workshop
on Catalytic Combustion, held May 14-15, 1980 at the Netherland Hilton Hotel in
Cincinnati, Ohio. Sponsored by the Combustion Research Branch of the EPA's
Industrial Environmental Research Laboratory — Research Triangle Park, the work-
shop served as a forum for the presentation of results of recent research in the
areas of catalytic combustion system components and applications, catalyst per-
formance, and alternative fuels use in catalytic combustors. The First and
Second Workshops on Catalytic Combustion were held at the Plantation Inn in
Raleigh, North Carolina on May 25-26, 1976 and June 21-22, 1977, respectively.
The Third Workshop on Catalytic Combustion was held at the Grove Park Inn in
Asheville, North Carolina on October 3-4, 1978. Summaries of the First and
Second Workshops and the Proceedings of the Third Workshop appear in EPA-600/
7-79-038, "Proceedings: Third Workshop on Catalytic Combustion."
Dr. John P. Kesselring, Acurex Corporation, acted as Workshop Coordinator,
and G. Blair Martin, Combustion Research Branch, Environmental Protection Agency,
was the Project Officer.
iii
-------
PROCEEDINGS: FOURTH WORKSHOP ON CATALYTIC COMBUSTION
(Cincinnati, Ohio, May 1980)
ABSTRACT
The proceedings document the major presentations at the Fourth Workshop
on Catalytic Combustion, held in Cincinnati, Ohio, May 14-15, 1980. Sponsored
by the Combustion Research Branch of EPA's Industrial Environmental Research
Laboratory — Research Triangle Park, the workshop served as a forum for the
presentation of results of recent research in the areas of catalyst performance,
components and applications of catalytic combustion systems, and the use of
alternative fuels in catalytic combustors. The workshop provided industrial,
university, and government representatives with the current state of the art in
the application of catalyst systems for pollution control and performance im-
provement. Applications include firetube and watertube boilers and gas tur-
bines for utility, industrial, automotive, and aircraft systems.
iv
-------
TABLE OF CONTENTS
Page
Session I — Components of the Catalytic Combustion System 1
"Autoignition in a Premixing-Prevaporizing Duct at Inlet Air Tempera-
tures up to 1000K," Robert R. Tacina 2
"Fuel Injector, Ignition, and Temperature Measurement Techniques
for Catalytic Combustors," S. J. Anderson, W. V. Krill, and J. P.
Kesselring ............................... 3
"Preliminary Design of Catalytic Combustors," William B. Retallick ... 35
"Catalytic Flame Stabilization," Moshe Lavid and A. E. Cerkanowicz ... 44
"Thermal Shock Resistant Catalytic Monoliths Using Foam Ceramics
Technology," William C. Pfefferle .................... 64
Session II — Catalyst Performance .................... 73
"Computed and Measured Emissions from the Catalytic Combustion of
Propane/Air Mixtures," F. V. Bracco, C. Bruno, Y. Yaw, P. M. Walsh. ... 74
"Modeling of Transient Operation of Catalytic Combustor," James S. T'ien. 117
"The Effect of Catalyst Length and Downstream Reaction Distance on
Catalytic Combustor Performance," David N. Anderson ...........
"CATCOM Catalyst 5 atm 1000 Hour Aging Study Using #2 Fuel Oil,"
I. T. Osgerby, B. A. Olson, H. C. Lee .................. 172
"Fuel-Nitrogen Conversion and Combustion Efficiency in a Catalytic
Combustor," Ronald D. Matthews, Mark L. Graham, Jeryl L. Lederman .... 216
"CATCOM Catalyst Axial Temperature Profile Measurements," I. T. Osgerby,
B. A. Olson, H. Lew, A. Cohn ...................... 244
"Development of Improved Catalyst Systems," J. P. Kesselring, W. V.
Krill, M. J. Angwin, and H. L. Atkins .................. 246
"Catalyzed Combustion of Lean Fuel/Air Mixtures," R. W. Schef er ,
F. Robben ................................ 277
-------
"Examination of the Nickel Catalyst Role in Rich Combustion,"
Gerald E. Voecks 297
Session III — Applications of Catalytic Combustion 313
"Prototype Catalytic Systems," W. V. Krill, J. P. Kesselring, S. J.
Anderson 314
"EPRI Stationary Gas Turbine Catalytic Combustor Development Program,"
Leonard C. Angello 335
"Catalytic Combustion for Gas Turbines," Hisashi Fukuzawa, Yoshimi
Ishihara 363
"Soot Reduction in Diesel Engines by Catalytic Effects," R. Sapienza,
T. Butcher, C. Krishna, J. Gaffney 380
"High Pressure Test Results," D. E. Carl, M. C. Krepinevich, I. T.
Osgerby 400
"Combustion Catalyst Study for Simulated Aircraft Idle Mode Operation,"
I. T. Osgerby, R. M. Heck, R. V. Carrubba, C. C. Gleason, E. J. Mularz . 415
"Experimental Evaluation of Catalytic Combustion With Heat Removal at
Near Stoichiometric Conditions," Daniel Bulzan 446
Session IV — Alternate Fuels and the Catalytic Combustor 449
"Fuel Nitrogen Conversion — The Impact of Catalyst Type," B. A. Folsom,
W. D. Clark, C. W. Courtney, M. P. Heap, W. R. Seeker 450
"Catalytic Combustion Characteristics of Low and Medium Btu Gas Fuels,"
H. C. Lee, I. T. Osgerby 475
"Catalytic Combustion of Alternative Fuels," Henry Tong, Edward K. Cliu,
Gerry C. Snow 509
"Gasification of a Heavy and a Distillate Fuel," E. J. Szetela, R. A.
Sederquist, J. A. TeVelde 535
"Catalytic Combustion of Heavy Partially-Vaporized Fuels," T. J.
Rosfjord 548
"Evaluation of a Novel Lean Burn Low NO Catalytic Combustion Concept,"
E. K. Chu, R. Chang, H. Tong * 560
vi
-------
SESSION I
COMPONENTS OF THE CATALYTIC
COMBUSTION SYSTEM
-------
AUTOIGNITION IN A PREMIXING-PREVAPORIZING DUCT
AT INLET AIR TEMPERATURES UP TO 1000K
By:
Robert R. Tacina
National Aeronautics and Space Administration
Lewis Research Center
Cleveland, Ohio 44135
Conditions were determined in a premixing-prevaporizing
fuel preparation duct at which autoignition (upstream burning)
occurred. An air blast type fuel injector with nineteen fuel
injection points was used that provided a uniform fuel-air
mixture. The range of inlet conditions were: inlet air
temperatures of 600 to 1000K, air pressures of 180-660 kPa,
fuel-air ratios from 0.009 to 0.070, and velocities from 3.5 to
30 m/s. The duct was insulated and the diameter was 12 cm.
Mixing lengths were varied from 16.5 to 47.6 cm. and residence
times ranged from 4.6 to 107 ms. The fuel used was diesel #2.
Results showed a strong effect of fuel-air ratio and pressure
on the conditions where autoignition occurs. The effect of
length or residence time on autoignition was not apparent in
these tests. A correlation of the conditions where auto-
ignition would occur which apply to this fuel injector over the
conditions tested is p 0 °'7 = 73 e T , where p is the
pressure in kPa, 0 the equivalence ratio, and T the
temperature in K.
-------
FUEL INJECTOR, IGNITION, AND TEMPERATURE
MEASUREMENT TECHNIQUES FOR CATALYTIC COMBUSTORS
By:
S. J. Anderson, W. V. Krill, and J. P. Kesselring
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
-------
ABSTRACT
As catalytic combustion systems approach the demonstration phase of
of development, special auxiliary components are required for fuel injection,
ignition, and catalyst temperature measurement. Following preliminary screen-
ing of potential concepts, several auxiliary systems were designed and tested.
Design details and test results are reported.
Liquid and gaseous fuel aerodynamic injectors and liquid fuel atomizing
nozzles were tested in both large and small duct sizes. Aerodynamic injectors
provided rapid and uniform gas mixing. However, they did not adequately vapo-
rize liquid fuels, which require spray nozzles for injection. To maintain
short mixing lengths in large ducts, multiple point injection was employed.
Among many catalyst lightoff concepts that were reviewed, the opposed
jet igniter and the aft-end torch were selected for detailed testing. The
opposed jet demonstrated excellent catalyst lightoff capabilities over a wide
range of mainstream stoichiometries and flowrates. It delivered preheated gas
temperatures as high as 1366K (2000°F) with both natural gas and diesel fuels.
In contrast, aft-end torch lightoff was possible only with gaseous fuels at
low gas velocities. This restriction considerably limits its application to
catalytic combustors.
Thermocouple probes and pyrometers were evaluated as alternative tem-
perature measurement techniques to thermocouples attached directly to catalyst
surfaces. Probes located immediately downstream of the catalyst were found to
be unreliable because temperature measurement depended in a complicated manner
on system geometry and exhaust gas velocities. Although inserting a probe
into the catalyst bed improved measurement accuracy, further testing is required
at gas velocities greater than 7 m/s. Testing showed that a single wavelength
optical pyrometer can accurately measure surface temperature within 10K (15°F)
of actual bed temperature if surface emissivity is known as a function of tem-
perature. However, interference in the sitepath (e.g., soot) adversely affected
measurement accuracy. To maintain accuracy when interference was a problem,
dual wavelength pyrometers were employed.
-------
INTRODUCTION
Catalytic combustion is an innovative technology capable of significantly
reducing NO formation while maintaining high combustion efficiency. However,
X
catalytic combustors require special ignition techniques and premixed fuel and
air. For catalyst lightoff, an ignitor must heat the catalyst to over 700K
in a controlled manner. Once lightoff is achieved, the ignitor must be easily
shut off. To effectively utilize all catalyst material and prevent excessive
temperatures due to locally enriched reactant mixtures, air and fuel should be
evenly distributed prior to entering the catalyst. Fuel injectors must provide
thorough and rapid mixing to minimize mixing length, prevent recirculation
which may lead to flameholding, and promote rapid prevaporization of liquid
fuels. Finally, catalyst temperature must be carefully monitored during
startup and steady state operation to maintain catalyst material within allow-
able operating limits. Correspondingly, a reliable system must be developed
to continuously measure catalyst surface temperature.
As part of the EPA's catalytic prototype system development program,
Acurex is investigating auxiliary system components capable of providing
acceptable fuel/air mixing, controlled-temperature ignition, and reliable
monitoring of catalyst surface temperature. This paper will describe the de-
velopment of gaseous and liquid fuel injectors, catalytic combustor ignition
systems, and catalyst temperature measurement techniques.
FUEL INJECTION CONCEPTS
Optimum catalyst performance requires uniform reactor inlet profiles
of temperature, fuel/air ratio, and velocity. This assures effective use of
-------
the entire catalyst area and prevents substrate damage caused by high tempera-
tures due to locally enriched fuel/air mixtures. The primary function of the
fuel injector is to introduce the fuel to the air stream in such a manner as
to provide complete mixing before entry into the catalyst. For liquid fuels,
the fuel must also be at least partially prevaporized. Both mixing and vapor-
ization occur in a flow region between the injector and the catalyst, and
proper fuel injection can greatly facilitate these processes.
Rapid mixing results from intimate fuel/air contact. This can be
accomplished by placing multiple injectors across an incoming air duct or
maintaining strong recirculation in the mixing zone. The latter technique is
often intolerable in catalytic systems, since flashback from the hot catalyst
surface can cause flameholding in the areas of recirculation. Therefore, in
systems with large sectional areas, multiple fuel injection points are pre-
ferred.
For systems operating on liquid fuels, the injector must also promote
rapid vaporization. It has been estimated in one study (Reference 1) that
90 to 95 percent of the fuel must be vaporized for stable catalyst operation.
However, tests at Acurex have shown that 70 percent vaporization may be accept-
able in some system applications. The rate of droplet vaporization is con-
trolled by the initial size of the droplet and the heat transfer to the drop
by convection from the surrounding air and radiation from the catalyst bed.
Total prevaporization can be achieved by allowing long mixing times to transfer
sufficient energy to vaporize the drop before entrance to the catalyst. How-
ever, the allowable mixing length is limited by autoignition delay times which
are particularly short in high pressure, high temperature systems. Therefore,
the alternative to long residence time is to produce a fine, uniform droplet
distribution over the cross-sectional area of the combustor mixing section at
the fuel injector.
Several fuel injection techniques exist which can provide rapid fuel/
air mixing and fine liquid fuel atomization. Aerodynamic fuel injectors for
both gaseous and liquid fuels and atomizing spray nozzles for liquid fuels
were selected as candidate injectors. Design details and test results as
they apply to both small and large scale combustors are dicussed below.
-------
AERODYNAMIC FUEL INJECTION
Aerodynamic fuel injectors promote fuel-air mixing by injecting fuel
into a relatively high velocity air stream. High velocities are achieved by
constricting the flow in a narrow duct, then gradually expanding the fuel-air
mixture to the catalyst front face. For larger systems with high air flow
requirements, an array of multiple ducts may be necessary to provide uniform
fuel-air distribution to the catalyst while maintaining high gas velocity.
In general, liquid fuels are injected at velocities below the air velocity to
increase atomization of the liquid fuel by droplet shear. In contrast, gaseous
fuels are injected at velocities comparable to the duct air velocity to mini-
mize recirculation regions in the mixture stream which could support flame-
holding.
Two types of aerodynamic fuel injectors, one for a small area duct and
one for a large area duct, were constructed and tested at Acurex. A layout of
the fuel injector for a small duct is shown in Figure 1. A total of six tubes
were oriented concurrent with the air stream: three small diameter tubes for
distillate oil and three larger tubes for gaseous (natural gas and propane)
fuels. Each of the fuel tubes supplies an equal area segment of the duct
cross-section. The tubes were sized such that the fuel injection velocity
is near the air stream velocity under most operating conditions, and several
inches of length in the flow direction allow flow recovery from upstream dis-
turbances prior to the injection point. Therefore, intimate fuel-air contact
is achieved without the formation of recirculation regions. The three tubes
for each fuel type are manifolded exterior to the test apparatus, and the
injector plate is bolted between flanges in the facility mixing section.
Expansion of the mixture occurs in a refractory-lined section with a total
cone angle of 0.087 radians (5 degrees).
Performance of the injector with gaseous fuels has been excellent. Fuel-
air mixing appears to be complete at the catalyst face (as indicated by cata-
lyst temperature variations) for a mixing length of 1.07 m (3.5 ft). Even at
conditions where incoming fuel and air are at greatly dissimilar velocities,
mixing appears complete and no flashback or flameholding problems have been
encountered.
-------
Operation on distillate oil with the same mixing length provides vary-
ing results. At lower air velocities, the fuel appears to be more heavily
concentrated at the lower half of the catalyst, even at preheats of 700K (800°F)
(well above the upper distillation point of the oils tested). This phenomenon
is attributed to rather large droplets coming off of the injector and gravitat-
ing to the lower half of the duct during mixing. Since shear force induced by
the air flowing over the tips of the injector ports is the only mechanism of
droplet atomization, the maldistributions experienced are consistent with poor
atomization at low flow velocities. At higher air velocities (usually in ex-
cess of the incoming fuel velocity as well), the distribution again appears
uniform. It therefore appears that finer atomization is required to meet the
wide range of flow conditions anticipated in systems applications.
For a large area duct, a multiple venturi fuel injector based on a con-
cept of Tacina (Reference 2) was fabricated for a model gas turbine combustor.
The fuel injector, capable of operation with both gaseous and distillate fuels,
is shown in Figure 2. This concept produces a number of high velocity air
streams, injects fuels into the high velocity regions, and expands the fuel-
air mixture into a large duct.
During testing of the injector assembly, flameholding resulted within
the expansion cones. Several features of the design may account for this dif-
ficulty. The downstream end of the injector cones is a close-packed hexagonal
array of seven tubes. The tested design did not attempt to eliminate the base
region between cones, and recirculation certainly existed in these regions.
The potential for flameholding in these regions is compounded by the introduc-
tion of fuel at the upstream end of the assembly in that a partially premixed,
partially prevaporized mixture can occur at the outlet of the cones.
Other features of the design may also result in operational problems.
A relatively long venturi cone is required to provide high velocities at the
fuel injection port, and a small cone angle is required to prevent boundary
layer separation on the internal walls. In addition, atomization of the dis-
tillate fuel results only from shear of the air over the fuel ports and may
produce relatively large droplets. Finally, great care is required to insure
uniform injection in all cones under varying fuel flowrates. Tacina and others
have revised their initial designs to address some of these problems. As a
-------
result of these problems, atomizing spray nozzle injectors were developed for
liquid fuels.
ATOMIZING SPRAY NOZZLE INJECTORS
To improve mixing and vaporization of liquid fuels, atomizing spray
nozzles were developed for small bench scale tests and for larger catalytic
system tests. Each application used an air-atomizing type of nozzle that pro-
duces a mean drop size of approximately 60 to 100 \im.
The single injector is currently used for catalyst screening tests in
an atmospheric test facility. Details of the fuel injector assembly are shown
in Figure 3. The system is oriented vertically and operates with both gaseous
and liquid fuels. Gaseous fuel and air are premixed and introduced through
four ports concentric about the spray nozzle. For liquid fuel injection, air
enters the same four ports with fuel sprayed axially into a conical venturi.
Further mixing occurs in a 6.35 cm diameter section below the venturi prior to
entering the catalyst. The nozzle assembly also includes a pneumatic on/off
valve and a coolant housing through which air is passed to prevent heatup of
the assembly by preheated air or radiation from the catalyst.
Operation of the fuel introduction system with gaseous fuels has been
totally acceptable for over 40 catalyst tests performed to date. Similarly
with diesel fuel, the fine atomization promotes rapid mixing and vaporization
over a wide range of velocities (1.5 to 7.6 m/sec, 5 to 25 ft/sec) through the
premixing section. Flashbacks occurred only when operating at low velocities
approaching the mixture flamespeed. The overall success of the injector in
achieving uniform mixing has made it possible to operate catalysts with par-
tially vaporized fuel in the air stream.
For systems that cannot accommodate long mixing lengths or pressure
losses associated with small ducts, a more sophisticated multiple point injec-
tion scheme is required. The concept, with the model gas turbine combustor
can, is shown in Figure 4. The injector is a three nozzle arrangement for dis-
tillate fuel injection in a 12.7 cm duct. The spray nozzles are mounted on
extensions approximately 5 cm away from the injector plate. The injector plate
is mounted between flanges of the test facility. Vaporization or mixing are
nearly complete in the allowed mixing length without autoignition difficulties.
-------
During initial testing of the multiple injectors, the nozzles' wide
spray angle caused fuel to impinge directly on the wall of the combustor can.
Some thermal NO was produced, apparently as a result of upstream burning of
X
the distillate fuel on the wall. This was confirmed by cold water model tests
using a clear acrylic combustor can. Fuel contact with the wall was suppressed
by reducing the nozzle spray angle from 77 to 15 degrees. Subsequent combus-
tion testing with a low spray angle was quite successful. The injection sys-
tem provided a uniform fuel mixture without upstream burning.
In summary, aerodynamic fuel injectors provide rapid and complete fuel-
air mixing of gaseous fuels over a wide range of flowrates. If system pressure
drop is not important, a single injector is sufficient. To minimize pressure
drop, multiple point injection may be necessary. Aerodynamic injectors do not
adequately atomize liquid fuels, however. Consequently, spray nozzles are
recommended for these applications. In addition, for large systems, multiple
point injection should be employed to reduce mixing length while maintaining
uniform fuel-air distribution.
COMBUSTOR IGNITION CONCEPTS
Catalyst lightoff occurs when a minimum catalyst temperature is achieved.
The lightoff temperature varies primarily with catalyst and fuel type and secon-
darily with stoichiometry and mass throughput. In general, temperatures in
excess of 700K (800°F) are necessary for lightoff. For example, noble metal
catalysts such as platinum and palladium require 617K to 784K (650°F to 950°F)
lightoff temperatures, while metal oxide catalysts such as nickel and cobalt
require 866K to 1367K (1100°F to 2000°F) temperatures. In early EPA testing,
lightoff temperature was achieved using electric air preheaters. However, air
preheat temperatures at this level are not availa'ble in boiler and gas turbine
systems. Consequently, alternate techniques for lightoff were investigated.
Table I lists advantages and disadvantages of possible catalyst ignition sys-
tems. Each system has potential application to utility and industrial boilers,
and stationary gas turbines. Two systems were chosen for test evaluation: the
opposed jet igniter and the aft-end torch. The opposed jet was selected due
to its simple operation, variable temperature control, and wide mixture ratio
10
-------
and flowrate blowout limits. The aft-end torch technique was chosen for its
simplicity.
OPPOSED JET IGNITER
During operation, a jet of premixed fuel and air (approximately 1 per-
cent of the total mainstream mass flow) is injected counterflow to the main
stream of reactants, producing a stagnation point flow condition. An upstream
spark plug ignites the jet and a bow flame front stabilizes in the jet region.
By consuming the mainstream reactants, the opposed jet acts as an aerodynamic
flameholder, providing a preheated fuel-air mixture to the catalyst located
downstream. When the desired ignition temperature is achieved, the jet fuel
and air are turned off, thereby extinguishing the flame. The jet is shown in
Figure 5 during operation. The opposed jet igniter was evaluated using both
gaseous and prevaporized liquid fuel in the mainstream. Test results are dis-
cussed below.
Operation with Gaseous Fuel
Optimum performance of the jet was identified initially without a cata-
lyst by varying mainstream and jet stoichiometry. Catalyst lightoff capability
was then evaluated with a catalyst. Although tests were conducted mainly with
natural gas in the jet and mainstream, similar results are achievable for
other gases with similar heating values (e.g., propane). Opposed jet stability
depended primarily on the relative magnitude of gas momentum in the mainstream
duct and the jet. Based on visual observation, the most stable jet operating
condition occurred at a jet-to-mainstream velocity ratio of 7.4 and a jet-to-
mainstream mass flowrate ratio of 0.01 to 0.02. Preheat temperatures delivered
by the opposed jet show small variations with jet stoichiometry and large vari-
ations with mainstream stoichiometry. At a nominal 105 MJ/hr (100,000 Btu/hr)
heat release rate with natural gas under lean mainstream conditions, the jet
operated at stoichiometries between 5.6 and 17.0 percent theoretical air.
Exit temperature varied from 28K to 56K (50 to 100°F) with changes in jet
stoichiometry. Smaller temperature variations occurred at decreasing main-
stream stoichiometries.
11
-------
The jet operated effectively over a wide range of mainstream stoichiome-
tries; specifically, 180 to 270 percent theoretical air. This yielded exit
temperatures of 1089K to 617K (1500 to 650°F). This temperature range appears
suitable for ignition of both noble metal and metal oxide catalysts. Figure 6
plots exit temperature as a function of mainstream stoichiometry for 10.9 per-
cent theoretical air in the jet. The opposed jet demonstrated excellent cata-
lyst lightoff capability under lean conditions with natural gas. A comparison
of an opposed jet lightoff sequence with a typical lightoff using preheated air
and fuel is given in Figure 7. As shown, the preheated air technique with a
noble metal catalyst required a bed temperature near 750K (900 F) for startup
under typical flowrate conditions with the lightoff rate controlled by the
stoichiometry of the reactant stream (two are shown). In contrast, the opposed
jet can operate with a reactant stream at ambient temperature while the heating
rate is controlled by jet and mainstream stoichiometry and the thermal mass of
the particular combustion system. For a specific application, the desired
heating rate is dictated by both catalyst material and other system startup
constraints.
Additional tests were run on the opposed jet using natural gas with the
mainstream under fuel-rich conditions. Although the flammability limit of
natural gas under rich conditions is approximately 57 percent theoretical air,
the jet igniter will not start between 57 and 80 percent. Above 80 percent,
violent explosions can occur. It appears that for safe operation, the opposed
jet igniter is not appropriate for ignition under rich conditions. For staged
combustors, where the first stage operates fuel rich, ignition can be done
under lean conditions followed by shutdown of the igniter and a switch to rich
operation.
Operation with Liquid Fuel
Based on successful experience with gaseous fuels, the opposed jet
igniter was further tested with liquid fuels. The igniter was added to the
primary components of the model gas turbine; namely, the multiple spray fuel
injector and combustor can. The opposed jet was located approximately equi-
distance within the 0.46 m (18-inch) distance between the fuel injector and
the catalyst front face. A spark plug mounted on the combustor can wall
12
-------
12.7 cm (5 inches) upstream of the jet serves to ignite the jet. After the
catalyst has attained necessary bed temperature (lightoff), the jet injection
tube can be withdrawn from the fuel-air stream to avoid interferences in the
mixing region.
To simplify manifolding of gas to the jet, the jet was redesigned to
operate solely on natural gas, rather than a fuel-air mixture as was used
previously for gaseous fuels. Although this approach did not restrict perfor-
mance of the jet, a much higher jet to mainstream velocity ratio (20 to 25)
was required to maintain jet stability.
Tests were conducted using No. 2 diesel fuel at a nominal 210 MJ/hr
(200,000 Btu/hr) heat release rate. This corresponds to a duct velocity of
approximately 0.38 m/sec (15 ft/sec). Fuel was prevaporized using air pre-
heated to 561K (550°F). The jet operated over a wide range of mainstream
stoichiometries; namely, 150 percent theoretical air to no fuel in the main-
stream. Below 150 percent theoretical air, flameholding developed on the
fuel injectors upstream of the jet. As shown in Figure 8, exhaust temperature
from the jet increased linearly from 1005 to 1422K (1350 to 2100°F) with de-
creasing mainstream stoichiometry. Similar to jet performance with natural
gas, this temperature range is suitable for ignition of both noble metal and
metal oxide catalysts.
Previously, catalytic combustion with liquid fuels was accomplished by
igniting the catalyst first with gaseous fuel followed by a switch to liquid
fuel. The opposed jet igniter allows lightoff directly with liquid fuel.
During testing of the model gas turbine, lightoff with diesel fuel was success-
fully performed 15 times using an Acurex chromium oxide catalyst on a Corning
MCB-1 substrate. The lightoff sequence, from jet ignition to catalyst stabil-
ization, can be tailored to meet system requirements.
Despite difficulties under rich conditions, the opposed jet demonstrated
excellent lightoff capability under lean conditions with both natural gas and
diesel fuels. Following successful preliminary testing, the jet was incorpo-
rated into both the atmospheric and high pressure test facilities at Acurex.
During 1 year of operation, the opposed jet has become a routine ignition
technique for both catalyst screening and system component development.
13
-------
AFT-END TORCH LIGHTOFF
In addition to the opposed jet, aft-end torch lightoff was evaluated
as a catalyst ignition technique for gaseous fuels. This approach is very
simple and only requires a small port in the exhasut downstream of the bed
for insertion of the torch. In operation, the air/fuel mixture passing
through the catalyst bed is ignited downstream using a propane torch. The
torch is then withdrawn. Gas velocities are set relatively low to allow
stabilization of the flame on the catalyst aft-end. As the catalyst is pre-
heated, the flame propagates upstream into the catalyst. After the bed is
fully active, fuel and air flow can be increased to the desired heat release
rate.
The aft-end torch was tested in the atmospheric pressure facility at
Acurex using natural gas. Lightoff was easily accomplished but complete activa-
tion of the bed required a lengthy 2 to 3 minutes.
Although simple in concept, aft-end lightoff has a very limited range
of applicability. During lightoff, the gas velocity must be low to allow
flame propagation back into the bed. Thus, a particular combustion system
must be capable of low flowrates to facilitate lightoff.
Among the possible techniques for catalyst lightoff, the opposed jet is
the preferred ignition system. It can operate over wide stoichiometric blowout
limits using both gaseous and liquid fuels and has performed many successful
catalyst lightoffs in duct sizes ranging from 3.8 to 12.7 cm (1.5 to 5 inches).
CATALYST TEMPERATURE MEASUREMENT TECHNIQUES
To maintain efficient, stable combustion and not exceed its material
temperature limitations, a catalyst must be operated within a relatively
narrow range of steady state temperatures. Thus, measurement of the catalyst
surface temperature is a key parameter in system control. The instrument used
for temperature measurement must be accurate within 28 to 56K (50 to 100°F),
maintain adequate life at low cost, and respond quickly to temperature tran-
sients. Furthermore, it should be simple to install and readily accessible for
easy maintenance.
14
-------
Currently, in-depth thermocouples attached directly to the substrate
are used to provide direct measurement of the catalyst wall temperature.
However, this technique is expensive (thermocouples are not reusable), rela-
tively unreliable for extended use, and requires complex installation. Con-
sequently, thermocouple probes (which measure gas phase temperature) and
optical pyrometers were investigated as alternate techniques for catalyst
temperature measurement. Both approaches use instruments external to the
catalyst, thereby facilitating installation and maintenance. However, using
these techniques may result in a loss of accuracy. Each approach is discussed
below.
THERMOCOUPLE PROBE
The first thermocouple probe configuration tested was a platinum/
rhodium (Type R) thermocouple located 0.64 cm downstream of a graded cell
catalyst bed with the junction on the central axis of the bed. The ratio of
thermocouple probe measurement (gas phase temperature) to in-bed thermocouples
(catalyst wall temperature) is plotted as a function of inlet gas velocity in
Figure 9. At all velocities, the downstream probe indicated a lower tempera-
ture than the bed surface temperature (which runs near the adiabatic flame
temperature). The closest agreement between the downstream and bed thermo-
couple measurements occurred at an inlet bed velocity of approximately
6.1 m/sec (20 ft/sec). This velocity corresponds to the baseline heat release
rate of 105.5 x 106 J/hr (100,000 Btu/hr).
The probe is heated by radiation from the catalyst surfaces and convec-
tion with combustion exhaust gases. It's cooled by radiation to cold walls
downstream of the combustor. A balance of these heat transfer mechanisms
determines the temperature measured by the probe. At low reference velocities,
radiative cooling dominates convective heating and results in a large differ-
ence between catalyst wall temperature and gas temperature measured at the
probe. As convective heat transfer increases with velocity, this temperature
difference decreases (see Figure 9).
At high reference velocities, however, temperature differences between
the bed and probe increase with increasing velocity despite high convective
15
-------
heat transfer to the probe. It appears that combustion is incomplete in the
bed, and additional fuel conversion occurs downstream of the probe. Therefore,
the gas temperature at the probe is again suppressed below the near-adiabatic
flame temperature at the catalyst wall.
Although a curve can be constructed for the data, certain system depend-
encies make it difficult to infer bed temperature from downstream measurements.
The correlation presented is expected to depend on such parameters as:
• Catalyst type and geometry — controlling in-bed and downstream
combustion efficiencies which determine exhaust gas temperature
• System geometries — controlling radiative and convective gains
and losses from the probe
Due to the complexity of this correlation, an aft-end temperature probe is
not a preferred technique for temperature measurement.
To minimize the differences in temperature measurement, the probe was
inserted between the medium and small cell segments of a catalyst bed. The
ratio of probe to bed temperature as a function of velocity is shown as the
upper curve in Figure 9. As indicated, the correlation between gas temperature
measured by the probe and the actual surface temperature was excellent. A
negligible difference in probe and bed temperature was measured at all gas
velocities tested. For the interbed configuration, it appears that radiation
from the catalyst surfaces in the front and rear segments dominates heat trans-
fer to the probe. However, additional testing of the probe at higher veloci-
ties is required.
Although thermocouple probes are easy to fabricate and install, their
reliability is uncertain (as with in-depth thermocouples) due to direct ex-
posure to high combustion temperatures. For this reason, thermocouple probes
are not a preferred technique for long-term catalyst temperature measurement.
PYROMETRY
Although an interbed probe may be an accurate technique for catalyst
temperature measurement, it may lack reliability for continuous use since
it depends on instrumentation in a high temperature region. In contrast, a
16
-------
pyrometer can be isolated from the combustion zone, enabling longer life with-
out loss of instrument reliability. However, a pyrometer requires a clear
visual path to the target and knowledge of the surface emissivity as a func-
tion of temperature.
Typically, the front (inlet) or back (exhaust) face of the catalyst
bed is accessible to sensing by a pyrometer. In order to avoid significant
infrared absorption by various hydrocarbon fuels, viewing the catalyst exit
face may be preferred, since combustion byproducts only weakly absorb infrared
radiation in the operating range of most pyrometers. Furthermore, the higher
temperatures at the aft end of the catalyst produce shorter wavelengths which
are capable of more accurate temperature measurement.
Two types of pyrometers are commercially available for test evaluation:
(1) single color instruments which determine temperature based on absolute
intensity of emitted radiation, and (2) dual color instruments which deter-
mine temperature based on the ratio of radiation intensity at two wavelengths.
The accuracy of each instrument was evaluated by comparing the temperature
measured by the pyrometers with in-depth thermocouples (R-type) over a 1200 to
1866K (1700 to 2900°F) temperature range. Tests were conducted using a Corning
proprietary monolith (MCB-11) at high temperatures (greater than 1590K) and
a platinum catalyst on DuPont alumina Torvex at lower temperatures. Test
results are discussed below.
For the single wavelength pyrometer (Thermodot Model TD-73), temperature
measurement was accurate within 27K (50°F) at an apparent surface emissivity
of 0.7. However, with a dirty viewing window and the same catalyst, apparent
emissivity dropped to 0.45 and measured temperature was 139 to 194K (250 to
350°F) below actual bed temperature. Test data are illustrated in Figure 10.
It appears that the accuracy of single color pyrometers depends strongly on
apparent surface emissivity. As shown above, emissivity depends upon the
integrity of the sight path (affected by infrared absorption by the quartz
viewing window and by obstructions such as dirt and soot).
In addition, depending upon the material, surface emissivity may change
with temperature. For example, although surface emissivity of the Corning
substrate remained constant (0.70) with temperature, emissivity of the DuPont
17
-------
Torvex substrate decreased from 0.48 to 0.40 with decreasing temperature.
Test data for the DuPont model are shown in Figure 11. The marked drop of
apparent emissivity at 1255K (1800°F) is probably due to nonuniform surface
temperature (cold spots) resulting from unstable combustion at low temperatures.
Despite emissivity variations with temperature, true catalyst tempera-
ture can be calculated from pyrometer readings based on a known emissivity-
temperature profile (Figure 11 for example). The relationship is given by
the following expression:
XT "^
C
where TT = true surface temperature (K)
T = temperature indicated by the pyrometer (K)
X = pyrometer wavelength (cm)
C = conversion constant =1.44 (cm-K)
e, = surface emissivity at pyrometer wavelength \ and surface
temperature T
Using test data, bed temperature was predicted within 10K of actual bed tem-
perature with this formula.
By measuring the ratio of radiation signals at two adjacent wavelengths,
two-color pyrometers are designed to accurately measure temperature regardless
of obstructions in the view path or changes in emissivity with temperature.
A dual wavelength instrument was used with the Corning substrate at high
temperature. Data are given in Figure 10. In this case, temperatures mea-
sured with the pyrometer were 55K less than actual bed temperatures. Closer
agreement may be possible if the unit is calibrated for the specific material.
Over the range investigated, measurement accuracy was independent of tempera-
ture, but further testing is required at lower temperatures.
Although single wavelength pyrometers can be adversely affected by
obstructions in the sight path (soot, dirty window, etc.), temperature mea-
surement depends in a simple and predictable manner on surface emissivity.
18
-------
In contrast, two-color pyrometers are not affected by obstructions, but depend
in a complicated manner on the ratio of surface emissivity at the two operating
wavelengths. In addition, a single color instrument grade pyrometer typically
costs $1500 compared with $2500 for a two-color pyrometer. Choosing the pre-
ferred type of pyrometer depends on the particular application. For systems
with a clear sight path, a single color pyrometer is adequate. However, for
combustors which produce some soot or must operate unattended for long periods,
a two-color pyrometer may be justified. In either case, pyrometers appear to
be the preferred method of catalyst temperature measurement. They are easy to
maintain, accurate, and should provide reliable, long term operation.
CONCLUSIONS
Aerodynamic fuel injectors, characterized by fuel injection into a high
velocity duct, provide rapid and complete mixing of gaseous fuels. However,
because aerodyanmic injectors rely strictly on droplet shear forces for atomiza-
tion, they do not adequately atomize liquid fuels for the variety of conditions
encountered in catalytic combustors. Atomizing spray nozzles, in contrast, do
provide rapid and complete fuel mixing. For large systems with short mixing
lengths, multiple injection points are necessary for even distribution of fuel
and air.
The opposed jet igniter is the recommended technique for catalyst light-
off. It is a versatile system capable of operating over a wide range of main-
stream gas velocities and stoichiometries. The jet has been tested successfully
using both natural gas and No. 2 fuel diesel fuel and is capable of delivering
controlled, high temperature exhaust products to the catalyst for lightoff.
Pyrometers can accurately measure catalyst temperature (within 30K) if
surface emissivity is known as a function of temperature. Because the sensor
is external to the combustion zone, it does not suffer reliability problems due
to direct exposure to high temperatures (as do thermocouple probes). Tempera-
ture measurement by thermocouple probes located downstream of the catalyst bed
depend in a complicated manner on gas velocity and system geometry. Conse-
quently, they are not an accurate method of temperature measurement. In
contrast, at the gas velocities investigated, interbed thermocouple probes do
19
-------
accurately measure catalyst temperature. Their performance in system applica-
tions requires further development, however.
Prototype combustion systems are currently being developed to demonstrate
the feasibility of catalytic combustion. Due to the uniquesness of these com-
bustors, special techniques are required for fuel injection, catalyst lightoff,
and temperature control. Substantial progress has been made in the development
of these systems. They will be further integrated with catalytic combustor
concepts as they are scaled up to the prototype demonstration phase.
20
-------
REFERENCES
1. Rosfjord, T. J. Catalytic Combustors for Gas Turbine Engines." AIAA
paper 76-46, Washington, D.C., January 1976.
2. Tacina, Robert R. Experimental Evaluation of Fuel Preparation Systems
for an Automative Gas Turbine Catalytic Combustors. DOE/NASA/1011-78/23,
June 1977.
21
-------
TABLE I. CATALYST IGNITION CONCEPTS
Igniter
Description
Advantages
Disadvantages
Packaged
burner
K3
Opposed jet
— Off-the-shelf burner
system
— Gas/No. 2 oil
— Flame safeguard and
fuel train
— Maximum 100 percent
excess air on gas yields
1370K (adiabatic)
— Maximum 300 percent
excess air on oil yields
950K (adiabatic)
— Installed in separate
firebox, with exhaust
ducted upstream of
catalyst
— Aerodynamic flameholder
— Jet directed axially up-
stream with premixed
reactants
— Jet/mainstream velocity
ratio =10
— Stable combustion possible
between 60 percent and
200 percent theoretical
air mainstream with a
stoichiometric jet
— At 200 percent theoretical
air mainstream and with a
stoichiometric jet, maximum
exhaust temperature is 980K
— Readily available com-
plete burner system
— Very reliable
— No engineering required
on burner hardware
— Rich lightoff may be
possible due to sepa-
ration of mainstream
and igniter stream
Possible control problem
during transition between
burner throttle-down and
catalyst lightoff
Can't leave on after
catalyst is ignited (to
increase catalyst per-
formance)
Long history of use as
experimental tool
Wide stoichiometry blow-
out limits
No control problems dur-
ing catalyst lightoff;
reactants immediately
exposed to catalyst after
jet is turned off
Rich lightoff probably
not possible
Jet may have to be water-
cooled
Flameholding may occur
behind jet upstream of
catalyst
-------
TABLE I. CONCLUDED
Igniter
Description
Advantages
Disadvantages
Electrical
resistance
preheated
air
Hydrogen
injection
Hydride
device
— Preheat combustion air
by electric heater
— Max temp = 810K
Metal monolith
Aft-end torch
— H2 gas is injected
upstream of catalyst
— H£ ignites catalyst at
room temperature
— ^-impregnated structure
(i.e., gauze) upstream
of catalyst
— Electrical current re-
leases H2, acts like
H2 igniter
Metal monolith heated
when electrical current
is applied
Gas torch ignites reac-
tants downstream of bed
Flame propagates up-
stream into the bed,
igniting the catalyst
— Simple operation
— Continuously preheated
air increases catalyst
performance
Simple concept and
operation
— Fairly simple concept
and operation
— Safer than IU injection
— Simple concept and
operation
— Safe; no flame or ancil-
lary ignition equipment
needed
— Simple concept and
operation
— Inexpensive
— Limited operating tem-
perature
— Long startup time
— Large size
— Expensive to operate
— Dangerous, difficult to
handle
— Unattractive to potential
users
— Experimental stage only
— Experimental stage only
— Catalyst-specific
— Required high temperature
electrical connectors
— No preheat capability
— Difficult to achieve
at high velocity
-------
3 gas
•injection
tubes
3 distillate
injection
tubes
Bolt
holes
3.81 cm diameter
Figure 1. Aerodynamic fuel injector.
-------
en
Figure 2. Multiple venturi fuel injector assembly.
-------
Nozzle
coolant
housing
Gaseous
fuel/air
ports,
4 places
Nozzle
assembly
Packing
ram
Expansion
insert
Mixing
section
Figure 3. Atmospheric test facility fuel injection system.
26
-------
NJ
Propane
injection
tube
3 pi.
Spray nozzle
ft
T^II
Injector
Nozzle
support
3 pi
Air
Lf
Figure 4. Multiple spray fuel injector.
-------
Flame front
-
Mainstream
reactants
Opposed jet
tube
-------
1000
900
OJ
C,'
800
700
600
1500
1400
1300
1200
HOC
1000
900
300
700
600
500
Preheat temperature = 478K (400;F)
Jet stoichiometry = 10.9 percent T.A.
_L
100 120 140 160 180 200 220 240 260 280 300
Main stream stoichiometry (percent theoretical air)
Figure 6. Opposed jet igniter — exit temperature versus mainstream stoichiometry.
-------
2500
1500
2000
1500
OJ
o
1000
500
1000
500
260% TA
Typical catalyst
lightoffs with
preheated air
305% TA
Jet! on
Time (min)
Figure 7. Lightoff time comparison.
-------
2200
2000
1800
Q.
1600
1400
1200
1000
3600
3400
3200
3000
— 2800
2600
~?2400
2200
2000
1800
1600
1400
1200
1000
Jet could
not be
extinguished
I
Adiabatic
flame
temperature
Face velocity
O 10.5 ft/sec
A 13.4 ft/sec
D 15.5 ft/sec
100 200 300
Mixture stoichlometry (percent theoretical air)
400
No flame
stabilizied
Figure 8. Gas turbine opposed jet results.
-------
0.9
O)
XI
o
i.
0.
0.8
0.7
O
Thermocouple
between
segments
0
,0 (3D
Downstream
thermocouple
10
40
(fps)
60
10
Reference velocity (m/sec)
15
Figure 9. Thermocouple probe comparison.
-------
2200
2000
o>
3
1C
OJ
O.
T3
O)
1.
1800
1600
i.
OJ
1400
1200
r 3800
3400
3000
2600
2200
1800
"• HOO
O Single X, c = 0.7
Two X, uncalibrated
Single A, E = 0.45
1800
2200
2600
3000
1200 1400 1600 1800 2000
Actual bed temperature (K)
Figure 10. Pyrometer temperature measurement comparison.
33
-------
0.50
0.49
0.48
0.47
+? 0.46
to
I/I
E 0.45
Ol
^ 0.44
0.43
0.42
0.41
0.40 -
1400
1500
1600
1700
1800
1900
2000
2100
2200 2300
1100
1200 1300
Actual bed temperature
1400
1500
Figure 11. Torvex emissivity changes with temperature (Pt catalyst).
-------
PRELIMINARY DESIGN OF
CATALYTIC COMBUSTORS
By:
William B. Retallick
Consultant
West Chester, PA 19380
ABSTRACT
At high temperatures the rate of surface combustion can be so fast
that the observed rate is simply the rate of mass transfer to the surface.
When the flow is turbulent, mass transfer is analogous to momentum trans-
fer, and the rate of momentum transfer is proportional to the pressure
drop. We can measure the pressure drop per length of honeycomb, and use
this measurement to calculate the length of honeycomb to give any desired
conversion.
In laminar flow, mass transfer occurs by diffusion alone, and the
diffusional process is not coupled to the pressure drop. Here, however,
we can calculate the length of honeycomb for the desired conversion di-
rectly from the velocity, the diameter of the channel, and the properties
of the fluid.
These calculations are quick and simple, but are accurate enough for
preliminary design.
35
-------
SYMBOLS
Re Reynolds number
Pr Prandtl number
Sc Schmidt number
f Friction factor
K Mass transfer coefficient, Length
Time
h Heat transfer coefficient, Heat
(Area)(Time)(Temperature)
KH Heat conductivity, Heat
(Length)(Time)(Temperature)
r Density, Mass
Volume
Cp Heat capacity, Heat
(Mass)(Temperature)
d Effective diameter of a channel, 4 (Sectional Area) = 4^
Wetted Perimeter a
v Velocity, Length
Time
A pf Pressure drop from fluid friction, Force
Area
gc Gravitational constant, (Mass)(Length)
(Force)(Time)2
T Temperature
D Diffusion coefficient (Length)2
Time
36
-------
THE TRANSFER UNIT
For both turbulent and laminar flow, the measure of the rate of trans-
fer is the length of the transfer unit. We will derive the length of the
transfer unit from the differential equation that describes the transfer
process. For mass transfer, this is:
d£ = - aKC
dL v
which leads to In C = - aKL
cinlet
when the length of reactor equals v
' Cinlet
This shows what one transfer unit does, and thereby defines the length of
one transfer unit. We can write similar equations for heat transfer, and
then summarize the results for both mass and heat transfer as :
nraass ~ —
Ka
These definitions are general, and hold for both turbulent and laminar
flow. It is not convenient to use these equations directly, because that
would require us to calculate the transfer coefficients h and K. For tur-
bulent flow, we will bypass the coefficients by using instead the friction
factor, which we can measure experimentally. For laminar flow, the co-
efficients are buried in the Nusselt number, which we calculate from a
mathematical model.
TURBULENT FLOW
In fast turbulent flow, when all of the transfer results from
37
-------
turbulent motion, and diffusional resistance is suppressed, the Reynolds
analogy applies.
In words, what the Reynolds analogy says is:
rate at which mass is transferred to the solid surface
total rate at which the component, in excess of the
inter facial concentration, flows past the surface
_ rate of heat transfer to the solid surface _
total rate at which heat, measured above the tempera-
ture of the surface, flows past the surface
_ rate of momentum transfer due to friction _
total momentum of the stream which flows past the
surface
Translated into symbols, the analogy becomes:
dC = dT = _ £Pf 2gc
AT v2
Introducing the transfer coefficients for mass and heat converts the last
equation into
v_ = *vCp = 2_
kmass ~ Tieat
which shows that the two transfer units have equal lengths when (and only
when) the Reynolds analogy applies. In the usual case where there is some
diffusional resistance, this resistance is accounted for by Colburn's modi-
fication of the Reynolds analogy, which leads to
y_(Pr)~2/3 = ^vCp (Sc)~2/3 = 2_
Ka ^~ha~ fa
or
= 2_(Sc) 2/3
fa
_ 2 (Pr) ^/3
fa
Now we will use our pressure drop data to obtain the friction factor
f and then Lmass andLheat- Ttie roost convenient way to get the data would
be to blow ambient air through the honeycomb, and to plot the pressure
drop per length against the velocity. For our present purpose, these data
are converted to a plot of friction factor vs the Reynolds number in the
38
-------
channels of the honeycomb. For channels having an irregular cross section,
the effective diameter d is
d = 4 (Sectional Area) = 4_
Wetted Perimeter a
Calculate the Reynolds number at the velocity and temperature expected
in the combustor, enter the plot with this Reynolds number, and pick off
the friction factor. Use this value of f to calculate Lj^gg and Lneat ^n
the combustor. Now we can calculate the length of honeycomb needed to give
any desired conversion, since C^n = e~N where N is the number of transfer
units . Cout
Another quantity that is useful for preliminary design is the number
of velocity heads lost to friction in the length of one transfer unit. The
velocity head is defined as v2 and it has the dimensions of energy per unit
2gc
mass of fluid. Rearranging the friction factor equation gives the number of
velocity heads used up in any length L
p 2 = 4fL = afL
d
substituting for L the values we just calculated for Lmass and Lheat gives
number of velocity heads = 2 (Sc) 2/3
per mass transfer unit
number of velocity heads = 2 (Pr) 2/3
per heat transfer unit
What makes these numbers so useful is that they are independent of the
velocity and of the diameter of the channel. We can estimate the number of
velocity heads needed to give any desired conversion, without measuring the
pressure drop.
LAMINAR FLOW
In laminar flow it is possible to calculate the lengths of the trans-
fer units directly from the velocity, the diameter of the channel, and the
properties of the fluid, without measuring the pressure drop. This is so
because the transfer occurs by diffusion alone, and the diffusional process
is not coupled to the pressure drop.
As a model for the calculations, assume laminar flow of a fluid in a
39
-------
circular channel with heat transfer to or from the wall. We define the
heat transfer coefficient from the relationship
KH Jr at the wall
Twall - Tbulk
From the model we can show that the Nusselt number = hd = 4.4. Likewise,
KH
for mass transfer, Kd = 4.4. The derivation is given by Holman, "Heat
D
Transfer," McGraw-Hill, 1976.
Now we have sufficient information to calculate Lmass and Lheat• F°r
circular channels, a = 4_, so that
d
Lmass = Y_ = vd ^ 1 vd2
Ka 4K 18 D
and
Lheat = v/c?Cp = v^?Cpd /x, 1 v x?Cpd2
'ha ~T4h18 'KH
These values of Lmass and Lheat were derived for circular channels, and we
need to estimate the error from applying them to channels of other shapes.
The Nusselt number is not very sensitive to the shape of the cross section;
it is 4.4 for a circle of diameter d and 4.1 for parallel plates at spacing
d. The difference is small, considering the great difference in shape.
We calculate the number of velocity heads lost to friction in the
length of one transfer unit, just as we did for turbulent flow:
number of velocity heads = 4f vd
per mass transfer unit 18 D
4_f (Re) (Sc)
18
number of velocity heads = 4f_ v r> Cpd
per heat transfer unit 18 ' KH
4_f (Re) (Pr)
18
Note that these expressions can be written so that they contain the prod-
uct of the friction factor and the Reynolds number. For laminar flow in
circular channels, this product is equal to 16, and the constant 16 comes
from the Poiseulle equation. For parallel plates the product is 24, when
the effective diameter for calculating both f and Re is the same as for
40
-------
the circle:
4 (Sectional Area) = d
Wetted Perimeter
The constants for other shapes should fall between 16 and 24.
The results of all of these calculations are summarized in Table I.
-------
TABLE I. SUMMARY OF THE CALCULATIONS
NJ
WHAT IS CONCENTRATION
BEING PER VOLUME
TRANSFERRED OF FLUID
Mass Weight or Mols
Volume
Heat /°cp T
Mass Above
Heat Above
1 Effective diameter d is 4
LAMINAR
DIMENSIONS OR
TRANSFER OF THE TURBULENT
COEFFICIENT COEFFICIENT FLOW
K Length Turbulent
Time
h Heat Turbulent
(Area) (Time) (Temp)
K Above Laminar
h Above Laminar
(Cross Sectional Area) = 4
LENGTH OF
ONE
TRANSFER
UNIT
d1 Sc 2/3
2f
d Pr 2/3
2f
1 vd2
18 D
1_ v/.Cpd2
18 Tfc
VELOCITY HEADS
USED UP IN
ONE TRANSFER
UNIT
2 Sc 2/3
2 Pr 2/3
4f Re Sc
18
4f Re Pr
18
Wetted Perimeter a
-------
CATALYTIC FLAME STABILIZATION
By:
Moshe Lavid and A. E. Cerkanowicz
Exxon Research and Engineering Company
Corporate Research-Technology Feasibility Center
P. 0. Box 45
Linden, New Jersey 07036
ABSTRACT
The Catalytic Flame Stabilization process is applied to aeropropulsion
afterburners combustion. For this application, a portion of the inlet fuel-
air stream passes through the porous flameholder wherein heterogeneous
reactions are promoted by catalytically coated surfaces. The partially
reacted mixture having higher temperatures, then passes into the aft region.
The other portion of the approaching fuel-air mixture flows around the
porous flameholder merging downstream with the previous stream. As a result
of the supplemental heat release due to the catalytic reactions, higher
wake temperature and improved combustion stability are obtained with
concurrent reduction in pressure drop.
In this concept the key role of the catalyst is to "bootstrap" the
conditions of temperature and concentration of partially reacted species
to levels favorable to stable efficient gas phase combustion. Consequently,
the catalytic unit need only process a portion of the reactant mixture and
need only provide for the initial stimulus.
This paper outlines the catalytic flame stabilization program undertaken
by Exxon. It is a comprehensive experimental research directed and guided
by a complementary analytical modeling effort.
43
-------
ACKNOWLEDGMENT
This work is supported in part by the Air Force Office of Scientific
Research, under Contract No. F49620-77-C-0085. Dr. B. T. Wolfson is the
Contract Monitor.
44
-------
INTRODUCTION
Combustion systems are limited by combustion associated phenomena such
as flammability, flame propagation, ignition and stable combustion, and by
the formation of combustion-related pollutants. Recent research has identi-
fied catalytic combustion as a potentially promising technique for recti-
fying these limitations. The technique involves the utilization of hetero-
geneous catalysis to enhance combustion processes and to broaden normally
encountered stability limits, particularly at fuel lean and low temperature
conditions.
In general, catalytic combustion is a concept wherein chemical reactions
initiated by a heterogeneous catalyst play an important role in the energy
release processes of a reacting fuel-air system. The effect of the solid
catalyst is predominant only during the early phases of the overall combus-
tion process. As fuel is oxidized and energy is released, the importance
of homogeneous reactions increases and eventually they become controlling,
and the balance of energy release is achieved through them.
A number of general potential applications have been considered.
Catalytic gas-turbine combustors can achieve stable combustion at lower
temperatures than that required to insure stable combustion in conventional
combustors. The lower temperatures will result in a substantial reduction
in NOX emissions (1) from the unacceptable levels produced in the primary
zone of the conventional combustors due to the high temperatures. Fuel-
air mixture that cannot be combusted efficiently and stably by ordinary
homogeneous means such as various types of low heating value fuels, can
be utilized by catalytically augmenting the combustion, (2). Catalytic
converters represent another commonly used application in the automotive
industry. In this application, the heterogeneous reactions dominate while
the gas phase reactions are of little importance (3,4). In some other
application, catalytic combustion is used as a means for flame stabilization.
45
-------
Catalytic flame stabilization differs from catalytic combustion in that the
heterogeneous reactions are only needed to promote the combustion process
to a point where normal gas phase reactions take over and become dominant.
This catalytic flame stabilization concept is particularly important to
aeropropulsion combustion; turbopropulsion mainburners, afterburners, duct
burners, and ramjet dump combustors.
This paper outlines the catalytic flame stabilization program currently
underway at the Corporate Laboratories of Exxon Research and Engineering
Company in Linden, New Jersey.
46
-------
PROGRAM OBJECTIVES
The objectives of the program are to develop a fundamental understanding
of the aerothermochemistry of catalytic flame stabilization and to establish
design aids to be used in subsequent exploratory development efforts.
Fundamental information includes: (1) characterization of the flow field
around a catalytic stabilizer, (2) determination of the role of the catalyst
in stabilizing the combustion process, (3) establishment of catalyst and
monolith material choices, (4) characterization of downstream combustion
zone processes including wake circulation, turbulent shear layer combustion
and flame propagation, (5) quantification of stability limits, and (6) de-
velopment of an analytical model of the catalytic flame stabilization process.
Design aids include correlations and models of the catalytic flame stabiliza-
tion process as well as guidelines for application related to fuel types,
fuel impingement, and gradients of fuel-air ratio and temperatures in the
approach flow.
The catalytic flame stabilizer is conceptually illustrated in Figure 1.
The conventional solid bluff-body flameholder, which stabilizes the flame
by producing a wake region behind it, is replaced by a porous device with
catalytically coated surfaces along its passages. The porous design permits
flow through, as well as around, the device resulting in less pressure drop
than a solid flameholder of equal cross-section area. However, the flow
through the porous catalytic unit has some adverse effect on the flame
stability. It tends to wash out the recirculating region which is needed
to provide flow residence times of the same order as the characteristic
chemical reaction times, and thus reduces flame stability. This difficulty
is alleviated by coating the surface of the passages with solid catalysts.
The key role of the catalyst is to "bootstrap" the conditions of temperature
and species concentration to levels favorable to stable, efficient gas phase
combustion. Thus, the catalyst offsets the problem associated with the
reduction in flow residence time, due to the flow stream through the
catalytic porous unit.
47
-------
PROGRAM APPROACH
The catalytic flame stabilization program encompasses experimental
work as well as analytical modeling effort.
The experimental portion of the program involves a detailed study
of the region about various flame stabilization devices. Figure 2
depicts some representative samples of the devices to be investigated.
Initially, a standard bluff-body flameholder (V-gutter) and a non-catalytic
monolith will be tested to establish an experimental baseline (A and B
in Figure 2). Subsequent test will employ catalytic monoliths or specially
fabricated flameholders which simulate catalytic stabilizing device
operation (C and D in Figure 2). In this latter case, a flameholder with
its downstream side sealed with a sintered metal plate will be used. It
is a simulated unit which has the bluff-body appearance of a catalytic
unit but also one in which we can vary the type of species and its concen-
trations. In other words, it is a diagnostic device capable of simulating
different modes of catalytic stabilization operation, by introducing into
the wake region various gas compositions at different temperatures, repre-
sentative of initial as well as partially oxidized fuels. This simulated
catalytic flameholder provides us with the capability of evaluating the
important combustion mechanisms in the catalytic flameholder under
well monitored and reproducible conditions independent of actual hetero-
geneous reactions. The experimental variables include fuel type, inlet
temperature and velocity.
The experimental work is also aimed at the examination of broad
variations in catalyst and monolith types. Catalysts will include from
various noble metals (especially platinum and palladium) either singly
or in combination with transition metal oxides (oxides of nickel, chromium,
etc). Substrates of different blockage ratios, configuration (triangle,
rectangular, circular, etc), material composition (ceramic or metal),
48
-------
and overall shape (triangular, disc, etc) are considered for evaluation.
Experimental direction and guidance will be provided by a complementary
analytical modeling effort described later in this paper.
EXPERIMENTAL APPARATUS
A combustion tunnel is currently under construction in Exxon Corporate
Laboratories. It is a very versatile apparatus designed to conduct combustion
experiments under continuous flow conditions. A diagram of the combustion
tunnel is shown in Figure 3. Essentially, it comprises of a horizontal
test section; 5" square and/or 6" cylindrical, equipped with transparent
side windows to allow visual observations and optical diagnostics. Regulated
and metered air is provided by a blower with a capacity of up to 2000 SCFM.
The air flows through a precombustor where it can be vitiated up to 1200°K,
and enters a settling chamber, housing a number of flow smoothing grids.
Then, through a transition piece, it flows into a mixing chamber where it
mixes with gaseous or liquid fuels. Before entering the test section^ it
passes through turbulence grids. The test section is designed to conduct
catalytic as well as radiative combustion augmented experiments. The
catalytic flame stabilization experiment is depicted in the first blow out.
The premixed fuel/air mixture is ^plit into two streams, one stream bypasses
the catalytic flameholder while the other one flows through it. The former
stream provides the stabilizing wake region, and the latter one improves
combustor stabilization due to the catalytically induced heat release, which
increases the heat content of the wake region. Concurrently, this flow con-
figuration reduces the pressure drop in the coirbustor.
The following diagnostics will be emphasized in the combustion tunnel:
LDV - upstream and downstream point velocity measurements
Schlieren - nonreactive/reactive flow patterns
Sampling probes/Gas analyzer - species concentrations
Thermocouples - temperatures
Pitot probes - pressure
In summary, the capabilities of the test facility are as follows:
• Continuous flow (not blow down)
49
-------
9 Cross section: 12.5 cm square, 15.0 cm round
• Velocity: 0-200 m/s
• Temperature: 300-1000°K (vitiated)
• Pressure: 101 kPa
• Fuel: liquid (Jet A, JP4) or gaseous (propane)
• Diagnostics: LDV, Schlieren, species, temperature, and pressure
CATALYTIC FLAMEHOLDER MODELING
The catalytic stabilization process can be analytically analyzed by
incorporating Exxon catalytic combustion model (5) with a bluff body flame
stabilization theory.
The flow pattern of the porous catalytic flameholder is depicted in
Figure 4. It is a simple engineering model aimed to elucidate the funda-
mental physical and chemical processes, and it is similar to the model
proposal by Mullins (6) for solid bluff body. In the model, the inlet
reactant stream mixes behind the bluff body with a partially combusted
gas stream at higher temperature, resulting in an increase in temperature
of the mixture. Part of this mixture enters into the recirculation region,
which causes an increase in residence time. During this period of time,
more fuel is combusted resulting in additional increase in temperature until
self-sustained combustion prevails. However, the catalytic porous flame-
holder has an additional flow through the catalytic passages which repre-
sents a variation from the solid bluff body model. This flow experiences
a temperature increase due to the catalytically induced reactions before
merging with the gas recirculation stream. As a result of this supplemental
heat input, higher wake temperatures and improved combustion stability are
achieved, with concurrent reduction in pressure drop. Alternatively, an
increase in area of the porous flameholder at constant pressure drop would
result in shorter overall combustor lengths.
The catalytic portion of the model utilizes Exxon catalytic combustion
model. The assumptions underlying the model are discussed in Reference 5.
It essentially employs one dimensional, steady-state, plug flow with global
reaction rates in each phase; heterogeneous and homogeneous. Some of the
50
-------
basic physical and chemical processes which are accounted for by the model
are illustrated in Figure 5. They include: (1) mass transport of fuel and
oxidant from the main stream to the surface as the flow proceeds through
the channel, (2) heat release on the surface caused by a global hetero-
geneous reaction, (3) heat transport by forced convection into the gas
phase, (4) axial heat conduction in the substrate, and (5) global homogeneous
reaction in the channel.
The model was used to simulate experimental results obtained by
Pfefferle et al (7). Figure 6 shows the effects of varying inlet gas
temperature on fuel conversion for operating conditions given in the legend.
The experimental data points are indicated by solid circles, while the solid
and dashed curves were obtained from computer solutions of the analytical
model.
Examining the results exhibited in Figure 6 three main features can be
identified:
1. Three operating regions of the catalytic system exist. Prior to "light-
off" (temperature less than 500 K) the system is limited by the slow
surface reaction rate. At intermediate temperatures (500-800 K) surface
reaction is fast but fuel conversion levels off at about 70% due to
diffusion limitations. The third zone (temperature greater than 800 K)
involves complete fuel conversion due to the contribution of gas phase
reactions which are initiated ("bootstrapped") by the surface reactions.
2. If no catalyst was present the inlet temperature would have to be
increased to 1050 K before appreciable fuel conversion occurs. With the
catalyst present, appreciable fuel conversion occurs at temperatures as
low as 500 K.
3. Our capability of modeling the complex behavior of the catalytic device
is illustrated by the good agreement between experiment and modeling.
Figure 7 illustrates the fraction of fuel conversion attributed to
catalytic and gas phase reactions as a function of reactor length or the
reaction temperature. It clearly demonstrates that the primary function of
the heterogeneous reactions is to "bootstrap" the mixture to conditions such
that homogeneous reactions become predominant. For this particular system,
51
-------
the catalytic reactor could probably be shortened by as much as 30%, allowing
the gas phase reactions to continue independent of any catalytic surface.
Typical results for catalytic performance at conditions similar to that
encountered in afterburners are presented in Figure {}• Increased fuel
conversion is observed at lower combustor velocities due to increased gas
residence time. As the velocity increases larger catalytic units are
required in order to obtain full conversion. The initial conversions
•observed at short catalytic length are due entirely to surface reactions.
At approximately 25% conversion, there is an. inflection point in the curve,
indicating rapid increase in conversion due to simultaneous occurrence of
gas phase reactions within the catalytic unitr
The analytical model of the catalytic flandholder will provide experi-
mental direction and guidance. It will assist in understanding the inter-
action of the various encountered processes. It will provide a more
detailed understanding of the experimental data, and will help predict
trends in combustor performances as a function of operating characteristics.
SUMMARY
To summarize, the major advantages of catalytic flameholder are:
1) it can perhaps act as a passive autoignition device
2) it can extend the stability range of an afterburner by allowing for
operation at inlet velocities and temperatures where conventional
flameholders begin to fail
3) it has the potential of reducing the length required to spread the flame
throughout the afterburner cross section by utilizing a porous bluff body
larger than a solid one for the same pressure drop
4) alternatively, it has the potential to reduce the pressure drop for the
same flame spreading rate.
52
-------
REFERENCES
1. Blazowski, W. S., and Walsh, D. E., "Catalytic Combustion: An Important
Consideration for Future Applications"; Combustion Science and Technology,
Vol. 10, pp. 233-244, 1975.
2. Pfefferle, W. C., "The Catalytic Combustor: An Approach to Cleaner
Combustion," Journal of Energy, Vol. 2, No. 3, pp. 142-146, 1978.
3. Young, L. C., and Finlayson, B. A. "Mathematical Models of the Monolith
Catalytic Converter: Part II. Application to Automobile Exhaust,"
AIChE Journal, Vol. 22, No. 2, pp. 343-353, Mar., 1976.
4. Hawthorne, R. D., "Afterburner Catalysts - Effects of Heat and Mass
Transfer Between Gas and Catalyst Surface," AIChE Symposium Series,
No. 137, Vol 70, pp. 428-438, 1974.
5. Cerkanowicz, A. E., and Stevens, J. G., "Case Studies in Simulation of
Novel Combustion Techniques", proceedings of the 1979 Summer Simulation
Conference, July 10-18, Toronto, Canada.
6. Mullins, B. P., "A Spontaneous Ignition Theory of Combustion Intensity
and Combustion Stability Behind a Baffle", Combustion Researches and
Reviews, pp. 58-71, Butterworths, 1955.
7. Pfefferle, W. C., et.al., "Catathermal Combustion: A New Process for
Low-Emissions Fuel Conversion," ASME Paper No. 75-WA/Fu-l.
53
-------
APPROACH CATALYTIC MIXING AND FLAME COMBUSTION
FUEL-AIR SECTION STABILIZATION PRODUCTS
MIXTURE SECTJON
Figure 1. Catalytic flame-stabilization concept.
-------
A. Standard V-Gutter
B, Non-Catalytic Monolith
Ol
C. Catalytic Monolith
D, Simulated Catalytic Stabilizer
Figure 2. Flame Stabilizing Devices.
-------
AIR
\f
BLOWER
^2000 SCFM
FLOW METER!
PRE-
COMBUSTOR
T.C. &
01 PRESSURE TAP
VITIATED
AIR
•*-
RADIATIVE
FLOW
CONTROL
-FUEL
•IGNITER
WEDG
CATALYTIC —I
CATALYTIC'
FLAME HOLDER
IGNITER
FUEL
SETTLING
CHAMBER
FLAME ARRESTER
MIXING CHAMBER
WINDOW
T.C. &
PRESSURE
TAP
EXHAUST
IGNITER
WATER/N?
INJECTION
SAMPLING PROBE
T.C. &
PRESSURE TAP
SCHLIEREN SYSTEM
TURBULENCE GRIDS
Figure 3. Combustion tunnel diagram.
TEST SECTION
DIAGNOSTICS
LDV
SCHLIEREN
SPECIES
TEMPERATURE
PRESSURE
-------
Catalytic
Bluff-Body
- W
--- T3 " V Vv w:
(HO) v"i
(Hi) ^^
•f " ttripcraturc
V » velocity
W » nass flov rate
(M) » ni>:in?, zone
(II) " heat release zone
Figure 4. Model for catalytic stabilization theory.
-------
Ul
00
ceramic or metallic substrate
mass transport
of fuel and 0?
to surface
fuel-air mixture
flowing through channel
- typically 0.16 to
0.32 cm 1n diameter
/ v
heat transport
between surface
, and gas
homogeneous combustion
may occur 1n gas phase
S
heterogeneous combustion
-- heat liberated at surface
catalytic
surface
,
axial heat conduction
In the substrate
Figure 5. Chemical and physical processes
modeled in catalytic combustion.
-------
CATALYTIC COMBUSTION
Ui
UJ
o
-I
UJ
o
u.
1.0
0.8
0.6
0.4
PRESSURE: 101 kPa
FUEL TO AIR
RATIO (wl): 0.018
FUEL: PROPENE
REFERENCE VELOCITY: 18.7m/s
Experimental Data
\
Catalytic awl
Homonsneous
Catalytic
Contribution Only
f
Homogeneous /
Contribution Only /
300 400 500
600
i
1
700 800 900 1000 1100
INLET TEMPERATURE (K)
Figure 6. Experimental data (o) and model-generated curves showing
the effect of inlet temperature on fuel conversion.
-------
PRESSURE: 101 kTa
INLET TEMPERATURE: 662 K
FUEL: PROPENE
REFERENCE VELOCITY: 18.3 w/s
FUEL/AIR RATIO: 0.0258 kg/kg
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
DIMENSIONLESS REACTOR LENGTH
700 900 1100 1200 1300 1400 J600
CORRESPONDING GAS FH4SE TEMPERATURE
Figure 7. Change in mechanism of fuel conversion
during catalytic combustion.
60
-------
Fuel: C-H, at T. . „ - 978 K, P. , - 101.3 kTa
3 6 inlet inlet
Catalyst: Platinum
Substrate: Channel Diameter - 0.32 era, Open Area = 0.93,
Thermal Conductivity • 16.76 W/mK
1.0
Citalyst Length (cm)
Figure 8. Typical catalytic reactor performance
at afterburner conditions.
61
-------
THERMAL SHOCK RESISTANT CATALYTIC MONOLITHS USING FOAM CERAMICS TECHNOLOGY
BY
William C. Pfefferle
Consultant
51 Woodland Drive
Middletown, New Jersey 07748
ABSTRACT
A review of the design considerations for thp use of ceramic materials
has shown that thermal shock tolerant catalytic combustor designs are feasible,
Several .such designs are proposed. Foam ceramics technology is most useful in
producing thermal shock resistant monolith catalysts based on such designs.
62
-------
INTRODUCTION
As is well documented, the high efficiency of a catalytic combustor as
compared to catalytic converters is the result of gas phase homogeneous
reactions induced by the catalyst. Thus, for efficient combustor performance,
monolith combustor catalysts must be able to operate at a temperature high
enough to promote thermal reactions, i.e. a temperature above 1800-2QOOF. To
achieve a reasonable turn-down ratio, operable catalyst temperature must be
even higher, for example 2500F or higher. For such high use temperatures,
ceramics are obvious materials of choice. Although certain metals can be
used in this temperature range, the high cost of the exotic materials required
is a serious drawback. The use of ceramics, on the other hand, imposes strin-
gent limitations on combustor design, and thus on catalyst configuration, because
of the need to limit thermal tensile stresses to relatively low values. There
would be no problem of course if stable zero expansion ceramics were available.
With available ceramics, the configurations required to minimize thermal shock
stresses are difficult or impossible to make using conventional monolith man-
ufacturing processes. Fortunately, ceramic sponge technology makes possible
catalytic monoliths of almost any desired configuration.
63
-------
CERAMIC MONOLITH DESIGN CONSIDERATIONS
Before discussing the advantages of sponge ceramics, it is well to review
the pertinent design considerations governing the use of ceramic materials. As
is well known, thermal shock failure is a major factor limiting the usefulness
of ceramic materials. For high temperature applications such as catalytic com-
bustor catalysts, it is often relatively easy to find materials which possess
satisfactory structural properties at relatively high use temperatures. Never-
theless, failures can occur at much lower temperatures during heating or cooling.
The explosive shattering of hot ceramic ovenware placed on a cold surface is a
common experience. A primary reason for this is that unlike metals, ceramics
possess almost no ductility except at very high temperatures and typically are
low in thermal conductivity and tensile strength. Only beryllia and silicon
carbide are comparable to metals in thermal conductivity. No polycrystalline
ceramic comes closer than about 25% of the tensile strength of readily available
metal alloys at temperatures below 2000F. Cermets, although far better in thermal
shock resistance than most ceramics, are subject to oxidation of the metallic com-
ponent and thus limited to lower use temperatures than available high temperature
ceramics.
The key to the successful use of ceramics, as with any material, is proper
design. Although ceramics are more susceptible to thermal shock damage than
most metals, even a metal structure can be readily destroyed by thermal shock.
Thus, care must be exercised in bringing large metal wheel gas turbines and metal
tube furnaces up to operating temperature. For ceramic combustor catalysts, good
design can permit rapid thermal cycling of even high expansion ceramics.
It should first be noted that if a structure is homogeneous, isotropic and
free of restraints against movement, then no stresses result from thermal expan-
sion. However, very large stresses can develop if the structure is kept from
expanding (or contracting) by internal or external restraints. Thus, a solid,
homogeneous ceramic rod will not suffer damage no matter how great the axial
temperature gradient if it is held at one end and heated or cooled only at the
opposite end. On the other hand, the same rod uniformly cooled at the sides will
rupture with rather modest radial gradients. If the surface is chilled rapidly
so that temperature equilibrium with the interior is not maintained, then the
64
-------
surface will be prevented from contracting by the hotter interior substrate.
Compressive stress in the interior balances tensile stress in the surface layer.
For elastic materials such as ceramics, the surface stress can be calculated
using the following equation;
& = E « a-AT/U -^0 Eq. 1
where,
E = modulus of elasticity,
a = coefficient of linear expansion,
X = Difference between surface temperature and
the mean temperature of the rod,
AA «= Poisson's Ratio,
(£ - Tensile stress at surface
For a typical polycrystalline alumina, E = 55x10 psi, a = 7x10 / C,
andytx = 0.25. Thus, if AT = 100°C, co"= SlOOOpsi compared to a tensile
streneth of only about 35,000psi at 300°C. With surface flaws, the effective
tensile strength could be lower than 10,000psi with failure resulting from a
temperature difference of as little as 10 or 20 C.
For many honeycomb catalysts, cell walls are thin enough so that temperature
o
differences between the surface and the interior are much less than 10 C under
most flow conditions. Thus, an alumina honeycomb with 10 rail walls would suffer
temperature differnces of no more than about 3°C within the cell walls for a step
change in air temperature of 500°C at a flow rate of twenty pounds per second per
square foot (i.e. at an air to surface heat transfer coefficient of 0.01 cal/sec/cn /
and an alumina thermal conductivity of 0.02 cal/sec/cm/°C). The maximum tempera-
ture difference would be less at lower flow rates or with thinner cell walls. For
a low thermal conductivity material such as zirconia, the temperature gradient
could be as high as 20°C. With sufficiently low gas flow rates, temperature
differences can be maintained within safe limits for any monolith substrate. A
graded cell catalyst system reduces temperature differences for a given flow rate.
Just as solid ceramic rods or monolith cell walls can be subjected to
damaging tensile stresses from temperature differences between the surface and
the interior, so honeycomb catalyst structures can be similarly damaged. If the
fuel air flow distribution is such that the interior flow channels are hotter or
65
-------
colder than those at the perphery, tensile stresses will result, even under
steady state flows. The maximum stress developed for a given temperature
difference increases linearly with catalyst diameter, Such stresses can be
reduced somewhat by use of a flexible cell structure such as Coming's flex-cell
honeycomb. However, the flex-walls of a flexible cell structure must be free of
cracks or other stress concentration features. The requirement for crack-free
walls is particularly hard to achieve with extrusion processes for any but the
smallest diameter extrusions.
In the axial or flow direction, a honeycomb catalyst structure can, in
principle, be designed so that there are no tensile stresses resulting from
axial temperature gradients. Avoidance of external restraints on movement in
the axial direction is quite simple. However, with a slip coated catalyst
stresses arise as a result of differences in expansion coefficients of the structurt
ceramic and the »lip coating. These stresses will cause loss of the catalyst
coating unless the coefficient of expansion of the catalyst layer is closely
matched to that of the substrate material. Note that an alumina catalyst would
be a poor choice for a coating on cordierite. However, any coating differ'!^
in composition from the support material will have a mismatch problem. Active
monoliths eliminate the need for a surface coating.
Differences in temperature profile between channels, such as from flow
differences or catalytically inactive areas, could also lead to destructive
tensile stresses. However, this should not be a major problem since the resulting
temperature differences would normally be small between adjacent channels.
Inasmuch as ceramics are brittle materials as compared to metals, it is
vital that honeycomb combustion catalysts be free of pre-existing strains,
cracks or other points of stress concentration. Thus, catalysts should be
formed to the correct size and shape rather than cut since cutting introduces
cracks. Similarly, a monolith is likely to be weakened if coated with any
material which can interact with the support structure at operating temperatures.
From the foregoing it can be seen that careful design is required if max-
imum performance is to be obtained from monolith combusotr catalysts. As
will now be shown, foam ceramics can play an important role in such designs.
66
-------
A DESIGN BASIS
From the foregoing discussion, it is evident that a ceramic monolithic
structure, such as a combustor catalyst, can withstand extreme variations in
temperature provided only that design of the structure, and that of the system
in which it is used, are such that there are essentially no restraints to move-
ment. It will now be shown that it is both possible and feasible to design such
structures and systems.
In the first place, it should be noted that external restraint should not
be a problem in view of the extensive art on the restraint-free mounting of
ceramics. However, internal restraints can be a problem without proper design.
Accordingly, the discussion will be limited to the control of internal restraint.
The appropriate restraint-free mounting is assumed.
The simplest form of a restraint-free design would be a combustor made up
of thin walled catalyst tubes placed perpendicular to the direction of flow.
If the walls are no thicker than about 10 mils there are no significant internal
restraints even for gas flow rates encountered in gas turbine combustors. Thus,
such a combustor could withstand thermal cycling without damage at least as well
as any conventional combustor. Even a thermal shock sensitive ceramic such as
stabilized zirconia could be used. Moreover, if the tubes were made of an
electrically conductive metal oxide, such as lanthanum chrome, the catalyst could,
in principle, be electrically heated to full operating temperature in milliseconds
without damage. This is because uniform heating does not lead to temperature
gradients. The only stresses are from inertial forces. Note that a conductive
coated tube would shatter under the same conditions.
A similar design is a combustor with vertically suspended U-tubes. Although
inferior, it is a good approximation to a restraint-free design. U-tube designs
are used in heat exchangers because the tube ends can be firmly secured. A
derivative of the U-tube design is the hemispherical or dished plate. Although
not completely restraint-free, a dished plate honeycomb exhibits far better thermal
shock resistance than flat honeycomb plates or cylinders, Another U-tubc desicn is
the split cylinder annular flow combustor. This type design is almost restraint-
free. Such dished shapes are readily fabricated only using foam ceramic tech»olop.y,
67
-------
FOAM CERAMICS
In any given structural configuration, foam ceramics have superior thermal
shock resistance. Accordingly, foam ceramics have long been used in applications
where this property is important. It should be noted that the thin walled spherical
cells of a foam can expand or contract more or less freely just as is the case for
a U-tube or dished plate. As a result, many foams can withstand direct flame
impingement without cracking and thus make excellent high temperature insulating
materials. Made from catalytic oxide, such foams represent almost the ideal form
for a combustor catalyst.
A major advantage of foam ceramics technology is the capability to produce
complex shapes. Although a number of companies have long had foam ceramics processes,
the process recently developed by Merano and Marcaccioli of Montedison (1) is most
interesting. In the Merano process, a polyurethane sponge may be converted, by
plasma spraying with an appropriate material, into a sponge of metal, cermet, or
ceramic. Cermet sponges of chrome, nickel arid alumina were made as were sponges
of pure alumina, of zirconia, and of catalytic substrates. It is claimed that a
zirconia sponge will withstand an oxyacetylene flame. Insley et al (2) of Champion
Spark Plug Company also have invented a method for converting ucethane aponges to
ceramic sponges. The group at Alfred University under W. Crandall have also developed
urethane to ceramics foam technology. These processes appear admirably suited to
producing thermal shock resistant catalytic monoliths based on restraint-free designs.
CONCLUSIONS
It has been shown that restraint-free designs for combustor catalyst monoliths
are possible. Such designs are in principle immune to thermal shock damage.
Monoliths based on such designs can be readily manufactured using existing foam
ceramics technology. Inasmuch as foam ceramics are inherently resistant to thermal
shock, it should be possible to produce monolith catalysts which can safely be
thermally cycled over wide temperature ranges at least as rapidly as can conventional
conbustors.
(1) U.S. Patent 4,083,905 (April 1978)
(2) U.S. Patent 4,076,888 (February 1978)-
68
-------
SESSION II
CATALYST PERFORMANCE
69
-------
COMPUTED AND MEASURED EMISSIONS FROM THE
CATALYTIC COMBUSTION OF PROPANE/AIR MIXTURES
By
F. V. Bracco, C. Bruno, Y. Yaw, P. M. Walsh
Department of Mechanical and Aerospace Engineering
Princeton University
Princeton, NJ 08544
ABSTRACT
The objective of this program is to collect detailed data within a
honeycomb catalyst for the development and testing of a two-dimensional model,
and to further current understanding of the fundamental subprocesses which
control catalytic combustion. Results are presented on the combustion of
propane/air mixtures in a Pt/Y-Al203/Cordierite honeycomb catalyst. The test
conditions were: inlet temperature, 650 K to 800 K; inlet pressure, 110 kPa;
inlet velocity, 10 to 40 m/s; C3Hs/air equivalence ratio, .19 to .32; and
H.O concentration, 1.2 to 1.7 mol%.
Efforts were made to obtain uniform profiles of temperature, velocity, and
C3Hg concentration across the inlet gas stream. Temperature and concentration
are uniform within ±1%, but velocity only within ±8%. The philosophy of the
program is that a detailed model of the catalytic combustor is necessary to be
able to draw consistent conclusions from the experimental data. The model
developed during the present investigation solves the 2-D, steady-state,
reactive flow in a cylindrical pipe with catalytic walls. When the experimental
wall temperature is used as boundary condition for the gas-phase equations, the
emission predictions are in reasonably good agreement with the measured ones.
The indications obtained from the model are that propane is oxidized via a multi-
step kinetic mechanism, and that, in the range of temperatures and equivalence
ratios explored, most of the fuel is burnt at the catalytic wall rather than in
the gas phase.
70
-------
CONTENTS
List of Figures
List of Tables
Symbols
Acknowledgments
Introduction
Equipment and Procedure
Data and Discussion
Theoretical Model and Results
Conclusion
References
71
-------
LIST OF FIGURES
1. Schematic diagram of the flow reactor.
2. Details of the catalyst test section.
3. Catalyst inlet uniformity measurements.
4. Exhaust gas composition downstream of the catalyst.
5. Exhaust gas composition downstream of the catalyst.
6. Exhaust gas composition downstream of the catalyst.
7. Exhaust gas composition downstream of the catalyst.
8. Temperature and composition as a function of T. , .
9. Temperature and composition as a function of $.
lOa. Temperature as a function of inlet velocity u .
lOb. Exhaust gas composition as a function of inlet velocity u...
11. Physical and chemical processes in a channel of a catalyst.
12. Coordinate system and variables in the model of a catalyst channel.
13. Theoretical profiles of unburned HC vs. r and x for 3-step (overall)
catalytic oxidation of C H .
J O
1.4. Theoretical profiles of unburned HC vs. r and x for 3-step (overall)
catalytic oxidation of C,H .
J O
15. Theoretical profiles of CO vs. r and x for 3-step (overall) catalytic
oxidation of C,H0.
3 O
16. Theoretical profiles of CO vs. r and x for 3-step (overall) catalytic
oxidation of C_H0.
J O
17. Pressure drop.
18. Unburned hydrocarbons.
19. Unburned CO.
20. CO- product.
21. Relative importance of gas-phase vs. surface reaction rates.
72
-------
LIST OF TABLES
1. Catalyst Properties.
2. Dimensional Governing Equations for Reacting Flow in a Circular Tube
with Catalytic Wall.
3. Chemical Reaction Rate Data Used in the Modeling of 3-Step Kinetics
of Propane Oxidation on Pt/Y-Al~0_/Cordierite Catalyst.
73
-------
SYMBOLS
C specific heat at constant pressure (J/kg»K)
2
D diffusion coefficient (m /s)
E activation energy (J/mol)
h enthalpy (J/Kg)
k rate constant (cm /mol«s); Thermal conductivity (W/m*k)
Le Lewis number = X/pDC
P pressure (Pa)
q heat flux (J/m-s)
Q heat of reaction (J/kg)
r radial coordinate (m)
R gas constant = 8.31 J/K-mol; pipe radius (m)
s pipe thickness (m)
t time (s)
T temperature (K)
u axial gas velocity (m/s)
v radial gas velocity (m/s)
W molecular weight
Y mass fraction
z axial coordinate (m)
v'jV*' stoichiometric coefficients
p density (kg/m )
a stress component (Pa)
<|> equivalence ratio
[ ] concentration (mol/cm )
-------
Subscripts
ad adiabatic (reaction temperature)
b backward reaction
f forward reaction
in condition at catalyst inlet
k species k
Si reaction H
out condition at catalyst outlet
r radial
ref reference condition (velocity)
w wall
z axial
75
-------
ACKNOWLEDGMENTS
This project is supported by the Air Force Office of Scientific
Research under Grant No. AFOSR-76-3052, Dr. B. T. Wolfson, Grant Monitor;
and by the Department of Energy under Contract No. DE-AC21-80MC14036,
Mr. T. O'Brien, Technical Project Officer.
The authors are indebted to Dr. W. C. Pfefferle for his many valuable
suggestions.
76
-------
INTRODUCTION
Although many important features of catalytic combustion have been inves-
tigated and identified, the relative importance of the relevant fundamental
physico-chemical processes still needs to be assessed. Especially interesting
is to find out the interaction between mass transport and heterogeneous and
homogeneous chemical kinetics. The design of large scale practical devices
for heat and power generation could therefore make use of a model which accu-
rately describes the system in terms of basic chemical and physical phenomena
and which relates performance to all the relevant design parameters. The goal
of our program is to develop a useful model for combustion in monolithic cata-
lysts and to obtain experimental data for testing the model and the predictions
of other investigators. The present work describes experimental results from a
preliminary study of C,H0 combustion in Pt/Y-Al_0,/Cordierite.
oo z j
77
-------
EQUIPMENT AND PROCEDURE
INLET CONDITIONS
A sketch of the experimental apparatus is shown in Figure 1. Preheated
air at a measured flowrate is supplied to a 690 mm long test section with 25.4 mm
square channel. A catalyst is placed with its downstream end 90 mm from the test
section outlet, and insulated from the wall by Fiberfrax paper. Details of the
catalyst section are shown in Figure 2. A fuel injector consisting of five
1.6 mm diameter tubes, each containing five 0.3 mm diameter holes, is located
440 mm from the catalyst inlet. A combination pitot tube and thermocouple is
mounted 200 mm from the catalyst inlet. In addition to measuring gas velocity
and temperature, the pitot tube is used to extract gas samples which are analyzed
to determine equivalence ratio. Pressure is regulated by a valve in the exhaust
pipe, and taps placed up and downstream of the catalyst are used to measure inlet
pressure and pressure drop. A mass flowmeter (Hastings Model AHL-100P with
H-3M/L-100 Transducer) measured the air flowrate. The water content of the inlet
air was measured using a semiconductor sensor (Thunder Model 2000 with BR-101B
probe) mounted in the airstream between the receiving tank and the heaters.
The inlet conditions used in the experiments are summarized below.
Inlet temperature (T. ) = 650 ± 13 K to 800 ± 16 K
Inlet pressure (P. ) = 110 ± 5 kPa
Inlet Velocity (u. ) = 10 ± 4 to 40 ± 7 m/s
C3Hg/air equivalence ratio (
-------
range of velocities used. Sufficient fuel/air mixing was obtained only by
placing baffles downstream of the fuel injector, and these affect the velocity
profile. By trying various configurations, the velocity uniformity was improved
while maintaining an even distribution of fuel. The arrangement used in the
present experiments consisted of two screens, each containing four 8 mm diameter
holes, placed 30 and 110 mm downstream of the fuel injector. Another screen
perforated with 1.6 mm diameter holes was 190 mm from the injector and 50 mm
upstream of the pitot tube. The resulting fuel distribution shows good uni-
formity over the measured range. Velocity profiles are less satisfactory
(range ± 6% over 50% of the channel width). Average reference velocities were
determined using the C H and air flowrates, inlet temperature and pressure, and
o o
the cross section area of the catalyst. The catalyst inlet velocity, taking
into account the fraction of open monolith area, is u. = (1.67 ± .06)u -.
r 'in ' ref
SUBSTRATE AND EXHAUST MEASUREMENTS
Substrate temperatures are measured by a method similar to that described
by Kesselring, Krill, and Kendall (1). Ni-Cr/Ni-Al thermocouples are fed
through the test section wall and into the ends of catalyst channels. The
lengths of wire inside the catalyst are covered by mullite insulator and both
ends of the channel sealed with ceramic adhesive. The lifetime of these thermo-
couples under test conditions is short (5-20 hr). Exhaust gas samples were
taken through an expansion quenched, water cooled, stainless steel probe mounted
in an elbow downstream of the test section. Exhaust gas temperature was measured
with a thermocouple mounted on the probe. CO and CO were determined by infrared
absorption (Horiba Model AIA-21), 00 by magnetic susceptibility (Scott Model 250),
and total hydrocarbon (HC, reported here as C ) by a flame ionization detector
J
(Scott Model 415). Pressure was measured inside the catalyst at 3 locations.
CATALYST AND FUEL
The catalyst was platinum supported on split cell corrugated Cordierite with
y-alumina washcoat. The overall dimensions of the monolith were 24 x 24 x 76 mm,
the open area was 64%, the channel cross-section area was 1.9 mm , and the plati-
7
num loading was 4.2 kg/m . The physical properties of the catalyst are listed in
Table 1. The sample was pretreated by burning propane for two hours with the
maximum substrate temperature at 1480 K. The fuel was natural propane, 96 mole9a,
nominal.
79
-------
DATA AND DISCUSSION
The combustor was operated over a range of equivalence ratios, inlet
temperatures and reference velocities at fixed inlet pressure. No attempt was
made to minimize unburned fuel, since the objective was to understand the re-
lationship between operating conditions and physico-chemical phenomena. Typical
exhaust gas composition results are shown in Fig. 4, 5, 6 and 7. Fig. 4 shows
CO formed by the C H breakdown, being oxidized to CO., in the exhaust pas down-
O O 4-
stream of the catalyst.
A primary objective was the measurement of gas temperature and composition
as functions of axial position inside a catalyst. This was accomplished by
drilling an 8 mm diameter hole partway through the catalyst along its axis.
The hole is lined with a ceramic sleeve to minimize the activity of the wall.
A combination gas sampling and thermocouple probe is moved along the catalyst
axis from downstream toward the bottom of the hole. If conditions are properly
adjusted, the measured gas temperature and composition at the bottom of the hole
are equal to their values at the same axial position in a catalyst with no hole.
The condition which is most affected by the presence of the hole and probe is
the gas velocity in the catalyst channels opening into the hole. The problem of
establishing the correct velocity has been solved by feeding pressure taps
through the side of the catalyst to measure the pressure at the bottom of the
probe hole.
The velocity is adjusted until the pressure is equal to its value at the
same axial position in a catalyst without a hole. Measured pressures show a
linear decrease with increasing distance from the inlet during propane combus-
tion in the Pt/Al90v/Cordierite catalyst. Another requirement for reliable gas
£* *_)
measurement inside a catalyst is that the substrate temperature be unaffected.
This condition can also be satisfied despite the changed gas velocity because
substrate temperature does not have a strong dependence on velocity. It is the
constraint of fixed substrate temperature which distinguishes this experiment
from one in which the overall length of catalyst is varied. In a catalyst of
given length, the substrate temperature profile is different from the profile in
an equivalent section of a longer catalyst. Direct measurements of fuel, oxygen,
and product concentrations and gas temperature as functions of axial position
inside a catalyst provide data for a rigorous test of a model for combustor
80
-------
operation. Calculations of species concentrations inside monolithic catalytic
combustors have been presented by Kelly, Kendall, Chu, and Kesselring (2) and
by Cerkanowicz, Cole, and Stevens (3), but an experiment with which to compare
these predictions had, to our knowledge, not yet been performed. Fig. 5 refers
to results obtained from a catalyst with a hole depth equal to half the length
of the catalyst. For the conditions of this experiment, fuel breaks down mono-
tonically while at the same time the CO formed oxidizes to CO- and decreases as
HC decreases.
Fig. 6 is interesting in the fact that while the UHC at the outlet keep
decreasing with X, CO has a relative maximum. This is an indication that CO may
be produced inside the catalyst at a rate larger than its rate of oxidation to
co2.
This suspicion is confirmed by the results of Fig. 7, in which probing
inside the hole reveals that not all the CO produced by propane pyrolysis is
immediately oxidized to C02, and that a maximum exists inside the catalyst
channels. More definite trends appear when temperatures and compositions are
plotted versus inlet parameters. Fig. 8 has as parameter the inlet temperature.
This has a strong effect on catalyst temperature and unburned HC at the outlet,
which decreases rapidly as Tin is raised. However CO behaves non-monotonically:
as Tin ls raised> CO emissions at the outlet at first increase and then decrease.
The larger the UHC at the outlet, the higher the CO peak corresponding to
HC breakdown to CO. From the figure it can be seen that the maxima of the rates
of HC disappearance and CO production occur at approximately the same axial
location.
The effect of varying equivalence ratio is shown in Fig. 9. has a very
strong influence on catalyst temperature, since it affects directly the adiabatic
flame temperature, to which the wall temperature is close. For low $ the UHC
are relatively large at the outlet, and the low exhaust temperature results into
low rates of HC breakdown into CO and CO oxidation. As is increased both HC
breakdown and CO oxidation proceed faster and faster: this seems to indicate
that the CO axial peak tends to move inside the catalyst channels as $ is raised.
81
-------
Figs. lOa and lOb show the effect of varying inlet velocity and therefore
residence time inside the channels. The wall temperature is only weakly affected
by changing u , since it is always close to the adiabatic flame temperature
corresponding to given T. and $, which is independent of velocity. Emissions,
ill i c L
instead, are strongly affected, and both DUG and CO become progressively larger
as u is increased. A maximum of CO at the outlet can be seen when u is suffi-
ciently high. This brief discussion of the experimental results points out the
complexity of the interaction between physics and chemistry inside of the
catalyst channels. The physical size of the probe used precludes a direct inves-
tigation inside an individual catalyst channel. What the probe can measure are
quantities averaged over several channels cross sections, and no indication of
radial diffusion effects can be obtained. A detailed description of these and
other effects may be supplied only by the mathematical model.
82
-------
THEORETICAL MODEL AND RESULTS
A sketch of the fundamental phsyico-chemical and fluid dynamics processes
in a catalyst channel is shown in Fig. 11. The purpose of the model is to pre-
dict the effect of these processes on the exhaust products and the substrate.
After reaching good agreement between predictions and experimental data, the
model was used to investigate the relationship between the different processes.
For simplicity the catalyst channels are approximated by cylindrical
channels. Preliminary results indicated that radiative heat transfer may be
important at the two ends of the channel and for about 6% of the channel length.
Therefore radiation is neglected in the model. With this simplification the
equations describing catalytic combustion and the coordinate system are shown
in Fig. 12 and Table 2.
These equations describe a two-dimensional, axisymmetrical, compressible,
fluid flow with arbitrary chemical reactions in the gas phase, and with heat
transfer and conduction in the substrate modeled assuming constant temperature
across the substrate thickness s.
The reaction rates used in this study are shown in Table 5. The 3-step
mechanism for propane oxidation in the gas phase was essential for a good agree-
ment with the experimental data. (No reasonable results could be obtained by
simply assuming C'3Hg to break down into CO and then oxidizing CO to C02).
This mechanism was obtained by Hautman et al. (4)(7) at Princeton and some slight
adjustments were made to the activation energies and preexponential factors of
reactions 1. and 2. These adjustments did not change the corresponding rates at
the average temperature of the gas inside the pipe.
To test the validity of the model and separate gas phase effects from sub-
strate effects, the gas phase equations were solved by uncoupling them from the
substrate energy equation. In this case the only extra variable unknown is TW,
which was taken from the experiments and imposed as boundary condition at the
wall. This procedure has as advantage the implicit inclusion of radiative
effects in the model (provided the gas is optically thin), since the experimen-
tal wall temperature does include the effects of radiation.
The basic structure of the computer program used to solve the set of
equations, "CATEACH", was based on the TEACH hydrodynamic code designed by
A. D. Gosman and coworkers at The Imperial College. Details of CATEACH can
be found in (5) and (6).
83
-------
Some of the theoretical results are shown in Figs. 13 through 16 [see (6)
for a detailed discussion]. Fig. 13 shows the monotonical decrease in HC con-
centration along the axis and from the axis to the catalytic wall. The rate of
HC breakdown and oxidation is a function of the radial mass transfer and of the
CO oxidation rate. Increasing the inlet velocity, as in Fig. 14, results in a
correspondent decrease in residence time and a slower HC disappearance. Con-
sistently, CO is still being formed at the outlet for this higher velocity case
(see Fig. 15), and would presumable peak in the exhaust as in Fig. lOa. For
the lower velocity case of Fig. 16, the residence time is sufficiently large
for CO to begin to be oxidized inside the channel. Thus a relative CO maximum
must occur, as in Fig. 5.
Overall comparison between theory and experiments is shown in Figs. 17
through 20. The pressure drop between catalyst inlet and outlet is important
for gas turbine applications. There is a certain amount of disagreement between
predictions and data; this is likely to be the effect of having approximated
the [roughly] trapezoidal shape of the channel cross section by a circular one.
The trends, however, are reproduced (see Fig. 17). Fig. 18 shows good agreement
in the UHC comparison. Less satisfactory (but consistent as far as the trends)
is the comparison for CO and CO- (see Figs. 19 and 20). It should be noted that
the outlet emissions predictions are averages over the cross section, while data
were taken with a probe hole slightly smaller than the channel size; part of
the disagreement may be also due to radial gradients effects on the probe. A
second cause may be a gradual loss of catalyst activity observed with increasing
total run time (aging).
Once the computer code has proved itself in predicting reasonable trends
over the range of experimental measurement, some conclusions can be drawn.
Fig. 21 shows the relative magnitude of gas phase vs. catalytic reaction rates.
Most of the HC is oxidized at the catalytic wall, but this tends to change as
the inlet temperature or the equivalence ratio is increased. The effect of
velocity is opposite to the effect of inlet temperature but not as strong.
84
-------
CONCLUSION
The agreement between mathematical model and experimental data for C HR
combustion in Pt/y-Al^O./cordierite monoliths is reasonably satisfactory.
Preliminary results exhibit several interesting features.
1. Substrate temperature is only weakly dependent on gas velocity.
2. Emissions depend strongly on all three parameters investigated.
3. Propane oxidation takes place via a multi-step mechanism involving
at least three overall steps.
4. While unburned hydrocarbon varies monotonically with the three
parameters investigated, CO emissions do not.
5. Most of the fuel is burned at the wall by catalyzed reactions
rather than in the gas phase.
Future work will include mathematical coupling of gas and substrate to
make the code independent of experimentally measured inputs.
85
-------
REFERENCES
1. Kesselring, J. P., W. V. Krill, and R. M. Kendall. Design Criteria for
Stationary Source Catalytic Combustors. Presented at the Second EPA
Workshop on Catalytic Combustion, Raleigh, NC, June 21-22, 1977.
2. Kelly, J. T., R. M. Kendall, E. Chu, and J. P. Kesselring. Development
and Application of the PROF-HET Catalytic Combustor Code. Paper 77-33,
Presented at the Fall 1977 Meeting, Western States Section, The Combustion
Institute, Stanford, CA, October 17-18, 1977.
3. Cerkanowicz, A. E., R. B. Cole, and J. G. Stevens. Catalytic Combustion
Modeling: Comparisons with Experimental Data. Presented at the ASME
Gas Turbine Conference and Products Show, Philadelphia, PA, March 27-31,
1977.
4. Hautman, D. J., Hydrocarbon Oxidation Analysis. Paper presented at the
Combustion Phenomena Contractors Review Meeting, U.S. Department of Energy,
Pittsburgh Energy Technology Center, Pittsburgh, PA, Sept. 19, 1979.
5. Gosman, A. D., and F. J. K. Ideriah, "TEACH" - A General Computer Program
for Two-Dimensional, Turbulent, Recirculating Flows. Dept. of Mech. Eng.
Report, The Imperial College, London, SW7, 1976.
6. Yaw, Y., MSc Thesis, Princeton University, to be published.
7. Hautman, D. J., Pyrolysis and Oxidation Kinetic Mechanism for Propane,
Ph.D. Thesis, Princeton University, 1980.
86
-------
Table 1. Catalyst Properties
SUBSTRATE: Cordierite, American Lava Corp., AlSiMag 795, split cell
length 0.0760 ± 0.0003 m_3
wall thickness 0.25 x10" m
open area 64% , _^
channels per unit area 0.34 x 10 m
open area per channel 1.9 x 10~^ nr 23
ratio surface area to total volume 2100 m /m
ratio surface area to gas volume 3260 m2/m3
channel hydraulic diameter (4/3260) 0.00123 m
bulk density 610 kg/m3
solid density 1700 kg/m3
safe operating temperature 1473 K
specific heat 800 J/kg-K _6 _j
coefficient of thermal expansion (linear, 294 - 1033 K) 3.8 x 10" K~
thermal conductivity of solid (572 K) 1.4 J/m.s»K
approximate channel cross-section is a trapezoid with base lengths .0011
and .0023 m, and height .0012 m
WASHCOAT: yalumina
loading 115-125 kg/m3
surface area (BET) (29.2-33.0) x 10 m /m
CATALYST: platinum
loading 4.2 kg/m0 423
surface area (CO chemisorption) 6 x 10 m /m
The y-alumina and platinum were applied by Matthey Bishop, Inc., Malvern, PA.
m
87
-------
Table 2. Dimensional Governing Equations for Reacting
Flow in a Circular Tube Kith Catalytic Kail.
3p 1 3pvr 3pu _ -
31 + r 3r 3z
Dpv _ j)p_ 1_ rr + rz
Dt = ~ 9r r 3r 3z
Dt
Ml
Dt
.
3z r
3z
= |E + A 3 (0 vr . o ur)
3t r 3r l rr rz 7
continuity
radial momentum
axial momentum
, ,
87 (0rzV + CzzU) - r -8T- - IT
energy
V/
r
i P^I
L „
p
3(P/P)
pk\k£1
w J
species
p = pTRE(Y./W. )
is. K.
3T
w
at
3 Tw 2k r3T,
TT - I IT [3?]D
3z w R
state
LE ok n f-M
— L KyKffl J1 I W I
\ I * tiL k V k
substrate
energy
88
-------
TABLE 3. CHEMICAL REACTION RATE DATA USED IN THE
MODELING OF 3-STEP KINETICS OF PROPANE OXIDATION
ON Pt/Y-Al203/CORDIERITE CATALYST
• C3H8
GAS
1 o2 -> | c
20, -»• 2CO
2H70
3. CO + 0 •* CO
REACTIONS
WALL
50_ -»• 3CO. + 4H_0
2. I, I.
2.
1
3. CO + 0 -»• C0
2H?0
RATES
GAS (mo I/ cm -s)
1. 1.4 x 108 exp(- 40608/RT)
2. 4.8 x 1014 exp (- 70000/RT)
WALL (mol/cm Pfs)
1. 1.09 x 109 exp(- 17600/RT)[C3Hg]
2. 107'04 exp(- 12000/RT)[C2H4]
3. 3.83 x 103 exp (- 24900/RTJ[CO]
5. 1.0 x 10 ° exp (- 30000/RT)
• [C0][02r25 [H20]'5
89
-------
Bypass Air
Thermocouple/
Sampling probe
Settling
Tank
Pitot-static/
Thermocouple
Probe
Figure 1. Schematic diagram of the flow reactor with
catalytic combustion test section.
90
-------
FLASHBACK
ARRESTOR
THERMOCOUPLE
PITOT/STA
THERMOCOUf
SAMPLING 1
<=i
AI203
TIC/ MONOLITI
?LE/ FLASHBA
^ROBE ARRESTO
FI nw .... »»-
0
i
i
H
CK
R
Pt/r-AI2 03/ THERMOCOUPLE/
CORDIERITE SAMPLING PROBE
MONOLITHv /
1
2 T^ 3
/
4
1 I
1
1
u
b Ic I
a 0.217 ± 0.001 m
b 0.0132 OR 0.0262 ± 0.0003 OR 0
c 0.001 TO 0.003
d 0.0760 ± 0.0003
e 0.0010 ± 0.0005
Figure 2. Details of the catalyst test section.
-------
26
25
24
23
10
15
20
25
T«55I°K
P * I ATM
111
J*
OUJ°
Q. —
U
540
530
..''"* .
m •
*
V= 8.6m/sec
P = 1 ATM
A T/T = 0.02
10 15 20 25
8
9000
8000
5 10 15 20 25
CROSS-SECTION POSITION (mm)
T = 538° K
P« I ATM
V = 5.8 m/sec
A C/C = 0.01
Figure 3. Catalyst inlet uniformity measurements.
92
-------
Run 11-11-C_H0
3 o
Uinlet = 15 m/s
-------
Run 11-13-CjHg
•J
o
o
(J
10
1 —
.1
,01
.001
.0001
°2
C°2 -
I CO '
—
-
" HC
-
—
-
DOO O O
*m
7v
V
V
D
D
hole
depth
D
1
o o o o o
A A A A A
V
V
V
o
pa
pi
X
rf
O
C
h-<
fD
1 1
SO 100
Distance from catalyst inlet (nun]
1SU
Figure 5. Exhaust gas composition inside and
downstream of the catalyst.
94
-------
Run 11-2-CJI
8
uinlet = 9'4 m/s
S> = .30
T. . = 750 K
inlet
•s.
o
0
u
10
.01 —
.001 —
.OuOl
°2 (
CO, ,
0,; ,
HC
1
D O O O O
^ A A A £
7 ^ V
V
V
3
1 1
50 100
Distance I'rom catalyst inlet
1SU
Figure 6. Exhaust gas composition downstream of
the catalyst.
95
-------
Run 11-15-C_H
uinlet = 6'4 m/S
$ = .28
T. . „ = 800 K
inlet
10 _
1 —
. 1
u
o
J-
o
o
,01 —
.001
,0001
°2
- C°2
-
HC
CO
—
r
200 o o
^A A
* V
V
7
n
D
1
o o o o o
A A A A
V
V
v
D
1 1
50 100
Distance from catalyst inlet (.mm)
ISO
Figure 7. Exhaust gas composition inside and
downstream of the catalyst.
96
-------
CATALYST
OUTLET
* I40O-
2
I
u
1200
lOOOf-
800
1
? f
kti •'-
*•
•
I ,
1
rry-j
I
u
«
o
.05 .10
AXIAL POSITION (m)
SOLID
GAS
O
Po'MOlSkPo
^=0.2610.05
99
T0(K)
650113
700114
800115
Figure 8. Temperature and composition as a
function of T. n ..
inlet
97
-------
CATALYST
OUTLET
.05 .10
AXIAL POSITION (m)
001
f.003
s
o
>? -002
.00!
0
jv v v
I
3 D D Q
I
°0
O
I A f, A I n
M ^ Ti ^J '
^
1
0
.003
.002
8
X .001
0
i i i
A o
-
.
& O
o A
o
O 4J 1
i p rf n i n /
.15
SOLID
GAS
V
D
O
Po*HO±5kPo
T0«750±I4K
:IS±5m/s
O.I9±.03
0.22 ±.04
0.25 ±.05
0.28 ± .05
Figure 9. Temperature and composition as a
function of .
98
-------
CATALYST
I^W
1300
* 1200
uj
§5 MOO
£ ooo,
£ '
UJ 900
800
700
' 1 1 C
f 1 t *• °
v I A ^ A « A ^
t
:
-
— TO
— i,
Sv y ,
1
.05 .10
AXIAL POSITION (m)
.15
SOLID GAS
v v
A A
• o
POS 110 ±5 kPo
T0«750ll4K
4«0.30t0.07
I0±4
2015
40±7
Figure lOa. Temperature as a function of inlet velocity u,.
99
-------
,004
.003 -
.002 -
.001
CATALYST
OUTLET
0 .05 .10
AXIAL POSITION (m)
.15
-------
r velocity
Uniform \ temperotion
Development length
Iconcentrotion \////////////
Radiation
Reactants
r velocity
,Local •( temperature
L concentration
~! convection ciSSfeTSf1™
*]diffusion ^L-J—diffusion j
^••^"surface reaction
L-^-- ^ \__re_oction
^77///777^/ / / ;//T\
ENTRANCE
"U Products
Radiation
conduction in solid
EXIT
Figure 11. Physical and chemical processes in a channel of catalyst.
-------
INLET
OUTLET
Figure 12. Coordinate system and variables in
the model of a catalyst channel.
102
-------
1.0
.8
LJ
5 .6
' eo'
X
O
UJ
X
o
LJ
CC
13
m
z
L=L .2
UrefB6m/s
^ = .3
Tjn«750°K
(Run 11-2)
RADIAL PROFILES
AXIAL PROFILES
'.610
.2
.4
.6
.8
1.0
r/R
Figure 13. Theoretical profiles of unburned HC versus r and x for
3-step (overall) catalytic oxidation of C-jHg.
103
-------
1.0
x/L=.125
Uref = 24m/s
.8
UJ
- .6
X*
u
•s.
»-
UJ
I-
o
O
LJ
Z
CE
CD
Z
D
.4
.2
j/R=0
RADIAL PROFILES
.2
.6 .8
1.0
r/R
1.0
Figure 14. Theoretical profiles of unburned HC vs. r and x for
3-step (overall) catalytic oxidation of C0H0.
J O
104
-------
i— Urefs24m/s
o
x
z*
o
o
<
ct
u.
V)
(/)
I
O
o
AXIAL PROFILES
Tin«750°K
(Run 11-4)
RADIAL PROFILES
Figure 15. Theoretical profiles of CO vs. r and x for 3-step
(overall) catalytic oxidation of
.
o
105
-------
x/L».771
m
i
o
b
O
o
Uref = 6.0m/s
-2)
«.947
AXIAL
PROFILES
r/R «0
«.610
«.947
0 .2 .4 .6 .6 1.0
K/L
Figure 16. Theoretical profiles of CO vs. r and x for 3-step
(overall) catalytic oxidation of C_H0.
J o
106
-------
THEORY VS, EXPERIMENTS
lOr-
T,n-750*K
Uref «9m/$
lOr—
8
Ap/p
%
Ap/p
.15 .19 .22 .25 .28
10
8
<£-.29 to.32
Tin«750«K
Ur€f»9m/t
J
650 700 750 800
Tin,INLET TEMPERATURE ,»K
• THEORETICAL
•EXPERIMENTAL
Ap/p
12 18 24
Figure 17. Pressure drop.
107
-------
o
1.0
.6
'
CO
X
.4
.2
0
.15
1.0
.8
THEORY VS, EXPERIMENTS
1.0
o
*•«••
o
X
bl
X
c? .4
.2
Tin « 750°K
Uref«9m/$
.19 .22 .25 .28
$«.29to.32
T,n.750'K
12 18 24
Uref(m/iec)
Uref
$-.28
650 700 750 800
T,n,INLET TEMPERATURE ,'K
•THEORETICAL
•EXPERIMENTAL
NOTE.U|NLET-UrefX
Totol oreo
Open oreo
Figure 18. Unburned hydrocarbons.
108
-------
1.0
.8
m
ii .4
.2
01-
.15
l.0r-
.8
IU
N
3
THEORY VS, EXPERIMENTS
T50»K
1.0
Ure
5
.6
o A
.19 .22 .25 .26
Uref • 9m/»
^•.28
650 700 750 800
Tln,INLET TEMPERATURE ,»K
^".29 to.32
Tin«750»K
• THEORETICAL
•EXPERIMENTAL
12
18
24
Figure 19. Unburned CO.
109
-------
THEORY VS. EXPERIMENTS
§
i
Ul
8
o
.15
Tto « 750"K
Urtf • 9m/»
5r—
JS 3
*ib
.19 .22 .25 .26
g
9m/s
1
50 700 75O 800
.INLET TEMPERATURE ,*K
5i—
UJ
o
*^
N
UJ
10
u
4>*. 29 to. 32
T,n-750-K
• THEORETICAL
•EXPERIMENTAL
12
18
24
Figure 20. (XL product.
110
-------
1.0
.8
.6
'GS
.2
THEORY VS. EXPERIMENTS
1.0
Tm-750-K
Uref »9m/j
.6
GS
.2
Uref » 9m/s
.15
1.0
.8
.6
.19 .22 .25 .26
^-.29 to. 32
Tin«750»K
650 700 750 800
TJn, INLET TEMPERATURE ,-K
• THEORETICAL
•EXPERIMENTAL
GAS PHASE REACTION
r6S'
SURFACE REACTION RATE(m°'* )
12 18
Figure 21. Relative importance of gas phase vs. surface reaction rates.
Ill
-------
MODELING OF TRANSIENT OPERATION
OF CATALYTIC COMBUSTOR
by
James S. T'ien
Department of Mechanical and Aerospace Engineering
Case Western Reserve University
Cleveland, Ohio 44106
ABSTRACT
A quasisteady-gas-phase and thermally-thin-substrate model is used
to analyze the transient behavior of catalytic monolith combustors in
fuel-lean operation. One particular concern for catalytic combustors
used in transportation engines is their transient response time. The
computed results indicate that fast response is favored by thin substrate,
short catalytic bed length and large channel diameter in high inlet-tempe-
rature operation. From combined steady and unsteady considerations, it is
possible to choose a catalytic combustor design which has reasonably fast
response (~1 second) and can satisfy the emission goals in an acceptable
total combustor length.
112
-------
ACKNOWLEDGEMENT
This project is supported by the Department of Energy under NASA
Grant NSG-3230 administered by NASA Lewis Research Center.
The author wishes to express his sincere thanks to Dr. David N.
Anderson, without whose advice and suggestions this paper would not be
possible.
113
-------
Nomenclature
A* crossectional area of one gas channel
A* substrate cross-sectional area associated with one gas channel,
see Fig. 1
Lf ' ' ,*
" = c>
VI
X Y/>»> ^o'^ H»~/J VH,"
-------
L* total combustor length (catalytic bed plus after-bed space)
required to meet emission goals
Le, Lewis number for hydrocarbon gas in air
Le- Lewis number for CO in air
m number of hydrogen atom in C H
n m
n number of carbon atom in C H
n m
Nu Nusselt number at x
x
Nu Nusselt number for fully developed flow
00
p = p*/p*(0,0), nondimensional pressure
q1 = q*/C*T*(0,0), nondimensional heat of combustion per unit mass
of C H in reaction A
n m
q9 = q*/C*T*(0,0), nondimensional heat of combustion per unit mass
of CO in reaction B
q = q*/C*T*(0,0), nondimensional heat of combustion per unit mass
j 3 p
of C H in reaction C
n m
q = q*/C*T*(0,0), nondimensional heat of combustion per unit mass
A 4 p
of CO in reaction D, same as q
r see Eq. (17)
UK.
S* circumferential length of one channel crossection (= ird*)
T = T*/T*(0,0), nondimensional gas temperature
T = 1L*/T*(0,0), nondimensional substrate temperature
s *
T* adiabatic flame temperature
3.X
T* combustor inlet temperature
t = t*/T*, nondimensional time
u = u*/u*(0,0), nondimensional flow velocity
W molecular weight of specie i
x = x*/L*, nondimensional axial distance measured from catalytic bed
entrance to downstream
115
-------
u* r reference gas velocity, measured at upstream of catalytic bed
Y. mass fraction of specie i
a* thermal diffusivity of gas
3 rate constant, see Eq. (22)
p* gas density
p* substrate density
s
T* characteristic substrate heat-up time, see Eq. (16)
r\ carbon-balanced efficiency
UD
HT combustion efficiency based on temperature difference
f J concentration, g-mole/c.c.
116
-------
I. Introduction
Due to the large thermal inertia of the substrate, the transient
behavior of catalytic combustors is expected to be different from that of
a conventional gas turbine burner. Non-steady operations are obviously
important to transportation engines since their power levels have to be
changed frequently. Transient may also be a concern with stationary gas
turbine applications, since the ignition/shutdown operation can produce
excessive thermal stress or thermal shock if the combustor is not designed
or operated properly and substrate can fail as a result [1].
Typical questions concerning transient operations include the combustor
response time, the type of response, the unsteady substrate temperature
history and the resultant thermal stress distribution. At the present time,
both experiment and theory on transient catalytic combustion are lacking.
A model is presented here attempting to analyze the problem theoretically.
This paper will proceed first to discuss the time scales associated with the
unsteady processes in a monolith catalytic combustor. Then a lean combustion
model will be presented for the limiting case of "quasisteady gas phase and
thermally thin substrate." The computed results will then follow.
II. Nonsteady Combustion Model
II.1 Consideration of Time Scales
For a monolith catalytic combustor with uniform cell distribution in
the plane normal to the approaching flow direction, if the combustor wall is
well insulated, then the study of whole reactor can be reduced to the con-
sideration of a single cell unit. Referring to Fig. 1, each cell unit.
consists of an open gas channel and the associated substrate volume.
Considering first the steady-state combustion, the relevant time scales
are the gas residence time in the reactor, the heat and mass transfer time
between the gas and the solid surface inside the channel and the gas-phase
and surface reaction times. When a transient is caused by an upstream para-
meter variation, then the new time scales involved include the input charac-
teristic time, the time for the temperature wave to reach the central plane
of the substrate and the time for heating-up (or cooling-down) the substrate.
Table 1 defines these characteristic times and their estimated magni-
tudes. It should be noted that some of the time scales vary along the
117
-------
channel of the reactor bed. For example, the heat/mass transfer time and
the substrate heat-up time are shorter at the flow entrance region than
those for far downstream, and the ones used in Table 1 are those based on
fully-developed flow profiles. However, the estimation from Table 1 does
serve the purpose of order-of-magnitude comparison. From this table we
see that the longest time is the solid heat-up time, being of the order
of seconds. If the time scale we are interested in is much greater than
the gas residence time (~15 msec), then the gas phase processes (including
heat/mass transfer and chemical reactions) can be regarded as in a quasi-
steady state. If furthermore the substrate half-thickness is smaller than
about .2 mm, then the substrate temperature distribution in the direction
normal to the channel axis can be regarded as uniform at any given time.
In this limiting case then, a "quasi-steady-gas-phase and thermally-thin-
substrate" model is applicable with the only transient process being the
substrate heating-up or cooling-down. In the following section, this model
will be described in mathematical form.
II.2 Quasisteady Gas Phase
Quasisteady gas phase implies that the differential equations for the
gas-phase processes are the same as those for the steady state. Only the
boundary conditions on the solid surface are different. There are several
steady-state catalytic combustor models in existence [2-5] and they show a
high degree of success in describing the events occurring in the combustor.
In particular, the model by Kelly et al [3] containing detailed chemistry
for methane/air reactions can predict pollutant emissions. Detailed kine-
tics, unfortunately, are not yet available for most other practical fuels.
Since emission characteristics are important considerations for catalytic
combustor design, in the absence of more detailed information, a model con-
sisting of several key semi-global chemical steps is proposed.
For fuel-lean catalytic combustion using nitrogen-free hydrocarbon
fuels, the NOX emission is negligible because of low flame temperatures.
The only pollutants needed to be considered are unburnt hydrocarbons (UHC)
and carbon monoxide (CO). For this reason, the two semi-global chemical
reactions used in the gas phase are:
118
-------
(A) CnH* + (ntf ) Oz — » n CO t £ H20
*<*«* '
- - c, f>
U)
(B) CO 1- i 0^ — * C02
In describing the flow in the reactor channel, we follow previous
investigators to use the plug flow approximation. For transient, in
general, the fluid properties are functions of x, the distance along the
channel axis and t, the time. For example, density of the gas will be
expressed as p(x,t). More detailed derivation of the gas-phase conserva-
tion equations can be found in Ref. 9. The nondimensional forms of these
equations are summarized in the following (subscript * denotes dimensional
quantity and without * denotes nondimensional quantities).
Continuity: pu = p(o,t) u(o,t) (3)
Momentum: p = p(o,t) (4)
Energy: f U - t JH ( T' Ts ) =
-------
Oxygen:
(o/o) WCO ' (8)
Equation of state: p = pT (9)
The reaction rates are given by
(10)
HO
The definition of B and B can be found in the nomenclature.
In deriving Eqns (5-8), heat conduction and mass diffusion in the
axial direction are neglected because the Peclet number based on typical
gas velocity (>10 M/sec) is much greater than unity [5]. The dimensionless
lateral heat and mass transfer coefficients J and J_., in Eqns (5-7) are
H Dl
derived from Nusselt number calculation in entrance flow in tube of constant
surface temperature [10]. Following Ref. 3, a modification is made at the
entrance point based on local stagnation flow estimation. The use of
entrance flow transport properties, rather than fully-developed constant
value, make a great difference in the temperature and species distributions
in the flow entrance region. More detailed discussion can be found in
Ref. 9.
The above system of equations can also be .applied to the after-bed
space where only gas-phase reactions occur. This is done by putting J and
J . equal to zero in Eqns (5-7) and by properly changing the mass flux per
unit area in Eq (3) due to area change.
II.3 Unsteady Solid with Thin-Wall Substrate
When the substrate wall is thermally thin, the solid temperature can
be regarded as a function of axial distance and time only, i.e., Tg(x,t).
The energy balance of the solid, including the solid surface, results in
120
-------
the following equation:
*l
J
(13)
In Eq (13) the heat conduction in axial direction in the solid is
neglected in comparison with the heat transfer rate across the gas channel.
A more detailed discussion of this approximation can be found in Ref . 9.
Two semi-global catalytic surface reactions are assumed. They are
(0 £„//„, t (v+~) 02 —-> >iCOz t f HZ0 (14)
with the corresponding surface reaction rate J* given by
i* - C;
and
(D) CO + 2 Qz -
with T* = r* frnl
Jtf ^ L^^J^ -
In surface reaction (C), the hydrocarbon is oxidized to form C02 and
H90, not CO such as in gas-phase reaction (A). This is based on the present
available experimental evidence that in oxygen-rich system, only C02» not
CO, is found on or close to the catalytic surface [6-8].
Eq (13) is nondimensionalized by first defining a proper reference time
scale T*.
tfcl
«,* (16)
and
.
t a t*/t
r = Nu* %*
121
-------
Eq (13) becomes
9Tj
TT l
(18)
Neglecting transient accumulation on the surface, the surface hydro-
carbon and carbon monoxide are given by
(19)
II.4 Initial and Boundary Conditions
The system of equations (3-12, 18-20) needs initial and boundary
conditions. Initial conditions (at t = 0) are required for T , (If )
,» S v tiL* S
(^PQ) as a function of x and upstream boundary conditions (at x = 0)
should be specified for^HC, ^CQ> ^Q2, T, p and u as a function of time.
These conditions depend on the types of transient input which can vary
with engine designs and mode of operations. In the present paper, two
simple types of transient are investigated. They are catalytic combustor
ignition and exponential change of fuel flow rate.
1. Ignition
Catalytic combustor ignition is modeled by specifying the substrate
temperature at time t = 0. This temperature can be the result of external
heating of the substrate, say, by a torch. The fuel and air are then
supplied at upstream. The model is intended to describe the events from
t = 0 to the final steady state. Mathematically, the conditions used are
122
-------
T (x,0) = constant (specified)
1fHC,s(x>°> = °
T(0,t) = 1
c(0,t) = 1 (21)
p(0,t) = 1
u(0,t) = 1
2. Exponential Variation of Fuel-Flow Rate
In this category, we start with a given steady state, the upstream
fuel-flow rate (or equivalence ratio) is then changed to another level in an
exponential fashion. We are interested in the transient combustor response
due to this unsteady upstream variation. Mathematically, the surface dis-
tribution of T (x,0), 'k (x,0) and \ (x,0) are given by the initial
S 0 rlLt 5 S O \s\j j S
steady-state computation. The upstream fuel/air equivalence ratio is des-
cribed by
(0,t) - 2 + (^ - <|>2) exp (-gt) (22)
where (j>1 = initial upstream equivalence ratio
4>2 = final upstream equivalence ratio
$ = rate constant
The corresponding upstream fuel variation, >Hr(^»t)» can ^e obtained
directly from Eq (22). The conditions for T(0,t), ^(O^)' ^O2(0>t)'
p(0,t) and u(0,t) are held constant as specified in Eq (21).
II.5 Numerical Solution Procedure
Although most dependent variables in this problem are function of both
x and t which, in general, results in partial differential equations, the
assumption of quasisteady gas phase and thermally thin substrate greatly
simplifies the mathematical property of the system and results in simpler
solution procedure.
123
-------
The gas-phase differential equations (5-7) have only first derivative
in x with t as a parameter, the solid equation (18) has only derivative in
t with x as a parameter; therefore, they can be integrated at each time (t)
or position (x) as ordinary differential equations using Runge-Kutta method.
The procedure is started by first integrating Eqns (5-7) forward in x from
upstream to the end of combustor for T,^RC,^CQ using the solid surface
quantities specified by the initial conditions, then Eq (18) is integrated
for one time step (At) to find Tg(x, At) and Eqns (19,20^ are solved alge-
braically to find^RC s(x,At) and ^co>s(*, At). These surface quantity
are substituted back to Eqns (5-7) to'start another cycle of integration.
III. Computed Results
III.l Steady State
Although the main objective of this paper is for transient response,
the unsteady calculation has to be based on a realistic steady combustion
model. Detailed steady-state predictions and their comparison with experi-
ment can be found in Ref. 9. Here several key features of the steady model
will be mentioned as they are important to the discussion of nonsteady
results.
Table II lists the values of chemical kinetic constants for reactions
A to D used in this paper. The hydrocarbon fuel chosen is propane. Its
semi-global kinetics are from Edelman with the pre-exponential factor
adjusted to match Dryer's data at 1000°K [12]. The CO oxidation data are
from Ref. 11. The activation energies for reactions C and D are obtained
from Refs. 8 and 13 respectively with adjusted pre-exponential factors.
It should be noted that the surface reaction rates contain information on
catalyst type and loading density as well as the washcoat properties;
therefore, they can vary from one combustor to the other. The purpose of
this computation is not to match a particular set of experimental combustor
data, but, rather, after choosing a reasonable set of kinetic constants,
to study the effects of other combustor design and operating parameters.
The design parameters considered are d*, channel hydraulic diameter,
A*/A*, the ratio of open-to-close areas, L*, the catalytic bed length and
s
L* , the total catalytic combustor length (catalytic bed length plus after-
EG
bed length) required to reach the emission goal in steady-state operations.
The emission goal is chosen to be 1.64 S/kg of fuel for UHC and 13.6 g/kg
124
-------
of fuel for CO [14]. The operating parameters include $, fuel/air equiva-
lence ratio, p*, pressure, T* , combustor inlet temperature and u* ., the
in ret
reference velocity of the approaching gas.
Figure 2 gives the profiles for one sample calculation. In this case,
the propane fuel is oxidized quickly on the catalyst surface upon entering
into the reactor channel and resulting in a high surface temperature, the
gas temperature is raised by heat transfer from the substrate, gas-phase
oxidation of propane is accelerated to form CO, and CO, in turn, is oxidized
by both gas-phase and surface reactions to C0?. Actually, two cases are
shown in Fig. 2. In one case, the reactor bed length is 4 cm, in the other,
8 cm. When the bed length is cut short at 4 cm, the surface reactions for
hydrocarbon fuel and CO stop there. This results in a higher CO concentra-
tion as shown by the dotted curve. The hydrocarbon curve also shows a dif-
ference after 4 cm but the difference is too small to be shown clearly in
this figure. The gas temperature is also lower in a portion of the combustor
for the 4-cm case.
Figure 3 gives the combustor length required to reach the emission goal
vs. catalytic bed length for three inlet temperatures. The three curves
have different fuel/air equivalence ratio so that their adiabatic flame
temperatures are approximately the same (1370°K < T*, < 1392°K).
af
The effect of the approaching flow reference velocity is shown in
Fig. 4. When L* and L* are divided by the reference velocities, the three
IjLr
velocity cases approximately collapse into one curve. Therefore the resi-
dence time should be a good correlation parameter for catalytic combustors.
III.2 Ignition Transient
One of the major concerns of the catalytic combustor application to
transportation gas turbine engines is that the response time might be long.
The computed ignition transient results to be presented are combustor
responses to a stepwise increase of fuel flow rate (zero to a fixed value).
Therefore, they can be used as a guidance in our search for "fast-response"
catalytic combustor designs.
Some indication of the order of magnitude of the response time can be
obtained from the definition of the characteristic time T* as defined in
eq. (16). Small T* requires small thermal inertia, A*p*C*. If the substrate
s s s
125
-------
density and heat capacity are fixed, small T* can be achieved with small
solid crossectional area A*, see Fig. 1 for definition of A*. Table III
s s
lists the values of T* for a range of cell diameter, d*, and open-to-close
area ratio, A*/A*. It should be noted that there are other parameters such
5
as catalytic bed length and channel diameter which will influence the
response time and they are not contained in T*. This will b.e discussed
later as we see the computed results.
The combustion efficiency at the end of the combustor will be used to
characterize the combustor response time. There are two ways to define the
combustion efficiency. One is the so-called "carbon-balanced" efficiency,
^CB ^*^ » where the combustion inefficiency is measured by the emission
levels of CO and UHC and the energy carried away by them. The other effi-
ciency is the ordinary one defined by the enthalpy rise across the com-
bustor divided by the chemical energy available. If the gas specific heat
is approximated as a constant, this efficiency, n , is given by
~ Tin^ where T is measured at the same location as the effi-
ciency. For a perfectly insulated catalytic combustor in steady state,
^CB and nT are the same' However, the two are different in transient
operations when there is a thermal lag in the substrate.
For the following transient ignition calculation, the initial surface
temperature condition is given by T (x,0) = 1, see Eq. (21). Also, the com-
P
bustor length has to be specified. From Fig. 3, we see that the required
combustor length is a function of catalytic bed length. For 1000°K inlet
temperature, two bed lengths, 4 cm and 8 cm, are chosen for transient
ignition calculation. Fig. 5 shows the combustion efficiencies as a function
of non-dimensional time. Looking at the curves for n (efficiency based on
temperatures), the shorter bed (4 cm) combustor has a faster response than
that of the longer bed (8 cm) . If we take the time for n to reach 70% as
an indication of characteristic response time, for the 4-cm bed, t = 1.1
Ch
and for the 8-cm bed, t = 2 ; namely, the response time of 4-cm bed
combustor is only half of that of the 8-cm bed combustor. From the consi-
deration of steady-state operation alone, one would choose the 8-cm cataly-
tic bed since according to Fig. 3, this produces the shortest required
combustor length ( 12 cm) . However, combining transient and steady-state
considerations, the 4-cm bed combustor may be a better choice as its response
126
-------
time is 50% of that of the 8-cm bed, with the total combustor length only
slightly longer (7.5%).
To understand why longer reactor bed results in a longer response time,
surface and gas temperature distributions are plotted in Fig. 6 as a function
of time. From the substrate temperature distribution we see that during
ignition transient, the solid temperature is higher in the flow entrance
region as a result of higher mass transport and surface reaction rates.
Comparing the solid and gas temperature histories for the case when catalytic
bed length is 8 cm, we see that at t = 0.5, the solid is transfering heat to
the gas in the upstream region in the bed. This raises the gas temperature
and accelerates the gas phase reaction so that at x* = 3.5 cm, the gas tempe-
rature becomes higher than the substrate temperature. Heat is then trans-
ferred from the gas to the solid in the downstream portion of the reactor
channel. If the heat loss rate is greater than heat generation rate due to
gas phase reactions, then the gas temperature will drop as shown in Fig. 6.
If we cut the reactor bed short at 4 cm, portion of this heat loss can be
eliminated and the downstream gas temperature (x* > 4 cm) will be higher
as indicated by the dotted curves in Fig. 6. This explains why a shorter
reactor bed can have a faster response.
The effect of catalytic channel diameter has also been studied. In
varying the diameter, the substrate cross-sectional area, A*, is kept con-
s
stant so that the characteristic time T* is constant. The lower graph of
Fig. 7 shows the variations of A*/A* and the minimum combustor length
s
required to reach emission goal. The increase of L* with d* is small over
EG
the range of diameters studied. This is partly because A*/A* is larger fpr
S
larger d* which leads to a lower gas channel velocity and a greater gas
residence time in the catalytic bed. The upper graph in Fig. 7 gives the
combustor response time as a function of channel diameter. The decrease
of response time with increasing diameter is quite drastic. Take 3.6 mm
channel, for example, the time to reach 70% efficiency is only one third
of that for the 1.8 mm channel. Translated into dimensional time using
the numbers in Table III, the response time for the 3.6 mm diameter chan-
nel combustor is about 1.5 seconds.
For the cases studied so far, the combustor inlet temperature is
1000°K. The computed results show that faster combustor response is
127
-------
favored by thin substrate, short catalytic bed and large channel diameter.
Shorter bed and larger channel, however, tend to increase the required total
combustor length. Therefore, a combined steady and unsteady consideration
is needed for a realistic design. In the calculations just presented, it is
possible to reduce the response time significantly with only slight increase
of the combustor length. However, we have not had enough time to perform
more computation in the low inlet temperature cases so it is not certain
at this time that all the trends we mentioned above will be the same for
low combustor inlet temperatures. In particular, the effect of channel
diameter is uncertain.
The response time has been found to be invariant with reference velo-
city if the gas residence times in the reactor bed and in the after-bed
space are unchanged.
III.3 Response to Exponential Decrease of Fuel Flow Rate
Here we study the combustor response from one steady combustion state
to the other due to an exponentially decreasing upstream fuel/air equivalence
ratio, . Mathematically, <|> as a function of time is given by Eq. (22).
In the computation performed, , is chosen to be 0.24 and 4> is 0.20 with
rate constant 6 as a variable. The combustor length is 25 cm long with a
10 cm bed and bed channel diameter of 1.8 mm. The reference velocity is
fixed at 10 M/S and the combustor inlet temperature is 800°K.
The ordinate in Fig. 8 is a temperature difference ratio,
|T - T±nj /[laf(t) - T±n^ , where Taf(t) is the adiabatic flame temperature
corresponding to the instantaneous upstream equivalence ratio <|>(t) and T
is the gas temperature at the combustor exit. When the transient is very
slow so that the whole combustor is in a quasisteady state (including the
substrate), this temperature difference ratio is the combustion efficiency.
For equal to 0.2 and 0.24, combustion efficiencies are nearly 100%. In
transient due to fuel flow rate reduction, this temperature difference
ratio exceeds one as shown in Fig. 8. This means that the gas in the bed
channel is receiving heat from the substrate in excess of the quasisteady
value. This of course is reasonable for the fuel reduction case because
of the thermal lag in the substrate. The departure of the temperature dif-
ference ratio from the steady-state combustion efficiency (one in Fig. 8)
128
-------
is a good measure of the extent of unsteadiness in the combustor.
Referring to Fig. 8, the curve with g = 100 peaks very early and is
a good approximation of the combustor response to a stepwise decrease of
fuel flow rate. The curve decays monotonically with time after the peak,
the decay rate is basically inversely proportional to the response time as
discussed in the ignition transient study. For slower transient input
(smaller 3), the peak is smaller and occurs at a later time which means
smaller transient effect but a longer transient duration as shown by the
various curves in Fig. 8.
More transient computation is being carried out and will be reported
at a later time.
129
-------
References
1. DeCorso, S.M. and Carl, D.E., "Structural Analysis of a Preliminary
Catalytic Ceramic Design," Proceedings: Third Workshop on Catalytic
Combustion, EPA-600/7-79-038 (1979).
2. Cerkanowicz, A.E., Cole, R.B. and Stevens, J.G., "Catalytic Combustion
Modeling: Comparison with Experimental Data," ASME paper (1977).
3. Kelly, J.T., Kendall, R.M., Chu, E. and Kesselring, J.P., "Development
and Application of the PROF-HET Catalytic Combustor Code," paper
presented at the 1977 Fall Meeting Western States Section, The
Combustion Institute.
4. Kendall, R.M., Kelley, J.T., Chu, E.K. and Kesselring, J.P., "An
Analysis of Catalytic Combustion in Monolithic Honeycomb Beds,"
Proceedings: Third Workshop on Catalytic Combustion, EPA-600/7-79-038
(1979)
5. Ablow, C.M. and Wise, H., "Theoretical Analysis of Temperature and
Composition in a Catalytic Monolith Reactor," ibid.
6. Schwartz, A., Holbrook, L.L. and Wise, H., "Catalytic Oxidation
Studies with Platinum and Palladium," Journal of Catalysts, 21, 199-207
(1971).
7. Anderson, D.N., "Preliminary Results from Screening Tests of Commercial
Catalysts with Potential Use in Gas Turbine Combustors. Part I. Furnace
Studies of Catalyst Activity," NASA TMX-73410 (1976).
8. Marteney, P.J., "Determination of the Rate of Heterogeneous Reaction on
Catalytic Surfaces," Proceedings: Third Workshop on Catalytic Combustion,
EPA-600/7-79-038 (1979).
9. T'ien, J.S., "Catalytic Honeycomb Combustor: "Steady-State Model and
Comparison with Experiment," to be presented as Paper No. 80-1289 at
the AIAA/ASME/SAE 16th Joint Propulsion Conference, June 30-July 2, 1980,
Hartford, Connecticut.
10. Ka,ys, W.M., Convective Heat and Mass Transfer, McGraw-Hill, New York,
1966.
11. Dryer, F.L. and Glassmap, I., "High Temperature Oxidation of CO and CH ,"
14th Inter. Symp. on Combustion, The Combustion Institute,
Pittsburgh, Pa., 1973.
12. Dryer, F.L. and Classman, I., "Combustion Chemistry of Chain Hydro-
carbons," in Alternative Hydrocarbon Fuels: Combustion and Chemical
Kinetics, Vol. 62, Progress in Astronautics and Aeronautics (1978).
Also discussion section, p. 295.
130
-------
13. Kuo, J.C.W. and Morgan, C.R., "Mathematical Models of the Monolith
Catalytic Converter," SAE Transaction, V. 80, 1971, #710289.
14. Anderson, D.N., "Performance and Emissions of a Catalytic Reactor
with Propane, Diesel, and Jet A Fuels," NASA TM-73786 (1977).
15. Anderson, D.N., Tacina, R.R. and Mroz, T.S., "Performance of a
Catalytic Reactor at Simulated Gas Turbine Combustor Operating
Conditions," NASA TM X-71747 (1975).
131
-------
Table I. Estimate of Transient Time Scales in Catalytic Monolith Combustor
Time
Definition
Estimated Magnitude
Gas residence time in
reactor bed
Gas residence time in
after-bed space
Heat/mass transfer time
between gas and solid
surface in channel
Chemical reaction times
(gas-phase and heteroge-
neous)
Time for temperature wave
to reach the substrate
center plane
Substrate heat-up
time
Transient input
time
L*/u* ,
ref
L*,/u* .
af ref
1 1
d*
Nu IT Act*(0,0)
d*
s
a*
s
. A*p*C*
1 1 s s s
Nu^ IT k*(0,0)
I 1/3 for exponential
I input; see Eq. (22)
3-15 msec
0-15 msec
.5 - 7.5 msec for
d* < 3.6 mm
Variable depending on the
local temperature and
reactant's concentrations
0.5 - 25 msec for
d* < . 2 mm
s
0.5 sec - 20 sec
variable, specified
132
-------
Table II. Chemical Kinetic Constants
Reaction (A)
Reaction (B)
Reaction (C)
Reaction (D)
C* = 1.5 x 105 C* = 0.71xl014
C* = 2.5x10'
C* = 10-
<* = 0.3
E* = 10
(kcal/mole)
E* = 17.8
(kcal/mole)
c*2 = 0.5
B2-l
0.25
E* = 24
(kcal/mole)
65 = 0.5
E2 = 40
(kcal/mole)
133
-------
Table III. Estimate of Substrate Heat-up Characteristic Time
T*
k*(0,0)
cal/cm sec°K
.000135
.000135
.000135
.000135
.000135
.000135
, A*p*C* n
_ 1 1 s s s 1
A*p*C*
s s s
~ Nu^ ir k*(0,0) 11.498 k*(0,0)
PS* 3
g/cm
2.5
2.5
2.5
2.5
2.5
7.1
C*
cal/g°K
.23
.23
.23
.23
.23
.145
d*
mm
1.4
1.8
2.4
3.0
3.6
1.8
A* /A*
2
2
3.56
5.56
8
5
As"
2
cm
.0077
.0127
.0127
.0127
.0127
.0051
T*
sec
2.85
4.7
4.7
4.7
4.7
3.38
.000135
(Kanthai metal)
2.5 .23
3.6
.051
19.0
134
-------
SUBSTRATE
CHANNEL
d*
////////////V 77///TV TV.
u;
REF
A*
Ms
AFTER-BED SPACE
Figure 1. Schematic Drawing of One Cell Unit
1,0
0
- 0,5
0
Figure 2. Detailed profiles showing the effect of catalytic bed length.
Operating conditions: p* = 3 atm, T*n = 1000°K, u*gf = 10 M/S,
= 0.15, d* = 1.8 mm and A*/A* = 2
s s
135
-------
20
u>
15
10
0
1370 K
-------
100
80
60
.CB
20
= IOOOK
irREF - 10
0
0
Figure 5. Combustion Efficiencies as a Function of Time during
Ignition Transient. p* = 3 atm, <}> = 0.15 and
A*/A* = 2
s
137
-------
oo
8
10
Figure 6. Gas and Substrate Temperature History during Transient Ignition for Two Bed Lengths.
Same Operating and Design Parameters as Those in Fig. 5
-------
-p
6-8
R
1,2
1,0
,8
,6
,2
20 r
(CM)
10 -
5
1 8
- 6
AVA;
- 2
Figure 7. The Effect of Channel Diameter on Combustor Response Time and
Required Combustor Length. p* = 3 a tin, T* = 1000°K,
u* = 10 M/S, catalytic bed length = 4 cm
139
-------
O
1,15 -
T- T
IN
- T
IN
1,10
1,05
1,00
10 CM BED
25 CM CQMBUSTOR
M
f * = 3 ATM
Figure 8. Temperature Difference Ratio vs. Time during an Exponential Reduction of Fuel/Air
Equivalence Ratio
-------
THE EFFECT OF CATALYST LENGTH
AND DOWNSTREAM REACTION DISTANCE
ON CATALYTIC COMBUSTOR PERFORMANCE
By:
David N. Anderson
NASA Lewis Research Center
Cleveland, Ohio 44135
ABSTRACT
A study was made to determine the effects on catalytic combustor per-
formance which resulted from independently varying the length of a catalytic
reactor and the length available for gas-phase reactions downstream of the
catalyst. Monolithic combustion catalysts from three manufacturers were
tested in a combustion test rig with no. 2 diesel fuel at an inlet-air
temperature of 1000 K, a pressure of 3 x 10 Pa, reference velocities of 10
and 15 m/s, and adiabatic combustion temperatures of 1200 to 1500 K.
Catalytic reactor lengths of 2.5 and 5.4 cm, and downstream gas-phase reaction
distances of 7.3, 12.4, 17.5, and 22.5 cm were evaluated. Measurements of
carbon monoxide, unburned hydrocarbons, nitrogen oxides, and pressure drop
were made.
The catalytic-reactor pressure drop was less than 1 percent of the
upstream total pressure for all test configurations and test conditions.
Nitrogen oxides and unburned hydrocarbons emissions were less than 0.25 g
NO /kg fuel and 0.6 g HC/kg fuel, respectively. The minimum operating
temperature (defined as the adiabatic combustion temperature required to
obtain carbon monoxide emissions below a reference level of 13.6 g CO/kg fuel)
ranged from 1230 K to 1500 K for the various conditions and configurations
tested. The minimum operating temperature decreased with increasing total
(catalytic-reactor-plus-downstream-gas-phase-reaction-zone) residence time but
was independent of the relative times spent in each region when the
catalytic-reactor residence time was greater than or equal to 1.4 ms.
141
-------
INTRODUCTION
The study reported here was performed to determine how the emissions and
pressure drop of a catalytic combustor are affected by changes in the length
of the catalytic reactor and in the length available for gas-phase reactions
downstream of the catalytic reactor.
A catalytic combustor contains three major components: (1) the fuel-
preparation region, in which fuel is injected and premixed with the inlet
airstream; (2) the catalytic reactor, in which both surface and gas-phase
combustion reactions are initiated; and (3) the downstream gas-phase reactor
in which gas-phase reactions continue after initiation in the catalytic
reactor. The importance of gas-phase reactions within the catalytic reactor
has been discussed in several reports (ref. 1-6). The study of reference 7
also showed that these reactions can be significant downstream of the
catalytic reactor. In that study the emissions levels were strongly dependent
on the distance downstream from the catalytic reactor to the gas sampling
probe. The observations were based on tests of a single length of catalytic
reactor at an inlet-air temperature of 800 K. The present paper reports the
effects of changing both the catalytic and the gas-phase reactor lengths at an
inlet-air temperature of 1000 K.
The work presented here was performed as support for the Department of
Energy's Gas Turbine Highway Vehicle Systems Project. Test conditions
simulated the combustor conditions in a regenerative automotive gas turbine
engine. The inlet-air temperature was 950 - 1000 K, inlet pressure, 3 x 10
Pa, reference velocity, 10 and 15 ra/s, and the combustor exit temperature was
varied from 1200 to 1500 K. Catalysts from three manufacturers were tested in
a 12-cm-diameter test duct with catalytic-reactor lengths of 2.5 and 5.4 cm.
The effect of gas-phase reactor length was determined by varying the axial
location of an exhaust-gas sampling probe from 7.3 to 22.5 cm downstream of
the catalyst exit plane. Emissions of CO, C02, UHC, and N0x were measured
as well as the catalytic-reactor pressure drop.
142
-------
APPARATUS
TEST RIG
The test rig is illustrated in figure 1. It was fabricated from 15.2-cm
(6-in nominal) diameter stainless-steel pipe. Carborundum T30R Fiberfrax tube
insulation with a 12-cm inside diameter was inserted inside the pipe. The use
of internal insulation served to maintain the inlet-air temperature near the
preheater discharge value and to reduce the heat loss from the test section.
Test-section metal temperatures were calculated to be less than 600 K.
The inlet air was indirectly preheated to approximately 1000 K for all
tests. This temperature was measured in a plane just upstream of the fuel-
injection location with an array of 12 Chromel-Alumel thermocouples mounted in
a flange. The thermocouples were positioned in the duct so that each measured
the temperature of an equal-area segment of the cross-section. The average
inlet temperature was then calculated as the numerical average of these twelve
thermocouples.
The test-section pressure was controlled by a back-pressure valve to
3 x 10 Pa for all tests. The test-section-inlet pressure was measured at a
tap located in the flange containing the inlet thermocouples. The airflow
rate entering the test section was measured by a standard AS ME orifice. Fuel
flow rate was determined with a Flo-tron linear mass flowmeter. This device
employs 4 matched orifices arranged to form a hydraulic Wheatstone Bridge
network. A constant-volume pump circulates flow through this network, and the
mass flowrate is related to the pressure drop across the network.
FUEL INJECTOR AND PREMIXING ZONE
A multiple-conical-tube fuel injector of the type developed by Tacina
(ref. 8, 9) was used; it is pictured in figure 2. Although it was designed
with two sets of fuel tubes for use with either gaseous or liquid fuels, only
liquid fuel (no.2 diesel) was used. The array of 21 conical tubes distributed
the airflow uniformly over the duct cross section. No. 2 diesel fuel was
introduced into the upstream end of each of these cones through a 0.5-mm-
inside-diameter fuel tube. The 21 fuel tubes were all the same length (25 cm)
to insure that equal fuel flow rates discharged into each of the air cones.
143
-------
Because of the good initial dispersion of fuel across the duct, excellent
fuel-air ratio uniformity is achieved with this type of injector. Tacina
(ref. 8) has reported that 23 cm downstream of the fuel-injector inlet the
fuel-air ratio profile was uniform to within _+ 10 percent of the mean. The
injector tested in reference 8 was designed with 75-percent blockage to the
airstream while the injector used in the present study had 87-percent block-
age; thus, the performance of the injector used here can be expected to be at
least equal to that of reference 8. In the present experiments, the mixing
distance from the point of fuel injection to the inlet plane of the catalytic
reactor varied from 27.5 cm to 32.5 cm depending on which catalytic reactor
was tested. Because the fuel-air ratio profile should already be nearly
uniform for even the shortest of the mixing lengthss additional length will
have little effect on the fuel-air distribution at the catalytic-reactor inlet
and, therefore, little effect on combustion performance.
The pressure drop across the fuel injector was measured with a differ-
ential pressure transducer connected between the inlet-pressure tap and one
located at the premixing-region thermocouple station.
A single Chrome1-Alumel thermocouple was inserted to a depth of about 1 cm
into the flow in the pretnixing zone downstream of the fuel injector. If
burning had occurred in the premixing zone, this thermocouple would have
triggered a relay to shut off the test-section fuel supply. No burning in the
premixer was observed during these tests.
CATALYTIC REACTORS
Catalytic reactors were prepared using catalyst elements from 3 manu-
facturers: Matthey Bishop, Inc. (MBI), Oxy-Catalyst, Inc. (OCI), and UOP,
Inc. (UOP). Reactors were either 2.5- or 5.4-cm long depending on whether one
or two catalyst elements were used. Figure 1 was drawn with a two-element
reactor in the test section. The two elements were each 2.5-cm long and were
separated by a 0.31-cm gap. The 2.5-cm-long reactors were made by removing the
upstream element from the two-element reactors.
The removal of an upstream element produced a slightly longer premixing-
zone length. For the MBI and OCI reactor tests, premixing-zone lengths of
27.5 and 30.3 cm were used with two-element and one-element reactors,
respectively. These lengths are shown in figure 1. Tests with the UOP
144
-------
catalysts were made with premixing-zone lengths of 32.5 and 35.3 cm for the
two-element and one-element reactors, respectively.
Table I provides a description of the catalytic reactors tested. The MBI
reactors used metal substrates. They were made by Matthey Bishop from
Fecralloy foil which was corrugated and wound into a cylinder. The resulting
2
cells had a sine-wave shape, a cell density of 62 per cm , and an open area
of 93 percent of the duct cross-sectional area. The MBI-2.5 reactor was a
single element of this substrate with a palladium catalyst. The MBI-5.4
reactor had a second, platinum-catalyst element placed in front of the
palladium-catalyst element. Both elements were previously unused.
The two Oxy-Catalyst elements were identical: both used a palladium
catalyst on a Corning cordierite square-cell substrate. The open area was 63
percent of the duct cross-sectional area and the cell density was 47 per
2
cm . The OCI-2.5 reactor used one of these elements and the OCI-5.4 used
two. These elements had also been unused before this testing.
The UOP elements also used the Corning cordierite square-cell substrate.
Both elements were identical with a proprietary noble-metal catalyst. These
elements had been furnace aged a total of 300 hours at a temperature of 1400 K
prior to testing.
Single platinum-vs-platinum-13-percent-rhodium thermocouples measured the
centerline gas temperature upstream of the reactor and in the gap between the
elements. The catalytic reactors were held in place by an array of 12
platinum-vs-platinum-13-percent-rhodium thermocouples at the downstream plane.
The pressure drop across the catalytic reactor was measured with a differ-
ential pressure transducer connected between the premixing-zone pressure tap
and one located 7.3 cm downstream of the catalytic reactor as shown in figure
1.
GAS-PHASE REACTOR
Downstream of the catalytic reactor the combustion products were sampled
with a single-point water-cooled probe at the duct centerline. The length of
the gas-phase reactor was determined by the axial position of the probe. Data
were obtained with the probe at each of four locations: 7.3, 12.4, 17.5 and
22.5 cm downstream of the catalytic reactor. The probe is shown in the
22.5-cm location in figure 1.
145
-------
The gas-sampling probe was constructed with a single 0.6-cm-inside-
diameter sampling passage. The sample line was electrically heated to
maintain a temperature of 410-450 K, and it was 18-m long and 0.5 cm in
diameter. The continuous-flow samples were analyzed to determine the con-
centration of nitrogen oxides using a chemiluminescent analyzer, unburned
hydrocarbons using a flame-ionization detector, and carbon monoxide and carbon
dioxide using nondispersive infrared analyzers.
Temperatures were measured in the gas-phase reactor with platinum-vs-
platinum-13-percent-rhodium thermocouples in each of the 4 gas-sampling planes.
146
-------
COMPUTATIONS
REFERENCE VELOCITY
The reference velocity was computed from the measured mass flow rate, the
2
average inlet-air temperature, the duct cross-sectional area (113 cm ) and
the test-section-inlet pressure (3 x 10 Pa).
EMISSION INDEX
The emissions were measured as concentrations in ppm by volume which were
converted to emission indexes using the expression
+
C x 1(T3
x
Mr
\
where
(E.I.) emission index of specie x, g /kgf ,
X XIU6 J.
C concentration of specie x, ppm V
X
f fuel-air weight ratio, (kg/s)f ^(kg/s) .
M molecular weight of specie x, g /mole
M molecular weight of combustion products, g , /mole ,
p ° "products products
COMBUSTION EFFICIENCY
The difference between the measured and equilibrium emissions of carbon
monoxide and unburned hydrocarbons represents available chemical energy which
has not been released in the combustion process. For the conditions of this
study, the equilibrium emission indexes of unburned hydrocarbons and carbon
monoxide are negligible. Futhermore, measured unburned hydrocarbons emissions
resulted in emission indexes less than 0.6 g HC/kg fuel and, therefore,
contributed little to inefficiency. Combustion efficiency was consequently
determined from the carbon monoxide emissions alpne:
147
-------
/(HV)CO \
EFF = 100 - 0.1 (E.I.)co
\(HV)fuel/
= 100 - 0.0235 (E.I.)CO
where
EFF combustion efficiency, percent
(HV) lower heating value of CO, 1.01 x 10 J/kg
Ou -*
(HV) lower heating value of no. 2 diesel, 4.30 x 10 J/kg
FUEL-AIR RATIO
The fuel-air ratio was determined both from the metered fuel flow and
airflow rates and by making a carbon balance from the measured concentrations
of CO, CO., and unburned hydrocarbons. The carbon-balance fuel-air ratio
had the advantage that it was the local fuel-air ratio at which the emissions
data were obtained.
ADIABATIC COMBUSTION TEMPERATURE
The adiabatic combustion temperature was calculated from the carbon-
balance fuel-air ratio using the equilibrium computer program of reference 10.
148
-------
RESULTS AND DISCUSSION
The carbon-balance fuel-air ratios were typically about 95 percent of
those determined from the metered fuel flow and airflow rates. This dif-
ference probably results from the use of a single gas-sampling probe at the
duct centerline, and it indicates that fuel-air premixing, although good, did
not result in a perfectly uniform fuel-air ratio over the entire cross
section. It also indicates that plug flow continued downstream of the
catalytic reactor so that there was little mixing across the duct. Thus, the
emissions measured by the probe were determined by local, rather than average,
conditions. The local adiabatic combustion temperature is that calculated
from the carbon-balance fuel-air ratio, and it will be used in the present-
ation of results.
No measurable temperature rise along the length of the gas-phase reactor
occurred, and the uncorrected catalytic-reactor exit temperatures measured at
the centerline were typically only about 25 K less than the adiabatic com-
bustion temperatures. The performance of the various lengths of catalytic and
gas-phase reactors will be compared on the basis of the measured emissions and
pressure drop.
EMISSIONS
Allowable emissions from passenger cars have been established by the 1970
Clean-Air Act Ammendments which specifies ultimate emissions standards of 0.25
g NO /km, 2.1 g CO/km, and 0.25 g HC/km. Emissions levels are determined by
X
collecting an exhaust sample during a standard metropolitan driving cycle
which simulates various transient operating conditions and includes few
periods of steady-speed motoring. Thus, there is no simple relationship
between the automotive emissions standards and the emissions measured during
steady-state tests, such as the ones performed in this study. However, the
emissions standards can be transformed into steady-state reference emission
indexes by multiplying the standards (in g/km) by an average cycle fuel
economy (in km/kg). For this study it was assumed that (Da gas-turbine
3 3
powered vehicle could achieve a fuel economy of 9.35 x 10 km/m (22
3
miles/gallon), (2) no. 2 diesel fuel density was 720 kg/m , and (3) the
149
-------
reference values should be based on only half the emissions permitted by the
standards. The third assumption provides some margin for unknown transient
and durability effects. The reference emission indexes were thus calculated
to be 1.6 g N02/kg fuel, 13.6 g CO/kg fuel, and 1.6 g HC/kg fuel. The CO
reference value represents about 0.32-percent loss in combustion efficiency,
and the HC level would result in an additional 0.16-percent loss. The NO
x
emissions in this study were never more than 0.25 g NO /kg fuel, and
X
unburned hydrocarbons were below 0.6 g HC/kg fuel. Therefore, only the carbon
monoxide emissions will be examined in detail.
The carbon monoxide emission indexes for each of the six catalytic
reactors are plotted in figure 3(a) - (f) as a function of the adiabatic
combustion temperature. The results for the four different gas-phase reaction
distances are given in each figure for reference velocities of 10 m/s (solid
curves) and 15 m/s (dashed curves). The combustion efficiency which cor-
responds with each CO emission index is shown as a second ordinate.
Results for the MBI-2.5 reactor are given in figure 3(a). Carbon monoxide
levels decreased nearly two orders of magnitude when the adiabatic combustion
temperature was increased by 150 - 200 K for all conditions. For both 10- and
15-m/s reference velocities, an increase in the gas-phase-reactor length from
7.3 to 22.5 cm also produced approximately a two-orders-of-magnitude re-
duction in CO at a fixed adiabatic combustion temperature. Emissions were
below the 13.6-g-CO/kg-fuel reference for adiabatic combustion temperatures
greater than 1265 K for the 22.5-cm-long gas-phase reactor at a reference
velocity of 10 m/s. When the velocity was increased to 15 m/s, the same
22.5-cm-long gas-phase reactor required temperatures of at least 1340 K to
achieve the reference level. For the 7.3-cm-long gas-phase reactor at 15 m/s,
a temperature of nearly 1500 K was required.
When the catalytic-reactor length was increased to 5.4 cm, the CO
emissions, shown in figure 3(b), were lower than those of the 2.5-cm-long
catalytic reactor at any given condition. For a 7.3-cm-long gas-phase reactor
and a reference velocity of 15 m/s, the adiabatic combustion temperature
required to achieve the reference CO level was reduced from nearly 1500 K for
the MBI-2.5 catalytic reactor (fig. 3(a)) to 1400 K. for the MBI-5.4 reactor
(fig. 3(b)). For the same 7.3-cm gas-phase reactor length at 10 m/s the
150
-------
MBI-2.5 reactor (fig. 3(a)) required an adiabatic combustion temperature of
1365 K to reach the reference level of CO emissions, while the M8I-5.4 reactor
(fig. 3(b)) required only 1345 K.
The same general trends were observed for the OCI-2.5 and OCI-5.4
catalytic reactors as shown in figure 3(c) and 3(d), respectively, and for the
UOP-2.5 and UOP-5.4 catalytic reactors as shown in figure 3(e) and 3 (f).
MINIMUM OPERATING TEMPERATURE
The minimum operating temperaure is defined as the adiabatic combustion
temperature at which all the emissions goals are met. As noted previously, in
this study the minimum operating temperature could be determined solely from
the CO emissions since unburned hydrocarbons were always well below the
reference level. The minimum operating temperature was determined for all of
the test configurations by noting the adiabatic combustion temperature at
which each of the CO curves of figure 3 crossed the reference emission index
level.
The effect of the total catalytic-plus-gas-phase-reactor length on the
minimum operating temperature for all six catalytic reactors can be seen in
figure 4. The open symbols and solid lines represent the minimum operating
temperatures of the 2.5-cm-long catalytic reactors, while the solid symbols
and dashed lines are for the 5.4-cm-long catalytic reactors. Results are
given at reference velocities of 10 m/s and 15 m/s for each catalytic-reactor
length.
From figure 4 it can be seen that within the experimental accuracy there
was no difference between the results for the MBI, OCI, and OOP reactors of
the same length. The fact that the UOP reactors, after having been aged for
300 hours at 1400 K, performed as well as the unaged catalytsts suggests a
potential for a long lifetime. The duration of these tests (typically 5-6
hours per reactor) was insufficient to determine if the MBI and OCI catalysts
would also be durable for long periods.
Figure 4 also shows that for each reference velocity and catalytic-reactor
length, an increase in the total catalytic-plus-gas-phase-reactor length
resulted in a lower value for the minimum operating temperature. For example,
for single-element reactors tested at a reference velocity of 10 m/s (solid
curve with open circles), an increase in total length from 10 cm to 13 cm
151
-------
reduced the minimum operating temperature from 1375 K to 1335 K. Similar
reductions were also observed with the 5.4-cm-long catalytic reactors (dashed
line, solid symbols) and at 15 m/s reference velocity (square symbols).
The reduction of 40 K in minimum operating temperature in the above
example was achieved by increasing the gas-phase-reactor length with a
constant catalytic-reactor length. If the total length had been increased
from 10 cm to 13 cm by increasing the catalytic-reactor length from 2.5 cm to
5.4 cm with a constant gas-phase-reactor length, the minimum operating
temperature for a reference velocity of 10 m/s would have decreased by 50 K.
(from 1375 K to 1325 K). Thus, the catalytic-length effect was slightly
greater than the gas-phase-length effect at a velocity of 10 m/s.
At a reference velocity of 15 m/s the catalytic-length effect on minimum
operating temperature was much greater than the gas-phase-length effect. The
increase in total length from 10 cm to 13 cm which was achieved by increasing
the gas-phase-reactor length while maintaining catalytic-reactor length
constant at 2.5 cm can be seen from figure 4 (solid curve, open squares) to
have reduced the minimum operating temperature by 50 K (from 1490 K to 1440
K). When the same increase in total length was obtained by increasing the
catalytic-reactor length from 2.5 cm to 5.4 cm while maintaining a constant
gas-phase-reactor length, the minimum operating temperature was decreased by
100 K (from 1490 K to 1390 K). At a velocity of 15 m/s, then, there was a
clear advantage, in terms of minimum operating temperature, to the use of a
longer catalytic reactor as opposed to the use of a longer gas-phase reactor.
Because residence time is a fundamental variable which includes both
length and velocity, it is of interest to analyze these results again on the
basis of changes in residence time. The minimum operating temperatures have
been plotted again in figure 5 to show the effect of the total combustion
residence time. The residence time was calculated from the total catalytic-
plus-gas-phase-reactor length by assuming that (1) the combustion temperature
increased linearly within the catalytic reactor so that the adiabatic com-
bustion temperature was reached at the exit of the first element for either
one- or two-element reactors, (2) the gas temperature throughout the rest of
the system was constant and equal to the adiabatic combustion temperature, (3)
the pressure was constant, and (4) the open area of the catalytic-reactor
channels was 93 percent of the duct cross-sectional area for the MBI reactors
152
-------
and 63 percent for OCI and UOP reactors. The residence times of the flow
through the various catalytic reactors at each of the two reference velocities
have also been indicated in figure 5 to assist in determining the effects of
catalytic-reactor residence time. The accuracy of the calculated residence
times was dependent on experimental measuring error as well as the validity of
these assumptions, and the values shown in figure 5 were estimated to be
accurate to within +^ 5 percent.
Figure 5 shows that the minimum operating temperature could be decreased
by providing more total combustion residence time. In addition to this
general trend, the results were segregated according to the catalytic-reactor
residence time into two curves. Catalytic-reactor residence times ranged from
0.9 ms to 4.1 ms for this experiment. The configurations which had the lowest
catalyst residence times (0.9 to 1.3 ms) also had the highest minimum operat-
ing temperatures. When the catalytic-reactor residence time was greater than
1.4 ms, minimum operating temperatures were about 50 K lower than those for
catalytst residence times of 0.9 to 1.3 ms. However, no further decrease in
minimum operating temperature was obtained when catalytic-reactor residence
times were increased to values as high as 4.1 ms at a constant total com-
bustion residence time. Thus, for these experimental conditions, gas-phase
residence time was as effective as catalytic residence time in determining
minimum operating temperature if the catalytic-reactor residence time was at
least 1.4 ms. By this time, apparently, gas-phase reactions were sufficiently
established within the catalytic reactor channels that additional catalyst was
unnecessary.
PRESSURE DROP
In addition to low-emissions capability, it is necessary for gas-turbine
combustors to operate with minimal pressure loss. The pressure-drop measure-
ments made in the present study do not necessarily provide values which would
be the same as those of a practical catalytic combustor; however, they do
indicate what contribution can be expected from the catalytic components. The
pressure drop as a percentage of the total upstream pressure is shown as a
function of the adiabatic combustion temperature for the MBI reactors in
figure 6(a), for the OCI reactors in figure 6(b), and for the UOP reactors in
figure 6(c).
153
-------
The measured pressure loss increased gradually with increasing adiabatic
combustion temperature. The two-element MBI and OCX reactors had pressure
losses which varied from 0.7 to 0.95 percent at 15-m/s reference velocity and
0.3 to 0.6 percent at 10 m/s. the UOP reactor produced lower values. The
losses measured across one-element reactors were typically about half the
values for two-element reactors* The flow in the catalyst channels was
laminar for all conditions studied and the measured effect of velocity on
pressure drop was consistent with a laminar-flow friction-loss mechanism
dominating over entry and exit losses. The pressure drop in the gas-phase
-3
reactor was calculated to be less than 2 x 10 percent for the test
conditions of this study.
The pressure drop across the multiple-conical-tube fuel injector was
typically about 1 percent at a reference velocity of 10 m/s and 2 percent at
15 m/s. These losses are higher than would be acceptable for many practical
applications; however, the fuel-injector blockage (87 percent) was greater
than was necessary. Tacina (ref. 8) achieved an excellent fuel-air ratio
profile from another injector of the same type which was designed with less
blockage (75 percent). He reported pressure losses to be only 0.25 percent at
a velocity of 10 m/s and 1 percent at 20 m/s.
For a low-blockage multiple-conical-tube fuel injector and a two-element
catalytic reactor, the combined pressure drop can be predicted to be about
0.75 percent at a reference velocity of 10 m/s and about 2 percent at 20 m/s.
Even at the higher velocity, this loss is acceptable for practial combustor
applications.
154
-------
CONCLUDING REMARKS
This study demonstrated that a 2.5-cm length of catalytic reactor followed
by a 7.3-cm length of gas-phase reaction distance is adequate to achieve
acceptable emissions at the following conditions: an inlet-air temperature of
1000 K, a pressure of 3 x 10 Pa, a reference velocity of 10 m/s, and an
adiabatic combustion temperature of 1400 K. Acceptable emissions could also
be obtained at adiabatic combustion temperatures lower than this by increasing
either the catalytic-reactor length or the gas-phase reaction distance
downstream of the catalytic reactor. In other words, the minimum operating
temperature (defined as the lowest adiabatic combustion temperature at which
the emissions are within acceptable levels) decreased with increasing total
combustion (catalytic-plus-gas-phase-reactor) residence time.
The relative time spent by the combustion gases in the catalytic and in
the gas-phase reactors was shown to affect the minimum operating temperature
only if the catalytic-reactor residence time was less than 1.4 ms. For
catalyst residence times greater than 1.4 ms, apparently, gas-phase reactions
were well established within the catalytic-reactor channels, and there was no
further contribution of surface reactions. Thus, a catalytic combustor de-
signed to operate at the conditions of this study would need to incorporate no
more catalytic-reactor length than was necessary to provide a 1.4-ms catalyst
residence time. Advanced automotive gas turbine engines may have combustor
inlet temperatures as high as 1340 K. For these inlet-air temperatures, even
less catalytic-reactor residence time may be required, and therefore, it may
be possible to increase reference velocities or reduce catalytic-reactor
length without penalizing emissions performance. Alternatively, lower
2
catalyst loadings or substrate cell densities below the 47-62 cells/cm used
in this study may be possible. Further evaluation is required to establish
the effect of higher inlet-air temperatures on the required catalytic-reactor
residence time.
The data reported were obtained at only one pressure, 3 x 10 Pa. While
this value is representative of part of the automotive-gas-turbine range (1.5
x 10 to 5 x 10 Pa), data are also needed at higher and lower pressures
before general conclusions can be drawn about required lengths of catalytic
reactors, catalyst loadings, or substrate geometries.
155
-------
Operation of catalysts at high temperaures for extended periods of time
will, of course, result in a loss of catalyst activity. Although one catalyst
included in these tests had received a 300-hour furnace aging at 1400 K and
showed no disadvantage relative to unaged catalysts, longer combustion aging
at higher temperatures will be experienced by catalysts used in catalytic
combustors. The subsequent loss of catalyst activity, the changes it will
produce in the performance of the catalytic reactor, and its effect on the
conclusions of this study cannot be predicted. As catalyst development
continues to produce catalysts capable of higher and higher maximum operating
temperatures, it is possible that little effect of aging would be seen during
the typical lifetime of an automobile. On the other hand, it may be necessary
to overdesign the catalytic reactor by making it longer than the initial
catalytic activity would necessitate.
Additional work is clearly required to complete the process of estab-
lishing design information for automotive gas-turbine catalytic combustors.
156
-------
REFERENCES
1. Pfefferle, W. C., R. V. Carrubba, R. M. Heck, and G. W. Roberts.
Catatherraal Combustion: A New Process for Low-Emissions Fuel Conversion.
ASME Paper 75-WA/Fu-l, 1975. 13 pp.
2. Wampler, F. B., D. W. Clark and F. A. Gaines. Catalytic Combustion of
C_H0 on Pt Coati
J O
14:25-31, 1976.
C_H on Pt Coated Monolith. Combustion Science and Technology,
J O
3. Cerkanowicz, A. E., R. B. Cole and J. G. Stevens. Catalytic Combustion
Modeling; Comparisons with Experimental Data. ASME Paper 77-GT-85, 1977.
8 pp.
4. DeCorso, S. M., S. Mumford, R. Carrubba, and R. Heck. Catalysts for Gas
Turbine Combustors - Experimental Test Results. J. of Engineering for
Power, 99(2): 159-167, 1977.
5. Pfefferle, W. C. The Catalytic Combustor: An Approach to Cleaner Combus-
tion. J. Energy, 2(3): 142 - 146, 1978.
6. Ablow, C. S., S. Schechter, H. Wise, and B. J. Wood. Contribution of Sur-
face Catalysis and Gas Phase Reaction to Catalytic Combustor Performance.
SRI International Corp., Menlo Park, California, 1979. 81 pp. Also
AFOSR-79-1146TR, Air Force Office of Scientific Research, AD-A078465.
7. T'ien, J. S. and D. N. Anderson. Gas Phase Oxidation Downstream of a
Catalytic Combustor. Paper presented at the American Chemical Society
Thirteenth Middle Atlantic Regional Meeting, West Long Branch, New Jersey,
March 20-23, 1979.
8. Tacina, R. R. Experimental Evaluation of Fuel Preparation Systems for an
Automotive Gas Turbine Catalytic Combustor. NASA TM-78856, 1977. 22 pp.
157
-------
9. Tacina, R. R. Degree of Vaporization Using an Air-Blast Type Injector for
a Premixed - Prevaporized Combustor. NASA TM-78836, 1978. 12 pp.
10. Gordon, S. and B. J. McBride. Computer Program for Calculation of Complex
Chemical Equilibrium Compositions, Rocket Performance, Incident and
Reflected Shocks, and Chapman-Jouguet Detonations. NASA SP-273, Rev.
1976. 145 pp.
158
-------
TABLE I . DESCRIPTION OF CATALYTIC REACTORS
Reactor
MBI-2.5
MBI-5.4
OCI-2.5
OCI-5.4
UOP-2.5
UOP-5.4
No. of
Elements
1
2
1
2
1
2
Reactor
Length,
cm
2.5
5.4
2.5
5.4
2.5
5.4
Upstream Element
Catalyst
Material
Platinum
Palladium
-
Noble Metal
Catalyst
Loading ,
kg/m->
5.3
3.6
-
*
Substrate
(see
Table la)
1
2
-
2
Downstream Element
Catalyst
Material
Palladium
Palladium
Palladium
Palladium
Noble Metal
Noble Metal
Catalyst
Loading,
kg/m3
7.1
7.1
3.6
3.6
*
*
Substrate
(see
below)
1
1
2
2
2
2
Manufacturer
Matthey Bishop
Matthey Bishop
Oxy-Catalyst
Oxy-Catalyst
OOP
UOP
Ul
VD
The MBI and OCI reactors were imaged.
The UOP elements had been exposed to 1400K air in a furnace for 300 hours.
* Proprietary
TABLE la. SUBSTRATE DESCRIPTION
Substrate
1
2
Manufacturer
Matthey Bishop
Corning
Material
Fecralloy
cordierite
Cell Shape
Sine wave
square
Cell Density,
cells/cm^
62
47
Open Area,
percent
93
63
-------
Ptvs. Pt/13%Rh
rC.A. THERMOCOUPLE
'//^THERMOCOUPLES
1 \ \
ALTERNATE
,*i SAMPLING
;// PROBE
LOCATIONS
CONICAL ;
TUBE /
FUEL
/ INJECTOR/
/ (FIG. 2)J
^ INLET
INSTRUMENTATION
PLANE
GAS
SAMPLING
PROBE
^CAT.
REACTOR
Figure I. - Test section. (Dimensions in cm.) 5. 4-cm-4ong catalytic reactor shown.
1HI C -78-1793
Figure 2. - Multiple conical tube fuel injector.
160
-------
REFERENCE GAS-PHASE REACTION LENGTH cm
VELOCITY,
m/s 7-3 I2-4 17.5 22.5
10
15
-O
-o a Q
98.0
99.0
99.5
99.7
99.8 -
99.9
99.95 [-
99.97 L
98.0 r
99.0
99.5
99.7
99.8
99.9
99.95
99.97
98.0
99.0
99.5
99.7
99.8
99.9
99.95 i
99.97
100
50
20
10
5
2
1
100
50
1 i_. ...._!_ 1 1 I
(a) REACTOR MB1-2.5; INLET-AIR TEMPER-
ATURE, 1000 K.
(b) REACTOR MBI-5. 4; INLET-AIR TEM-
PERATURE, 1000 K.
2 20
z
o
X
o
10
5
2
1
100
50
20
10
5
2
REFERENCE
EMISSION
INDEX,
13.6gCO/kg FUEL
1 1 1__.
(c) REACTOR OCI-2.5; INLET-AIR TEMPER-
ATURE, 1000 K.
I I ?J X J
(d) REACTOR OCI-5. 4; INLET-AIR TEM-
PERATURE. 1000 K.
ll 1 ..._ -1- 1W _LX\_L J
1200 1250 1300 1350 1400 1450 1500 1200 1250 1300
ADIABATIC COMBUSTION TEMPERATURE. K
i J
1350 1400 1450
(e) REACTOR UOP-2.5; INLET-AIR TEMPER-
ATURE, 950 K.
(ft REACTOR UOP-5.4; INLET-AIR TEM-
PERATURE, 950 K.
Figure 3. - Carbon monoxide emissions. Pressure, 3x10^ Pa.
161
-------
1500
CH
=
I —
f.
1400
1300
1200
CATALYTIC
REACTOR
MBI-Z5
MBI-5.4
OCI-2.5
OCI-5.4
UOP-Z5
UOP-5. 4
INLET -AIR
TEMPER-
ATURE,
K
1000
950
950
REFERENCE VELOCITY,
m/s
10 15
— O — O —
— -0- — — • —
— 9 — — 2 —
T T ~
— p — p-r-
10 12 14 16 18 20 22 24 26 28
CATALYTIC + GAS-PHASE REACTOR LENGTH, cm
Figure 4. - Effect of combustion length on minimum oper-
ating temperature. Pressure, 3x10^ Pa.
162
-------
0
D
•
•
Pa.
163
-------
REFERENCE
VELOCITY,
m/s
10
15
MBI-2.5
OCI-2.5
O
D
MBI-5.4
OCI-5.4
1.0 i-
.8
.6
.4
.2
Q-'
O
1.0
.8
.6
.4
.2
(a) REACTORS MBI-2.5 AND MBI-5.4; INLET-AIR TEMPER-
ATURE, 1000 K.
*•
-#*£r-
?150
i
i
1200 1300 1400 1500 1550
ADIABATIC COMBUSTION TEMPERATURE, K
(b) REACTORS OCI-2.5 AND OCI-5.4; INLET-AIR TEMPER-
ATURE, 1000 K.
REFERENCE
VELOCITY,
m/s
UOP-2.5 UOP-5.4
1200 1300 1400 1500
ADIABATIC COMBUSTION TEMPERATURE. K
1550
(C) REACTORS UOP-Z5 AND UOP-5.4; INLET-AIR TEMPER-
ATURE, 950 K.
Figure 6. - Catalytic reactor pressure drop. Pressure,
3X105 Pa.
164
-------
CATCOM* CATALYST 5 ATM 1000 HOUR AGING STUDY
USING #2 FUEL OIL
By:
I. T. Osgerby
B. A. Olson
H. C. Lee
Engelhard Industries Division
Engelhard Minerals and Chemicals Corporation
Menlo Park, Edison, New Jersey 08817
' Program supported by NASA-Lewis
under Contract NAS 3-19416
R. Priem, Contract Manager
ABSTRACT
Previous publications have described the combustion process designated
catalytically supported thermal combustion, which combines the features of
catalytic combustion with those of thermal or homogeneous combustion. The
process, characterized by extremely low pollutant emissions, in particular
NOX emissions, has been demonstrated over a wide range of operating
conditions of pressure, temperature, reference velocity and fuel types of
interest to commercial applications ranging from process heat generation to
gas turbines.
The durability of a proprietary Engelhard combustor catalyst has been
recently demonstrated, under a NASA contract, at five atmosphere pressure,
complementing a previous 1000 hour durability study at one atmosphere.
Both of these studies were conducted at about 640°K air preheat temperature
at a reference velocity of about 14 m/sec~l. The adiabatic flame temperature
of the fuel/air mixture was about 1530°K. Performance of the catalyst was
determined by monitoring emissions, thermocouple temperatures and catalyst
pressure losses, and by examining the physical condition of the catalyst core
at the conclusion of the test. Parametric studies were carried out at the
beginning and end of the durability test and periodic CO activity tests
were carried out to monitor catalytic performance.
The catalyst proved to be capable of low emissions operations after
1000 hours diesel fuel aging, however, more severe deactivation occurred
in the five atmosphere test. This was attributed to a loss in kinetic
(ignition) activity. Reasons for this loss in kinetic activity are suggested
and discussed in terms of observed performance relative to one atmosphere
test performance.
* CATCOM is a trade name of Engelhard Minerals and Chemicals Corporation.
165
-------
INTRODUCTION
The development of the process by which fuel and air is combusted in
a catalyzed monolith to yield ultra low emissions of carbon monoxide (CO),
unburned hydrocarbons (UHC) and nitrogen oxides (NOX) via catalytically
supported thermal combustion was pioneered by Engelhard Industries
(References 1 through 7). Many short duration tests (1-50 hours) have been
carried out using a variety of gaseous (References 9, 10, 12, 13, 15, 17,
18-23, 25-30, 33, 34) and liquid fuels (References 8-11, 13, 15-20, 24, 28,
31, 32, 35, 36, 38) both in Engelhard's R&D laboratories (References 9, 12,
23, 29-36) and prototype systems (References 23, 29, 35, 36) and government
(References 8, 10, 11, 15-20, 28) University and industrial (References
13, 24-27, 38) laboratories. The proven feasibility of the process has been
enhanced by investigations either carried out or supported by NASA
(References 8, 10, 11, 15-20, 28) the Electric Power Research Institute
(References 31, 32, 27, 40) and EPA (References 41-43). These test have
demonstrated the unique capabilities provided by combustion catalysts
(CATCOM*) in providing high efficiency low pressure loss combustion with low
emissions of CO, UHC and NOX, otherwise not generally attainable (Reference 11)
An outstanding requirement for these combustion catalysts is, of
course, long term durability. Durability is essential for the CATCOM*
catalyst to have commercial value since long periods are desired between
aintenance shutdowns of process equipment and ready acceptance of this new
technology by certain industries is not yet apparent.
='-CATCOM is a tradename of Engelhard Minerals and Chemicals Corp.
166
-------
Engelhard Industries, a division of Engelhard Minerals & Chemicals
Corporation, has pioneered the development of durable CATCOM* catalysts.
Early investigations (1970-1975) were internally funded, resulting in
proven durability of >1000 hours at design operating conditions on propane
and #2 distillate fuels on the first generation catalyst: DXA-111.
Deficiencies in off-design perfornxance (increased reference velocities
(>15 m sec~l), decreased adiabatic flame temperatures (< 2300°F), led to
the development of more stable catalyst systems.
In support of the E.R.D.A. highway vehicle gas turbine engine
development program, a test program was initiated with Engelhard Industries,
to demonstrate the durability of selected CATCOM* catalysts and catalyst
supports and to demonstrate the practicality of catalytically supported
thermal combustion. The program, funded by E.R.D.A. and managed by
NASA-Lewis, was conducted under NASA contract NAS3-19416. Two CATCOM*
catalysts (DXB-222 and DXC-532) were successfully tested for over 1000
hours on //2 distillate fuel at one atmosphere as reported in References
21 and 22.
Further refinements were incorporated into new CATCOM* catalysts
resulting in a more reproducible system designated DXE-412. This catalyst
was proven to have equal or better performance to DXB-222 at one atmosphere
and was selected as the catalyst to be tested at five atmospheres in a
continuation of NASA contract NAS3-19416. The highlights of the results of
this contract investigation are presented in this paper.
*CATCOM is a tradename of Engelhard Minerals and Chemicals Corp.
167
-------
EXPERIMENTAL PROCEDURES AND TEST PROGRAM
The testing program selected in this study was structured to serve the
following purposes:
1. Test the long-term durability of catalysts under realistic
combustion conditions.
2. Determine an acceptable low emissions operating range for the
catalysts.
In addressing these requirements, a test sequence was set up, which
focused mainly on the maintenance of low emissions during a 1000 hour life
test with #2 diesel. This life test provided a means of measuring either
abrupt or long-range changes in a catalyst core's overall performance. To
measure more subtle changes in the activity of the catalyst core, periodic
activity tests were conducted at 250 hour intervals using carbon monoxide
as a fuel.
The acceptable low emissions operating range for a CATCOM* catalyst
was determined using #2 diesel over a range of selcted variables. The
catalyst core designated as DXE-442 (Table I) was tested in this experimental
program. The catalyst was prepared on a monolith support.
The specifics of the overall testing sequence can be represented
schematically as follows:
*CATCOM is a tradename of Engelhard Minerals and Chemicals Corp.
168
-------
Virgin Catalyst
24 hour break-in period
Initial CO activity test
Initial set of parametrics to
map performance range
CO Activity Test
1000 hour life test with #2 diesel fuel
and CO activity test every 250 hours
Final set of parametrics to map
performance range changes
The initial 24 hour break-in period (see Table II for conditions) of
each catalyst core was required to remove the initially high activity and
stabilize the catalyst's activity level. This prevents potential problems
with interpretation of experimental results due to an artificially high
initial activity.
1. Life Test with #2 Diesel
The main purpose of this study was to evaluate the effect of pressure on
the durability of a catalyst core at combustion conditions for extended periods
of steady-state operation. The life test conditions were otherwise identical
to the steady-state operating conditions of the previous one atmosphere test.
The test fuel was #2 diesel. This fuel contains additives and impurities that
provide a sound test for determining the effects of potential poisons on the
catalyst core. Table III contains the analyses of the fuel used in the life
test.
169
-------
The steady-state conditions for the 1000 hour life test are shown in
Table II. These life tests were conducted on a continuous basis when
possible. During shutdown, a house air purge was maintained over the test
catalyst cores.
The maintenance of low emissions performance during the life test is
considered the prime critetia for determining changes in a specific catalyst
core's performance. This life test also provides a measure of the physical
durability of the catalyst core support material. The results of this test
provide information on long term durability of CATCOM* catalyst cores.
2. Carbon Monoxide Activity Test
In conducting the CO activity test, the life test conditions are
discontinued and the CO activity test conditions are then established. The
catalyst core is not disturbed and remains inside the test rig reactor.
This test consists of measuring the response of CO conversion to
increases in the air preheat temperature. The carbon monoxide test gas was
purchased from Scientific Gas Products and is rated chemically pure (C.P.)
with a 99.7% minimum purity specification. The remaining tests parameters
are held constant and are shown in Table IV. Typical responses for this
test are shown in Figure 1.
The response curve in Figure 1 provides two important facts about the
catalyst cores' activity. The CO ignition temperature, defined as the lowest
air preheat temperature required to obtain a measurable temperature rise
*CATCOM is a tradename of Engelhard Minerals and Chemicals Corp.
170
-------
across the catalyst core, is a measure of Intrinsic catalytic activity.
Increases in the ignition temperature reflect possible changes in the energy
of activation or pre-exponential factor using the Arrhenius rate expression
as the typical reaction rate model. The sloping portion of the response curve
in Figure 1 indicates that in this temperature range the rate of reaction is
kinetically controlled. The region showing a constant conversion level as a
function of temperature is considered to be mass transfer controlled. A mass
transfer controlled reaction is independent of the catalyst and temperature,
and is only a function of the apparent geometric surface area of the catalyst.
If the mass transfer conversion declines with catalyst age, then the apparent
mass transfer area is decreasing.
The utility of this CO activity test is that it provides a means to
monitor subtle changes in catalyst core activity without disturbing the
catalyst test piece in the reactor.
3. Parametric Testing
In order to define the low emissions operating range for test catalysts,
a series of parametric studies was carried out over the following operating
range:
Variable Range
Pressure 1 x 10^ - 5 x 105 N/M2
Air Preheat Temperature 630 - 810°K
Adiabatic Flame Temperature 1306 - 1533°K
Reference Velocity 14 - 26 M/S
171
-------
The parametric tests were conducted initially at the start-of-life
testing to determine the range of operation of a virgin catalyst, and at
the end-of-life testing to determine if changes in activity had occurred
after 1000 hours.
TEST RIG MODIFICATIONS
In order to carry out the experimental program the test rig constructed
for the one atmosphere life test had to be modified to allow safe, unattended
operation at five atmospheres pressure. The test capabilities of the unit
previously described in reference 21 had to be broadened. A simplified
schematic of the modified NASA test rig is presented in Figure 2.
The equipment modifications can be broken down to three major categories:
A. Safety System
B. Process Control System
C. Fuel Presentation System
Details of the reactor/controls sytem can be obtained from references 21 and 39.
The experimental reactor used in testing the catalyst cores at combustion
conditions is constructed of corrosion-resistant heavy-walled Inconel 601 pipe
and designed for long term endurance testing at 1533°K and 5 x 105 N/M2 (5 atm)
and short term testing at 1533°K and 10 x 105 N/M2 (10 atm). The reactor
was instrumented for measurement of:
catalyst core inlet and outlet temperature
catalyst core pressure drop
• catalyst core inlet pressure
catalyst core emissions
172
-------
The pressure drop apparatus consisted of a manometer with pipe tap
locations upstream and downstream of the catalyst core in accordance with
ASME recommended practice (reference 44).
The range of operating conditions of the test rig is summarized in
Table V. Detailed operating procedures used for 5 atmosphere diesel fuel
life testing are given in reference 39.
Emission samples were all taken with a 0.00635M (1/4") diameter
water-cooled sample probe. The sampling train adhered to SAE Standard
ARP-1256. The description of each individual analytical instrument and
standard operating procedures are the same as reported in reference 21.
Fuel Presentation System
In the course of carrying out the experimental program usin^ #2 diesel
fuel, it was found that high reactor inlet temperature shutdowns occurred
repeatedly at 5 atmosphere life test conditions upon extended operation
(overnight) with a fuel injection system essentially identical to that used
for 1 atmosphere endurance testing. To improve fuel vaporization and
mixing. A Delavan air assist siphon fuel nozzle (#30610-2) in the
horizontal inlet section of Unit 6 was used to generate 10-40 micron diameter
fuel droplets.
173
-------
Test Performance of the New Fuel Injector
After making the necessary equipment changes, a continuous run of
90 hours (3.8 days) was made with a duplicate catalyst at 5 atmosphere life
test conditions. No pre-ignition or any other safety shutdown occurred using
the new air-assisted siphon fuel nozzle. To determine the flashback limit,
the inlet temperature at 5 atmosphere life test conditions was increased
gradually from 633°K by increasing the air preheat furnace temperatures. At
approximately 823°K, the inlet temperature started to climb steadily as
flashback occurred. The experiment was repeated and the same result was
seen again. At the normal operating inlet temperatures, well below 823°K
no flashback or preburning was observed with the new fuel injector.
EXPERIMENTAL RESULTS - PERFORMANCE OF CATALYST CORE DXE-442
Life Testing with #2 Diesel Oil
The most useful technique for monitoring performance of a catalyst
during life testing is to monitor daily the input variables and output
responses from the test. A control plot of the most critical observations
then provides a method for detecting trends which may indicate a degradation
in the performance of the catalyst or problems in test rig operation.
Control charts were maintained on a daily basis for the following:
- exhaust gas composition (CO, UHC, NOX, C02, 02)
- pressure drop
- air preheat temperature
- catalyst core outlet temperature
174
-------
Perhaps the most significant control plot in the present life test is
the exhaust gas analysis and, in particular, the emissions of CO, UHC and
NOX (see Figure 3.)- The gas compositions are reported as measured in the
exhaust, which was nominally 12.5% oxygen. These plots provide a measurement
of stable operation during life testing. Major events occurring in the life
test are shown in Figure 4. A summary of performance measurements before
and after life testing is shown in Table VI.
Carbon Monoxide Emissions
The CO emissions were constant around 30 Vppm for the initial 200 hours
on stream and gradually increased during life testing as shown in Figure 3a.
Two high CO emissions data points greater than 2000 ppm were measured after
848 hours on stream. However, after the final CO activity test at 1014 hours
on stream, the CO emissions were significantly lower (525 Vppm) under life
test conditions. The CO emissions approached initial performance levels
after the catalyst had been maintained around 673°K in the presence of air
for two days while a thermocouple was repaired. After this, low CO
emissions (less than 200 Vppm) were maintained for 22 continuous hours at the
end of the life test.
Unburned Hydrocarbon Emissions
The unburned hydrocarbon emissions exhibit the same trend as the CO
emissions (shown in Figure 3b).
Nitrogen Oxides Emissions
The NOX emissions varied from 3.5 Vppm to 6.5 Vppm during life testing
as plotted in Figure 3c. Low NOX emissions were observed (4.8 Vppm)
throughout the life test.
175
-------
Downstream Temperature Profiles
Maximum outlet temperature measured at 0.1080 M downstream of the
catalyst core was around 1475°K and this outlet temperature corresponds to
94% adiabaticity.
Diesel Parametric Tests
In order to define the low emissions operating range for the initial
and final catalyst performance after 1000 hours on stream, a series of
parametric tests were carried. Table VII gives initial and final parametric
test results for the comparison runs, and detailed results are given in
reference 39.
Comparison plots showing effects of inlet temperature, reference velocity
and pressure on combustion efficiency, are presented in Figures 5, 6 and 7,
for the initial (after 63 hours on stream) and final (after 1014 hours
on stream) parametric tests over the range of adiabatic flame temperature
used.
Carbon Monoxide Activity Test
Carbon monoxide activity tests were performed periodically at
24 hours, 206 hours, 507 hours, 729 hours and 1014 hours on stream during
life testing. Carbon monoxide conversion was measured under test conditions
listed in Table IV. The test procedures are detailed in references 21 and 39.
CO conversion data are plotted versus air preheat temperature in Figure 3.
Activity Test Procedure Modification
Since poor results of the previous standard CO activity test were
obtained at 729 hours on stream, a different approach was tried to obtain
mass transfer limited CO conversion levels by starting the test from a high
temperature (turndown) and with CO fuel on continuously.
176
-------
High CO conversion data denoted by closed symbols in Figure 9 were
obtained with fuel on continuously and mass transfer limited CO conversions
of approximately 40% were achieved. Standard CO activity test results are
shown in Figure 9 for comparison.
DISCUSSION OF TEST RESULTS
The primary criteria for evaluation of catalyst performance consisted
of experimentally determining if the range of low emissions operation and
physical durability of the catalyst core could be maintained after 1000 hours
of aging with #2 diesel fuel at 5 atmospheres pressure. Secondary criteria
considered were the maintenance of carbon monoxide activity and range of
low emissions performance in diesel parametric tests before and after 1000
hours aging at 5 atmospheres pressure.
Life Test Results
Emissons Performance:
During the 1000 hour durability life test of the catalyst core,
DXE-442, at 5 atmospheres pressure, emissions were as follows:
Initial after After After
Emissions Measured after 63 hours 1014 hours 1062 hours
UHC (as C3) :
CO :
NOX
Combustion efficiencies were calculated from carbon balance
(Reference 31) in the exhaust gas stream and plotted in Figure 10.
A high combustion efficiency was maintained around 99.95% during the
initial 200 hours of the life test and a gradual degradation in the
177
0
35
5.7
146
2420
5.6
0
35
4.3
-------
catalyst performance was noted for the remainder of the life test.
Combustion efficiency dropped to 95% at 848 and 1014 hours on stream. There
is no evidence for abnormal contaminants in #2 diesel fuel which had been
used during the particular period. Final emissions levels obtained after
1062 hours on stream suggest that some sort of reversible catalyst
deactivation may have occurred during the life test.
Compared to the one atmosphere life test results (references 21, 22)
the deactivation of the catalyst was more severe at 5 atmospheres pressure.
The test catalyst was subject to five times the total mass throughput,
thus any decline in catalyst performance caused by poison contaminants in
the fuel or by attrition of active catalytic components would have been
accelerated.
At the end of the life test, catalyst performance was partially restored
by passing preheated air in-situ over the catalyst at 673°K for a two day
period. Thereafter, high combustion efficiencies were observed as shown in
Figure 10. This unexpected result may have been caused by partial catalyst
regeneration: 1) removal of fuel deposited poisons from the catalyst
surface, or 2) alteration of the active catalyst surface by some other
mechanism.
NOX emissions were fairly stable and ranged from 4-5 ppm by volume.
Oxides of nitrogen are generated entirely from fuel bound nitrogen.
Significant thermal NOX generation is unlikely at the operating temperatures
used in a well mixed CATCOM* system.
*CATCOM is a tradename of Engelhard Minerals and Chemicals Corp.
178
-------
Physical Durability of DXE-442 Catalyst:
Photographs of the inlet and outlet ends of the test catalyst used for
life testing are presented in Figure lla. Physical damage was limited to
minor breakage frotj catalyst handling during removal of the catalyst substrate
and reloading it into the test reactor.
A photograph of the catalyst test configuration is shown in Figure lib.
As shown, the two test segments are separated by 1/4" spacers and secured in
the catalyst holder using high temperature cement and Fibre Frax packing
around the circumference of the substrates.
Carbon. Monoxide Activity Testing
The CO activity test is intended to detect differences in effective
catalytic surface area by measuring mass transfer limited CO conversion
levels at high flow conditions. Stabilized CO conversion of approximately
40% were obtained at 206 and 507 hours on stream.
From the data taken after 729 and 1014 hours on stream, it was
apparent that these conversion levels could not be achieved under normal
activity test conditions. It was found, however, that 40% CO conversion
levels could be reached by starting up at high temperature and leaving feed
CO on while decreasing the air preheat temperature. Lower CO conversion
levels were measured when the activity test was done by the standard
procedure of increasing the air preheat temperature as illustrated in Figure 9.
The observed hysteresis phenomenon has been well described for
heterogeneous reaction systems for CO oxidation over various metal component
catalysts (reference 45). Although mass transfer limited conversions in the
179
-------
range of 40% could be achieved after 1000 hours on stream, it is apparent
that more severe catalyst deactivation had occurred during the 5 atmosphere
life test than observed in any 1 atmosphere life test conducted to date,
including that carried out using a similar DXE-442 catalyst preparation.
Diesel Parametrics
Effect of Air Preheat Temperature Before and After Life Testing
Effect of air preheat temperature on combustion efficiency is shown
in Figure 5, a plot of the negative logarithm of complement combustion
efficiency versus adiabatic flame temperature at 633°K and 723°K air
preheat temperatures.
Since combustion efficiency obtained with the initial catalyst core
was not sensitive to inlet temperature at a low adiabatic flame temperature,
there was no apparent kinetic activity limitation in the overall reaction.
Above an adiabatic flame temperature of 1450°K, combustion efficiency was
not affected by adiabatic flame temperature
However, the performance of the catalyst had changed after 1014 hours
on stream as shown in Figure 5. A kinetic effect was noted at low adiabatic
flame temperatures and the catalytically supported homogeneous combustion
zone was reduced at high adiabatic flame temperatures. This was attributed
to partial deactivation. Thus, combustion efficiency is more sensitive to
adiabatic flame temperature after 1014 hours on stream, and higher fuel/air
ratios are required to maintain efficient, low emissions performance. After
1014 hours on stream, a minimum adiabatic flame temperature to obtain
combustion efficiency greater than 99% is interpolated to be 1513°K from
Figure 5 at 633°K air preheat temperature and 14 M/S reference velocity.
180
-------
Effect of Reference Velocity Before and After Life Testing
Combustion efficiency is a strong function of reference velocity as
shown in Figure 6, a plot of the negative logarithm of complement combustion
efficiency versus adiabatic flame temperature at 14 M/S and 26 M/S reference
velocities. Combustion efficiency substantially declined as reference
velocity was increased up to 26 M/S for both the initial and final catalyst
performance.
Initial combustion efficiencies greater than 99% were achieved over
the entire range of adiabatic flame temperatures at 14 M/S reference
velocity. An adiabatic flame temperature of 1553°K was required to obtain
combustion efficiencies greater than 99% at 26 M/S reference velocity.
After 1014 hours on stream, an adiabatic flame temperature required
for 99% combustion efficiency is extrapolated to be 1597°K at 26 M/S
reference velocity from Figure 6.
Effect of Pressure Before and After Life Testing
As shown in Figure 7, a plot of the negative logarithm of complement
combustion efficiency versus adiabatic flame temperature, at 1 x 105 N/A2
and 5 x 10^ N/M2 pressures, shows that pressure is the least sensitive
variable, although slightly higher combustion efficiencies were obtained at
5 x 105 N/M2 pressure and the higher adiabatic flame temperatures. The
same trend was observed at the beginning and after 1014 hours on stream.
Low Emissions Operating Range
The low emissions operating regions have been defined as those with
combustion efficiency performance greater than 99% as shown in Figure 12.
The boundary for the required adiabatic flame temperature was formulated for
181
-------
a high combustion efficiency at a given reference velocity by interpolating
parametric test results assuming a linear relationship between each data
point in the plot of logarithm of complement combustion efficiency versus
adiabatic flame temperature (shown in Figures 5 and 6).
As illustrated in Figure 12, the region of low emissions operation has
been narrowed after life testing at 5 atmospheres pressures. Inlet
temperature in the range of 633°K to 723°K does not affect the initial
boundary but is significant after 1014 hours on stream.
CONCLUSIONS AND RECOMMENDATIONS
The following conclusions and recommendations have been made from the
experimental test results obtained under this contract:
1. The catalyst core tested, DXE-442, maintained low emissions
for 1000 hours of continuous operation at 5 atmospheres with
//2 diesel fuel. Initial and final emissions were:
After After
Initial 1014 hours 1062 hours
Unburned Hydrocarbons (03, Vppm) 0 146 0
Carbon Monoxide (Vppm) 30 2420 35
Nitrogen Oxides (Vppm) 5.7 5.6 4.3
2. The DXE-442 catalyst core tested maintained its physical integrity
throughout the 1000 hour, 5 atmosphere life test at operating
temperatures characteristic of catalytically-supported thermal
combustion operationg temperatures (1533°K) and withstood the
thermal shock of numerous intentional and unintentional startups
and shutdowns without any evidence of failure.
182
-------
3. NOX emissions measured were consistently low throughout
endurance and parametric testing and can be maintained at
4-5 Vppm.
4a. It has been experimentally determined that the operating
conditions of the 5 atmosphere test were apparently more
severe,in terms of maintaining catalyst activity,than those
of the previous test conducted at 1 atmosphere pressure.
4b. The modified CO activity test data indicated that adequate
mass transfer surface was retained, but kinetic (ignition,
light off) activity had been severely deteriorated. This is
presumed due to exposing the inlet section of the catalyst
to sustained high temperatures (caused by preburning) and the
malfunctioning fuel preparation method, prior to the successful
modifications.
5. The observation that combustion efficiencies exceeding 99.5%
could be achieved at the end of life testing, although levels
as low as 95% had been measured, indicates that at least part
of the catalyst deactivation was reversible.
6. It is recommended that either the test be repeated with a new
catalyst using the successfully modified reactor, or carry out
further development work to improve the catalyst high temperature
capability (1644-1700°K) and to improve catalyst life.
183
-------
REFERENCES
1. Pfefferle, W. C., "Two-Stage Combustion Process", U. S. Patent
3,846,979 (11/12/74).
2. Hindin, Saul G., "Method for Catalytically Supported Thermal
Combustion", U. S. Patent 3,870,455 (3/11/75).
3. Pfefferle, W. C., "Method and Furnace Apparatus", U. S. Patent
3,914,090 (10/21/75).
4. Pfefferle, W. C., "Catalytically Supported Thermal Combustion",
Belgium Pat. 814,752, November 8, 1974 and U. S. Pat. 3,938,961,
December 30. 1975.
5. Pfefferle, W. C., "Method of Operating Catalytically Supported
Thermal Combustion System", U. S. Patent 3,940,923 (3/2/76).
6. Hindin, Saul G., and Pond, George R., "Method of Combustion Using
High Temperature Stable Catalysts", U. S. Patent 3,966,391 (6/29/76).
7. Dettling. J. and Hindin, S. G., U. S. Patent 3,956,188 (1976) and
U. S. Patent 3.945,946 (1976), Hindin, S. G. and Pond, G. R., U. S.
Patent 3,966,391 (1976).
8. Blazowski, W. S. and Bresowar, G. E., "Preliminary Study of the
Catalytic Combustor Concept, As Applied To Aircraft Gas Turbines",
AFAPL-TR-74-32, 1974.
9. Pfefferle, U. C., Heck, R. M., Carrubba, R. V. and Roberts, G. W.,
"CATATHERMAL Combustion: A New Process for Low-Emissions Fuel
Conversion", ASME-57-Wa/FU-l, 1975.
10. Anderson, D. N., Tacina, R. R. and Mroz, T. S., "Performance of a
Catalytic Reactor at Simulated Gas Turbine Combustor Operating
Conditions", NASA-TMX-71747, 1975.
11. Blazowski, W. S. and Walsh, D. E., "Catalytic Combustion: An Important
Consideration for Future Application", Comb. Sci, and Tech., Vol. 10,
pp 233-244, (1975).
12. Carrubba, R. V., Chang, M., Pfefferle, W. C. and Polinski, L. M.,
"Catalytically-Supported Thermal Combustion For Emissions Control",
Paper presented at the Electrical Power Research Institute NOX
Control Technology Seminar, San Francisco, Calif., Feb. 6, 1976.
13. DeCorso, S. M., Mumford, S., Carrubba, R. V., and Heck, R.'M.,
"Catalysts for Gas Turbine Combustors - Experimental Test Results",
Presented at ASME Gas Turbine Division Conference, New Orleans, La.,
March, 1976.
184
-------
14. Anderson, D. N., "Preliminary Results from Screening Tests of
Commercial Catalysts with Potential Use in Gas Turbine Combustors.
Part I: Furnace Studies of Catalyst Activity", NASA TM X-73410, 1976.
15. Anderson, D. N., :Preliminary Results from Screening Tests of Commercial
Catalysts with Potential Use in Gas Turbine Combustors. Part II:
Combustion Test Rig Evaluation", NASA TM X-73412, 1976.
16. Rosfjord, T. J., "Catalytic Combustors for Gas Turbine Engines",
AIAA pages 76-46, Washington, D. C., (Jan. 1976).
17. Anderson, D. N., "Emissions and Performance of Catalysts for Gas
Turbine Catalytic Combustors", NASA TM X-73543 (1977).
18. Anderson, D. N., "Effect of Catalysts Reactor Length and Cell Density
on Performance", Presented at the 2nd Workshop on Catalytic
Combustion, 21-22, June, (1977).
19. Anderson, D. N., "Emissions and Performance of Catalysts for Gas
Turbine Catalytic Combustors", ASME Paper No. 77-GT-65, March, 1977.
20. Anderson, D. N., "Performance and Emissions of a Catalytic Reactor
With Propane, Diesel and Jet A Fuels", NASA TM-73786, Sept., 1977.
21. Heck, R. M., "Final Report Part I. Durability Testing at 1 atm of
Advanced Catalysts and Catalyst Supports for Automotive Gas Turbine
Engine Combustors", NASA CR-135132, June, 1977.
22. Heck, R. M., Chang, M. and Hess, H., Mroz, T., "Durability Testing of
Advanced Catalysts and Catalyst Supports for Gas Turbine Engine
Combustors", AIChE Symposium Series No. 188, Vol. 75, 1979.
23. Flanagan, P., Norster, E. R., Carrubba, R. V. and Heck, R. M. "Method
for Effecting Sustained Combustion of Carbonaceous Fuel", U. S.
Patent 4,118,171 (10/3/78).
24. Pogson, J., Mansour, M. N., "Catalytic Combustion of No. 6 Fuel Oil",
3rd Workshop on Catalytic Combustion, 3-4 October 1978.
25. Marteney, P. J., "Determination of the Rate of Heterogeneous Reaction
on Catalytic Surfaces", loc. cit.
26. Schefer, R. W., Robben, F. S., Cheng, R. K., "Catalyzed Combustion of
H2/Air Mixtures in a Plate Plate Boundary Layer", loc. cit.
27. Walsh, P. M. Santavicca, D. A., Kim, B., Bracco, F. V., "Progress of
Reaction in a Honeycomb Catalyst: CO/Air Combustion", loc. cit.
28. Anderson, D. N., "Effect of Inlet Temperature on the Performance of
a Catalytic Combustor", NASA TM-78977 (1978).
185
-------
29. Flanagan, P., Carrubba, R. V. and Norster, E. R., "The Development
of a Gas Turbine Combustor Utilizing a Catalysts", Journal of
Energy, 1979.
30. Osgerby, I. T., Heck, R. M., Carrubba, R. V., and Bunker, W. W.,
"Investigation of Process and System Design Variables for
CATATHERMAL Combustion of Low BTU Gas", ASME 79-GT-66.
31. Lew, H., Olson, B. A., Dickson, W. H., Heck, R. M., and DeCorso, S. M.
"Experimentally Determined Catalytic Reactor Behavior and Analysis
for Gas Turbine Combustors", ASME 79-GT-150.
32. Osgerby, I. T., Olsen, B. A., Lew, H. and Conn, A. "CATCOM
Catalyst Axial Temperature Profile Measurements". Presented at 4th
Workshop on Catalytic Combustion, May, 1980.
33. Lee, H. and Osgerby, I. T. "Catalytic Combustion Characteristics of
Low and Medium BTU Gas Fuels", loc. cit.
34. Osgerby, I. T., Gleason, G. L. and Mularz, E. J. "Combustion Catalyst
Studies For Simulated Aircraft Idle Mode Operation", loc. cit.
35. Osgerby, I. T. and Norster, E. R. "Development of a Prototype Industrial
CATCOM Burner for Process Heating". To be published.
36. Osgerby, I. T. "Performance Characteristics of a Prototype Industrial
CATCOM Burner For Process Heating". To be published.
37. Angello, L. "The EPRI Stationary Gas Turbine Catalytic Combustor
Development Program:. 4th Workshop on Catalytic Combustion, May 1980.
38. Carl, D. and Osgerby, I. T. "High Pressure CATCOM Test Results",
loc. cit.
39. Olson, B. A., Lee, H. C., Osgerby, I. T., Heck, R. M. and Hess, H.
"Durability Testing at 5 atmospheres of Advanced Catalysts And Catalyst
Supports for Gas Turbine Engine Combustors" NASA-CR- 1980.
40. Lew, H. G., et al., "Ceramic Turbine Components R&D. Part 3. Ceramic
Combustor Design and Analysis Final Report" (EPRI Project 421-1).
(March 1979).
41. Kesselring, J. P., Krill, W. V. and Kendall, R. M., "Design Criteria
For Stationary Source Catalytic Combustors", Paper No. 77-32, Western
States Section, The Combustion Institute, October (1977).
42. Krill, W. V., et. al., "Catalytic Combustion for Gas Turbine
Applications", ASME Paper No. 79-GT-188, (1979).
186
-------
43. Krill, W. V., et al., "Catalytic Combustions for System Applications",
ASME Paper No. 79-HT-54 (1979).
44. Bean, H. S. (ed.) "Fluid Meters: Their Theory and Applications",
6th Ed., 1971.
45. Hlavacek, V. and Votruba, J. "Experimental Study of Multiple Steady
States in Adiabatic Catalytic Systems", Chemical Reaction Engineering.
II, American Chemical Socity, Washington, pp. 545-558, 1974.
187
-------
TABLE I. PROPERTIES OF TEST CATALYST CORE
Catalyst Identification
Catalyst Components
Support
Channel Density
Nominal
Actual
Hydraulic Diameter
Open Fraction
Bulk Surface Area
Length
DXE-442
Palladium
Zircon Composite
256 Channels/in2
144-159 Channels/1" Score
9.754 X 10"4 M
65.5%
2683 M2/M3
1524 M
TABLE II. LIFE TEST CONDITIONS
Catalyst Core Dimensions
Fuel Type
Air Flow
Fuel Flow
Air/Fuel
Inlet Temperature
Adiabatic Flame Temperature
Inlet Pressure
Reference Velocity
Space Velocity at NTP
0.0254 M diameter X 0.1524 M length
n Diesel Oil
1.42 X 10"2 Kg/S
3.73 X 10"4 Kg/S
38/1 (Kg/Kg)
633°K
1533°K
5 X 105 N/H2
14 M/S
142 M3/S - M3 Cat.
188
-------
TABLE III. ANALYSES1 OF #2 DIESEL FUELS USED FOR CATALYST LIFE TESTS
Test
Batch #1 Batch *23 Batch #34
Gravity, API @ 60°F (289°K)
Flash Point, °K
Pour Point, °K
Uater and Sediment, Vol. %
Ash, Wt. %
Colour, ASTM
Distillation Temperature, °K
Initial
10;;
50%
90S
End Point
Recovery, X
Carbon/Hydrogen Atomic Ratio
Heating Value, Joule/Kg (Gross)
Viscosity,SU5 at 311°K
Sulfur, WtS
Nitrogen, ppm
Phosphorous, ppm
Lead, ppm
33.8
351.0
261
0.1
-Mil-
454.3
487.6
532.1
575.4
608.8
99
0.535
4.57X107
34.6
.03
SCO
-
1
33.9
348.8
258
-Ni 1 -
Trace
L 2.5
461.0
489.9
534.3
583.2
613.2
98
0.602
4.57X107
35.0
.11
<10
< 1
< 1
34.0
333.2
250
< .05
.016
-
440.0
483.2
531.0
577.5
593.2
96
0.529
4.55X107
35.3
.13-. 14
600
< 1
2.4
1
Analyses performed by Saybolt and Co.
2Batch #1 used between 0 and 206 hours on stream.
Batch -fi used between 206 and 825 hours on stream.
4Batch #3 used between 825 and 10G5 hours on stream.
189
-------
TABLE IV. CARBON MONOXIDE ACTIVITY TEST CONDITIONS
Fuel Type C.P. Grade CO
Air Flow 10.55 X 10'3 Kg/S
Fuel Flow 4.09 X 10"5 Kg/S
Air/Fuel 258/1 Kg/Kg
o
Inlet Temperature Varied up to 733 K
Inlet Pressure 1 X 105 N/M2
Reference Velocity at 633 K 36 M/S
Space Velocity at NTP 125 M3/S-M Cat.
TABLE V. OPERATING RANGES FOR UNIT 6
Automatic Control Operation:
Air Flow 1.85 X lo"3 to 5.9 X 10'3 Kg/S
Air Preheat Temperature Up to 810°K
Fuel Flow (Liquid) 6.7 X 10"5 to 67 X 10"5 Kg/S
Reactor Pressure 1.0 X 105 N/M2 to 5 X 105 N/M2
Adiabatic Flame Temperature* Up to 1533°K
Fuel Type #2 Diesel
Manual Operation:
Air Flow 1.85 X 10"3 to 66.7 X 10'3 Kg/S
Air Preheat Temperature Up to 810°K
Fuel Flow (Gaseous) 1.7 X 10~4 to 17 X lo"4 Kg/S
Reactor Pressure 1.0 X 105 to 5.0 X 105 N/M2
* o
Adiabatic Flame Temperature Up to 1533 K
Fuel Type C.P. Carbon Monoxide. #2 Diesel
*
For conditions of 90» reactor adiabaticity, adiabatic flame
temperature may be increased to 1570°K without damage to
reactor walls.
190
-------
TABLE VI. COMPARISON OF PERFORMANCE DATA AT START AND END OF LIFE TEST
Hours on Diesel Fuel
Air Flow Rate (Kg/sec)
Fuel/Air Ratio (by weight)
o
Temperature ( K)
Preheat Air
Reactor Outlet
Adiabatic Flame
Pressure
Inlet (N/M2)
Drop (N/M2)
Loss ( ?. )
Reference Velocity (M/Sec)
Heat Release Rate
(Joules/sec, M3, N/M2)
Combustion Efficiency ( % )
"^Emissions (Vppm)
CO
UHC (as C )
J
NO
Start
63
1.43 X 10'2
.02647
533
H63
1539
5.1 * 105
6767
i ***
i . w* i
u.o
1136
99.95
30
0
5.7
End
1014
1.43 X 10'2
.02686
633
1463
1539
5.0 * 105
9473
T O T
1 • Ww
14.2
mi
94.97
2420
146
5.f
Results After Final
CO Activity Tests
at 1062 Hours
1062
1.43 X lO'2
.02647
633
1473'
1539
5.0 * 105
9812
- o~
14.2
1151
99.94
35
0
A. 3
Catalyst Core Dimension, Nominal 0.0254 M (Diameter)
* 0.1524M (Long)
+A11 emissions measured with water cooled sample probe located at
0.102 M downstream of catalyst core.
191
-------
NJ
TABLE VII. DIESEL PARAMETRIC TEST RESULTS
_L Combustion Efficiency
Run
iiu-.ber
146-1
145-2
146-3
*
146-4
146-5
145-6
146-7
146-8
146-9
146-10
145-11
146-12
146-13
Pressure
(Atm)
5
5
5
5
5
5
5
5
1
1
1
5
5
Air Preheat
Temperature
CO
360
360
450
450
450
360
360
360
360
360
360
360
360
Flaiiie
Teir.perature
CO
1093
1176
1093
1176
1260
1093
1176
1280
1093
1176
12CO
1093
1176
Ratio
(W./H.)
0.0210
0.0236
0.0183
0.0213
0.0241
0.0210
0.0236
0.0270
0.0210
0.0236
0.0270
0.0210
0.0236
Kcierence un
Velocity CO
(M/Sec) (p;rni)
14
14
14
14
14
26
26
26
22
22
22
14
14
345
>5COO
35
3300
300
>5000
25
2200
15
93
> 5000
5000
>sooo
500
1100
> 5000
>5000
> 5000
>5000
700
2200
2G3
>sooo
2S08
i bb ion:
UIIC
(ppm]
0
310
4
0
115
0
0
_
900
0
15
-
-
n
203
3
2io
> i
"Ox -
1 fppr.i)
3.6
3.5
5.1
3.2
3.7
4.3
5.1
_
4.0
2.8
4.7
-
-
5.2
5.3
4.0
ti
( % J
Initial
99.20
99.90
99.21
99.94
99.97
49.1*
84.87
99.12
36.8*
61.3*
98.70
99.33
99.94
Final
G4.6 *
91.53
73.1*
91.19
99.82
25,5*
25.4 *
97.33
26.2 *
23.2 *
91.03
32.0 *
92.97
* Estimate based on % Adiabaticity = (Insured temperature rise/Theoretical tenpcrature rise)
+ Large type denotes emissions at end of life tect.
Sr:all type denotes emissions at beginning of life test.
-------
100
500
600
100
80 -
60 •
UJ
_l
o
CO
cr
LU
o
CJ
o
20 -
MASS-TRANSFER-
CONTROLLED REGION
KINETIC CONTROLLED REGION
IGNITION TEMPERATURE
0 i ,-= — , r
200 300 400
CATALYST INLET TEMPERATURE (°C>
Figure 1. Typical response obtained from carbon
monoxide activity test.
193
-------
COMBUSTIBLE
DETECTOR
EXHAUST
FAN
OUTLET
TEMP,
OUTLET
TEMP,
HEATED SAMPLE
LINE FOR
EMISSIONS
LEGEND:
rZ^ MOUNTED ON PANEL
UNIT MOUNTED
FI) FLOW INDICATOR
FR) FLOW RECORDER
FLOW RECORDER CONTROLLER
P PRESSURE MEASUREMENT
PRESSURE RECORDER CONTROLLER
™ INSULATION
NOTE:
1. CO FOR ACTIVITY TEST
COOLER
Figure 2. Schematic of modified test rig,
194
-------
VO
10,000 .
5,000 :
1,000 -
o_
a.
GO
CD
CO
co
100
50
10
0
NOMINAL TEST CONDITIONS
FUEL = #2 DIESEL
REFERENCE VELOCITY = 14 M/S
FEED F/A WEIGHT
INLET PRESSURE
AIR PREHEAT TEMP, =
CATALYST CORE
DIMENSIONS
,0263 (G/G)
5 X 105 N/M2
633°K
= .0245M I
X .1524M L
OPEN SYMBOLS DENOTE DATA AFTER
48 HOURS, 673°K AIR SOAK,
T
L
PREHEATED AIR (673°K) WAS.
PASSED OVER THE CATALYST
BED IN TWO DAYS,
0
200 400 600 800 1,000
HOURS ON #2 DIESEL FUEL
Figure 3a. Carbon monoxide emissions control chart.
-------
1,000
500
GO
a.
Q_
GO
CD
GO
GO
CD
PQ
CD
Q±
100
50
10
5
NOMINAL TEST CONDITIONS
FUEL
REFERENCE VELOCITY
FEED F/A WEIGHT
INLET PRESSURE
AIR PREHEAT TEMP,
CATALYST CORE
DIMENSIONS
= #2 DIESEL
= 14 M/S
= ,0263 (G/G)
= 5 X 105 N/M.2
= 633°K
= ,0254M 6
X .1524M L
OPEN SYMBOLS DENOTE DATA
TAKEN AFTER 48 HOURS,
673°K AIR SOAK,
PREHEATED AIR (673°K)
WAS PASSED OVER THE
CATALYST BED FOR 2 DAYS
0
200
400 600
HOURS ON #2 DIESEL FUEL
1,000
Figure 3b. Hydrocarbon emissions control chart.
-------
16
14 •
12 '
's
a.
Q_
GO
2 8 -
GO
GO
^^ n •
w
X
CD
4 '
2 •
n
NOMINAL TEST CONDITIONS
FUEL
REFERENCE VELOCITY
FEED F/A WEIGHT
INLET PRESSURE
AIR PREHEAT TEMP,
CATALYST CORE
DIMENSIONS
s
•«J •
GfiP
©e®
0 200
= #2 DIESEL LEGEND
= ^f, / ^ ° N0 MISSIONS
= ,0263 (G/G)
= 5 X 105 N/M2 • NOX EMISSIONS
= 633°K
= .0254M i X .1524M L
o
•
• • o 0
•
o o @
0
0©
400 600 800 1000
HOURS ON #2 DIESEL FUEL
Figure 3c. Nitrogen oxide emissions control chart.
-------
INITIAL
PARAMETRIC TEST
FINAL
0 HR
hSEE
NOTE(l)
#1 CO
ACTIVITY TEST
200 MRS 400 HRS 600 HRS 800 HRS
PARAMET
1000
SEE
NOTE(2)
VO
00
PERIOD OF REACTOR
MALFUNCTIONS '
(HIGH INLET TEMP,
SHUT-DOWNS)
#3 CO
ACTIVITY TEST
#4 CO
ACTIVITY TEST
NEW FUEL INJECTION
JOZZLE INSTALLED
UC TEST
HRS
#5 CO
ACTIVITY TEST
HIGHEST
CO EMISSIONS
(25 VPPM)
LIFE TEST
RESUMED
#2 CO
ACTIVITY TEST-
TOTAL 1065 HOURS
ON STREAM
* NOTES (1) 24 HOURS BREAK-IN PERIOD
(2) MORE THAN TEN HIGH INLET TEMPERATURE
SHUTDOWNS OCCURRED DURING THE PERIOD
FROM 66 TO 206 HOURS,
Figure 4. Event chart for the life test.
-------
Test Conditions
en
O
Run No. 146-1 through 5
Fuel
Reference Velocity
Inlet Pressure
Air Preheat Temperature
A
O
#2 Diesel
14 M/S5
5 x0io N/r
633 K
723°K
Opon symbols: Initial Parametric Data
Closed symbols: Final Parametric Data
.9999
J3 .9990
o
-H .9900
o
c
QJ
c
o
JD
O
O
3 -9000
Figure 5.
1400 1500
Adtabatic Flame Temperature (°K)
Effect of air preheat temperature on combustion
efficiency before and after life testing.
199
-------
en
o
Test Conditions
Run No. 146-1,2,6,7 and 8
Fuel :
Inlet Pressure :
Air Preheat Temp. :
Reference Velocity
#2 Diesel 2
5 X 10b N/M
633°K
A : 14 M/S
O : 26 M/S
Open Symbols: Initial Parametric Data
Closed Symbols: Final Parametric Data
3 L_
.9999
.9990
d .9900
o
p-
u
c
cu
*^-
u
LU
c
o
O
_i .9000
1400 1500
Adiabatic flame temperature (°K)
Figure 6. Effect of reference velocity on combustion efficiency
before and after life testing.
200
-------
Test Conditions
P~
i
cn
O
Run No. 146-6 through 11
Fuel #2 Diesel
Reference Velocity 22-26 M/S
Air Preheat Temperature 633 K 2
Inlet Pressure Q : 1 X 10c N/M?
A : 5 X 10 N/M
Open Symbols: Initial Parametric Data
Closed Symbols: Final Parametric Data
.999
J .990
.1 .900
u
c
O)
c:
o
E
O
CJ
0
1400
1500
Adiabatic Flame Temperature ( K)
Figure 7. Effect of pressure on combustion efficiency before and
after life testing.
201
-------
100
80
60
2 40
nu
20
0
200
LEGEND
0 24 HOURS AGING
O 206 HOURS AGING
& 507 HOURS AGING
O729 HOURS AGING
HOURS AGING
500
600
RUN CONDITIONS
REFERENCE VELOCITY =
FEED CO
PRESSURE
CATALYST CORE
DIMENSIONS
700
36,5 M/S
(AT 633°K)
4000 VPPM
1 x 105 N/M2
•0254M i
X .1524M L
Q Q-
300 400
AIR PREHEAT TEMPERATURE )°C)
500
Figure 8. Carbon monoxide activity test response during life
test of catalyst core DXE-442.
202
-------
100
RUN CONDITIONS
REFERENCE VELOCITY =36,5 M/S (AT 633°K)
FEED CO = 4000 VPPM
PRESSURE = 1 X 105 N/M2
CATALYST CORE
DIMENSIONS = 0.254M 6 X .1524M L
^ 1014 HOURS AGING (STANDARD PROCEDURE; TEMPERATURE
INCREMENTALLY INCREASED)
V 1014 HOURS AGING (MODIFIED PROCEDURE; TEMPERATURE
INCREMENTALLY DECREASED)
500 600 700 (°K)
T
I
80
60
CO
o
CO
00
20
CJ
0
200
300 400
AIR PREHEAT TEMPERATURE (°C)
Figure 9. Carbon monoxide activity responses after 1014 hours
using catalyst core DXE-442.
203
-------
INLET ENDS
UPSTREAM
S/N 4070 J-12
DXE-442
DOWNSTREAM
S/N 4070 J-l
DXE-442
OUTLET ENDS
UPSTREAM DOWNSTREAM
S/N 4070 J-12 S/N 4070 J-l
DXE-442 DXE-442
Figure lla.
Photographs of catalyst core DXE-442
after 1000 hours life testing.
205
-------
NJ
O
3 -
2 -
1
0
#1 CO ACTIVITY TEST
| (24 HRS)
T
®cS®®1
_
#2 CO ACTIVITY TEST
e (206 HRS)
©
i i
©
—
PERIOD OF
JEACTOR MAL
FUNCTIONS
(HIGH
INLET TEMP,
SHUT DOWN),
O~
#3 CO ACTIVITY TEST | J
, (507 HRS)
#4 CO ACTIVITY TEST
I (729 HRS)
.
© :
:
© f b J
0
#5 CO ACTIVITY!
TEST
(1014 HRS) L -1
O €
(PREHEATED AIR (673°K)
V
JEW FUEL f
INJECTOR
INSTALLED, i
200 400
HOURS ON
^AS PASSED OVER THE
;ATALYST BED FOR 2 DAYS,
, , , . i
~""
—
-
—
z
,9999
,—,
,sr-
,9990 U
LU
U_
LL_
1 i 1
L 1 _'
,9900 2
i —
GO
PQ
O
,9000
0
600 800 1000
#2 DIESEL FUEL
Figure 10. Response of combustion efficiency during five atmospheres life testing.
-------
NASA LIFE TEST 5 ATM CATALYST
AFTER 1065 HOURS
FLOW
[TW1
1 y 21
liiiiiliiliiliiitliiiiiiliiiiiliiiiiliiiiliiiil
l"(j) x 3" LONG
UPSTREAM
S/N 4070 J-12
DXE-442
1" $ x 3" LONG
: DOWNSTREAM
S/N 4O7OJ-I
DXE-442
Figure lib. Photographs of catalyst core DXE-442
after 1000 hours life testing.
206
-------
'10
30
CO
20
LU
LL
LU
ce:
10
0
TIN: AIR PREHEAT TEMPERATURE
INOTE:
LOW EMISSIONS OPERATION
(COMBUSTION EFFICIENCY>99%)
IS ACHIEVED BELOW INDICATED
LINES,,
1450 1500 1550
ADIABATIC FLAME TEMPERATURE (°K)
Figure 12. Low emissions operating condition range,
207
-------
FUEL-NITROGEN CONVERSION AND COMBUSTION
EFFICIENCY IN A CATALYTIC COMBUSTOR
By:
Ronald D. Matthews
Mark L. Graham
Jeryl L. Lederman
Combustion Science Laboratory
Department of Mechanical Engineering
University of Texas
Austin, TX 78712
This work supported, in part, by a DOE Institutional Energy Grant to the Uni-
versity of Illinois.
ABSTRACT
This manuscript reports the initial stages of the eventual development of
a phenomenological model that can be used to predict fuel-NO emissions from
X
catalytic combustion systems. The effects of radial heat loss rate and inlet
velocity on combustion efficiency and fuel-NO conversion were experimentally
A
investigated. The radial heat loss rate is of interest because envisioned
practical catalytic combustors incorporate concentric annular airflow. Because
of the different inlet velocities used by various researchers, this parameter
is also of interest.
An experimental catalytic combustion system was designed and constructed
which allowed the radial heat loss rate to be varied from 0-15% of the energy
release rate. The inlet velocity was varied over the range 5-25 cm/s. The
equivalence ratio and volumetric energy release (adiabatic flame temperature)
were decoupled using argon as a diluent. The equivalence ratio was varied
from 0.5 to 1.25 and volumetric energy release cases of 6.2, 8.3, and 10.6
kcal per mole of reactant mixture (T^ - 1400, 1700, 2000K) were used.
208
-------
It was found that radial heat losses up to 12% of the energy release rate
could significantly affect the temperature distributions in the catalytic com-
bustor. This, in turn, can decrease the combustion efficiency and will prob-
ably decrease the NO yield (resulting in, presumably, higher unburned fuel-N
concentrations in the exhaust). The effect of radial energy loss becomes more
pronounced as the mixture gets leaner. Higher inlet velocities also increase
the average combustor temperature, resulting in higher combustion efficiencies
and NO yields. Increasing the volumetric energy release also increases the
combustion efficiency and the NO yield.
This project was supported, in part, by a Department of Energy Institu-
tional Energy Grant to the University of Illinois at Urbana-Champaign. This
work was carried out at the Combustion Research Facility, Department of Me-
chanical Engineering, University of Illinois between May 21, 1978 and De-
cember 21, 1979.
209
-------
ACKNOWLEDGMENTS
The authors would like to thank William F. Hammett who set up the experi-
mental apparatus and obtained the initial data (W. F. Hammett, A. Study of
Catalytically Stabilized Combustion at Equivalence Ratios Including Stoichio-
metric, MS Thesis, University of Illinois, August, 1978). We are also tre-
mendously grateful to Che-Kwan Shum, Ph.D. candidate in Aerospace Engineering
at the University of Texas, for developing and running the computer graphics
programs.
210
-------
INTRODUCTION
The use of catalytic combustors in systems that operate with near-
stoichiometric mixtures has been the subject of recent research interest
(References 1-7). One of the several advantages of catalytic combustors is
that they have exhibited a very wide fuel tolerance, being capable of burning
fuels as disparate as low-Btu gas (References 5, 6, 8-11), distillate and
residual fuel oils (References 7-9, 11), and possibly even pulverized coal
(Reference 13). It is probable that those systems that are required to
operate near stoichiometric because of system efficiency considerations will
also be required to use non-traditional fuels such as these in the future.
Unfortunately these fuels generally contain significant concentrations
of bound nitrogen (fuel-N). Investigations of fuel-N in gas phase combustion
systems have shown that fuel-bound-nitrogen is oxidized to NO (fuel-NO ) at
X X
up to 100% efficiency while molecular nitrogen is oxidized (to thermal-NO )
at less than 1% efficiency (References 14-16). Most of the combustion modifi-
cations that are successful in controlling thermal-NO adversely affect fuel-
NO . Water injection, increased mixing intensity, and inert addition all in-
X
crease fuel-NO (References 17-19). Flue gas recirculation may increase
(Reference 20), decrease (Reference 21), or have no effect (Reference 22) on
fuel-NO . Only staged combustion (References 22, 23) and catalytic combustion
(References 1-6) have shown the potential to simultaneously control both
thermal-NO and fuel-NO . In fact, the combination of these two combustion
x x
modifications - staged catalytic combustion - appears to be one of the most
promising techniques available for decreasing pollutant emissions while main-
taining or improving combustion efficiency.
211
-------
A portion of the results of catalytic combustion fuel-NOx studies is
shown in Figure 1, which exhibits the effect of equivalence ratio on fuel-
NO yield (the mass ratio of nitrogen in the products as NO to fuel-N in
x x
the reactants). These results agree qualitatively but the quantitative dis-
crepancies are significant. This paper reports an initial effort to examine
the reasons for these discrepancies with an eventual goal in mind of develop-
ing a phenomenological model that can be used to predict fuel-NO^ emissions
from catalytic combustion systems.
OBJECTIVES
This report is a condensation of the Master's theses of Mr. Graham
(Reference 24) and Mr. Lederman (Reference 25). They investigated the effects
of radial heat loss and reference (inlet) velocity on combustion efficiency
and fuel-N conversion in a catalytic combustor. This was the initial stage
of an explanation of the differences in the data of various researchers as
depicted in Figure 1. The eventual goal of the project is the development of
a phenomenological model that can be used to predict fuel-NO emissions from
X
catalytic combustors.
The researchers whose results were used to construct Figure 1 had several
experimental conditions in common and several more that were different. All
used a ceramic monolith supporting a platinum catalyst. All but Pogson and
Monsour (Reference 7) used NH as the model fuel-N compound. (It has been
shown that fuel-NO conversion is only weakly dependent on fuel-N species
X
in both gas-phase and catalytic combustion systems, (References 1, 3, 14)).
Matthews and Sawyer (Reference 1) used fuel-N fractions of 1.75 wt % and
5.0 wt %, Chu and Kesselring (Reference 3) used 2.0 wt %, and Folsum et al
(Reference 6) used 0.2 wt % and 1.0 wt %. (The effect of fuel-N fraction
is shown in Figure 2. The data of Chu and Kesselring is for different condi-
tions than those discussed here, including a different catalyst). All but
Matthews and Sawyer used graded cell catalysts. Matthews and Sawyer did not
use air preheat although the other researchers did. Matthews and Sawyer re-
port a maximum bed temperature of HOOK for an adiabatic flame temperature of
2000K while Folsum and coworkers report holding the bed temperature near
their adiabatic flame temperature, 1473K. Chu and Kesselring report a bed
212
-------
temperature of 1367K and Pogson and Mansour report the bed temperature ranging
from 1340 to 1540K. Matthews and Sawyer decoupled the equivalence ratio and
adiabatic flame temperature by using an oxygen/argon mixture as the oxidant.
Folsum et al and Chu and Kesselring accomplished this by using an air/nitrogen
mixture. Pogson and Mansour did not decouple these parameters but did use
argon or nitrogen dilution in some cases to insure that the temperature never
exceeded the failure temperature of the quartz housing, 1600K.
The most important differences in the operating conditions used by these
researchers are considered to be inlet velocity (and therefore space velocity
and energy release rate), radial heat transfer boundary condition, and fuel
characteristics. Pogson and Mansour used #6 fuel oil, Folsum and coworkers
used a simulated low-Btu gas consisting of CH,, H , and CO, and the other
researchers used C0H . The reference velocities used ranged from 5 cm/s by
J o
Matthews and Sawyer to 3 m/s by Folsum et al and about 6 m/s by Chu and
Kesselring. The velocity in the research of Pogson and Monsour varied from
2.3 to 4.6 m/s. The radial heat transfer boundary condition varied from
backheating (heat addition) by Folsum et al to free convection energy loss
in the work of Matthews and Sawyer and insulation by the other researchers.
This radial boundary condition is of interest because envisioned practical
catalytic combustors (References 2, 26-28) have concentric annullar airflow
around the catalytic combustor. This, in turn, might affect temperature
profiles and combustion efficiency. The purpose of the research reported in
this manuscript was to investigate the effects of inlet velocity and radial
energy loss on fuel-nitrogen conversion and combustion efficiency. It is
believed that these factors are the major parameters of interest in the de-
velopment of a phenomenological model.
EXPERIMENTAL APPARATUS AND CONDITIONS
The experimental apparatus is similar to that used by Matthews and
Sawyer (Reference 1). The flowrates of C_H , Q^, Ar, an Ar/NH3 mixture,
and the coflowing air (used in the radial heat loss cases) were controlled
using calibrated rotameters. Unburned hydrocarbons were measured using an
FID, NDIR's were used to detect CO and CO , and chemiluminescent analysis was
used to quantify NO. It was found that the low third body quenching efficiency
of Ar offset the high quenching efficiencies of C02 and H20, yielding a
213
-------
maximum error of 6% due to the relative quenching efficiencies of product
third bodies in chemiluminescent analysis (Reference 29).
The combustor was a 51 mm diameter single grade cordierite monolith
with an a-alumina washcoat and a 0.187 wt % Ft.loading. Upstream from the
combustor was a blank ceramic monolith which served as a heat shield to
protect glass beads which were used to flatten the velocity profile. The
beads, heat shield, and combustor were housed in a stainless steel tube.
Outside of this SS housing was either 1) 6 inches of Babcock and Wilcox
K30 insulating firebrick with a mean thermal conductivity of 0.288 W/(m-K),
or 2) in the radial heat loss cases, a coflowing annular air stream which
passed through a conical housing.
Combustion occured at atmospheric pressure and the reactants were not
preheated. Inlet velocities of 5, 10, and 25 cra/s were investigated. The
Op/Ar oxidant mixture was used to decouple the equivalence ratio and the
equilibrium adiabatic flame temperature (References 1, 24, 25). The
equivalence ratios studied ranged from = 0.5 (100% excess air) to = 1.25
(80% theoretical air). Adiabatic flame temperatures of 1400, 1700, and 2000K
(corresponding to volumetric energy release rates of 6.2, 8.3, and 10.6 kcal
released per mole of reactant mixture) were studied. Radial heat loss rates
were varied from practically zero for the insulated cases up to 15% of the
energy release rate.
RESULTS AND DISCUSSION
The effects of radial heat loss and inlet velocity on fuel-N conversion
and combustion efficiency were investigated for various equivalence ratios
and adiabatic flame temperatures. The combustion efficiency is defined by:
n = AE ./AE ,
c act theo
where AE is the actual energy release rate and is determined from
act
AEact = * nPhP
where h is the enthalpy of each product species at the combustor exit temper-
ature and n is the number of moles of each product species formed per mole
of fuel burned (obtained from the exhaust gas analysis data, accounting for
214
-------
those species measured on wet and dry bases, and assuming that the total
number of moles of products formed per mole of fuel burned, E np> is
negligibly different in the actual case than for equilibrium reaction at the
combustion temperature). The last term, Z HRh , is the enthalpy of the
reactants at room temperature and reduces to the enthalpy of formation of
propane. The theoretical energy release is determined from
AEtheo - E 1phP - E VR
where n* is the number of moles of each product for equilibrium combustion at
the average temperature along the axial centerline for the insulated (no
radial energy loss) case corresponding to the particular equivalence ratio,
volumetric energy release rate, and inlet velocity of interest. The product
enthalpy term, hi, is determined at 298K. Note that the combustion efficiency
is similar to the enthalpy efficiency, r\ , except that the denominator in the
definition of ru is the Lower Heating Value of the fuel. That is, the heat
n
of reaction, AlL~Qfi, is the denominator of both n and TU, but r\ uses
K/yo C n C
equilibrium products and n assumes complete combustion. As expected, it was
n
found that there was no difference in r\ and nu for lean combustion and only
L n
a small difference for rich combustion up to = 1.25. The combustion effi-
ciency is presented in this paper because it is the more physically significant
of the two.
Figure 3 shows the radial temperature distributions for five axial loca-
tions in the combustor. Separate graphs for lean, stoichiometric, and rich
operation are presented. The data for Figure 3 was obtained with the combus-
tor insulated. The peak temperature occurs at = 1.25, is about 50K lower
for the <|> = 1.0 and = 0.75 cases, and drops an additional 100K for <{> = 0.5
even though the equilibrium volumetric energy release (and therefore the
adiabatic flame temperature) is constant for all four cases. As expected,
the radial temperature profiles are relatively flat. However, the temperature
of the upstream face decays rapidly near the edge of the combustor. The tem-
perature differential between the center of the upstream face and the outer
edge is not affected by the equivalence ratio but the rate of temperature de-
cay and the portion of the leading edge affected increases as the mixture
gets leaner. For <^ 1.0, the temperature at x/L = 0.2*5 also decays over the
outside 25% of the radius. Therefore, even for an insulated combustor,
215
-------
portions of the temperature distribution appear to be affected by heat trans-
fer losses.
Figure 4 demonstrates the effect of radial heat loss on the temperature
distributions for the maximum radial heat loss operating condition. The
radial heat loss was obtained by setting the coaxial air flowrate and then
varying the mixture ratio in the combustor. Because the combustor temperature
was dependent on (even though 4> and the volumetric energy release had been
decoupled), the radial heat loss rate varied slightly from one equivalence
ratio to the next. It is apparent from Figure 4 that radial energy losses
significantly affect the axial and radial temperature distributions. In fact,
only for rich operation is the peak temperature located at the center of the
leading edge of the catalyst. For the <)> = 0.5 case, the radial profiles 3/4
up and at the trailing edge are not affected but the temperatures throughout
the rest of the catalyst are significantly decreased and show significant
radial temperature gradients, especially at the axial plane 1/4 up from the
upstream face. With radial heat loss, the upstream face evidences the lowest
temperatures, rather than the highest temperatures as was observed for the in-
sulated combustor. For = 0.75, the peak temperature is depressed about 200K
and there are significant temperature gradients on all axial planes except the
downstream face and there is a uniform 50K temperature decrease on this plane.
The same general trends are evident for the stoichiometric and rich operating
conditions, with lower temperatures throughout the catalyst, especially the
upstream face, and steep radial temperature gradients except for the downstream
face. The rich data shows the least effect of radial heat loss, with only
small differences at x/L « 0.75 and x/L = 1.0 (downstream plane).
Figure 5 shows the effects of radial heat loss in a slightly different
manner. Here the combustion temperature is taken as the average of the five
axial temperatures at any given radial location. It is seen that radial heat
loss depresses the average temperature throughout the catalyst and increases
the average radial temperature gradient. In most cases, the magnitude of
radial heat loss is not as significant as the fact that there is heat loss.
This is most apparent for the leanest operating condition.
The effect of radial heat loss on combustion efficiency is shown in
Figures 6 and 7. Figure 6 shows the characteristic increase in combustion
216
-------
efficiency with increasing equivalence ratio. However, f| continues to in-
crease past $ = 1.0. This is probably due to the previously noted fact that
radial heat loss had the least effect on the fuel rich case. This trend is
more visually apparent in Figure 7, which is a graph of relative heat loss
rate versus combustion efficiency for the four equivalence ratios studied.
For = 1.25, increasing heat loss depresses the combustion efficiency only
slightly. For stoichiometric and lean mixtures, 3-5% heat loss depresses t|
by 5-20% and approximately doubling the radial heat loss does not result in a
further decrease. (In fact, for $ - 0.75, the efficiency appears to increase
again. This fact is not readily explainable, but is probably due to a defi-
ciency in the data.) Note that the temperature profiles of Figure 5 also
showed the largest change with initial heat loss with generally smaller
changes for a further increase in the radial heat loss rate.
The data of Figures 3-6 were obtained with an inlet velocity of 10 cm/s
and a volumetric energy release of about 10.6 kcal per mole of reactant mix-
ture (i.e., T._ = 2000K). Figure 8 demonstrates the effect of inlet velocity
and volumetric energy release on the average axial centerline temperature and
on the combustion efficiency. The data for Figures 7-11 were obtained with an
insulated combustor. As evident in Figure 7, the combustion temperature in-
creases significantly with increasing inlet velocity for all cases except for
fuel rich combustion at a volumetric energy release of 6.8 kcal/gmole (T =
AJJ
1700K). The combustion efficiency also increases with increasing velocity ex-
cept for the stoichiometric and fuel rich cases at T = 2000K. Attempts to
fllJ
obtain data at 25 cm/s were unsuccessful because of unstable operation.
Figure 9 is a plot of r| versus for the two inlet velocities and three
volumetric energy release cases studied. It was found that the most stable
combustor operation was obtained with the highest volumetric energy release,
i.e. TAD = 2000K. The 2000K data shows an increase in n with increasing inlet
velocity. The 10 cm/s data indicates an increase in n with increasing volu-
metric energy release (T.D).
Figure 10 illustrates the effect of equivalence ratio on the NO yield.
The characteristic decrease of NO yield from stoichiometric operation to fuel
rich operation was observed (see Figure 1). More importantly, it is seen
that increasing the inlet velocity will increase the NO yield for fuel-lean
217
-------
conditions, thereby partially explaining the differences in the results of
Matthews and Sawyer and those of Chu and Kesselring. This is probably due, at
lease in part, to the higher bed temperatures obtained at higher velocities,
as seen in Figure 5. It is also apparent that increasing the volumetric en-
ergy release (T.n) increases the NO yield for both lean and rich operation.
This is probably also partially due to the higher bed temperatures noted in
Figure 5. Although exhause NH- was not specifically measured, it is assumed
that lower NO yields correspond to higher exhaust NH3 levels.
Unlike the data presented in Figures 6 and 7, the efficiencies and NO
yields presented in Figures 8-10 are not averaged over the cross-sectional
exit area of the combustor. This was not thought to be necessary because the
radial temperature and species profiles for the insulated combustor were not
appreciable. However, as shown in Figure 11, it was determined that radial
gradients in the NO yield were apparent for stoichiometric and fuel-rich com-
bustion. This effect would probably be even more pronounced for a combustor
with significant radial heat losses.
SUMMARY AND CONCLUSIONS
The effects of radial heat loss and inlet velocity on combustion effi-
ciency and fuel-nitrogen conversion in a catalytic combustor were investigated.
Volumetric energy release rates of approximately 6.2, 8.3, and 10.6 kcal/
gmole of reactants (T^. - 1400, 1700, and 2000K) were studied and the fuel-air
equivalence ratio was varied from 0.5 to 1.25.
Radial heat loss was investigated because envisioned practical catalytic
combustors incorporate concentric airflow around the combustor. It was found
that radial heat losses up to 12% of the energy release rate can significantly
affect the temperature distributions in the combustor. This, in turn, can de-
crease the combustion efficiency and will probably decrease the NO yield (re-
sulting in, presumably, higher unburned fuel-N concentrations in the exhaust).
The effect of radial heat loss on temperature and combustion efficiency becomes
more pronounced as the mixture gets leaner.
It was also determined that increasing the inlet velocity will also in-
crease the combustor temperature. This was expected because the energy re-
lease rate and inlet velocity are directly coupled. An increase in the com-
218
-------
bustion efficiency and the NO yield was noted with increasing inlet velocity
and increasing volumetric energy release (adiabatic flame temperature).
This manuscript reports the initial stages of the development of a phe-
nomenological model which can be used to predict the fuel-NO emissions from a
X
catalytic combustion system. More experimental research is necessary before
the model can be formulated. Most notably, it is necessary to: 1) investigate
the effect of inlet velocity over a wider range, up to the order of 5 m/s,
thereby 2) increasing the bed temperature up to the adiabatic flame tempera-
ture. In order to explain Figure 2, it will be necessary to 3) vary the
fuel-N concentration for various bed temperatures and equivalence ratios.
Other parameters of interest include the 4) catalyst support configuration,
5) catalyst species, and 6) combustor operating pressure.
219
-------
REFERENCES
1. Matthews, R.D., and R.F. Sawyer. Fuel Nitrogen Conversion and Catalytic
Combustion, Paper No. 77-40, Fall Meeting, Western States Section/Combus-
tion Institute, Lawrence Berkeley Laboratory Report LBL-6396, Berkeley,
CA, 1977.
2. Krill, W.V., J.P. Kesselring, and E.K. Chu. Catalytic Combustion for Gas
Turbine Applications. ASME Paper 79-GT-188, presented at the ASME Gas
Turbine Conference, San Diego, CA, March 12-15, 1979.
3. Chu, E.K., and J.P. Kesselring. Fuel-NO Control in Catalytic Combustion.
X
In: Proceedings - Third Workshop on Catalytic Combustion. Environmental
Protection Agency Publication EPA-600/7-79-038, 1979. pp. 291-330.
4. Krill, W.V., and J.P. Kesselring. The Development of Catalytic Combustors
for Stationary Source Applications. In: Proceedings - Third Workshop on
Catalytic Combustion, Environmental Protection Agency Publication EPA-600/
7-79-038, 1979. pp. 259-290.
5. Folsum, B.A., C.W. Courtney, and M.P. Heap. The Effects of LBG Composition
and Combustor Characteristics on Fuel NO Formation. ASME Paper 79-GT-185,
x
presented at the ASME Gas Turbine Conference, San Diego, CA, March 12-15,
1979.
6. Folsum, B.A., C.W. Courtney, and M.P. Heap. Environmental Aspects of Low
Btu Gas-Fired Catalytic Combustion. In: Proceedings - Third Workshop on
Catalytic Combustion. Environmental Protection Agency Publication EPA-
600/7-79-038, 1979. pp. 345-384.
7. Pogson, J., and M.N. Monsour. Catalytic Combustion of No. 6 Fuel Oil.
In: Proceedings - Third Workshop on Catalytic Combustion. Environmental
Protection Agency Publication EPA-600/7-79-038, 1979. pp 111-138.
220
-------
8. Pfefferle, W.C., R.M. Heck, R.V. Carrubba, and G.W. Roberts. Catathermal
Combustion: A New Process for Low Emission Fuel Conversion. ASME Paper
75-WA/Fu-l, 1975.
9. DeCorso, S.M., S. Mumford, R.V. Carrubba, and R.M. Heck. Catalysts for
Gas Turbine Combustors - Experimental Test Results. ASME Paper 76-GT-4,
1976.
10. Carrubba, R.V., and I.T. Ogersby. Catalyst Design Studies in Low Btu
Gas Combustion. Presented at the Third Workshop on Catalytic Combustion,
Asheville, NC, October 3-4, 1978.
11. Osgerby, I.T., R.V. Carrubba, R.M. Heck, and W.W. Bunker. Investigation
of Process and System Design Variables for Catathermal Combustion of
Low Btu Gas. ASME Paper 79-GT-66, presented at the ASME Gas Turbine
Conference, San Diego, CA, March 12-15, 1979.
12. Anderson, D.N. Performance and Emissions of a Catalytic Reactor with
Propane, Diesel, and Jet A Fuels. Presented at the Fall Meeting,
Western States Section, the Combustion Institute. Palo Alto, CA, 1977.
13. Pfefferle, W.C. Liquid Droplet Heating and Vaporization in the Catalytic
Combustor. ASME Paper 79-HT-52, presented at the ASME/AIChE 18th
National Heat Transfer Conference, San Diego, CA, August 6-8, 1979.
14. Sarofim, A.F., and R.C. Flagan. NO Control for Stationary Combustion
X
Sources. Progress in Energy and Combustion Science, 2: 1-25, 1976.
15. Sawyer, R.F., N.J. Brown, R.D. Matthews, M.C. Branch, and S.M. Banna.
The Formation of Nitrogen Oxides from Fuel Nitrogen. Electric Power
Research Institute Publication EPRI 223-1, 1976.
16. Matthews, R.D. The Nature and Formation of Nitrogenous Air Pollutant
Emissions from Combustion Systems. Lawrence Berkeley Laboratory Report
LBL-6850, 1977.
17. Wilkes, C., and R.H. Johnson. TIS Report 74-GTD-75, Technical Informa-
tion Service, General Electric, 1974.
18. Fenimore, C.P. Effects of Diluents and Mixing on Nitric Oxide From
Fuel-Nitrogen Species in Diffusion Flames. Sixteenth Symposium
(International) cm Combustion, the Combustion Institute, Pittsburgh,
221
-------
1065-1071, 1977.
19. Appleton, J.P., and J.B. Heywood. The Effects of Imperfect Fuel-Air
Mixing in a Burner on NO Formation from Nitrogen in the Air and in the
Fuel. Fourteenth Symposium (International) in Combustion, the Combustion
Institute, Pittsburgh, 777-786, 1973.
20. Brown, T.D., E.R. Mitchell, and G.K. Lee. Low NO Combustion: The Effect
X
of External Flue Gas Recirculation on the Emissions from Liquid Fuel Com-
bustion. Combustion Institute European Symposium. Academic Press, New
York, 487-492, 1973.
21. Martin. G.B., and E.E. Berkatl. An Investigation of the Conversion of
Various Fuel Nitrogen Compounds to NO in Oil Combustion. Presented at
X
the AIChE National Meeting, Atlantic City, NJ, August, 1971
22. Turner, D.W., R.L. Andrews, and C.W. Siegmund. Influence of Combustion
Modification and Fuel Nitrogen Content on Nitrogen Oxides Emissions from
Fuel Oil Combustion. Combustion; 21-30, 1972.
23. Gibbs, B.M., F.J. Periera, and J.M. Beer. The Influence of Air Staging
on the NO Emissions from a Fluidized Bed Coal Combustor. Sixteenth
Symposuim (International) on Combustion, the Combustion Institute,
Pittsburgh, 461-474, 1977.
24. Graham, M.L. Effect of Radial Heat Loss on Efficiency of a Catalytic
Combustor. MS Thesis, University of Illinois, Urbana, IL, 1980.
25. Lederman, J.L. An Investigation of Parameters Affecting Fuel-NO
X
Emissions from Catalytic Combustion Systems. MS Thesis, University of
Illinois, Urbana, IL. 1979.
26. Flanagan, P., E.R. Norster, and R.J. Carrubba. Development of a Gas
Turbine Combustor Utilizing a Catalyst. £. Energy, 3(2): 75-81, 1979.
27. Anderson, D.N., R.R. Tacina, and T.S. Mroz. Catalytic Combustion for
the Automotive Gas Turbine. NASA TM X-73589, presented at the Fourth
International Symposium on Automotive Propulsion Systems, Washington,
D.C., April 17-22, 1977.
222
-------
28. Enga, G.E. Catalytic Combustion in Actual Engines: A Summary of Engine
and Rig Tests. In: Proceedings - Third Workshop on Catalytic Combustion.
Environmental Protection Agency Publication EPA-600/7-79-038, 1979.
pp. 491-512.
29. Matthews, R.D., R.F. Sawyer, and R.W. Schefer. Interferences in the
Chemiluminescent Measurement of NO and N0» Emissions from Combustion
Systems. Environmental Science and Technology, 11(12): 1092-1096, 1977,
223
-------
120
i i
Chu 8 Kesselring 2%
on 8
Monsour
Motthews a
Sawyer 5%
Folsum et al 0.2%
Matthews a
Sawyer 1.75%
olsum et al 1%
I
0.8 1.0 1.2
Equivalence Ratio
FIGURE 1. Experimental results from fuel-NO studies using catalytic combus-
tion systems.
224
-------
c
o
c
o
32
O>
CD
T3
'x
O
80
60
40
owo
o
£ 20
0
0
I \ I
0=0.32 Chu a Kesselring
0=1.03
Matthews & Sawyer
A
= 0.75
Matthews 8 Sawyer
8
Fuel Nitrogen Concentration(wt% of Fuel)
FIGURE 2. Effect of fuel-N concentration on fuel-NOx production. The data of
Chu and Kesselring is for different conditions than those of Figure
1 as discussed in the manuscript. Specific differences include a
1478K bed temperature, use of a proprietary UOP catalyst, and 3.1
m/s inlet velocity
225
-------
•/. RflOIR!
OISTRNCE FROM CENTER
100
X RROim
OISTRNCE FRCM CENTER
25 50 7
X RROIRL DISTRNCE FROM CENTER
00
JS 50 15
RflDIRL OISTRNCE FROM CENTER
100
FIGURE 3. Radial temperature distributions at five axial locations for an in-
sulated combustor. Bottom face » downstream plane of combustor.
'IN
100 cm/s, T
AD
2000K, 0% fuel-N.
226
-------
so ~5T
/: RADIAL DISTANCE FROM CENTER
100
RADIAL DISTANCE FROM CENTER
100
2S SO IS
/: RAOIflL DISTANCE FROM CENTER
« sb
•/. RADIAL DISTANCE FROM CENTER
too
FIGURE 4. Radial temperature distributions at five axial locations with ap-
proximately 10% radial heat loss. Upper face • upstream plane of
combustor. V •= 10 cm/s, T - 2000K, 0% fuel-N.
227
-------
X RflDIflL QISTRNCE FROM CENTER
i 0
so "15"
•I. RRDIRL DISTRNCE FROM CENTER
1 0
g
UJ-"
EC
»-o
Zw.
I
RRDIflL DISTRNCE FROM CENTER
100
1-1.25
O -
B - B/i
« 507?
V. RflOIRL OISTflNCE FROM CENTER
100
FIGURE 5. Radial temperature distributions of the average axial temperature
for three radial heat loss operating conditions, V „ = 10 cm/s,
T = 2000K, 0% fuel-N.
IN
228
-------
o
o.
"2L
LU
Its-
UJ
CO
o
CJ
m -
Q/F. = 0.0 7.
Q/E = 3. 1 2
Q/E = 9. 3 X
0.25 0. bO 0.75
EQUIVRLENCE RnTIO
i. oo
1. 25
FIGURE 6. Effect of equivalence ratio on radially averaged combustion effi-
ciencies for three diff
TAD »= 2000K, 0% fuel-N,
ciencies for three different radial heat loss modes. VIN - 10 cm/s,
229
-------
1. 25
1. 00
0. 75
0. 50
3 6
RELRTIVE HEflT LOSS
9
(X)
12
FIGURE 7. Effect of radial heat loss on radially averaged combustion effi-
ciencies as a function of equivalence ratio. V
AD
2000K, 0% fuel-N.
IN
10 cm/s,
230
-------
UJ
1C
1*1
ECui
UJ
0-
m
3
coo
noo K
s ib
INLET VELOCITY (CM/SEC)
o.
T - 2000 K
+ - I- 1.10
• - I- I.CO
« - I- Q-7S
a - I- 0. bO
INLET VEI.OCITT (CM/SEC)
Eg-
to
"
T - 1100 K
+ - I- 1.
A - I- 1.
o - i- o.
s -.0
INLET VELOCITY ICM/SEC)
IS
in
1
o
T - 2000 K
+ - I- 1. 10
• - !• i-ss
o - i. o.ls
o - i- o. so
INLET VELOCITY ICM/SEC)
IS
FIGURE 8. Effect of inlet velocity on the average axial centerline tempera-
ture and the axial centerline combustion efficiency for an insulated
combustor and a mixture containing 2% fuel-N.
231
-------
a
CJ
CJ
LU
i—i
CJ
I—I
LL
LU
CQ
a
u
X -
o
a
v
v
v
v
v
10
10
5
5
5
CM/S,
CM/S,
CM/S,
CM/S.
CM/S,
T
T
T
T
2000
1*700
2000
noo
1400
K
K
K
K
K
0.25 0. SO 0.75
EQUIVRLENCE RRTIO
1. 00
1. 2!
FIGURE 9. Effect of fuel/oxidant equivalence ratio on the axial centerline
combustion efficiency as a function of inlet velocity and volu-
metric energy release for an insulated combustor and a mixture
containing 2% fuel-N.
232
-------
o.
CO
o
I—I
CO
cc
UJ
C3-.
LJ
X
Q
I
UJ
i—i
a
a_
•4
a_
CM
m
CD
w ~
v
V
V
V
V
5 CM/S, T
5 CM/S, T
5 CM/S, T
10 CM/S, T
1C CM/S, T
14CO
1700
2000
1700
2000
0.25 0.50 0.75
EQUIVniENCE RRTIO
i. oo
1. 25
FIGURE 10. Effect of equivalence ratio on conversion of 2% fuel-N to NO for
various operating conditions of an insulated combustor.
233
-------
CO
O
i — i
CD
cc
UJ
a_
CD
a
C_J
X
a_
a
_i
UJ
m -
CD -
I _
(I-
(! =
(1 =
0. SO
0. 75
1. CO
1. 10
ti—
-©-
25 50 75
RRDIflL DISTflNCE FRQM CENTER
100
FIGURE 11. Radial gradients in the NO yield for an Insulated combustor.
VIN " 10
' TAD * 2000K> 2%
234
-------
CATCOM* CATALYST AXIAL TEMPERATURE
PROFILE MEASUREMENTS
By:
I. T. Osgerby
B. A. Olson
H. Lew
A. Cohn, Electric Power Research Institute
Engelhard Industries Division
Engelhard Minerals and Chemicals Corporation
Menlo Park, Edison, New Jersey 08817
ABSTRACT
Studies have been conducted by Engelhard under a Westinghouse
subcontract (No. 51-9-644006) funded by EPRI (Contract No. RP421)
to experimentally determine the performance characteristics of
CATCOM* catalysts under simulated gas turbine combustion conditions
using #2 diesel fuel. A series of axial catalyst monolith temperature
profile measurements were made for two lengths of catalyst using
small diameter, high temperature, sliding thermocouples positioned
in sealed monolith channels. Process variable effects of air preheat
temperature, adiabatic flame temperature (based on fuel/air ratios),
catalyst inlet reference velocity, and catalyst length were determined.
Combustion efficiencies greater than 99% were achieved at adiabatic
flame temperature of 2150°F (1176°C) or greater and NOX emissions
averaged less than 3 ppm -by volume. Catalyst durability for short time
periods was demonstrated up to operating temperatures of 2600°F (1427°C)
and at intentional upset conditions of fuel nozzle failure and partial
blockage of the catalyst in>et face.
*CATCOM and CATATHERMAL are tradenames of Engelhard Minerals and
Chemicals Corporation.
235
-------
DEVELOPMENT OF IMPROVED CATALYST SYSTEMS
By:
J. P. Kesselring, W. V. Krill, M. J. Angwin,
and H. L. Atkins
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
ABSTRACT
The performance of a graded cell catalyst and a single cell catalyst of
identical composition has been compared. The graded cell system achieved nearly
three times the maximum throughput prior to surface blowout, and also showed
significantly improved axial temperature uniformity. The development of cata-
lytically active monolithic substrates was initiated with tests of five special
formulations and three commercially available monoliths. All showed good com-
bustion performance, but thermostructural weakness at high temperatures remains
as a problem needing further investigation,
Work supported by U.S. Environmental Protection Agency Contract 68-02-3122.
236
-------
INTRODUCTION
Limited fossil fuel supplies and environmental constraints make low
emissions and high thermal efficiency necessities for combustion systems.
Catalytic combustion is actively being investigated as a means of achieving
these goals. Nearly all steady-state combustion systems — including commer-
cial and industrial boilers and furnaces, home heaters, and gas turbines—can
be modified or redesigned to incorporate catalytic combustors.
Catalytic combustors typically use ceramic or metallic honeycomb mono-
lithic supports with many small diameter cells. As shown in Figure 1, premixed
fuel and air flow through the cells, where a catalyst coating on the interior
surface promotes conversion of fuel and air to combustion products. Near the
cell entrance, where most of the gas is at low temperature, gas-phase chemical
reactions are unimportant. In this region, catalytic wall chemical reactions
control heat release. Further down the channel, where wall reactions have pre-
heated the gas to high temperature, gas-phase reactions become active. In
this region fuel i?. rapidly consumed by a "flame type" phenomenon which con-
trols the amount of unburned hydrocarbon emissions that escape from the system.
Under normal operating conditions, wall and gas-phase reactions are
active, and very little unburned hydrocarbon escapes the bed for lean and stoi-
chiometric initial mixture ratios. However, experimental observations have
shown that above a certain mass flow limit, small increases in flowrate cause
an abrupt rise in unburned hydrocarbon emissions. The abruptness of the in-
crease indicates that a "flame type" phenomenon has been extinguished. This
condition, called breakthrough, represents an upper mass throughput for low
unburned hydrocarbon emissions.
Increasing the mass throughput in a catalytic bed to levels much above
the breakthrough point can cause the front of the bed to become cool. It has
237
-------
been experimentally found that small increases in mass throughput, once the
front end of the bed has become cool, can cause the cool region to spread
downstream. At this point, all wall reactions are extinguished and the entire
bed becomes cold. This condition, called blowout, represents the maximum mass
throughput for heterogeneous combustion.
Blowout and breakthrough for catalytic combustors depend on a number of
parameters, including initial fuel/air ratio and temperature, honeycomb con-
figuration and dimensions (cell diameter, void fraction, bed overall diameter),
length of monolith, and bed and catalyst materials. As described in Refer-
ence 1, the use of a catalytic combustor system which has high mass throughput
(high blowout limit) and low emissions (no breakthrough) can be constructed by
joining two or more bed segments in series. The first segment has cells large
enough to prevent blowout but small enough to convert sufficient fuel to meet
the preheat/blowout requirement of the second bed segment. The second segment
has smaller diameter cells to convert more of the fuel and further heat the
bulk gases. The last segment has very small diameter cells and the entering
gas preheat is sufficient to light off the homogeneous reactions. This system,
called the graded cell catalyst, is depicted in Figure 2.
While the graded cell catalyst system is capable of achieving high
throughput with low emissions, the question of catalyst life remains. Experi-
mental results (Reference 2) have shown that maintaining a finely dispersed
catalyst coating on substrate materials at high temperatures is difficult, and
that the interaction between metal oxide catalysts and substrate materials at
high temperatures can result in severe thennostructural weakening of the sub-
strate. Therefore, the desirability of fabricating substrates directly from
materials which include active oxides and which have adequate thermostructural
strength is apparent. These systems, called active monoliths, are postulated
to have improved life while operating at combustor temperatures.
The remainder of this paper describes graded cell and single cell cata-
lyst comparative testing and formulation and testing of active monolith systems.
GRADED CELL AND SINGLE CELL CATALYST TEST RESULTS
The high throughput performance of graded cell catalysts had previously
been documented (Reference 3), but a direct comparison between identical graded
238
-------
cell and single cell catalyst configurations had not been conducted. In order
to compare the performance of the two systems, a series of comparative tests
were run. Two catalyst configurations were supplied by UOP, Inc. on DuPont
Torvex alumina honeycomb substrates. The graded cell system consisted of
three 2.54-cm segments with cell sizes of 6.4 mm, 4.8 mm, and 3.2 mm. The
single cell system used three 2.54-cm segments with a 3.2-mm cell size. Both
systems were washcoated and catalyzed by UOP with a proprietary catalyst.
Pretest analysis of the single cell model indicated a surface area and
dispersion of 4.88 m2/g and 1.22 ymoles H_/g, respectively, while the graded
cell indicated 5.36 m2/g and 17.63 ymoles/g catalyst. Although initially the
graded cell had a much higher dispersion than the single cell, following
5 hours of aging, dispersion for both configurations should be essentially the
same. This was verified by post-test analyses which indicated greatly reduced
dispersions of 0.39 and 1.3 for the single and graded cell monoliths, respec-
tively.
A comparison of the graded cell and single cell catalyst configurations,
characterizations, and combustion performance on natural gas fuel is given in
Table I. Both systems were operated in an identical manner during testing,
which included 5 hours of steady-state operation under lean conditions at
1560K temperature, followed by maximum throughput tests. As a result of this
testing, the following was noted:
• At a standard heat release rate of 105.5 MJ/hr, emissions from the
graded cell and the single cell monoliths were approximately the
same.
• Volumetric heat release at maximum throughput for the graded cell
was almost three times higher than for the single cell system
(7.4 x 106 J/hr-Pa-m3 versus 2.6 x 106 J/hr-Pa-m3).
• Under fuel-rich conditions, both reactor types maintained combus-
tion at preheat temperatures as low as 320K. However, hydrocarbon
emissions from the single cell configuration were an order of mag-
nitude higher than for the graded cell catalyst, as shown in Fig-
ure 3.
239
-------
• At the standard heat release rate of 105.5 MJ/hr, the graded cata-
lyst had a much more uniform axial temperature profile than the
single cell system
The HET code (Reference 1) was used to predict surface temperature pro-
files in the graded cell and single cell reactors, and the results are shown
in Figure 4. Excellent agreement between code prediction and surface tempera-
ture as measured by in-depth thermocouples is noted. The more uniform tempera-
ture profile for the graded cell system is important, since in order to avoid
poisoning effects, the catalyst should operate at as high a temperature as
possible over its entire length. As shown in Figure 4, the single cell reactor
runs a significant risk of being poisoned, while the graded cell reactor has a
high, relatively uniform temperature profile. In addition to running the risk
of poisoning, the single cell reactor is also operating in an unstable mode.
A slight change in inlet conditions could move the reaction zone downstream in
the combustor, eventually resulting in blowout.
These comparative tests verify that combustors with high volumetric heat
release rates can best be achieved with catalysts in the graded cell configura-
tion with no penalties in combustible or NO emissions. In addition, the high
uniform temperature of the catalyst walls throughout the length of the graded
cell model may have important implications in avoiding catalyst poisoning or
liquid fuel deposition on the walls.
ACTIVE MONOLITH CATALYSTS
High temperature combustion catalysts have many requirements placed on
their performance. They must be active in promoting the oxidation of hydro-
carbons and carbon monoxide, be capable of high temperature use without melt-
ing or losing activity, maintain structural integrity in combustion conditions,
and have good thermal shock resistance to withstand repeated thermal cycling.
Currently available materials have difficulty in meeting all these requirements.
Since no ideal high temperature catalysts are available, choosing a
material for high temperature catalytic ur.e requires a series of assumptions
and approximations. The first step is to test materials which are active at
240
-------
low temperatures and seem capable of tolerating high temperatures. Some of
these materials have been tested under combustion conditions, and include:
• Platinum and other noble metals applied to high temperature sub-
strates, such as alumina or zirconia
• Base metal oxides applied to high temperature substrates
Both types of catalysts showed good initial high temperature activity,
but lost activity and structural stability after prolonged high temperature
combustion. The noble metals tend to agglomerate or volatilize, while the
base metals tend to interact with the substrate, destroying its structural
stability.
To solve these problems, the active monolith concept was developed.
In this type of catalyst, an active metal oxide or mixed metal oxide is incor-
porated into the structure of the monolith when the monolith is formed. This
type of catalyst solves the problems inherent in the previously tested types
of catalysts. It depends on base metals, thus eliminating the problem of noble
metal agglomeration or volatilization. By incorporating the base metal directly
into the monolith, interactions are controlled during the formulation steps and
further interaction (destruction) should not occur during test conditions.
Active monoliths will have metal ions that can move readily between
valence states; therefore, they are nonstoichiometric or defect structures.
It is therefore reasonable to choose an active monolith where the defect
structure is incorporated into a well-known refractory structure. One example
of this type of system is MgAl^O, eNiAl20,, where nickel is incorporated into
the spinel structure. It is also important to use phase diagrams to choose
high temperature ceramic formulations which will not melt at the intended use
temperature, and which will have reduced volatility of the active metal oxide,
thereby extending the lifetime.
Based on these configurations, five active "monoliths" were prepared
by Trans-Tech, Inc. to Acurex specifications. Three were prepared by pressing
and sintering oxide powders into disks approximately 5.7 cm in diameter and
3.2 cm long and drilling 18 to 30 0.64 cm holes axially. The other two were
bundled sets of hollow tubes. The surface areas of these dense materials were
assumed to be negligible. Table II lists the active monolith materials tested.
241
-------
ACTIVE MONOLITH TEST RESULTS
Although all the active monoliths were produced with a single cell size
and were limited in length and surface area, they showed good catalytic activ-
ity. Test results are summarized and compared in Table III.
For the three monolith disks, bed temperatures were undoubtedly higher
than those recorded as the catalyst face was visibly cooler around the holes
which had been plugged with ceramic cement to secure the bed thermocouples.
The flow of mainstream air and fuel was also blocked on the monoliths at the
bed perimeter, since there were no holes closer than 0.5 cm from the edge.
Model EPA-041 was a MgAl-0,'NiAl_0, composition. Combustion was stable
with natural gas under fuel-lean conditions down to a minimum preheat of 436K.
Blowout of the catalyst bed did not occur at the highest throughput attained,
849,000 hr space velocity. Combustion was relatively unstable with natural
gas under fuel-rich conditions, operating only between preheat temperatures
of 546K (523°F) and 621K (658°F). On diesel fuel, lean, EPA-041 sustained
combustion to a minimum preheat of 587K. During maximum throughput, the bed
was not blown out even at a space velocity of 1,152,000 hr"1. Throughout the
testing, CO emissions were below 50 ppm, and NO ranged from 1 to 30 ppm. After
10 hours, the center of the bed had cracked, and the thick outer ring had frac-
tured cleanly into a few large chunks. The original blue color and smooth sur-
face appearance were unchanged.
Model EPA-042 was a LaAlO.,• LaCrO- composition. This catalyst performed
well on natural gas (rich and lean) and diesel (lean) fuels. Emissions of NO
did not exceed 20 ppm. On natural gas, the minimum preheat was 464K, both
lean and rich. At a throughput of 874,000 hr space velocity, the bed was
not blown out. Operating lean on diesel fuel, the minimum preheat was 589K,
nearly the same as for EPA-041. Blowout occurred at a space velocity of
1,124,000 hr~ . Post-test, the monolith retained its original red-brown color
and chalky surface texture. As with EPA-041, structural breakdown occurred
with moderate fracturing of the thicker outer section and greater fracturing
of the thinner center sections.
EPA-045, the third Trans-Tech monolith model, was a MgA^O, «Fe 0,
composition. Operating on natural gas, lean, the minimum preheat for this
242
-------
catalyst was 533K and the maximum space velocity was 874,000 hr (without
experiencing blowout). Minimum preheat on diesel, lean, was 397K. This was
considerably lower than the minimum preheat temperatures for diesel fuel from
earlier tests, where the required preheat temperature was at least 550K. NO
emissions during this test increased slightly in the low preheat range, 533K
to 397K, but did not exceed 30 ppm. EPA-045 was the least durable of the
three Trans-Tech monoliths tested, cracking more severely during testing. The
color changed from a dark brown with light brown patches to light brown on the
exposed surfaces. This indicates the probable change in iron oxidation state
to the catalytically active MgFe,0, .
The first set of tubes fabricated by Trans-Tech was a MgAl^-MgCr^
composition, Model EPA-046. This lavender-colored material was formed into
5.1 cm long tubes having nominal dimensions of 0.64 cm OD and 0.32 cm ID.
Forty-four tubes, bundled and wrapped in Fiberfrax insulation, were supported
vertically within a holder tube on a disk of alumina Torvex 2.5 cm long by
5.0 cm diameter, honeycombed with 0.48 cm diameter cells. Three tubes in the
center of the bed were cemented together to form a cavity for the bed thermo-
couple.
Performance of EPA-046 on natural gas, lean, was good. CO and NO emis-
sions were at or below 18 and 10 ppm, respectively, from preheat temperatures
of 658K to ambient. There was no blowout of the bed up to a throughput of
8.8 x 106 J/hr-Pa-m3. However, on natural gas, rich, and on diesel fuel, bed
uniformity could not be maintained. From initial temperatures of 1644K to
1700K, the bed would cool and extinguish rapidly, beginning at the front face.
There was no loss in catalytic activity after 5 hours of testing. Twelve of
the tubes were broken; the rest were recovered undamaged.
The second set of Trans-Tech tubes (EPA-048) was of MgAl^'Fe^ in a
3:1 mole ratio with 40 percent Cr as a cermet. The tubes were 0.95 cm OD by
0.32 cm ID and were instrumented and supported in the same manner as EPA-046.
On natural gas, the surface temperature was maintained below 1589K. Minimum
preheat temperatures were 367K, lean, and 583K, rich. Blowout occurred at
8.8 x 10s J/hr-Pa-m3. Also on natural gas, the minimum operating temperature
was investigated by increasing theoretical air at a fixed preheat of 533K.
243
-------
Stability became marginal below 1256K (1800 F) as partial darkening occurred
on the front face of the tubes.
To maintain combustion on diesel fuel, the surface temperature was
raised to 1644K. Minimum preheat temperature observed was 683K. Operation
was relatively poor, and due to difficulties in relighting the catalyst,
diesel testing was terminated. During the 5-hour test period, combustion
efficiency was good and NO emissions were less than 10 ppm. After testing,
all the catalyst tubes were fractured. The tube ends of the front face of
the bed had light brown streaks while the rest of the catalyst remained the
original dark brown. As noted with EPA-045 (Trans-Tech MgAl 0,'Fe_0, in 9:1
mole ratio), the color change indicates a probable change in oxidation state
of the iron present.
Several of the Trans-Tech active monoliths were subjected to SEM and
X-ray analysis. On model EPA-041 (MgAl^'NiAl^), a comparison between
pretest and combustion-tested models showed significant grain size growth.
The pretest surface is shown in Figure 5 and the combustion surface is shown
in Figure 6. Comparison of the three elemental analyses in these figures
shows all active elements to be present in the same relative quantities. No
significant depletion of nickel has occurred as with coated nickel catalysts,
and the composition is an excellent candidate for long life applications.
The LaAl03'LaCr03 catalyst (EPA-042) was tested for 8 hours with six
combustion cycles. A comparison of the pretest surface and X-ray analysis of
Figure 7 with the post-test combustion surface and analysis of Figure 8 shows
the surface has remained planar, and the elemental composition is relatively
unchanged.
On EPA-045, the pretest MgAl 0,*Fe 0, showed some initial lighter
"rusty" areas, which were tentatively identified as MgFe^O,. After combustion
testing, the dark MgAl.O,'Fe_0, turned considerably lighter. Post-test scan-
ning electron micrographs and elemental analysis showed that the post-test com-
bustion sample resembled the rusty pretest samples. A change of valence of the
iron apparently took place during combustion. The iron changed from a mixture
of +2 to +3 valence (Fe 0.) to all +3 valence (MgFe 0,). This change can
j *4 £- "
easily take place in the high temperature oxidizing conditions that occur
during combustion.
244
-------
The pretest analyses of the two surface oxidation states are shown in
Figure 9. The pretest rusty area showed a somewhat smaller granularity. The
post-test combustion sample closely resembles this rusty area in surface
appearance (Figure 10). The composition also resembles that of the rusty pre-
test sample, appearing somewhat depleted in iron. It appears that a valence
change in the iron takes place under combustion conditions.
The final model, EPA-048, was also analyzed pre- and post-combustion.
Of all models tested, it showed the least physical or compositional change.
The original structure of the surface is maintained (Figure 11 compared to Fig-
ure 12), and the active elements are all present in the same relative amounts.
These analyses of active monolith surface composition show the enhanced
material stability that can be achieved over that of coated catalysts. In
general, the surfaces may sinter, but none of the active elements are lost.
All of the materials analyzed are fully suitable for further development from
the material stability standpoint. Differences in activity and structural
characteristics were apparent, however, that make certain compositions appear
more attractive.
All of the Trans-Tech fabricated models showed good catalytic activity
as evidenced by the ability to run at low surface temperatures and with typical
variations in preheat and mass throughput when compared to coated catalysts.
As with other oxide catalysts tested, a higher level of preheat energy is re-
quired for ignition, but combustion efficiency was high at steady state oper-
ating conditions. NO levels were low for all five models tested.
X
Thermostructurally, the material formulations appear promising. The
crude, drilled monoliths fractured most readily at the thickest sections, but
not as badly as with many commercial monoliths tested under this program. It
might be expected that fabrication of these new Trans-Tech materials in mono-
lithic form would have better structural characteristics due to the more uni-
form web thicknesses and lower material densities that could be achieved.
BARE MONOLITH TEST RESULTS
Due to the success of the Trans-Tech active monolith models and the
noted continued activity of several coated catalysts that appeared to have
245
-------
no active catalyst remaining on the surface, it was suspected that several
commercial substrate materials may have limited activity of their own. The
three substrates tested in graded cell sets were:
• Corning yttria stabilized zirconia (MCB-10), model EPA-035
• Corning zirconia spinel (MCB-12), model EPA-043
• DuPont alumina Torvex, model EPA-047
All have essentially no surface area but were found to be catalytically active
to some degree. A summary of test results of the three models is given in
Table IV.
The Corning model EPA-035 had a pretest surface area measured as 0 m2/g
as expected. No additional pre- or post-test measurements were made on the
three substrate materials. Since no active coating was applied to the sub-
strates and surface degradation of the monolith was not expected to be sig-
nificant, combustion aging was omitted.
Shortly after lightoff of EPA-035, it was observed that nearly all of
the cell walls in the front of the bed had cracked. Also, the monolith had
moved downstream within the combustor tube until it rested on the downstream
thermocouple. Since the bed was badly fractured, it was not repositioned.
The structure did remain intact for the duration of the testing. Later, the
thermocouple in the rear of the bed ceased functioning.
Throughout the first 5 hours of testing on natural gas, EPA-035 was
uniform and NO emissions were low. Maximum CO under lean conditions was
X
26 ppm with negligible unburned hydrocarbons. Minimum preheat was 478K rich
and 372K lean. The blowout point was beyond the capacity of the facility at
a volumetric heat release rate of 3.86 x 106 J/hr-Pa-m3. All natural gas data
were taken at a bed temperature near 1561K. The catalyst was also tested on
diesel fuel. Stable combustion under lean conditions required an observed
minimum preheat of 550K and a bed temperature of 1644K. At the minimum pre-
heat value of 550K, the bed apparently went out due to insufficient prevaporiza-
tion of the fuel.
Test model EPA-035 exhibited excellent combustion characteristics in
the mid-temperature range. Similar results are expected at high temperatures,
246
-------
since the substrate is rated at a use temperature above 2144K. The thermo-
structural characteristics of the material are extremely poor, however. Early
fracturing did not affect testing since the catalyst was held sufficiently to
prevent shifting of the pieces, but total fracturing occurred when the bed was
removed from the facility.
EPA-043, Corning MCB-12, was described as zirconia spinel. It was
catalytically active at bed temperatures of 1556K to 1700K. NO emissions
were low, below 10 ppm, except during the maximum throughput test with diesel
fuel. CO emissions were £30 ppm. Minimum preheats required for combustion
on natural gas were 399K lean and 537K rich. The maximum throughput attained
with natural gas was dictated by facility flowrate limitations. On diesel
fuel, lean, the minimum preheat was 603K, and blowout occurred at a throughput
of 5.3 x 105 J/hr-Pa-m3. Following testing, the front (large cell) section
was moderately fractured, and the middle section very mildly fractured. The
rear segment was undamaged. The substrate had also changed to a somewhat
whiter appearance, similar to that noted for post-test MCB-10.
Test model EPA-047 consisted of uncoated DuPont alumina Torvex. With
natural gas, lean, combustion was stable and efficient at bed temperatures of
1533K to 1822K. Minimum preheat temperatures were 394K, lean, and 567K, rich.
At the facility maximum volumetric heat release rate of 5.9 x 106 J/hr-Pa-m
blowout did not occur. The minimum preheat on diesel fuel, lean, was 682K.
Blowout on diesel fuel occurred at the facility throughput limit of 7.5 x 10
J/hr-Pa-m3 volumetric heat release rate. Three and one-half hours of combus-
tion testing resulted in delamination of the Torvex cell walls.
Each of the uncoated substrates exhibited some catalytic activity when
operated in the medium temperature range (1478K to 1756K). The activity
appeared to be somewhat less than that of the Trans-Tech produced materials
as evidenced by minimum required operating temperatures and relatively little
flexibility in preheat temperatures for the bare substrates. All models pro-
duced low NO emissions and were quite combustion-efficient at the velocities
x
tested. Therefore, applications may be possible in systems where sufficient
preheat can be provided during ignition to initiate reactions.
247
-------
Clearly, however, all three monoliths have structural limitations.
Cracking of the Corning MCB-10 was severe, and the material is judged to be
unacceptable. Similarly, delamination of the DuPont alumina would be unaccept-
able in application. The Corning MCB-12 exhibited mild fracturing and may have
potential for material improvement. Resolution of the structural problems for
any of these substrates could lead to potential applications as active monolith
catalysts.
CONCLUSIONS
Based upon the comparative testing of the graded cell and single cell
catalyst systems, the graded cell catalyst has shown significant improvements
in both maximum throughput capability and temperature uniformity within the
bed. While excellent performance of graded cell catalyst systems (both noble
metal and oxide) have been achieved, the problems of thermostructural degrada-
tion and chemical interaction have led to the concept of an active monolith
catalyst, in which the active element is an integral part of the substrate.
Testing of early prototype active monoliths, including bare monoliths of com-
mercially available materials with no special active elements, have shown good
activity and combustion performance. Improvements in manufacturing techniques
and material formulations are required, however, to further develop adequate
structural life of these systems.
248
-------
REFERENCES
1. Kendall, R. M., et al. An Analysis of Catalytic Combustion in Monolithic
Honeycomb Bed. Proceedings: Third Workshop on Catalytic Combustion,
EPA-600/7-79-038, February 1979, pp. 197-238.
2. Kesselring, J. P., et al. Design Criteria for Stationary Source Catalytic
Combustion Systems. EPA-600/7-79-181, August 1979.
3. Krill, W. V. and Kesselring, J. P. The Development of Catalytic Combustors
for Stationary Source Applications. Proceedings: Third Workshop on Cata-
lytic Combustion, EPA-600/7-79-038, February 1979, pp. 239-268.
249
-------
TABLE I. COMPARISON OF SINGLE CELL AND GRADED CELL CATALYST PERFORMANCE
Parameter
Graded Cell
Catalyst
Single Cell
Catalyst
Catalyst
manufacturer
Cell geometry
Substrate
Surface area
(m2/g)
Pre-test
Post-test
Dispersion
(moles H2/g)
Pre-test
Post-test
Operating
temperature (K)
Maximum heat
release rate
(J/hr-Pa-m3)
Typical emis-
sions (ppm)
NOX
CO
HC
UOP
(3) 25.4-mm long
segments of 6.4,
4.8, and 3.2-mm
cell size
DuPont alumina
5.36
0.986
17.63
1.30
1561
7.4 x 106
UOP
(3) 25.4-mm long
segments of 3.2-mm
cell size
DuPont alumina
4.88
0.05
1.22
0.39
1561
2.6 x 106
1
30
12
2
60
35
250
-------
TABLE II. ACTIVE MONOLITH TEST CONFIGURATION
Composition
Configuration
LaAlO.- LaCrO_
with
40% Cr (cermet)
Cylinder, drilled
holes axially
Bundled hollow
tubes
251
-------
TABLE III. COMBUSTION PERFORMANCE OF ACTIVE MONOLITHS
Test
Model
Test
M d**
Measure*
Bed
Temperature
Lean Minimum Preheat
Temperature
o
v '
Maximum Volumetric
Heat Release Rate
MJ/hr-Pa-m3
(MBtu/hr-atm-f t3)
Natural
Gas
Diesel
Fuel
Natural
Gas
Diesel
Fuel
Ul
EPA-041
EPA-042
EPA-045
MgAl?0 -NiAl 0, 10 1367-1561 436 (325)
(2000-2350)
LaAlO .LaCrO
EPA-046 MgAl20 -MgCr 0,
8 1367-1500 464 (375)
(2000-2240)
MgAl 0 -Fe 0, 10.5 1506-1561 533 (500)
J (2250-2350)
5 1617-1711 354 (178)
(2450-2620) (ambient)
EPA-048
MgAl70,-Fe 0 -40% Cr 5 1033-1644 367 (200)
^ ^ (1400-2500)
587 (596) 12 (33)* 22 (59)*
589 (600) 14 (39)* 21 (56)
397 (255) 13 (34)* 18 (48)
--- 8.8 (24)*
--- 8.8 (24>
Below blowout point
Actual temperatures somewhat higher than those measured
-------
TABLE IV. COMBUSTION PERFORMANCE OF BARE MONOLITHS
to
l_n
OJ
Test
Model
EPA-035
(MCB-10)
EPA-043
(MCB-12)
EPA-047
(Torvex)
Test Temperature
Hours K
(°F)
7.5 1506-1867
(2250-2900)
5.5 1556-1700
(2340-2600)
3.5 1533-1822
(2300-2820)
Lean Minimum Preheat
Temperature
K
Natural
Gas
372
(210)
399
(258)
394
(250)
Diesel
Fuel
550
(530)
603
(625)
682
(768)
Maximum Volumetric
Heat Release Rate
MJ/hr-Pa-m3
(MBtu/hr-atm-ft3)
Natural
Gas
3.86
(10.5)*
3.68
(10.0)*
5.89
(16.0)*
Diesel
Fuel
5.34
(14.5)
7.73
(21.0)
Post-Test
Condition
Poor
Fair
Poor
Below the blowout point
-------
Surface reaction
control led
Gas phase reaction
control led
\
^conduction
radiation
Figure 1. Physical events occurring in one cell of a honeycomb
monolith catalvtic combustor.
-------
Figure 2. Graded cell catalyst segments. The large cell piece is placed at
the bed inlet, and the small cell piece at the bed exit.
-------
5.
CL
(Sl
c
o
c/i
i/l
Single eel 1
'reactor
r~ 1 f\ '
o
-Q
u
o
"O
Graded cell
reactor
10'
100
200
300
400
500
600
300
400 500
Preheat temperature (K)
600
Figure 3. Hydrocarbon emissions comparison for two catalyst
geometries at rich conditions.
256
-------
Single cell reactor
• 140,000 hr'1 data
Graded eel 1 reactor
D 184,000 hr"1 data
140,000 hr" prediction —- 184,000 hr"1 prediction
1600
1500
1400
21 1300
3 1200
I 1100
1000
900
800
700
600
Preheat
temp
i!
-2
2 4
Bed depth (cm)
Figure 4. Comparison of graded cell and single cell reactor
temperature profiles with HET code predictions.
257
-------
a. Surface appearance at 1000 X magnification
b. Elemental analysis of area a
Figure 5. MgAl^O -MiAl-CL pretest surface analysis,
258
-------
a. Surface appearance at 500X magnification
b. Surface appearance at 2000X magnification
Figure 6. (above left, above right, and
opposite)
MgAl-CL'NiAl-CL post-test combustion analysis
c. Elemental analysis of above area
-------
a. Surface appearance at 1000X magnification
b. Elemental anal/sis of area a
Figure 7. LaAlO-LaCrO, pretest surface analysis
260
-------
:
3
a. Surface appearance at 500X magnification
b. Surface appearance at 2000X magnification
Figure 8. (above left, above right,
and opposite)
LaAlO -LaCrO- post-test combustion analysis,
c. Elemental analysis of above area
-------
R '
a. Normal surface appearance at 1000X
magnification
b. Elemental analysis of
area a
c. Rusty surface appearance at 1000X d. Elemental analysis of
magnification area c
Figure 9. MgAl20.'Fe 0. pretest surface analysis
262
-------
a. Surface appearance at 500X magnification
Figure 10. Cabove left, above right,
and opposite)
MgAl20.-Fe 04 post-test combustion analysis
b. Surface appearance at 2000X magnification
c. Elemental analysis of above area
-------
:
-
a. Surface appearance at 500X magnification
Figure 11. (above left, above right, and
opposite)
MgAl204«Fe 0 -40% Cr pretest analysis.
b. Surface appearance at 2000X magnification
c. Elemental analysis of above area
-------
a. Surface appearance at 500X magnification
b. Surface appearance at 2000X magnification
Figure 12. (above left, above right, and
opposite)
MgAl20.-Fe 0-40% Cr post-combustion test analysis
c. Elemental analysis of above area
-------
CATALYZED COMBUSTION
OF LEAN FUEL/AIR MIXTURES
By
R. W. Schefer and F. Robben
Lawrence Berkeley Laboratory
Berkeley, CA 94720
ABSTRACT
A study has been made of the combustion characteristics of lean fuel/air
mixtures flowing over heated catalytic and non-catalytic surfaces. The objec-
tive of the investigation was to develop a better understanding of the effect
of a catalyst on the combustion process as the lean flaramability limit of
various fuels is approached. The fuels studied were H-, Co^e* anc* C3H«' ^as
phase combustion ignition temperatures were determined over a range of
equivalence ratios from boundary layer density profiles obtained using
Rayleigh scattering. Surface reaction ignition temperatures and surface heat
rates were also determined for the above fuels. A comparison of the results
for a vacuum deposited platinum surface with results for a relatively "non-
catalytic" quartz surface showed that the gas phase ignition temperature is a
function of catalyst surface reactivity. At lower temperatures the quartz
surface was essentially noncatalytic and resulted in gas phase ignition
temperatures up to 100 K lower than with the platinum surface. At increased
surface temperatures the catalytic activity of the quartz surface approached
that of the platinum and gas phase ignition temperatures were identical for
both surfaces. A detailed study of C_H0 oxidation on a platinum surface
J O
showed that for temperatures less than 670 K, CO is the primary oxidation
product and for temperatures greater than 870 K, complete oxidation to C02
occurs.
This research was supported by the U. S. Department of Energy, Division
of Transportation Energy Conservation under Contract W-7405-ENG-48.
266
-------
SECTION 1
INTRODUCTION
Lean combustion shows considerable promise as a means of reducing pol-
lutant formation in combustion systems (1). However, the extent to which this
concept can be effectively utilized is limited by the combustion characteris-
tics of a particular fuel as its lean flammability limit is approached.
Catalyzed combustion has been demonstrated to be effective in extending the
stable burning region well below conventional lean gas phase flammability
limits while at the same time providing low levels of unburned hydrocarbons
and CO (2, 3). While preliminary tests have been quite promising, a number
of aspects of catalyzed combustion are either not well understood or the
necessary data are lacking. In the development of an optimum catalyst design
for lean combustion applications it will be necessary to 1) develop a greater
understanding of the role of the catalyst in the combustion process and 2) to
determine high temperature catalytic surface reaction mechanisms and reaction
rate data. It is the goal of this investigation to provide some of this
necessary data for future catalytic combustion design development.
In the present study the surface ignition and gas phase ignition charac-
teristics under conditions approaching the lean flammability limits of several
fuel/air mixtures were investigated for boundary layer flow over a heated surface.
The fuels used in the study were H0, C0H,, and CQHQ. To better determine the
/ Z O JO
effect of the catalyst on the combustion process both a catalytic platinum
surface and a relatively noncatalytic quartz surface were utilized in the
study. Results are also presented on a detailed study of lean C^Hg oxidation
over a platinum surface.
267
-------
SECTION 2
EXPERIMENTAL PROCEDURE
Details of the experimental system used in the present investigation
have been described in References 4 and 5. Briefly, the configuration consists
of a quartz plate with vacuum deposited platinum heating strips mounted as one
wall of a 5 cm square, open, atmospheric jet of premixed fuel and air.
The gas flow is parallel to the plate surface and results in the formation of
a velocity and thermal boundary layer in the region near the wall surface.
Control of the surface temperature is provided by five platinum heating strips
which have been vacuum deposited on the quartz plate and can be electrically
heated to temperatures approaching 1400 K. The heating strips are oriented
perpendicular to the flow and are of varying widths to improve temperature
control near the plate leading edge. The extent to which gas phase combustion
and surface reaction occurs can thus be controlled through variations in plate
surface temperature and fuel/air mixture ratio.
Surface temperatures greater than 1000 K were measured using a disappear-
ing filament type optical pyrometer and an emissivity of 0.65 for the
platinum surface. Surface temperatures below 1000 K were determined from the
measured resistivity of the platinum strips and, independently, with a labora-
tory built infrared detector. Using this technique plate surface temperatures
could be measured to an estimated accuracy of ± 15 K.
In the present investigation the platinum catalytic surface was provided
by orienting the quartz plate so that the side with the platinum heating
strips was directed toward the flow. To obtain a "noncatalytic" surface the
plate was reversed so that the uncoated quartz surface was directed toward the
flow.
268
-------
The presence of gas phase combustion in the boundary layer was deter-
mined from total gas density profiles measured using Rayleigh scattering.
As discussed in Reference (5), the boundary layer density profiles are self
similar along the plate when no gas phase combustion is present. However,
heat release due to gas phase combustion results in density profiles charac-
terized by a more positive curvature in the high temperature region near the
plate surface and an increase in thermal boundary layer thickness. To deter-
mine the gas phase ignition temperature the equivalence ratio was fixed at a
desired value and the plate surface temperature was increased in steps of
25 K. After each increase in surface temperature the system was allowed to
come to steady state and the boundary layer was traversed using Rayleigh
scattering to measure the resulting gas density profile. A comparison of
these profiles with profiles for no fuel added at the same surface temperature
then made it possible to determine that temperature, defined as the gas phase
ignition temperature, at which heat release due to gas phase combustion
resulted in a detectable departure from self-similar behavior. The results
presented in this paper were based on an analysis of density profiles at a
distance of 50 mm from the heated wall leading edge since at distances
greater than this, fluctuations in the jet mixing layer become a significant
source of disturbance to the boundary layer flow.
To determine steady state surface energy release rates the plate tempera-
ture was initially adjusted to the desired value with no fuel present in the
flow. This was done by varying the power input to the individual heating strips.
Typically strips 1 and 2 required greater power inputs for a constant plate
temperature due to the higher heat transfer coefficient near the heated wall
leading edge. Fuel was then added to the flow, which increased the surface temper-
atures to some higher nonuniform value due to surface energy release. The
power inputs to the individual strips were then decreased until the desired
temperature was again reached. The resulting difference in power input to
each strip, with and without fuel addition, is equal to the average heat
release due to surface chemical reaction for that particular strip provided
heat losses due to other sources remain constant. For a given plate tempera-
ture radiative losses are constant. Experimental measurements and numerical
calculations have also shown that the boundary layer temperature profiles are
269
-------
increased to a value of 1220 K at 4> = 0.1. This behavior was characteristic
of all fuels investigated, that is a weak dependence of ignition temperature
on equivalence ratio at higher equivalence ratios and a rapid increase in
ignition temperature as the equivalence ratio was reduced.
With the quartz surface at lower surface temperatures, ignition occurs
at approximately 100 K lower than with the platinum. The differences in
ignition temperature decreases as the equivalence ratio is reduced until at
an equivalence ratio of 0.1 identical ignition temperatures of 1220 K are
achieved with both surfaces. Similar results were obtained with C0H0 and
J o
C2H,. In both cases at an equivalence ratio of 0.9 ignition with the platinum
surface occurred at approximately 1170 K and with the quartz surface it
occurred at surface temperatures approximately 100 K lower. At an equivalence
ratio of 0.5, where the ignition temperature was 1220 K for both fuels with a
platinum surface, a difference in ignition temperature of only 25 K was
observed in comparing the quartz and platinum surfaces.
This behavior is consistent with an increase in catalytic activity of
the quartz surface at higher temperatures. Numerical calculations for H,,/air
mixtures flowing over a catalytic surface (7) have shown that catalytic
reactions at the plate surface, particularly surface oxidation of the fuel and
radical recombination, result in a surpression of gas phase reactions in the
boundary. Thus, the increase in gas phase ignition temperature with respect
to the platinum surface at higher surface temperatures would seem to imply an
increase in catalytic activity which, at a temperature of 1220 K, equals
that of the platinum in the case of KL. To verify this, surface heat release
rates were measured for H?/air at an equivalence ratio of 0.2 flowing over a
quartz surface. Assuming a one step model for surface reaction in which HL
a *•
and 02 react to from H20 as a product (At^ = 2.38 x 10 J/kg-mole) values
for the surface reaction were calculated. Those results are shown as a
function of surface temperature in Fig. 5. Note that the surface reaction
rate for quartz has been normalized by the reaction rate for H2/air mixtures
on a platinum surface obtained previously (8). At temperatures less than
870 K no surface reactivity was measured for the quartz surface. However,
as the surface temperature was increased above 870 K, a gradual increase in
surface reactivity relative to platinum was observed, followed by a rapid
272
-------
increase at temperatures greater than 1170 K. At a surface temperature of
1220 K the catalytic activity of the quartz surface was equal to that of the
platinum. Thus, it appears that the observed behavior in gas phase ignition
characteristics for the platinum and quartz surfaces is due to the increased
activity of the quartz surface at higher temperatures.
Propane Surface Oxidation Studies
Typical experimentally measured surface energy release rates for C0H0/air
-5 O
are shown in Figs. 6 and 7 for plate temperatures of 670 K and 870 K and an
equivalence ratio of 0.2. The free stream velocity was 1.5 m/s. As mentioned
previously, the data points represent average values of the energy release rate
over the individual heating strips and not local values. It can be seen that the
maximum heat release rates occur at the plate leading edge and then rapidly
decrease as one moves downstream. As discussed by Rosner (9), this behavior
indicates that mass transfer effects are important under the above conditions.
For viscous flow over a flat plate with reactions occurring at the surface,
the transport of reactant species from the free stream to the plate surface
varies with streamwise distance from the plate leading edge. Near the leading
edge where the boundary layer is thin, convective and diffusive transport
rates to the surface are rapid and the surface reaction rate approaches its
kinetically limited value. Farther downstream the boundary layer becomes
considerably thicker and transport rates to the surface are insufficient to
prevent local reactant depletion near the surface. Under these conditions
the surface reaction rate is limited by the maximum rate at which reactants
can be transported to the surface.
Also shown are predicted local surface heat release rates calculated
using a computational scheme developed previously (7). Briefly, the program
uses a finite difference approach to solve the two dimensional boundary layer
equations for flow over a flat plate surface with simultaneous gas phase and
surface reactions. In the present application, no gas phase reactions are
assumed to occur and the program is used to describe the convective/diffusive
transport of reactant and product species through the boundary layer to and
from the plate surface. Surface reactions are included in the calculations
in the form of a species boundary condition at the surface in which the rate
273
-------
of diffusion of a particular specie to the surface is equal to the surface
reaction rate. The rate expression used to describe C ~HQ oxidation at the
plate surface is based on the work of Hiam, et al. (6) and is given by
Rg = 1.1 x 107 [H2] exp (-17000/RT) kg m°leS (1)
m s
Q
where [H_] is the concentration of H_ at the surface (kg moles/m ). Initial
modelling calculations were based on the assumption of a one step reaction in
which coHg and 0- react at the surface to form C02 as the product. These
results are indicated by the solid lines in Figs. 6 and 7. As can be seen in
Fig. 7, the predicted and experimental results agree well at temperatures
greater than 870 K. However, at 670 K it was found that the predicted heat
release ratios were approximately a factor of 2 higher than those measured
experimentally. If it is assumed that only partial oxidation of C0HQ to CO
J O
occurs at the surface at lower temperatures then, as shown by the dashed line
in Fig. 6, good agreement is obtained with experimental results. These
results agree qualitatively with the measurements of Marteney and Kesten (7)
in which it was found that the ratio of C02 to CO in the product species from
a catalytic reactor (utilizing a Matthey Bishop manufactured catalyst
material) increased with increasing exhaust temperature.
An Arrhenius plot of experimentally measured average surface energy
release rates for C_,Hg as a function of inverse surface temperature for
heating strips 1 and 2 is shown in Fig. 8. Predicted average surface heat
release rates for strips 1 and 2 assuming C02 (solid line) and CO (dashed
line) as the product species are also shown. Once again the transition from
C02 to CO as the product of surface oxidation is apparent as the surface
temperature is decreased from 870 K to 670 K. The effect of diffusion
limited behavior on apparent activation energy is also shown. The slope of
each curve at a given temperature is equal to E./R, where E. is the apparent
A A
activation energy for surface oxidation of C0H0 and R is the gas constant.
J O
At higher surface temperatures the apparent activation energy is significantly
less than the true activation energy. As the surface temperature is reduced
the apparent activation energy approaches the true activation energy as the
surface reaction rate becomes kinetically limited.
274
-------
SECTION 4
SUMMARY
The effect of catalytic surface activity on gas phase and surface
reaction ignition temperatures was studied for H_-, C_H0-, and C0H,-air
2. Jo / O
mixtures flowing over a platinum and a quartz surface. It was found that
the gas phase ignition temperature is a function of catalyst surface reactiv-
ity. That is at lower temperatures the quartz surface was essentially non-
catalytic and resulted in gas phase ignition temperatures up to 100 K lower
than with the platinum surface. At increased surface temperatures the
catalytic activity of the quartz surface approached that of the platinum and
gas phase ignition temperatures were identical for both surfaces. A detailed
study of C3Hg oxidation on a platinum surface showed that for temperatures
less than 670 K, CO is the primary oxidation product and for temperatures
greater than 870 K, complete oxidation to CO- occurs.
275
-------
REFERENCES
1. Sawyer, R. F. and Blazowski, W. S. Propulsion Effluents in the
Stratosphere, CIAP Monograph 2, Chapter 3. Department of Transportation
Report DOT-TST-75-52 (1975).
2. Pfefferle, W. C. The Catalytic Combustor: An Approach to Cleaner
Combustion. J. Energy 2, 3, 142 (1978).
3. Blazowski, W. S. and Walsh, D. F., Comb. Sci. Technonl. 10, p. 233
(1975).
4. Schefer, R. W., Robben, F., and Cheng, R. K. Catalyzed Combustion of
H2/Air Mixtures in a Flat Plate Boundary Layer. Proceedings of the
Catalytic Combustion Workshop, Ashville, North Carolina (Oct. 1978).
5. Schefer, R. W. Catalyzed Combustion of H^/Air Mixtures in a Flat Plate
Boundary Layer. I. Experimental Results. To be published in Comb.
and Flame (1980).
6. Hiam, L., Wise, H., and Chaikin, S. Catalytic Oxidation of Hydrocarbons
on Platinum. J^ Catal. _10, 272 (1968) .
7. Schefer, R. W. Catalyzed Combustion in a Flat Plate Boundary Layer. II.
Numerical Calculations. Western States Section/The Combustion Institute,
Stanford, CA (1977).
8. Schefer, R., Cheng, R., Robben, F., and Brown, N. Catalyzed Combustion
of H2/Air Mixtures on a Heated Platinum Plate. Western States Section/
The Combustion Institute, Boulder, CO (1978).
9. Rosner, D. E., Convective Diffusion as an Intruder in Kinetic Studies of
Surface Catalyzed Reactions, A.I.A.A.J. 2, 593 (1964).
276
-------
1.0
Q.8
C3H8
0.6
ro
^4
VJ
0.4
0.2
0.0
I
400 500
600
700
(K)
800
900
1000
Figure 1. Surface reaction ignition temperatures for CoHg and
platinum surface. U =1.5 m/s .
on a
-------
to
«vl
00
0.4
0.3
0.2
0.1
0.0
800
SiO,
SURFACE
REACTION
900
GAS PHASE
COMBUSTION
I
I
I
1000
1100
1200
I
1300 1400
(K)
Figure 2. Gas phase ignition temperatures for I^/air combustion. U^ = 1.5 m/s.
-------
1.0
0.9
VO
0.8
0.7
0.6
0.5
SURFACE
REACTION
SiO
GAS PHASE
COMBUSTION
Pt
0.4
0.3
I
900
1000
1100
1200
(K)
1300
1400
Figure 3. Gas phase ignition temperatures for C HQ combustion. U - 1.5 m/s.
Jo °°
-------
1.0
00
O
0.9
0.8
0.7
0.6
0.5
SURFACE
REACTION
SiO
GAS PHASE
COMBUSTION
Pt
0.4
0.3
I
I
900
1000
1100
1200
1300
1400
(K)
Figure 4. Gas phase ignition temperature for C2H, combustion
U = 1.5 m/s.
OO
-------
N5
C»
1.0
0.8
0.6
0.4
0.2
600
800
1000
1200
(K)
Figure 5. Nondimensionalized surface reaction rate constant as a function of
surface temperature for H£ oxidation on a quartz surface.
4> = 0.2, U = 1.5 m/s •
-------
o
,—I
X
CO
s
E
"-3
CO
o-
K)
00
KJ
0.5
10
20
30
40
x (mm)
Figure 6. Surface heat release rate as a function of distance along the plate for C3Hg/air
combustion on a platinum coated plate. Tg = 670 K, Vm = 1.5 m/s, = 0.2. Solid
line indicates numerical results for oxidation to CC>2; dashed line indicates results
for oxidation to CO.
-------
CO
GO
x (mm)
Figure 7. Surface heat release rate as a function of distance along the plate for
combustion on a platinum coated plate. Ts = 870 K, Um = 1.5 m/s, = 0.2.
line indicates numerical results for oxidation to CO,
for oxidation to CO.
Solid
dashed line indicates results
-------
ic
10
0.5
1.0
1.5
2.0
103/Tg (if1)
2.5
3.0
3.5
Figure 8. Surface heat release rate as a function of inverse temperature for
C3Hs/air combustion on a platinum coated plate. $ = 0.2, U^ « 1.5
m/s. Solid line indicates numerical results for oxidation to CO?;
dashed line indicates results of oxidation to CO.
284
-------
EXAMINATION OF THE NICKEL CATALYST ROLE IN RICH COMBUSTION
by
Gerald E. Voecks
Jet Propulsion Laboratory
California Institute of Technology
4800 Oak Grove Drive
Pasadena, California 91103
and
George Gavalas
Charunya Phichitkul
California Institute of Technology
Pasadena, California 91103
285
-------
EXAMINATION OF THE NICKEL CATALYST ROLE IN RICH COMBUSTION
Rich combustion has been shown to be maintained successfully with a
nickel catalyst in various experiments conducted at JPL. However, the
activity of these alumina supported catalysts was found to degrade slowly but
consistently with time and reduce the performance in extent of conversion and
maximum throughput. To better understand how to maintain catalyst activity and
reduce the severity of catalyst degradation, we have begun to study the rich
combustion of methane by nickel on alumina. Based on thermodynamic equilibrium
and experimental evidence, a mixture of methane and air containing 30% of
theoretical stoichiometric air will produce the maximum amount of hydrogen and
carbon monoxide without carbon formation and at a temperature near 1000°F.
Since the reaction conditions were to be carefully controlled to permit
kinetic studies, an alumina reactor tube 61cm long with a 0.5cm I.D. was used
and coupled with a traversing thermocouple located at the centerline. This
reactor was mounted inside an insulated furnace to allow isothermal operation.
Methane, oxygen and nitrogen were independently controlled to alter the
air/fuel ratio and throughput (space velocity) over a range of constant
catalyst bed temperatures.
Initial tests were conducted with the Girdler Catalyst Company catalyst
G-56B. The catalyst was ground and sieved to a 60-80 mesh and a catalyst bed
was packed to a length of 2cm in the center of the reactor. These tests were
designed to determine the temperatures at which controlled conversion could
be maintained for kinetic measurements at air/fuel ratios near 6.0. The
"light-off" temperature at which high methane conversion took place was found
to be at approximately 650°C (Figure 1). Once high conversion was achieved,
its extinction temperature limit where methane conversion ceased extended far
below the light-off temperature. More importantly, however, it was found that
two reactions were taking place at temperatures below where light-off had
occurred. Of the methane which reacted initially at the "pre-light off"
condition (i.e., below 650°C), partial oxidation to carbon monoxide was not
observed, but rather, complete oxidation to carbon dioxide took place (Figure
2). Following "light-off", carbon monoxide was the predominant carbon oxide
product with some carbon dioxide also formed (approximately 80/20 C0/C0~).
This compares favorably to the thermodynamic equilibrium ratio of 86/14.
During the subsequent reduction in operating temperature from 710°C to 500°C,
it was found that the methane oxidation reaction continued below the light-off
temperature to a greater extent than initial pre-light off. The rates of the
two competing reactions (or series of reactions) to form carbon monoxide or
carbon dioxide changed, reaching a crossover point at approximately 550°C. A
similar phenomenon has been observed during tests in a full sized catalytic
286
-------
rich combustor (methane flows of up to 5 Ibs/hr through a catalyst bed 3"
long x 3.6" diameter). It appears that the catalyst surface is altered
during the reaction sequence. Although this phenomenon is not clearly
understood and will be studied later in more detail, the primary direction at
this time was to examine more closely the catalyst activity and its
degradation in the kinetics controlled regime, i.e., pre-light off
temperatures, where no carbon monoxide is formed and only partial methane
conversion occurs.
Because the G-56B catalyst (25 wt.% Ni) was too active to control the
reaction at temperatures higher than 650°C, in-house catalysts were prepared.
Nickel was impregnated into a low surface area alpha alumina support in order
to minimize the activity degradation due to spinel (nickel aluminate)
formation. The catalyst preparation procedure consistently followed involved
the time controlled soaking of the alumina particles in a nickel nitrate
solution, filtering, drying at 120°C and finally calcination. The same
operating procedure of mixing oxygen, methane and nitrogen independently was
used. Examination of the catalyst during deactivation was pursued further
because of a similar occurrence which had been observed when commercial
nickel on alumina catalysts were used as the catalyst in the larger unit
described above.
The effect of catalyst calcining was the first parameter examined. As
shown in Figure 3 the relative difference in initial activity between 850°C
calcination in air for three hours and 400°C for one hour prior to reaction
is nearly an order of magnitude. After a period of five hours continuous
operation, the difference in activity had been reduced to half the
difference observed at one hour of operation. To explain the relatively
stable but less active character of the calcined catalyst which had undergone
the high temperature treatment, the catalyst was reduced prior to operating
under rich combustion conditions. Figure 4 shows that the relatively stable
activity is altered and, further, that it also decreases in activity with
time of operation and with increasing temperature of operation. The fact
that the stable activity which the calcined catalyst initially possessed
could be enhanced by reduction indicated that the catalyst activity
responsible for producing the same products, C02 rather than CO, may be due
to two different catalyst features. In observing the effect of cycling the
catalysts, shown in Figure 5, it can be seen that an irreversible process is
taking place each time because the same level of activity cannot be regained.
In an attempt to understand why this is the case, the effects of reduction
time and temperature and also of oxygen were examined. It was anticipated
that perhaps crystallite growth could explain part of the irreversible
decline in activity. It appears that if crystallite growth is taking place,
it continues even in the presence of hydrogen because, as seen in Figure 6, a
longer reduction time results in approximately the same loss in activity as
if the catalyst was converting methane. Reduction at higher temperature
(Figure 7) had no beneficial effect on changing the irreversibility of the
catalyst activity decline. A 110 degree Centigrade increase in reduction
temperature produced only about the expected activity recovery from the
consistantly used 705°C reduction temperature. A change in the oxygen
concentration during reaction while maintaining isothermal reaction
conditions did not affect the rate of activity decline (Figure 8).
237
-------
A closer examination of the surface species was expected to lend better
Insight into the catalyst alteration taking place. Nickel spinel does not
form to any appreciable extent from nickel on alumina below 600°C according
to various authors. Since the methane oxidation reactions being examined
here are taking place at temperatures about 600°C, it is anticipated that
this mode of deactivation is possible. However, two interesting points in
this work are notable. First, use of alpha alumina as a support should
reduce the rate of nickel oxide-alumina reaction to form spinel. Second, the
relative ease of reducibility of the surface active catalyst species
indicates that spinel is not responsible for the majority of the activity
observed. This can be seen in Figure 9 where a prepared nickel spinel was
found to have a specific activity two orders of magnitude lower than the
nickel/alumina catalysts. It did, however, possess the same property of
increasing activity following reduction. Further, from X-ray diffraction
data, the active calcined catalyst was found to contain no spinel but simply
nickel oxide and alpha alumina. It is expected that the activity associated
with the calcined nickel/alumina catalyst is due to nickel-nickel oxide and
that the slow loss in activity is due more to the crystallite growth of
nickel oxide than to spinel formation at these temperatures.
In order to discount any major role of nickel spinel activity in methane
oxidation at "pre-light-off" temperatures, a series of nickel spinels were
formed. The ratio of the tetrahedral and octahedral crystal structures were
changed to determine if one may possibly have been responsible for enhanced
activity. Table 1 shows the comparison of the spinels prepared and Figure 10
illustrates the relative activities of each. In can be seen that the effects
of crystal structure of nickel spinel has little to do with the initial
activity observed in the nickel alumina catalysts.
In summary, the selective inactivity of nickel supported on alpha
alumina to maintain partial oxidation of methane to carbon monoxide at
temperatures below "light-off" has been demonstrated. Examination of this
type catalyst under various pretreatment conditions seems to indicate that
the surface nickel-nickel oxide species is undergoing a loss in activity
probably due to crystallite growth. The possibility of nickel spinel being
an active catalytic species is nil, regardless of the crystal form in which
it may be. Further work is being conducted on improving the understanding of
the catalysts involved, the deactivation process and the kinetics of
reaction.
288
-------
TABLE 1. EFFECT OF CALCINATION ON SPINEL PHYSICAL PROPERTIES
CATALYST
T-375
SP1
VD
SP2
SP3
SP4
SUPPORT
—
T-375
T-375
T-375
T-375
CALCINATION
INAIR
—
950° C,
8 hours
1200° C,
50 hours
1370° C,
50 hours
1550° C,
50 hours
XRD
3-Aj02 03
SPINEL
+
A4203
SPINEL
SPINEL
SPINEL
Aj2203
PERCENTAGE
Ni ++ (tet)
IN SPINEL
—
12
24
29
35
SPINEL
CRYSTAL
SIZE (A)
—
250
360
630
630
BET AREA
-------
NJ
VO
O
80
o
£ 60
to
0
0
;£* 40
o
20
1 W) ' «(5)
/7. (3)6** >"
(7) • ./^
^S x s^
s' x
s' x (8) \
S^* (9)
x (10)
x (11)
X
(13)
(1) (2)
• •
I 1
SV - 100, 000 HR"1
02/C - 0. 7 (35% THEORETI CAL Al R)
N2 - 58ft BY VOLUME
G-56B CATALYST
• INCREASING TEMPERATURE (1-5)
x DECREASI NG TEMPERATURE (6-13) -
^
-
1 1 1
500 600 700 800
FURNACE TEMP (°C)
Figure 1. Multiple steady state of rich methane oxidation.
-------
80 -
o
CO
O
O
O
CD
0£
s
tfi
60
20
CO
sv
o2/c
N«
100,000 HR"1
0.7 (35% THEORETICAL AIR)
58% BY VOLUME
G-56B CATALYST
(6)»
(7) INCREASING TEMPERATURE (1-5)
DECREASING TEMPERATURE (6-13)
_L
500
600 700
FURNACE TEMP (°C)
Figure 2. Oxygenated carbon product distribution.
800
-------
tsi
VO
K5
..
3 *"
= ^ 2
•
ID
i— i
X
o
a.
ZD
tO
o 1
o x
<
0
-i — r 1 1 , ,
A Nl 2 CATALYST AFTER CALCINATION
• Nl 2 CATALYST PRIOR TO CALCINATION
o2/c -
N
N2
T
SV -
* m
A A A A
1 1 1 L__
0.7 (35% THEORETICAL AIR) -
5^5. BY VOLUME
705 °C
200, 000 HR"1
• .
A
1 1
0
12345
TIME(hr)
Figure 3. The effect of calcination on catalyst activity.
-------
3.5
O x x
O x
• . ,691C
O
O
O
O
O
x x 7Q3uc
o 717°C
N3
VD
U)
0
0
THE CATALYST WAS REDUCED
AT 700°C FOR 1 HOUR IN
H2 PR I OR TO ACTUAL RUN
Nl 2 CATALYST
REACTION CONDITION:
02/C =0.75 (40%THEORETICAL AIR )
BALANCE
,-1
SV - 200,000 HR
R - RATE OF REACT I ON PR I OR
TO REDUCTION
I i
1
TIME(hr)
Figure 4. Effect of temperature on deactivation
-------
VO
imJ
1.2
..
O
0>
O «*
!_^ L1
tr\
o
i— i
z 1
0
£
1 0.9
CO
O
°^r 0.8
O
u_
£ 0.7
0.6
0.5
*
—
—
—
1
•
•
\
i
•
.
•
•
<
•X-
1 1
*•
9
•
Nl 2 CATALYST
REACTION CONDITION:
OJC - 0. 75 (40fr THEO-
RETICAL AIR)
N2 - BALANCE
T * 710 °C
SV * 200, 000 HOUR"*1
•REDUCED INH2 AT 705 QC
2
3
4 5
TIME(hr)
8
10
Figure 5. The effect of alternating reduction and oxidation on catalyst activity.
-------
£ 1.4
o
X
O
£
CO
O
O
^
O
LL.
O
^~~
—
*
—
-
1
•
•
•
9
(
1
*
'
•
^
o
.9
<
1
REDUCTION*
PERIOD
i
i i i
Nl 2 CATALYST
REACTION CONDITION:
OJC • 0. 75 (40ft THEO-
L RETICALAIR)
N0 - BALANCE
2
T - 710 °C ~
-i
SV - 200, 000 HOUR
'REDUCED INH2AT705°C
•
I I I
8 10
TIME(hr)
12
Figure 6. The effect of reduction time on cayalyst activity.
-------
1.4
1.2
ho
CO
o
o
0.8
a:
REDUCE
IN
H2AT
705°C
i t
i
REDUCE
IN
H2AT
815°C
1 1 |
Nl 2 CATALYST
REACTION CONDITION:
OJC - 0. 75 (4055. THEO- _
RETICALAIR)
N2 = BALANCE
T - 710 °C
. SV - 200, 000 HOUR"1
•
1 1 |
2
3
Figure 7.
TIME (hr)
The effect of reduction temperature on catalyst activity
8
-------
1.5
of
OS!
1
r - i i i i i
Nl 2 CATALYST
A ® 02
~ x 02
A Q2
*•* Ro =
f CH4 =
* & N2 =
1
A ®
&
= 11% BY VOLUME
= 29% BY VOLUME
= 17% BY VOLUME
RATE OF CH4 CONSUMPTION
AFTER 5 HOURS ON STREAM
25% BY VOLUME
BALANCE
250, 000 HOUR'1
® &
THE CATALYSTS WERE REDUCED AT 700°C
FOR Ihr IN50%H2/N2PRIORTO
ACTUAL RUN
ii ill __l
0
12345
TIME(hr)
Figure 8. The effect of oxygen on deactivation
-------
NJ
vo
00
* w
o
i? o
en o
^
0
»— i
X
1 6
i
to
o 4
0 ^
o
u_
O
0
_
L
^
REDUCED
INH2AT
700°C
«h
1 1 1 1 | I "
SP2
REACTION CONDITION:
OJC - 0.6
^
N2 • BALANCE
T - 705°C
SV - 60, 000 HOUR"1
* •
—
—
1 L 1 I 1 i
234567
TIME(hr)
Figure 9. The effect of reduction on spinel activity.
8
-------
NJ
vo
vo
v
O
03
1 r
esj
^ 4,
2
1
Q_
CO
0
°^ 21
0
1 {
n
I
• : a -AI203
02/C - 0.7 (35% THEORETICAL AIR) D: SP1
N2 - BALANCE A: SP2
SV - 60,000hr~1 O: SP3
> ^ ^ ^ ^.
3n n n r- 1
s A A A .... w*i
1 1 1 1 1 1 1
0
3 4 5
TIME(hr)
Figure 10. Relative activities of spinel and
8
-------
SESSION III
APPLICATIONS OF CATALYTIC COMBUSTION
301
-------
PROTOTYPE CATALYTIC SYSTEMS
By:
W. V. Krill, J. P. Kesselring, and S. J. Anderson
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
ABSTRACT
The development of catalytic combustion systems is continuing toward
the prototype demonstration phase. Three combustor concepts have been de-
veloped to the system integration stage: (1) a small gas turbine combustor,
(2) a watertube boiler concept, and (3) a firetube boiler burner. The results
of concept testing are reported.
The model gas turbine combustor shows continued promise for low NOX
emissions with gaseous and distillate fuels. Greatest development difficult
ties are associated with introduction of the premixed fuel/air mixture and
its interaction with catalyst lightoff systems. An integrated system has been
developed, including a multiple nozzle, atomizing injector and an opposed jet
lightoff burner. Testing of the concept is nearing completion to show its
transient and steady state capablities.
The watertube boiler concept uses direct radiative transfer to water-
tubes in the combustion region. Structural problems of the radiative zone
are currently being addressed, and final integration of the concept will fol-
low. Thermal NO emissions are typically less than 2 ppm.
X
The firetube boiler burner also utilizes radiative heat transfer from
a fiber matrix burner to the wall of the firetube. The matrix burner operates
at a surface temperature below 1644K at low excess air levels to control the
formation of thermal NOX. Pad material screening tests have been conducted,
and a mockup burner test is in preparation. Continuing program work will
focus on prototype development of the watertube and firetube boiler concepts.
Work supported by the U.S. Environmental Protection Agency Contract 68-02-3122.
302
-------
INTRODUCTION
The advancement of catalytic combustor concepts toward the system
demonstration phase is continuing. Through a series of contracts with the
Environmental Protection Agency, Acurex Corporation has continued the develop-
ment of several system concepts. These concepts are being further developed
under the current EPA contract 68-02-3122: (1) a model gas turbine combustor,
(2) a watertube boiler system, and (3) a firetube boiler burner. This paper
presents a brief history of each concept and progress since the Third Workshop
on Catalytic Combustion in October 1978.
The development of each of the three combustor types has been supported
by other activities of the program. Catalyst development has continued to im-
prove performance and operability of each of the combustors. The separate
development of auxiliary systems has allowed fuel injection, lightoff, and
temperature measurement problems to be solved. Data analysis and fundamental
studies have provided greater ability to predict reactor performance. Finally,
technology transfer through report publication, technical panel meetings, and
workshops has allowed valuable interchange with the user industries. This
multi-level approach is continuing, focusing on the prototype demonstration
of watertube and firetube boiler systems.
MODEL GAS TURBINE COMBUSTOR
The demand for low N0x gas turbine combustors is increasing. Potential
applications include utility power systems, pipeline pumping, transportation
vehicles, and future combined cycle processes. As a result, government and
private development activities have greatly increased. Among selective cata-
lytic reduction, thermal DeNO , and NO scrubbing, catalytic combustion may
x x
prove to be the most cost effective NO control technique.
303
-------
COMBUSTOR DESIGN
Under early work with fuel-lean combustors, Acurex developed a model
gas turbine can combustor (Reference 1). The combustor consisted of a stain-
less steel can, multiple cone fuel injector (based on a concept developed at
NASA, Reference 2), and the catalyst element. The concept was tested under
steady state operating conditions in both Acurex and Pratt and Whitney test
facilities. Fuels used were natural gas, propane, and No. 2 oil at test pres-
sures from 0.101 to 0.808 MPa (1 to 8 atmospheres). Very low thermal NO emis-
A
sions (less than 5 ppm) were achieved throughout the test series. The combustor
was also tested with nitrogen-doped fuels to investigate fuel nitrogen conver-
sion to NO . Lean conversion rates were generally high (greater than 50 per-
X
cent) at all test conditions.
The greatest system operational problem was injection of fuel to obtain
a well-mixed, prevaporized reactant stream at the catalyst. The injector design
resulted in flameholding under some operating conditions that would be encoun-
tered in machine applications. Therefore, the range of operability required
improvement.
A second distillate fuel injector was designed and constructed. The
injector incorporated three air-assist atomizing spray nozzles spraying through
a perforated air distribution plate. The combustor injector assembly is shown
in Figure 1. Cold flow and combustion tests were run on the configuration to
verify the injector performance. Flameholding caused by the fuel introduction
to the air stream was alleviated following optimization of the spray pattern
into the combustor.
To further advance the combustor design, a lightoff system was incor-
porated to simulate typical turbine ignition and startup requirements. The
advanced concept is shown in Figure 2. The system consists of an opposed jet
burner, initiated by a spark plug. The opposed jet contains premixed gaseous
fuel and air delivered counterflow to the main stream of reactants through a
small tube. In application, the jet would be ignited during low air flow con-
ditions and produce energy to accelerate the machine and preheat the catalyst
to its lightoff temperature. As the machine approaches a loaded condition,
the jet is extinguished by depletion of its flow, and combustion occurs totally
304
-------
in the catalyst. To date, the overall concept has been tested to simulate
startup without the catalyst element to investigate operating characteristics
of the fuel injector and opposed jet.
TEST RESULTS
Tests were conducted with premixed air and natural gas in the opposed
jet and atomized diesel fuel in the mainstream air. The lightoff temperature
that could be achieved was determined by varying the amount of diesel flow
and hence the mixture adiabatic flame temperature. Tests were run at atmos-
pheric pressure with a mainstream reference velocity of 4.6 m/sec (15 ft/sec)
and a preheat temperature of 547K (525°F).
The test data is shown in Figure 3. As the mixture stoichiometry
(including jet and mainstream fuels) is decreased toward stoichiometric, the
lightoff temperature increases. The increasing adiabatic flame temperature of
the mixture is shown for comparison. The actual temperature achieved at a
given stoichiometry is dependent on the degree of mainstream fuel consumed.
Less fuel is consumed near stoichiometric conditions due to increasing diffi-
culty of burning liquid fuel droplets at 547K preheat. Increasing the preheat
would result in greater fuel consumption and higher lightoff temperatures at
a given mixture stoichiometry.
At 400 percent theoretical air, the combustor is operating with only
jet fuel at a temperature of 1005K (1350°F). Above 400 percent, flame cannot
be stabilized in the combustor. At 100 to 150 percent theoretical air, the
majority of the fuel is diesel. Below 150 percent theoretical air at the pre-
heat and velocity conditions described, the stabilizied flame could not be
extinguished upon depletion of the jet flow.
The wide range of jet operability should allow smooth turbine startup
in applications. At low air flow, the jet fuel would be introduced and ignited.
During acceleration (and increasing compressor discharge temperature), main-
stream fuel would be increased to the catalyst lightoff temperature and the
jet extinguished. Final acceleration and loading could then be achieved cata-
lytically by further fuel increases. Testing of the startup and steady state
operation of the combustor with a catalyst is in progress.
305
-------
WATERTUBE BOILER SYSTEM
The effect of NO control on boiler systems is even more significant
X
than for gas turbines. Whereas stationary gas turbine combustors contribute
approximately 3 percent of the total U.S. NO inventory, packaged and utility
Ai
boilers are estimated to contribute 20 percent and 48 percent, respectively
(Reference 3). Clearly, controls developed for watertube and firetube boilers
are an important factor in reducing this country's NO emissions.
A
SYSTEM DESCRIPTION AND EARLY RESULTS
A catalytic watertube boiler system has been devised and extensively
tested. Combustion occurs on the exterior surface of catalyzed cylinders in
crossflow to the premixed reactant stream. Energy is transferred to water-
tubes by radiation from the combustion surfaces.
The concept is shown in Figure 4. A stoichiometric fuel/air mixture
is fed to the radiative section which contains a close-packed array of catalyst
elements and watertubes. The mixture is partially combusted by the catalyst
which is kept at a low surface temperature by radiation heat loss to the water-
tubes. The combustion products and remaining unburned fuel and air are then
passed to a downstream catalytic adiabatic combustor to complete combustion
reactions. A final convective section extracts energy from the fully combusted
gases. Both catalyst sections operate well below the maximum use temperature
of the catalyst supports — the radiative section by radiative cooling and the
adiabatic section by dilution of the fuel/air mixture with exhaust products
from the radiative section.
The radiative section has been constructed and tested independently of
the downstream adiabatic combustor and convective sections. In early testing,
the staggered tube arrangement (separate catalyst tubes and watertubes) was
used. Later tests were conducted with a concentric tube design. The results
for the staggered tube configuration were reported in Reference 4. The radia-
tive section was run with natural gas at varying stoichiometry (40 to 220 per-
cent theoretical air) heat release rate (1.09 x 10a to 3.44 x 108 J/hr), and
preheat to establish its combustion efficiency, thermal and fuel N0x emissions,
306
-------
and heat transfer characteristics at varying conditions. The catalyst was a
platinum preparation on alumina tubes
At stoichiometric conditions, the combustion efficiency of the stag-
gered tube combustor was approximately 37 percent. This result was slightly
lower than the anticipated level of 50 percent necessary to avoid flame tem-
peratures above the temperature capability of the adiabatic section. The
thermal NO emissions were very low, however, never exceeding 2 ppm at these
X
conditions. To evaluate the conversion of fuel nitrogen to NO , natural gas
X
was doped with ammonia (NH-). A minimum conversion level of 20 percent was
measured at approximately 55 to 60 percent theoretical air.
These early combustion test results suggested several design modifica-
tions for improvement of the radiative section. Closer spacing of the catalyst
tubes would improve the combustion efficiency. It would also be desirable to
isolate the cold surfaces of the watertubes from the combustion zone to avoid
fouling when firing with fuel oils. To meet these requirements, a concentric
tube design was completed where the catalyst tube surrounds the watertube with
an air gap between the two. Therefore, combustion energy generated on the
catalyst tube outside surface would be transferred by conduction to its inner
surface and by radiation across the air gap to the watertube.
CONCENTRIC DESIGN TEST RESULTS
The concentric configuration was first tested under conditions similar
to those for the staggered arrangement with a platinum catalyst on alumina
tubes. Combustion efficiency was markedly improved, reaching a high of 93 per-
cent fuel conversion at stoichiometric conditions as shown in Figure 5. Thermal
NO emissions were again less than 2 ppm, and fuel nitrogen conversion was
A
below 30 percent under fuel-rich conditions. Tests were not performed with
fuel oil to verify that capability, however.
Despite the improved combustion performance of the concentric watertube
design, structural problems were encountered. The configuration develops a
substantial temperature gradient across the walls of the ceramic tubes. The
temperature difference results in nonuniform radial expansion and, ultimately,
fracturing of the material. The stresses that would be encountered in
307
-------
application were calculated as 198 x 106 Pa (28,700 psia) to 257 x 106 Pa
A
(37,300 psia), clearly greater than the 190 x 106 tensile strength of alumina
tubes at 1367K (2000°F).
Only one other ceramic can be made with adequate strength and is
appropriate for combustion applications. Silicon carbide not only has a high
strength of over 400 x 106 Pa at the above conditions, but it has a higher
thermal conductivity than alumina and therefore results in lower temperature
gradients and developed stresses. As a result, tubes of silicon carbide were
purchased from the Norton Company in the proper dimensions and coated with a
chromia (Cr^) and platinum catalyst preparation for an additional test series.
Testing of the concept is in progress.
FIRETUBE BOILER BURNER
A large percentage (approximately 50 to 65 percent) of the commercial
size boiler market is composed of firetube boilers. These smaller units
account for over 30,000 gas-fired boilers and an equal number of distillate
oil-fired boilers in the 3.2 x 109 J/hr (3 x 106 Btu/hr) size alone (Refer-
ence 5). In fact, of all commercial and industrial size boilers in the field,
the majority are firetubes. These applications are principally gas- and
distillate oil-fired.
x
Thus, firetube boilers represent another area for significant NO
reduction. A catalytic burner would be especially attractive if it were
retrofitable into the large number of boilers already installed or at least
presented only minor changes to current designs.
BURNER DESIGN CONCEPTS
A number of catalytic burner designs were conceptualized and reviewed
with major firetube boiler manufacturers. The simplest and most acceptable to
the manufacturing community utilizes a fiber matrix material that .is molded
into a cylindrical surface burner for insertion into the firetube.
The burner concept is shown schematically in Figure 6. A fuel/air
mixture is passed into the central region of the cylindrical fiber matrix.
308
-------
The cylinder construction is best described by Figure 7. The matrix material
is approximately 2.5 cm (1 inch) in thickness, is closed with similar material
at the downstream flow end, and is sealed at the upstream end by a combustion
chamber flange. The matrix material is rigid, is supported at the upstream
end, and has internal flow passages as shown in Figure 7, maintaining concen-
tricity with the boiler firetube.
In the described configuration, premixed fuel and air are forced to
flow into an annulus between the fiber matrix and an internal cylinder. The
fuel may be gaseous or vaporized liquid. The mixture then flows through the
matrix and is ignited at the outer cylinder surface to start the burner. A
heterogeneous surface reaction zone is established on the cylinder outer sur-
face, within a depth of only a few millimeters of the matrix thickness. The
majority of the matrix thickness remains at the temperature of the incoming
reactants and promotes no additional reaction.
Heat is transferred directly from the matrix outer surface to the fire-
tube as shown in Figure 6. The transfer is primarily radiative although con-
vective transfer also contribues. This heat transfer limits the matrix
surface temperature to less than 1644K (2500 F), the temperature limitation of
the material, although the adiabatic flame temperature of the reactive mixture
exceeds 2200K. The heat is transferred through the metal firetube wall to the
boiler water to heat the water or raise steam. Flue gases are forced to the
end of the firetube and out to the additional firetube passes.
FIBER MATRIX EXPLORATORY TESTS
Catalytic fiber matrix materials have demonstrated low NO emissions
X
and high combustion efficiency with gaseous fuels (Reference 6). However, the
optimum matrix material, appropriate heat transfer configuration, and ability
to operate on distillate fuel required additional study. A series of tests on
fiber matrix disks was conducted to evaluate these parameters.
Figure 8 shows the matrix surface temperature of the combustion zone as
a function of theoretical air for three values of face velocity. Lower face
velocities result in lower surface temperatures at low excess air levels (100
to 130 percent theoretical air) where firetube boilers normally operate. This
309
-------
low surface temperature is advantageous in reducing NO emissions as shown in
Figure 9. Emissions of less than 10 ppm at 115 percent theoretical air are
possible. Tests also indicate high combustion efficiency (low CO and HC)
can be achieved simultaneously with low NO emissions.
The low surface temperatures of Figure 8 also demonstrate the effective
radiative heat transfer away from the combustion zone. The indicated tempera-
tures of 1478K to 1644K (2200°F to 2500°F) at 100 percent theoretical air are
within the use temperature capabilities of the alumina silicate fibers and
far below the adiabatic flame temperature of 2228K. From these tests, the
optimum flow velocity for the selected matrix material appears to be 10 to
15 cm/sec (1/3 to 1/2 ft/sec).
The catalytic burner was also operated on diesel fuel. To achieve
stable combustion with similar results to natural gas, it was necessary to
partially vaporize the oil. This was accomplished by preheating the air
stream to 492K (425°F) prior to admitting the fuel. Current firetube boiler
designs operate without air preheating systems. Therefore, operation of the
matrix burner with fuel oil in a firetube boiler will require additional
development of an air preheating technique.
These flat disk burner tests identified the operating requirements of
the matrix material. To further develop the concept, a cylindrical burner as
in Figure 7 was constructed for testing in a simulated firetube. Results of
the tests are pending. The fiber matrix burner does appear attractive for
further development and should be applicable to many firetube designs with
only minor modifications.
CONCLUSIONS
The development of catalytic gas turbine, watertube boiler, and firetube
boiler combustors is rapidly progressing. The potential for near-term reduc-
tion of thermal NO emissions with clean fuels is promising.
X
An integrated model gas turbine combustor has been tested to simulate
both startup and steady state operating conditions, following refinement of
310
-------
fuel injection and lightoff techniques. Only further refinement of the concept
is required for small machine applications. The concept may also be scalable
to larger combustor sizes.
The watertube boiler concept shows promise for further development
despite being a radically different design from present systems. Production
units would probably differ further from those tested. Stress analyses on
the catalyst supports determine silicon carbide as the most appropriate ceramic
in the preferred concentric tube design. Ongoing testing will verify the mate-
rial recommendations and lead to testing of the integrated system concept.
The firetube boiler burner is a retrofitable design for gas fired units.
For distillate fired firetubes, the fiber matrix concept requires preheated
air for acceptable operation. Detailed tests on fiber materials have shown
low NO emissions due to surface cooling with concurrent low CO and hydro-
X
carbon emissions. A cylindrical burner of the most acceptable material is
being fabricated for testing.
The development of burners for these three classes of equipment addresses
a large portion of the NO emissions in this country. Further EPA program de-
X
velopment will focus on prototype demonstration of catalytic boiler systems.
311
-------
REFERENCES
1. Kesselring, J. P., et al. Design Criteria for Stationary Source Catalytic
Combustion Systems. EPA-600/7-79-181, August 1979.
2. Tacina, Robert R. Degree of Vaporization Using an Airblast Type Injector
for a Premixed-Prevaporized Combustor. NASA TM-7883b, August 1978.
3. Mason, H. B. and Waterland, L. R. Environmental Assessment of Stationary
Source NOX Combustion Modification Technqiues. Proceedings of the Second
Stationary Source Combustion Symposium. EPA-600/7-77-073a, July 1977.
4. Krill, W. V. and Kesselring, J. P. The Development of Catalytic Combustors
for Stationary Source Applications. Proceedings: Third Workshop on Cata-
lytic Combustion. EPA-600/7-79-038, October 1978.
5. Evaluation of National Boiler Inventory. EPA-600/2-75-067, 1975.
6. Blair, Martin G. Prototype Surface Combustion Furnace Evaluation. EPA-
600/7-79-038, October 1978.
312
-------
OJ
M
LJ
Air
3 places
Oil
3 places
r
5.563
dia
ref
4 a
25.75 ref
_, ,
Thermocouple
leads
3/16" dia. propane
injection tube
3 pi
?lEf
ffFL
Figure 1. Model gas turbine combustor assembly.
-------
U>
I--
js
Figure 2. Model gas turbine combustor with opposed jet startup system.
-------
220
2000
1800
g 1600
1400
1200
1000
Jet could
not be
extinguished
100
Adiabatic
flame
temperature
200
Mixture stoichiometry (%)
300
No flame
tabilizied
400
Figure 3. Gas turbine opposed jet results
-------
Steam
drum
Mud
drum
Catalyst
coated cylinder
Watertube
Radiative heat
transfer section
Monolith bed
- Adiabatic
combustor
Convective heat exchanger
Figure 4. Radiative watertube boiler concept.
-------
100
a;
OL
O)
•t—
D
O
90
80
70
J3
O
60
50
70
80
90 100 110
Theoretical air (percent)
120
130
Figure 5. Concentric watertube boiler combustion efficiency.
-------
Flue gas
Fuel/air
UJ
M
00
Boiler water
J
Fiber matrix
•*Q 1
^-^
//// //////////////*/////////
)
Firetube surface
Figure 6. Firetube boiler catalytic burner concept.
-------
u>
(-•
VO
Fiber matrix
radiative
burner
Firetube
combustion
chamber
Fuel/air
mixture
Figure 7. Catalytic firetube boiler burner.
-------
1600
1400
Face velocity
O 0.61 m/sec
O 0.30 m/sec
A 0.15 m/sec
3
4->
-------
Face velocity
O 0.61 m/sec
O 0.30 m/sec
A 0.15 m/sec
0
100 110 120
130 140 150 160 170
Theoretical air (percent)
190 200
Figure 9. Fiber matrix NO emissions.
X
321
-------
EPRI STATIONARY GAS TURBINE
CATALYTIC COMBUSTOR DEVELOPMENT
PROGRAM
By
LEONARD C. ANGELLO
ADVANCED POWER SYSTEMS DIVISION
ELECTRIC POWER RESEARCH INSTITUTE
PALO ALTO, CALIFORNIA 94303
ABSTRACT
The EPRI Stationary Gas Turbine Catalytic Combustor Development
Program is a multi-year, multi-phase major contract effort. The
primary program objectives are to design, develop and demonstrate
a full-scale, field tested stationary gas turbine combustor,
employing catalytic combustion technology for reduced NO
X
emissions. The program focus is the development of a practical
and reliable catalytic combustor for current and near-future
stationary gas turbine designs fired with low nitrogen (>500
ppmw) distillate fuels. The purpose of this paper is to present
the overall program approach and details of the Phase I program
plan.
322
-------
INTRODUCTION
The control of nitrogen oxide (NOV) and unburned hydrocarbon (UH)
A
emissions from stationary gas turbines is of major concern to
California utilities. Projections based on current California
NOX emission standards for utility boilers (Rules 67, 475 and
475.1) indicate that in the future even more stringent NOX
control may be required for stationary gas turbines. Although
intensive efforts have already been made by utility gas turbine
manufacturers to reduce NO emissions, current approaches show
J\
limitations in meeting projected California emission standards.
The catalytic combustor represents a potentially attractive
approach for meeting future emission standards due to its ability
to operate at high thermal efficiencies with very low levels of
pollutants (NO , UH, CO, and smoke). Typical NO emission for
A A
current stationary gas turbines burning petroleum distillate
fuels are about 200 ppmv at base load operation. Initial
laboratory studies (References 1 and 2) have shown that the
catalytic combustor operating on similar fuels could achieve NO
A.
emissions of less than 20 ppmv at the same steady-state operating
condition.
The above studies, as well as others reported in the literature,
have recognized the potential of the catalytic combustor for
achieving substantial reductions in NO emissions. However,
before the emission advantage of the catalytic combustor can be
realized in commercial stationary gas turbines engines, it must
be demonstrated that a practical and reliable catalytic combustor
design is possible.
Although the scope of the current program is restricted to
analytical and experimental work with low nitrogen distillate
fuels (<500 ppmw), it is recognized that catalytic combustion
offers the potential for improved combustor performance on low or
medium Btu gas derived from coal gasification. For medium Btu
323
-------
gas, NO emissions could be markedly reduced and with low Btu
j\.
gas, additional stability and improved part load performance
could result. It is possible that if work under the current
program is successful, further work on coal derived medium and
low Btu gases may be appropriate.
PROGRAM DESCRIPTION
The catalytic combustor development program described below has
been formulated with primary consideration for the development of
a practical and reliable catalytic combustor design. During this
program a practical, full-scale, field tested, catalytic
combustor capable of extremely low NOX emissions will be
designed, developed and demonstrated. The catalytic combustor
developed under this program will be designed for commercial
stationary gas turbine engines fired with low nitrogen (<500
ppmw) distillate fuels. The full-scale demonstration of the
operational and performance capability of a practical and
reliable catalytic combustor design over the entire range of
stationary gas turbine operating conditions is the ultimate goal
of this effort. This effort is planned to be conducted in three
phases:
Phase I - Preliminary Design and Development
Phase I of the overall development plan involves the preliminary
design, development, and demonstration of a full-scale, single
can, catalytic combustor. Based on a thorough review of current
catalytic design information, a single catalytic combustor design
concept having the highest practical design potential for full-
scale developmert will be determined. Detailed analytical and
experimental evaluation of the selected design concept' over a
specified stationary gas turbine operating cycle will be
performed. Special instrumentation will be used to measure
combustor efficiency, pressure losses, temperature distribution,
and emissions.
324
-------
Initial testing will provide the data base for design improve-
ments. Revised designs will be further evaluated as part of the
development process for refining the initial catalytic combustor
design. This experimental program is expected to provide the
design data base for specification of a pre-production prototype
catalytic combustor to be evaluated in Phase II.
Phase II - Prototype Design and Evaluation
Phase II of the overall program address the bill of material
design, fabrication, and long-term endurance evaluation of a
single can, catalytic combustor. Sufficient full-scale rig
testing will be performed to adequately assess major durability
and performance characteistics of the pre-production prototype
design. Special emphasis will be placed on evaluating and
optimizing emissions and performance characteristics, durability,
engine compatibility, and complexity. Improvements based on
these evaluations will be incorporated into the final prototype
design. This experimental effort will provide the necessary data
base and technology acquisition for specification to be used in
the Phase III engine demonstration and field test.
Phase III - Final Design and Field Test
The objective of the Phase III effort will be the verification of
the final prototype combustor design in an integrated stationary
gas turbine for further refinement toward commercial accepta-
bility. A full set of prototype combustors will be fabricated
and installed on a commercial engine. Demonstration testing will
be performed to verify the performance, operating, and durability
characteristics obtained in Phases I and II. Successful
completion of the Phase III effort will demonstrate that
development of the catalytic combustor is complete and ready for
commercial application.
325
-------
Program Schedule
The overall planned program schedule is shown in Figure I. The
Phase I effort was initiated on June 25, 1980, with issuance of
EPRI RFP 1657. Contract initiation was accomplished on March 24,
1980, with completion of Phase I scheduled to occur within 24
months of contract signing. Implementation of Phases II and III
will be determined at a future date. For planning purposes, it
is expected that Phase II and III will each be approximately 18
months in duration and that the entire development program
(Phases I, II, and III) will be completed within five years of
the Phase I program initiation.
PROGRAM OBJECTIVES AND GOALS
The Phase I technical effort will employ existing catalyst/
support technology to design, develop and demonstrate a
practical, full-scale, single can, catalytic combustor capable of
estremely low NOX emissions. The catalytic combustor developed
under this effort will be sized for existing or near future
stationary gas turbines in the 30 to 100 MW power range. Design
goals for this effort include the following:
o NOX <20 ppmv corrected to 15% 02 at Base Load
o Smoke < SAE Number 2 at Base Load
o Combustion Efficiency
Eta > 99.9 at Base Load
Eta > 99.5 at Spinning Idle
Eta > 99.1 at all other operating conditions
326
-------
o Combustion total pressure losses (Ap/p%) <5% over all
operating conditions
o Combustor life > 2000 hours (Peaking Application)
Combustor designs developed under this program will be based on
burning low nitrogen (<500 ppmw) distillate fuels and may employ
full utilization (all fuel passed through the catalyst) or
partial utilization catalytic devices. Designs may employ pilot
burners, fuel staging or variable geometry technology. Design
emphasis will be on system integrity, reliability, and ease of
maintenance. Methods of determining catalyst flame detection and
failure detection will also be addressed.
CATALYTIC COMBUSTOR DESIGN CONSIDERATIONS
Over the past five years, a growing body of technical literature
has developed, describing low emission catalytic combustion
programs carried out by a number of government agencies and
industrial gas turbine manufacturers. These tests have confirmed
the potential for achieving extremely low levels of nitrogen
oxide emissions and have affirmed the feasibility of the
catalytic combustor concept for gas turbine application. From a
develoopment point of view, however, the concept is in its
infancy.
In previous investigations pipe flow rigs have been used to
explore catalyst steady-state performance at or close to a
hypothetical full power design point condition. Furthermore, for
reasons of experimental control, preheated air streams and an
extended mixing length were used. In contrast, a practical
stationary gas turbine combustor must start cold and operate over
a wide range of fuel/air ratios, inlet temperatures and pressure
conditions. Furthermore, stable gas turbine operation is
required at high combustion efficiencies over a wide range of
fuel turndown conditions. Finally, a practical combustor must be
327
-------
of realistic size and geometry to meet design requirements and of
sufficient durability to meet warranty requirements.
From the preceding, it is evident that translation of the
exceptional catalytic combustion performance obtained for
premixed fuel/air mixtures in steady-state test rigs into
practical system hardware operating over the entire range of
required stationary gas turbine conditions is a formidable
task. Such a task shall require technical innovation based on
the comprehensive understanding of both catalyst and combustor
design constraints and capabilities. The following discussion is
provided as a brief summary of the principle design problems
which must be resolved prior to the demonstration of a practical
catalytic combustion system.
The Startup and Loading Sequence of a Catalytic Combustion System
Although catalytic combustion has excellent thermal and emissions
performance at the design point (full load) condition, at off-
design point conditions, the reaction temperature may be too low
to sustain stable combustion. This is because the fuel/air ratio
turndown is limited to 1.5:1 when diluent gas is not used. To
alleviate this problem, modifications to the gas turbine are
necesary .
According to conventional combustor design, the percent of air
distribution within the combustor is kept constant at all
times. When the fuel flowrate (or load) varies, the location of
the primary combustion zone inside the combustor also varies.
Hence, the only control parameter during the entire sequence is
the fuel flowrate.
On the other ha1"^ , if a catalytic combustor were implemented, the
combustion zone is fixed by the physical location of the
catalyst. In order for the system to operate over the entire
startup and loading sequence, the catalytic combustor must be
328
-------
designed to sustain combustion over the required fuel/air ratio
turndown, or the fuel/air ratio iin the catalytic combustor must
be maintained constant by redistributing the air or fuel flow
during the sequence. Alternatively, a conventional combustor may
be used ahead of a catalytic element.
Catalytic Combustor Durability
For practical purposes, a minimum requirement for catalyst life
is 2000 hours for peakload units and 8000 hours for baseload
units. However, existing life test data are not sufficient to
make a realistic assessment on practical catalytic combustor
life. Hence, there is a need for more experimentation in this
area.
Another area of concern is the integrity of ceramic monolith
support materials may fracture for the following reasons:
o Thermal shock as a result of cyclic operation
o Material fatigue as a result of mechanical vibrations
The failure of the monolith support material will not only
shorten the combustor life, but pieces may also damage the tur-
bine blades, and allow the combustion process to move into the
turbine area causing blade failure,etc.
The catalyst performance may be degraded for the following
reasons:
o Long operaing hours
o Cyclic operation
o Catalyst poisoning
329
-------
The first effect has been investigated in a preliminary manner by
work sponsored by the NASA Lewis Research Center (Reference 3).
However, further testing which simulates the gas turbine
operating mode is required.
Catalyst Lightoff System
Catalytic combustor ignition systems differ from those of
conventional combustors. Catalytic lightoff with air preheat is
possible, but the compressor outlet temperature at startup does
not meet the 600°F to 1000°F requirements for lightoff. The
spark ignition system currently employed by most gas turbines may
overheat the catalyst during abrupt lightoff of the fuel/air
mixture. The catalyst requires a more gradual lightoff
sequence. A pilot flame might provide such a lightoff technique.
The pilot would ideally be a long, internally spark-ignited pilot
burner. Upon ignition, it would produce a long and wide flame
directly across the incoming fuel and air mixture. This would
act as an aerodynamic flame-holder and as an ignitor, the
fuel/air mixture would be at least partially combusted, thereby
raising the bulk flow temperature. The reactive mixture would
then complete the combustion in the catalyst as a result of its
increased temperature.
Probable Damage to Turbine by Catalyst Substrate Failures
During cyclic operations, ceramic catalyst substrates may
fracture due to thermal stresses. If such an incident occurred,
part or all of the combustion process would move to the turbine
section, thereby overheating the turbine components.
Furthermore, the substrate fragments could be dragged by the high
velocity air and impinge on the turbine blades. Depending on the
velicity and the size of the fragments, severe turbine blade
damage could occur.
330
-------
In the design of a catalytic combustor, provision should be
included for fuel shutoff in case of catalytic substrate
failures. This will prevent combustion occurring in the turbine
section. Also, the potential for damage to the turbine as a
result of fragment impingement should also be estimated prior to
the application of catalytic combustors to the gas turbine.
Fuel/Air Preparation for Catalytic Combustors
The use of monolithic substrates as catalyst supports neces-
sitates nearly complete premixing prior to entry of the fuel and
air into the catalyst channels. Little can be done to correct a
fuel-air maldistribution once inside the catalyst. Since for
adiabaic systems the operating temperatures of the catalyst
approximates the adiabatic flame temperature of the mixture, it
is important that the combination of inlet temperature and fuel
concentration profile is more or less uniform. Maldistribution
of fuel will cause excessive emissions due to lean fuel mixtures
in some portions of the combustor cross section, while in those
regions where high fuel concentration leads to high adiabatic
flame temperature, catalyst damage can occur. Uniform premixing
of the fuel and air prior to entry into the catalyst eliminates
these potential problems, while yielding a combustor temperature
pattern that, with minor adjustments, is ideal for gas turbine
engines.
The question of auto-ignition must also be addressed. Since a
uniform mixture is desired from the catalyst performance
standpoint, it is important that sufficient residence time
upstream of the catalyst is provided consistent with gas turbine
shaft length constraints. However, it must be realized that as
mixture pressure and temperature are increased, the auto-ignition
delay time decreases. At high temperatures the ignition delay
time may be reduced sufficiently to approach the time necessary
for premixing and vaporization; under such conditions preignition
may occur. For conventional distillate fuels, the ignition
331
-------
delay time at 700°F (a reasonable catalyst inlet temperature)
ranges from approximately 3 seconds at 2 atmospheres to 500 msec
at 10 atmospheres pressure. These times are generally ample for
premixing, provided the premixing gas is without stagnant
regions. At higher pressures, the margin between preignition and
premixing times is less and it is necessary to design for minimum
mixing time. In this connection it should be noted that rapid
mixing may be obtained through the provisions of pressure loss
and of the optimum number and distribution of fuel injectors.
The vaporization time for liquid fuels may be comparable with
mixing time, if sufficiently small droplets are produced. In
addition it should be noted that preignition is not necessarily
catastrophic, provided local temperatures do not become
excessive. The occurrence of preignition as described would
cause a significant loss of machine power due to an increase in
catalyst pressure drop.
Combustor Pressure Losses
Finally the question of overall pressure loss must be
addressed. Combustor pressure loss will derive from fuel-air
premixing and catalyst pressure loss. Both premixing and
catalyst pressure loss may be decreased by reducing the inlet
velocity to the lowest acceptable value. Combustor design
optimization involves minimizing pressure loss consistent with
two compensating factors:
o The residence time for feed preparation and presentation to
the catalyst must be less than the ignition delay time.
o Low velocities lead to undesirable large combustor cross
sectional areas.
In conventional combustors the trend towards lower velocities and
higher surface areas cause cooling problems. With the catalyst
system the decrease in peak operating temperature results in less
cooling problems than is the case with conventional combustors.
332
-------
PHASE I TECHNICAL EFFORT
The Phase I technical effort consists of eight tasks covering
(1) Conceptual Design, (2) Definition of Design Criteria, (3)
Design Analysis, (4) Preliminary Design, (5) Design Optimi-
zation Plan, (6) Fundamental Studies for Design Verification,
(7) Full-Scale Combustor Evaluations, and (8) Prototype
Design Specification. The planned Phase I program schedule
is shown in Figure 2 and the overall Phase I work flow
diagram is contained in Figure 3. The complete Phase I work
breakdown structure is shown in Figure 4. Descriptions of the
individual Phase I tasks are given below:
Task 1 - Conceputal Design
During this task, a specific stationary gas turbine engine design
and operating cycle will be identified as a basis for a conceptual
catalytic combustor design. The selected engine design will be
representative of a current or near future stationary gas turbine
in the 30 to 100 MW power range.
A thorough review of all available information on catalytic
combustor design will be conducted. This review will be the
basis for establishment of the various critical design
requirements including minimum catalyst preheat temperature,
maximum catalyst bed temperature, range of allowable reference
velocities, pressure drop characteristics, catalytic bed length,
startup requirements, and dilution air requirements.
Based on the above review, a catalytic combustion design concept
or concepts will be identified for further development. Layout
drawings of the identified design concepts(s) will be prepared.
These drawings will define the entire combustor system in
sufficient detail to define all major assemblies, including fuel
333
-------
injectors and cooling liners and the dimensions of all passages
and chambers. The catalytic element or elements of the design
shall be characterized as the size, shape, and major
constituent. The intent shall be to define the combustor design
in sufficient detail to conduct the analyses and evaluations
specified in the following section.
Task 2 - Definition of Design Criteria
The conceptual catalytic combustor design(s) previously
identified will be analyzed to assess (1) potential for meeting
the prescribed emission and performance goals, (2) durability and
time-between-overhaul predictions, (3) combustor performance and
degradation potentials, (4) system costs, (5) system retrofit-
ability, (6) system safety, and (7) system operating environment.
Task 3 - Design Analysis
The conceptual catalytic combustor design(s) previously
identified will be analyzed to assess (1) potential for meeting
the prescribed emission and performance goals and (2) feasibility
for meeting engine installation and operating goals. In these
studies, the identified catalytic combustor concept(s) will be
defined and analyzed in detail. Aerothermodynamic analyses will
be evaluated to calculate the detailed pressure/velocity and flow
distributions throughout the combustor. Heat transfer analyses
will be conducted to determine cooling liner flow distributions
and metal temperatures. Vibrational analyses will be conducted
to determine the effect of gas turbine vibrational amplitudes and
frequencies on catalytic combustor hardware durability and
reliability. Mechanical analyses will be conducted to define
fabrication/assembly details, structural stresses, weights, and
estimated life. In addition, relevant design considerations
regarding start capability, turndown capability, reliability, and
maintainability will be qualitatively evaluated as part of the
design analyses performed under this task.
334
-------
Task 4 - Preliminary Design
Upon completion of conceptual design analysis, a detailed
preliminary catalytic combustor design will be prepared. This
design will include (1) detailed layout drawing of the refined
catalytic combustor design with enough auxiliary views to show
all major dimensions, location and size of combustor air dilution
holes and how all the parts are assembled, and (2) a design
summary which describes the features, intended operating mode,
results of design calculations for sizing the flow elements and
predicted emission/performance characteristics. Mechanical
layout studies will also be conducted to define the combustor
fabrication assembly and mounting arrangements for the end
application engine design.
Task 5 Design Optimization Plan
Based on the selected preliminary design, a detailed design
optimization test plan will be prepared. This plan will describe
the critical path development needed to translate the prepared
preliminary combustor design into practical system hardware.
Areas where additional experimental data are needed to obtain a
fundamental understanding of pertinent design principles will be
identified and specific experiments to gain this information
shall be described. In particular, the optimization plan will
define a catalyst/support program to evaluate catalytic combustor
failure modes with respect to catalyst poisoning effects,
monolith durability and turbine damage potential. The EPRI
Project Manager will review and approve the final version of the
Design Optimization Plan prior to initiation of Task VI,
Task VII, and/or Task VIII.
335
-------
Task 6 - Fundamental Development Testing for Design Verification
Under this task, fundamental/empirical development experiments
will be conducted to provide the required design data identified
in the Task V optimization plan. Experiments performed under
this task will be at full pressure and may involve both full-
scale and subscale evaluations.
Experiments conducted under this task will be closely coordinated
with those performed under Task VII, Full-Scale Combustor
Evaluation. The primary objective of these fundamental/empirical
studies is to guide the full-scale catalytic combustor hardware
development. In each of the tests conducted under this task,
individual elements of the combustor will be evaluated to answer
specific questions with regard to design feasibility. Based on
those tests, preferred designs may be substantiated, modified or
eliminated as technically unfeasible.
Areas of design anticipated to require fundamental/empirical
development under this task include the following:
!• Fuel/Air Preparation System Design - Initial emphasis will be
placed on determining the degree of vaporization and pre-
mixing required for the catalytic design selected. Based on
this data, fuel injectors will be designed, fabricated, and
evaluated on their ability to provide the required fuel/air
distribution.
2. Ignition/Control System Design - Fundamental data will be
obtained to allow design and development of an ignition/
control system capable of meeting the startup and loading
sequence normally required of generation equipment used by
the electric utility industry.
3. Liner Cooling Development - Fundamental data will be acquired
to evaluate the effect of steady-state and transient response
336
-------
on liner cooling schemes. Based on this data required liner
cooling schemes will be fabricated and evaluated.
4. Catalyst Packaging Design - Data will be obtained to ensure
the adequacy of the catalyst encapulation design. Fabrica-
tion of the final encapsulation design will be prepared and
used in subsequent material durability evaluations.
5. Material Durability Evaluations - Based on existing catalyst
technology, a single catalyst/substate design will be
selected for long-term durability evaluation. A cyclic
durability evaluation will be performed on the selected
catalyst/substrate design at full pressure. Each cycle will
be of suitable duration to establish steady state operation
at each of the following test conditions: cold start,
spinning idle, 50% base load, 70% base load, base load, and
peak load. The total number of test cycles shall be
sufficient to simulate 2000 hours of engine life (peaking
application).
In addition, a mechanical vibration test will be performed.
In these tests catalytic combustors fabricated of the
selected substate material will be evaluated at vibration
amplitudes and frequencies equivalent to those of an actual
gas turbine operation. This test shall be performed at full
pressure and shall be made to simulate 2000 hours of actual
engine life (peaking application). The intent of this test
shall be to demonstrate the mechanical integrity and inherent
attrition resistance of the selected substate material over
the expected life of the combustor.
Task 7 - Full-Scale Combustor Evaluation
Based on the preliminary design and the results of the funda-
mental development work performed, one or more full-scale
catalytic combustors will be fabricated for test and evalua-
337
-------
tion. All evaluation and demonstration testing will be conducted
in full pressure, reverse flow test rigs capable of handling max
peak operating conditions of temperature, pressure, and mass
flow. All testing will be performed with non-vitiated air.
Test conditions will include, but not be limited to, the
following:
o Average concentration of oxygen (02) in the exhaust gas
(% C>2, dry basis)
o Average concentration of carbon monoxide (CO) in the exhaust
gas (ppmv CO, dry basis)
o Average concentration of unburned hydrocarbons (UHC) in the
exhaust gas (ppmv UHC, wet basis)
o Average concentration of carbon dioxide in the exhaust gas
(% CO2, dry basis)
o Average concentration of nitrogen oxides NO in the exhaust
A.
gas (ppmv NO , dry basis)
X
o Average SAE smoke number
o Total pressure loss over the combustor (strain gage
transducers)
o Air flow rate (ASME sharp-edged orifice)
o Average inlet total temperature (thermocouple)
o Average inlet total and static pressure (strain gage
transducers)
o Inlet air humidity (LiCl type meter)
338
-------
o Fuel injector pressure drop (presure transducer)
o Combustor exit temperature (thermocouple)
o Fuel flow rate (turbine flow meter)
o Combustion wall temperatures (thermocouple)
o Catalyst bed temperature (thermocouple)
Based on the above data, the following combustor performance
characteristics will be determined:
o Ignition and load acceleration characteristics
o Fuel/air ratio over the operating range (fuel and air flow
measurement basis and emission measurement basis)
o Combustion efficiency (enthalpy and emissions basis)
o Exit pattern factor
o Combustor Pressure loss
The above data will be analysed and recommendations prepared for
future design improvements. These analyses and recommendations
will be reviewed with the EPRI Project Manager for redesign
approval. Based on the EPRI Project Manager's approval, the
recommended design modification will be made and the retesting of
the improved combustor design will be initiated. No more than five
combustor redesigns will be allowed under this program.
339
-------
Task 8 - Prototype Design Specification
Upon successful demonstration of the optimized full-scale
catalytic combustor, a final bill-of-materials combustor design
specification will be prepared. This specification will include
a detailed prototype layout drawing, a detailed design summary
and the full identification of all fabricated materials.
A detailed Phase II test plan will also be prepared under this
task. Requirements for the Phase II test plan will include
definition of the following parameters:
o Test Schedule
o Data Requirements
o Location of Measurement Instrumentation
o Data Analysis Procedures
CONCLUDING REMARKS
Phase I of the EPRI Stationary Gas Turbine Catalytic Combustor
Development Program is now in progress. The Phase I contract was
awarded to the Westinghouse Electric Corporation in March of
1980. As of the date of this publication, work has begun on
Tasks 1 and 2 of the Work Breakdown Structure shown in Figure 4.
Progress to date under Task 1 has identified the Westinghouse
Model 501D as the engine/combustor design envelope to be used for
the conceptual design studies. Figure 5 is a cut-away view of
the Westinghouse Model 501 gas turbine. The Model 501 turbine is
a 12:1 compression ratio engine with a 2100°F turbine -inlet
temperature. It consists of 17 stages axial compressor, 16 can-
type combustors and a 4 stage turbine. Each combustor can is
approximately 20 inches in diameter. The baseload power output
340
-------
of the Model 501 is 94 MW at a heat rate of 10,660 Btu,kWh.
Figure 6 depicts a conceptual design arrangement of a
catalytic combustor for the Model 501 turbine. The princi-
pal design features of this arrangement includes 1) a conven-
tional fuel nozzle preburner, 2) multiple secondary fuel
injection, and 3) a downstream catalytic element. The purpose
of the preburner is to provide for catalyst lite-off and load
range capability. The multiple fuel injection allows rapid
fuel vaporization and mixing in the fuel preparation zone up-
stream of the combustion catalyst. The monolithic ceramic
element provides minimum pressure loss through the catalyst.
In conclusion, the EPRI Stationary Gas Turbine Catalytic
Combustor Development program represents a comprehensive
development program to demonstrate catalytic combustion
technology for reducing NOx emission levels in current-design
utility gas turbines. The program will examine the durability
and reliability of this technology and its successful comple-
tion will lead to the eventual commercialization of catalytic
combustion technology for utility gas turbine systems.
341
-------
REFERENCES
Blazowski, W. S. and Bresowar, G. E., "Preliminary Study of
the Catalytic Combustor Concept as Applied to Aircraft Gas
Turbines," AFAPL-TR-74-32, June 1974.
Decorso, S. M., Mumford, S., Carubba, R. V., and Heck, R. M.,
"Catalysts for Gas Turbine Combustor—Experimental Test
Results," Journal of Engineering for Power, Vol. 99, No. 2,
April, 1977.
Heck, R. M., Chang, M., Hess, H., and Carrubba, R. V.,
"Durability Testing at One Atmosphere of Advanced Catalysts
and Catalyst Supports for Automotive Gas Turbine Engine
Combustors," Part I, NASA CR-135132, June, 1977.
342
-------
EPRI Stationary Gas Turbine Catalytic Combustor Development Program
Overall Program Schedule
Description
Phase 1: Preliminary Design and Development
Phase II: Prototype Design and Evaluation
Phase III: Final Design and Reid Test
1980
1234
1981
1234
1982
1234
1983
1234
1984
1234
1985
1234
US
-P-
FIGURE 1
-------
U)
J^
-IS
was
Ret. No.
EPRI Stationary Gas Turbine Catyalytlc Combustor Devlopment Program
Major Milestone Chart
Task
Description
110000 Task I: Conceptual Design
120 000 Task II: Definition of Design Criteria
130000 Task III: Design Analysis
140000 Task IV: Preliminary Design
150000 TaskV: Design Optimization Plan
160000 Task VI Fundamental Development
Testing for Design
Verification
170 000 Task VII: Full Scale Combustor
Evaluation
180 000 Task VIII: Prototype Design
Specification
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
1MO 000
Program Management
FIGURE 2
-------
EPRI Stationary Gas Turbine Catalytic Combustor Development Program
Phase I Work Row Logic Diagram
Contract
Award
Definition
of Design
(Acceptance)
Criteria
• Performance
• Controllability
• Reliability
• Maintainability
• Degradation
•Cost
• Retrofltability
•Operation
Specifications
Task II
Design
Analysis
•Catalyst and
Substrate
• Emissions/
Performance
• Aerothermo
• Heat Transfer
• Vibration
•Mechanical
• Starting;
Control
• Reliability and
Maintainability
• Turbine Damage
Potential
Task III
TaskV
Fundamental
Development Testing
For Design
Verification
EPRI
Decision
Point
Approval
• Fuel/Air Preparation
Systems
• Ignition/Control Systems
• Liner Cooling
• Catalyst Packaging
Design Development
• Material Durability
-Catalyst
-Thermal Cycle
-vibration
Phase II
Prototype
Design and
Evaluation
Task VI
EPRI
Decision
Point
Approval
FIGURE 3
-------
T.»,v
Preliminary Desigr
140 OOO
-
DMiQ^Ul^
Detailed Desiyn L
Deajfln Optimization
Plar
150 OOC
-
Idem ideal*
Development
151 Oft
Development T
Fuel -Air
Sysse
161
Vepa-atipr
TI Test*
000
161 100 SuB-Scaie Combusto- Tesla (Wh
161 2OO t*ou\» Spray Tt»§(g (C)
161 300 Full Sea* GomDuttcx Teats (C)
Tas*. Viu
Prototype De&gr1!
Specification
180 OOC
-
Prototype C-ombu
Specif ica
161 X
Pfiasa
Prototype Co
Test Ran
182 000
EPRI Stationary Gas Turbine Catalytic Combustor Development Program
FIGURE 4
-------
-
W-501
FIGURE 5
-------
A Conceptual Design Arrangement of a Catalytic Ceramic Element in a
Stationry Gas Turbine
Conventional
Fuel Nozzle
I |
Primary Air
- u
Fuel Preparation
Zone
=4 ,
—
•MMM»
n-?\
V
Multiple Fuel Nozzles
Catalytic Ceramic Element
FIGURE 6
-------
CATALYTIC COMBUSTION FOR GAS TURBINES
By:
HISASHI FUKUZAWA AND YOSHIMI ISHIHARA
ENERGY AND ENVIRONMENT LABORATORY
CENTRAL RESEARCH INSTITUTE OF ELECTRIC POWER INDUSTRY
KOMAE-SHI, TOKYO, JAPAN
ABSTRACT
A national project for research and development of a high
efficiency gas turbine began in Japan in 1978. The objective of
the project is to raise the overall thermal efficiency of a
combined-cycle power plant system using a high-efficiency gas
turbine and a steam turbine. The target efficiency of 55% (LHV)
is expected to be reached by 1984.
The CRIEPI research and development program on catalytic
combustion for gas turbine combustors includes the development
and evaluation of catalysts, determination of optimum operating
conditions for combustibility and flue gas composition, and
development of a model catalytic combustor. The last stage of
the CRIEPI program is to develop a prototype of a low-NOx
catalytic combustor for a high-efficiency gas turbine.
A new experimental apparatus for catalyst evaluation at
atmospheric pressure was constructed in November 1979, and the
characteristics of catalytic combustion were investigated using
in-house manufactured catalysts. The results were as follows:
(1) Thermal NOx formation was reduced to as low as below
20 ppm at temperatures up to 1,500° C.
(2) Complete catalytic combustion was achieved at above
1,000° C.
(3) The volumetric heat release rate increased with the
temperature and fuel rate, and more than
100 x 106 kcal/h.atm.m3 was obtained.
349
-------
INTRODUCTION
A national project for research and development of a high
efficiency gas turbine of 100 MW to be operated at 1,500° C under
55 atmospheres was begun in 1978. The objective of the project
is to achieve an overall thermal efficiency of 55% (LHV) by a
combined-cycle power plant system, shown in Fig. 1.
Heat-resistant materials such as alloys and ceramics, and
elements technology for the turbine blade cooling system, com-
pressor, combustion, turbine and controlling system have been
investigated according to the R&D schedule shown in Table I.
The design and construction of a 100 MW gas turbine pilot
plant to be operated at 1300° C and 55 atmospheres has now been
incorporated into this project by the Ministry of International
Trade and Industry (MITI).
The CRIEPI R&D program on catalytic combustion was intended
to connect with the national project on the development of a high
efficiency gas turbine for NOx control. The program includes
experiments on pressurized catalytic combustion at 10 at-
mospheres in 1980 and the development of a model catalytic
combustor in 1981. The objective of this program is the applica-
tion of catalytic combustion technology to a high temperature
gas turbine combustor to eliminate NOx formation.
CRIEPI R&D PROGRAM ON CATALYTIC COMBUSTION FOR GAS TURBINES
Promotion of a national project for R&D of a high efficiency
gas turbine was begun by the Agency of Industrial Science and
Technology of the Ministry of International Trade and Industry in
350
-------
1978. CRIEPI is participating in this project by conducting some
of the research on combined-cycle systems, heat-resistance mate-
rials, and cooling systems.
The CRIEPI R&D program on catalytic combustors for gas tur-
bines is being promoted as part of the national project starting
in 1980. We are conducting research in the following areas
according to the R&D schedule shown in Table IE.
1. Development and selection of catalysts
(1) High operational capability
(2) Combustion rate of fuels with each catalyst
(3) Catalyst life at high temperatures
2. Operating conditions and flue gas composition
(1) Catalytic combustion at atmospheric pressure
(2) Catalytic combustion at pressurized atmosphere
3. Model catalytic combustor tests for gas turbines
(1) Development of a model catalytic combustor
(2) Describing characteristics of a model catalytic
combustor
(3) Design of a full-scale catalytic combustor
Today we are involved in the development and evaluation of cata-
lysts for a catalytic combustor for a high efficiency gas tur-
bine. We are constructing another experimental apparatus for
pressurized catalytic combustion at 10 atmospheres.
351
-------
EXPERIMENTAL CATALYTIC COMBUSTION UNDER ATMOSPHERIC PRESSURE
EXPERIMENTAL APPARATUS
The schematic figure of the experimental apparatus used in
this investigation is shown in Fig. 2. The apparatus consisted
of an air heater, a catalytic reactor, a gas cooler and analyzers
for some contents of flue gases.
LPG was supplied from a cylinder, premixed with preheated
air and sent into the reactor at a uniformly controlled velocity.
The premixed fuel was oxidized completely on the catalyst sur-
face, settled in the middle of the reactor, and expected tem-
peratures v/ere obtained by controlling fuel/air ratios. Gas
temperatures were determined at positions upstream, downstream,
and on the surface of the catalyst.
The maximum air flow rate was 120 Nm3/h and the maximum LPG
fuel consumption was 2 Nm3/h. The inside diameter of the reactor
was 9.4 cm and the maximum catalyst size 9.0 cm in diameter and
15.0 cm in length. Flue gas analyses of NOx, THC and CO were
made separately by NDIR method analyzers.
Most of the catalysts used in these tests were monoliths
82 mm in diameter and 75 mm in length, including 25 mm long seg-
ments. The monoliths consisted of a 2.8 mm square cell mullite
or cordierite substrate washcoated with V-alumina. Thirty kinds
of catalysts were loaded on them.
THERMAL NOx FORMATION
The maximum inlet temperature of the gas turbine is expected
352
-------
to be increased to 1,300° C in 1981 and to 1,500°C in 1384 by
the national project on high-efficiency gas turbines. Therefore,
catalytic NOx reduction FGT equipment will be installed for
reducing NOx, assumed to be formed at several hundreds ppm.
With the objective of eliminating the FGT equipment, we
have started a R&D program in catalytic combustion technology.
In experiments in the first stage of the program, we have
confirmed the low thermal NOx emission as described in previous
papers.
NOx concentration which corrected to 0% excess air increased
while the combustion gas temperature was raised, as shown in
Fig. 3. As the figure also shows, NOx concentrations, which
varied somewhat with the different catalysts, were as low as
20 ppm at below 1,500° C.
UNBURNED HYDROCARBONS AND CARBON MONOXIDE EMISSION
Small amounts of unburned hydrocarbons (THC), which were
analyzed as total hydrocarbons, were emitted in the combustion
temperature range from 900 - 1,100° C, but not at above 1,100° C
as shown in Fig. 4. Carbon monoxide (CO) was also emitted at
20 ppm or less in the temperature range from 900 - 1,100° C,
not at above 1,100° C. But both concentrations of THC and CO
increased greatly at temperatures below 900° C.
As a result, a combustion gas temperature of approximately
1,000° C or more was required to effect stable catalytic combus-
tion using LPG fuel. The temperature varied slightly with
different catalysts.
COMBUSTION TEMPERATURE
In the catalytic combustor the premixed LPG fuel gas is
oxidized on the preheated catalyst surface without an igniter.
The preheated temperature required for platinum catalysts was
above 330° C.
353
-------
Combustion temperature was controlled by controlling the
preheated air temperature and the fuel/air ratio . The catalyst
surface temperature obtained was approximately 50 - 160° C higher
than the combustion gas temperature at above 1,200° C, as shown
in Fig. 5. Further, the combustion gas temperature agreed with
the adiabatic combustion temperature, as shown in Fig. 6.
VOLUMETRIC HEAT RELEASE RATE
The volumetric heat release rate was raised as both the flow
rate of premixed fuel and the combustion gas temperature
increased.
In catalytic combustion it is possible to burn a large quan-
tity of fuel because the combustion rate is extremely fast. That
is, the maximum volumetric heat release rate obtained in these
tests was 137 x 10^ kcal/h.atm.m3 at combustion gas temperatures
under 1,500° C, 190% of theoretical air, 30.9 x 10^ h~l of space
velocity. This value indicates the possibility of burning a fuel
corresponding to 60 MW of electrical power generation with a
thermal efficiency of 40%, that is, about 13 t/h of oil as a fuel.
Figure 7 shows that the volumetric heat release rate will be
raised extensively as the premixed fuel gas flow rate increases.
The above value of the volumetric heat release rate indicates the
catalyst volume is small enough to settle in the combustor
designed for high-efficiency gas turbines.
354
-------
REFERENCES
1. Krill, W.V. and Kesselring, J.P., "The Development of Cata-
lytic Combustors for Stationary Source Applications,"
Proceedings: Third Workshop on Catalytic Combustion,
Asheville, N.C., Oct. 1978, EPA-600/7-79-038 Feb. 1979.
355
-------
Stack
Waste Heat Boiler
Fuel
Electricity
Generator
High Efficiency Gas Turbine
Steam
Turbine
Steam
Electricity
Generator
Rgure I. Schematic Figure Combined-Cycle Power
Plant System Having High Efficiency Gas Turbine
-------
Reactor
Gas
Cooler
U)
To Stack
Catalyst
NOx analyzer
THC analyzer
CO analyzer
02 analyzer
Vaporizer
Air Heater
Blower
Air
LPG
Figure 2. Schematic Figure of Apparatus for
Catalytic Combustion under Atmospheric Pressure.
-------
"8
QE
50
40-
o
Q.
Q.
.30
So
§8 2O-
O
O
10
0
SV, 25.3 x|Q4h"'
Catalyst
o C - 102
* C - 104
° C - 105
1000 1100 1200 1300 1400 1500 1600
Combustion gas temperature,°C
Figure 3. Thermal NOx Formation
358
-------
o
CD
o
O
r*
Sd
g
E'E
•£•£
CD CD
°i
o o
O x
xo
HZ
CO
THC
NO
/
800 1000 1200 1400 1600
Combustion gas temperature, °C
Figure 4. Temperature Zone of Unburned
Hydrcxxntons and Carbon Monoxide
Formation.
359
-------
o
"O
II
"2"o
i-s.
|§
§10
a
M — O">
JU§
-2 E
0 0
OO
1 OUU
1600
UOO
1200
1000
«nn
C-II2 Catalyst
Gas flow rate, IOONm3/h
°xo SV, 25.3x|04rTl
No o Catalyst surface temp.
/v \ A Combustion gas temp
*A O
h X\
vx
XX
N^
- "^V--
\
\
— \
i i i i i
150
200
250
300
350
Theoretical air, %
Figure 5. Effect of Theoretical Air on
Combustion Temperature.
400
-------
2000r
o
o
1800-
g- 1600
o 1400
.o
1 1200
JD
O
O
1000
1000 1200 1400 1600 1800 2000
Adiabatic combustion
temperature,0 C
Figure 6. Relation between Combustion
Temperature and Adiabatic
Combustion Temperature.
361
-------
I40r
120
•5 100-
ro
E
"5
o
O)
-------
Table I. RaD Schedule of Natioal Project for High
Efficiency Gas Turbine.
"\^
RaD of heat-resistant
materials and elements
technology
100 MW Prototype
Plant tests
1978
'////,
1979
'////,
Calen
I960
/ / / / /
dar Ye
1981
iar
1982
/////,
\
/////
Design
1983
'////.
> i
/ / / / s
Construct-
ion
1984
'////
>
Y////A
OperatJonj
-------
Table II. CRIEP Schedule for Low NOx
Emission Catalytic Combustor Program.
u>
— - — ________
Development and selection
of catalysts
Operating conditions and flue
gas composition
Development of model catalytic
combustor for gas turbine.
1979
'////,
'////.
Calendar
I960
'////,
' / / / /
Year
1981
'////,
/ / x
Y //
1982
/ ////
-------
SOOT REDUCTION IN DIESEL ENGINES BY CATALYTIC EFFECTS
BY
R. Sapienza, T. Butcher, C. Krishna, J. Gaffney
Brookhaven National Laboratory
Upton, New York 11973
ABSTRACT
Recent tests at Brookhaven National Laboratory indicate that both small
additions of alcohols to the fuel and the presence of platinum surfaces in the
combustion chamber can reduce soot emissions in a diesel engine. These tests
were conducted over a limited range of operation in a single cylinder CFR en-
gine. Most of the testing was done using pure cetane as a fuel at constant
speed and load. Possible major features of the reaction mechanisms for both
fuel additives and surface catalyst effectiveness are presented.
365
-------
ACKNOWLEDGEMENT
We are indebted to T. O'Hare, W. Marlow, M. Sansone, J. Barry, J. Hurst,
R. Klemm, and C. Waide for their assistance and support. We also wish to
thank I. Tang, H. Munkelwitz, F. Salzano, R. Whisker, and A. Berlad for help-
ful discussions. This research was carried out at Brookhaven National Labora-
tory under contract with the U. S. Department of Energy.
366
-------
INTRODUCTION
Diesel car sales are climbing in the United States and some manufacturers
are predicting that by 1985 25% of the new cars made in the U.S. may be diesel
powered (1). The diesel engine*s proven fuel economy and lower emissions of
unburned hydrocarbons and carbon monoxide makes it a viable alternative to a
gasoline automotive plant, but the inherent production of particulate matter
(soot) threatens the expanded use of the diesel engine.
Current diesel cars emit about 13 Ibs. of particulate matter for every
10,000 miles - 50 times the gasoline auto. High concentrations of these emis-
sions may have a synergistlc effect when combined with other carcinogens in
the urban environment (2) and this problem may overshadow the diesel's fuel
savings.
A variety of methods have been proposed over the years for reducing soot
emissions (3). These methods include engine modifications such as fuel injec-
tion optimization, combustion chamber shapes and turbocharging; fuel modifica-
tions such as the use of water injection, fuel additives, and fuel fumigation;
and more recent suggestions which stress exhaust treatments including "trap
oxidizers" similar to catalytic converters and various types of replaceable
filters. Fuel blending such as methanol-diesel emulsions, as well as separate
introduction of methanol, have also been tried.
In the diesel engine the air charge is compressed to a high temperature
and pressure, proceeding fuel injection and combustion begins spontaneously
shortly after. The rate of burning and heat release is controlled by the
rate of fuel air mixing. This mixing-limited operation leads to not only many
of the diesel1s desirable characteristics but also to hot, fuel rich zones
despite the air charge being in excess of the stoichiometric requirements.
Cracking occurs, leading to soot formation in these fuel rich zones. This can
367
-------
be perceived as an autothermal cracking process in which heat, from combustion
of the fuel, serves to crack the remainder endothermally (similar to the manu-
facture of ethylene and acetylene via the pyrolysis of hydrocarbons (A)).
Carbon nucleus >soot
decomposition
Auto thermal cracking
Fuel * C9Kj cyciization ^ aromatics ^polynuclear
aromatics
pyrolysis
polymerization - _.
- ^ olefins
&
dienes
All of the soot formed is not emitted as some of it is combusted in later parts
of the engine cycle. It has long been recognized that the engine maximum load
is limited by excessive smoking at overall air fuel ratios still well above
stoichiometric.
Since a precise mechanism of soot formation during combustion in a diesel
engine is not completely defined, a tentative approach was formulated relating
the partial combustion of liquid hydrocarbons to the initiation of soot forma-
tion in the diesel combustion process. There is ample evidence that soot formed
in combustion processes consists of a mixture of surface and gas formed carbons
(5), therefore, the influence of both homogeneous and wall reactions in the
combustion chamber were considered in order to define the fuel and engine pro-
perties which contribute to the formation of particulate matter.
EXPERIMENTAL
Complete combustion must be attained if soot is to be avoided and the in-
troduction of either oxygen carriers or catalytic surfaces, which would in-
crease the probability of contact between fuel and oxygen could influence the
formation of soot in the combustion chamber. The addition of hydroxyl (OH)
368
-------
carriers to the fuel was felt to be a direct way to verify or illustrate the
proposed chemistry. The possible catalytic influence of wall and valve mate-
rials to promote the surface ignition of carbon and soot precursor was also
investigated.
EXPERIMENTAL DETAILS
A standard Waukeska CFR engine coupled to a D.C. dynomometer was used in
the tests. Cetane was used instead of diesel oil as fuel for obvious chemical
reasons and n-butanol was used as the fuel additive. The engine was operated
at the ASTM Cetane number test conditions making a concerted effort to keep
the engine operating variables constant. Typical engine operating conditions
were: speed, 900 rpm; fuel flow, 13 ml/min., compression ratio, 19:10, (see table
for each test).
Sampling of the engine's particulate effluent was done by drawing 15.6
1. p.m. through a 47mm Millipore FA filter directly from the exhaust flow,
(see Figure 1). This procedure minimized distortion of the sample and main-
tained the integrity of the structurally-weak filter with no particulate
break-off. The filters were equilibrated at constant humidity and weighed to
determine the soot by difference. Duplicate samples were reproducible to better
than 10% and at least two samples were taken to each data point.
After the cetane/n-butanol mixtures were studied, the engine was dismantled
and platinum was deposited onto the piston crown and valve faces by vacuum sput-
tering to a thickness of approximately 6,000 & with an MRC model SEM-8620 RF
Bias sputter-etch module. Prior to insertion in the chamber the components
were cleaned mechanically in a glassblaster and washed with acetone and alcohol
as a final step. Sputter etching was not possible.
RESULTS AND DISCUSSION
Use of Alchols
The technology of alcohol fuels utilization in diesel engines is not nearly
as advanced as that for spark ignition engines. Since our tentative approach
to homogeneous soot formation may be conveniently formulated in terms of the
free radical theory of the cracking of hydrocarbons (6), alcohol could suppress
the formation of soot by providing species which remove or retard radical fuel
pyrolysis intermediates.
369
-------
Alcohols Action
1. Radical Retardation
a. Transfer agent
P« + RCH2OH - * PH + R*CHOH
P« + R*CHOH — — > PH + RCHO
P» + RCHO - *RCO + PH
b. Copolymerization agent
• •
P» + R CO >P-CR
?H
-R
H
2. Precursor Trap
a. reaction with acetylene
OH
HC=CH + RCHO — 4 H2C=C-C-R
fi
b. inhibits diene cyclization reactions
c. reaction with aromatic radicals
OH
C—
+ RCHO - > (oj ^A
The observed data as presented in Figure 2, reveal a rapid and then grad-
ual decrease in soot with n-butanol addition with an optimum 30-40% reduction
at ^ 7% n-butanol addition. Similar results were observed when using conven-
tional diesel fuel and ethanol mixtures (Figure 3).
Although recent tests using methanol and ethanol/diesel emulsions have
shown particule reduction (7,8), to our knowledge this represents the first
systematic investigation of specific alcohol addition versus reduction of soot
particulates in a diesel engine. These results seem to support our concept of
controlling the gas-phase mechanism of soot formation utilizing the radical
retarding effect of alcohols.
370
-------
Use of Metals
Catalytic combustion can promote the oxidation of long chain hydrocarbons
and soot precursors minimizing polymerization reactions which result in soot
formation.
The mechanism by which metal containing fuel additives reduce smoke is
obscure (9), but British Petroleum tests suggest that successful soot suppres-
sants lower the soot ignition temperature while other ineffective additives^
yield little change in this temperature (10).
Alkaline-earth metal fuel additives such as barium salts affect homogene-
ous soot formation in a manner similar to that proposed for alcohols. These
metals which reduce soot in all oxygen-fuel ratios, act by the gas-phase cata-
lysis of the decomposition of hydrogen or water vapor to yield hydroxyl
radicals (11).
Other effective metal fuel additives such as molybdemum, tungsten and
chromium perform only at high oxygen-fuel ratios, (>.9) suggesting a different
mechanism for their action (11). It is our contention that these metals act as
surface catalysts in the combustion process.
Early exploratory work on surface combustion has shown that hot surfaces
have the capability of accelerating combustion at the boundary region between
gaseous and solid phases (12). Considering that the nature of the reactor wall
has a significant effect on the rate of carbon formation and, the composition
of the exit gas in a cracking reactor (13), the oxidation properties of the
combustion chamber surface could determine the quantity and quality of the
particulate emissions in a diesel engine.
Since an oxidation catalyst serves as the source of oxygen transferred
to the reactant molecule, the most important property generally ascribed to a
combustion catalyst is its ability to dissociate oxygen molecules into adsorbed
oxygen atoms. If the formed oxide can be reduced by the hydrocarbon to a lower
oxide or the metallic state, the metal will function as an active catalyst.
We wish to propose a novel yet simple mechanism which would provide a
unified picture capable of explaining or predicting catalyst behavior. The
basis for this hypothesis was generated from various studies in slow combustion,
371
-------
which neglected wall effects in elucidating the mechanism and understanding
the chemistry of combustion (14).
This proposal starts with the assumption that oxygen must become activated
upon a surface, M (step ^). This may be the metal in catalyzed combustion or
even the walls of the reaction vessel. It has been found that reaction initia-
tion in the absences of walls may be severe (15) and most radical mechanisms
use the walls to convert intermediates to products (6).
The carbonaceous material (e.g., ethylene in diagram) reacts with the sur-
face oxygen forming a saturated intermediate (B) which undergoes carbon-carbon
bond cleavage to give the oxymethylene species, (C). This would be associated
with the induction period of combustion, and would be an exothermic reaction.
This species would explain why formaldehyde, which would chemisorb on a surface
in this manner, is responsible for branching (F) in the combustion of many
simple fuels (16). Similarly, saturated hydrocarbons could form acetaldehyde as
an intermediate (I). The addition of formaldehyde or acetaldehyde to a hydro-
carbon combustion reaction should eliminate the induction period and indeed this
is the case in the oxidation of ethane (17). The possible reaction paths of
the oxymethlyene intermediate result in the various products observed in the
slow combustion of ethylene (16).
Carbon dioxide arises from the reaction of the oxygen coordinated carbon
monoxide with surface oxide (£). This step is consistent with the mechanism
of carbon monoxide oxidation (18). Carbide or soot formation occurs (H) when
the adjacent surface oxide is not present (e.g., thru reduction). If the oxide
site were available soot formation would be reduced.
An interesting feature of this scheme is that it is closely related to the
recently proposed oxide Fischer-Tropsch (F-T) mechanism (19). Since both pro-
posals assume similar reaction intermediates correlations of metal catalysts
should be noted. Indeed, metals that form strong oxides are not good F-T
catalysts and will inhibit the oxidation of carbon. This relationship is
particularly attractive in the examination of controlled combustion for chemical
conversion.
In view of the well known combustion catalytic effects of platinum (20)
the coating of the combustion chamber surfaces with platinum was tried.
372
-------
The platinum coated combustion surface produced ^ 40% reduction in soot
with no alcohol addition. Addition of alcohol to the platinized reactor was
found to further reduce soot with reduction up to ^ 80% being measured (see
Figure 4).
A.
JL-
£.
SL-
E_.
F.
£.
JL-
X-
0 0
2
oo oo
M-M + CH.-CH- >ft —ft
2 2
CH CH CH0
1 2 1 2 11 2
9 9 ?
M M — * 2 M
CH c
II /
00 0
M ft ^ ^1 + M +
CH
M > CH 0 + M
CH2 CH2— CH2 ^.
II? 1
0000
2
^
0 0 C
ft + M — >ft + ft
0 0
1 f
CH3 CH2 CH2 CH2 . . . + M - M — —
CH-CH CH CH . ..
0 J Z
II
M + £-
H00
2
0
M
+ M
Saturated Molecule
> CH3CH2 CH2C1
9 9
M M
373
-------
The cyclic oxygen transfer mechanism appears to offer reasonable explanations
of these results since the free energy decrease on reduction of iron oxides
with carbon (or hydrocarbons) is less than that for the reduction of platinum
oxide. Because platinum will markedly lower the ignition temperature of
carbon (21) and promote ultralean combustion of hydrocarbons, (20) soot and
unburned fuel reaching the platinized combustion chamber surface could be
partially or completely oxidized leading to reduced particulate emissions.
Heterogeneous reactions of this sort will be limited by the amount of carbon
reaching the surface by molecular and turbulent transport but may produce
products which also affect the homogeneous chemistry. That is, reaction inter-
mediates initially formed at the surface, breakoff to take part in homogeneous
reactions. This chemistry would be similar to the action of alcohol described
earlier.
After eight hours of continuous operation the platinum coated surface was
lost, and the observed soot emission found to be identical with the original
iron surfaced piston and valves. Although the precise mechanism by which the
platinum was removed is unknown, we believe that surface impurities such as,
iron carbon and oxide may have caused problems in the platinum adhesion since
no stringent cleaning of the piston crown was performed prior to coating.
Thermal expansion differences or excessive vibration would also have deleterious
effects (22).
The short service life may also be ascribed to penetration by poisons or
oxygen through the catalytic lining with resulting pitting or corroding of the
iron alloy composing the piston. To overcome this a hard, smooth, non-catalytic
and non-absorbtive cap of low corroding properties (e.g., chrome-iron) should
be interposed between the catalyst and the coated body (23) .
CONCLUSION
This work towards reducing soot emission from diesel engines suggests that
both homogeneous and heterogeneous approaches are feasible. The homogeneous
approach consisted in mixing small quantities of alcohol with the fuel and the
heterogeneous approach consisted of coating parts of the combustion chamber
(bounding) surface with an oxidation catalyst like platinum. Both approaches
produced significant reductions in soot emissions. These tests were conducted
374
-------
over a small range of operation in a single cylinder engine (a CFR type), with
most of the testing done using pure cetane as fuel at constant speed, and load.
Although the engine and test conditions are not representative of an automotive
diesel engine, the results may be useful to assess the feasibility of soot re-
duction in automotive and truck diesel engines, using these methods.
The major features of the reaction mechanisms proposed for both fuel
additive and surface catalyst effectiveness could be applied to various com-
bustion problems.
375
-------
REFERENCES
1. Goen, R. L. and Ivory, M. E. Diesel Cars In the United States. TID-28735,
1978.
2. French, C., Monaghan, M., Fresse, P. G. A Study of the Diesel as a Light-
Duty Power Plant, EPA-PB 236-896 (NTIS), 1974.
3. Chalmers, J. F., Fullenwider, E. D., Minton, M. G., and Tomlinson, J. C.
Review of the Research Status on Diesel Exhaust Emissions. ATR-78(7716)-3,
The Aerospace Corporation, 1978.
4. Kamptner, H. K., Krause, W. R., and Schilken, H. P. High Temperature
Cracking. Chemical Engineering, 73:93-98, 1966.
5. Palmer, H. B. J. Chem. Phys., 87, April 1969.
6. Rice, F. 0. The Thermal Decomposition of Organic Compounds from the
Standpoint of Free Radical, I. Saturated Hydrocarbons, J. Amer. Chem. Soc.,
53:1959-1972, 1931.
7. Havemann, H. A., Rao, M. R. K., Natarajan, A., and Narasimhan, T. L.
Alcohol with Normal Diesel Fuels. Gas and Oil Power, 50:15-19, 1955.
8. Lawson, A. and Last, A. J. Development of an On-Board Mechanical Fuel
Emulsifier for Utilization of Diesel/Methanol and Methanol/Gasoline Fuel
Emulsions in Transportation. Proceedings of the Third International Sym-
posium in Alcohol Fuels Technology, Asilomar, California, 1979.
9. Lahaye, J. and Prado, G. Mechanisms of Carbon Black Formation. Chemistry
and Physics of Carbon, P. L. Walker, Jr. and P. A. Thrower, eds., Marcel
Dekker, Inc., New York, 1978 pp.167-294.
10. McConnell, G. and Howells, H. E. Diesel Fuel Properties and Exhaust Gas -
Distant Relations. SAE 670091, 1967.
11. Cotton, D. H., Friswell, N. J., and Jenkins, D. R. The Suppression of Soot
Emission from Flames by Metal Additives, Comb. Flame, 17:87-98, 1971.
12. Bone, W. A. Surface Combustion with Specific Reference to Recent Develop-
ment. Gas Journal, p.423, 1923.
376
-------
13. Trlmm, D. L. The Formation and Removal of Coke from Nickel Catalyst.
Cat. Rev. Sc. Eng., 16:155, 1977.
14. Bone, W. A. and Townend, D. T. A. Flame and Combustion in Gases. Long-
mans Green, London, 1927.
15. Boudart, M. Chemical Kinetics and Combustion. Proceedings of the Eight
Symposium (International) on Combustion, The Wilkams and Wilkins Co.,
Baltimore, Maryland, 1962, pp.43-50.
16. Minkoff, G. J. and Tipper, C. R. H. Chemistry of Combustion Reactions.
Butterworths, London, 1962.
17. Lewis, B. and VonElbe, G. Combustion, Flames, and Explosions in Gases.
Academic Press, New York, 1951.
18. Happel, J., Kiang, S., Spencer, J. L., Oki, S., and Knatow, M. A.
Transient Rate Tracer Studies in Heterogeneous Catalysis: Oxidation of
Carbon Monoxide, J. Catalysis, 50:429-440, 1977.
19. Sapienza, R. S., Sansone, M. J., Spaulding, L. D., and Lynch, J. F.
Novel Interpretation of Carbon Oxide Reductions. Fundamental Research in
Heterogeneous Catalysis, 3:179, 1979.
20. Pfefferle, W. C. The Catalytic Combustor: An Approach to Cleaner Combus-
tion. J. Energy, 2:142-146, 1978.
21. Robell, A. J., Ballou, E. V., and Boudart, M. Surface Diffusion of Hydro-
gen on Carbon, J. Phys. Chem., 68:2748-2753, 1964.
22. Roessler, W. U., Weinberg, E. K., Drake, J. A., White, H. M., and luara,
T. Investigation of Surface Combustion Concepts for NO Control. EPA
650/2-73-014, U. S. Environmental Protection Agency, Research Triangle
Park, North Carolina, 1973.
23. Prentice, J. U.S. Patent 1,869,0771, 1932.
377
-------
TABLE I
SOOT EMISSION MEASUREMENTS AS A FUNCTION OF
ADDITION OF N-BUTANOL TO A CONVENTIONAL (IRON) CFR ENGINE
PERCENT
ALCOHOL
ADDITION
0
1
2
5
8
10
20
SOOT (mg)a
1.15
1.00
.94
.86
.77
.76
.70
ENGINE OPERATING CONDITIONS
HP
2.8
3.0
2.9
2.9
2.7
2.6
2.4
RPM
890
895
925
880
880
890
900
INJECTIOND
ADVANCE
8.4
8.7
8.8
7.8
7.5
8.2
7.4
IGNITION^
DELAY
5.2
4.7
4.7
4.6
4.7
4.7
5.3
Soot samples taken for 30 seconds (see experimental section for details).
In degrees.
378
-------
TABLE II
SOOT EMISSION MEASUREMENTS AS A FUNCTION OF ADDITION OF N-BUTANOL TO CFR
ENGINE WITH PLATINUM COATED PISTON CROWN AND VALVE FACES
PERCENT
ALCOHOL
ADDITION
0
1
5
10
20
50
Oc
2c
15C
SOOT (mg)a
.66
.62
.57
.55
.31
.28
1.13
.96
.76
ENGINE OPERATING CONDITIONS
HP
2.8
2.7
2.6
2.5
2.1
1.9
2.5
2.6
2.3
RPM
930
890
890
920
885
930
915
910
885
INJECTION0
ADVANCE
9.2
8.3
8.2
8.3
7.4
7.2
8.2
7.9
7.5
IGNITION0
DELAY
4.7
4.7
4.9
5.2
5.2
5.7
4.5
5.0
5.2
Soot samples taken for 30 seconds (see experimental section for details)
In degrees.
"Platinum coating lost from surfaces.
379
-------
COLLECTION
FILTER
SURGE
19UMW
JTANK
TO
PUMP
FIGURE 1, SAMPLING SYSTEM
380
-------
10
O 20k A
O
UJ
QL
O 30
X
O
LJ
^ 40
50
CETANE ASELNE
J L
5 10 15 20
% VOLUME BUTANOL
FIGURE 2. BUTANOL ADDITION TO CETANE
381
-------
00-
10
20
Q
LJ
or
o
H
X
52
LJ
30
40
50
DIESEL BASELINE
O ETHANOL
D ETHANOL (INCREASED HP)
A BUTANOL
V CETANE
5 10
% VOLUME ADDITION
FIGURE 3. DIESEL FUEL DATA
382
-------
• CONVENTIONAL PISTON
O PLATINUM COATED PISTON
10 15 20 25 30 35 40
PERCENT VOLUME n-BUTANOL
45 50
FIGURE 4. EFFECT OF CATALYST (PLATINUM)
383
-------
HIGH PRESSURE TEST RESULTS
By
D. E. Carlll), M. C. Krepinevlch^, I. T. Osgerby(3)
(1) COMBUSTION TURBINE SYSTEMS DIVISION
WESTINGHOUSE ELECTRIC CORPORATION
Concordvllle, PA 19331
(2) ADVANCED COAL CONVERSION DEPARTMENT
WESTINGHOUSE ELECTRIC CORPORATION
Madison, PA 15025
(3) ENGELHARD INDUSTRIES DIVISION
ENGELHARD MINERALS & CHEMICALS CORP.
Edison, NJ 08817
ABSTRACT
The Westinghouse/Engelhard program with the objective of developing
a full scale prototype catalytic combustor requires judicious use of
sub-scale test facilities. On this basis the Westinghouse test program
at the Waltz Mill facility was established. The specific areas of
effort involving design information are those of the evaluation and
development of both conceptual fuel preparation zones and catalysts.
Selected accomplishment thus far have been the following:
Advanced catalytic combustor test technology to 15 atmospheres.
This has included extensive rig and instrumentation develop-
ment spanning a period of approximately 18 months.
Recorded excellent emission of 12 ppmv CO and 1 ppm N0x at 190
psig, 2200°F outlet temperature (CATATHERMAL operation).
384
-------
HIGH PRESSURE TEST RESULTS
INTRODUCTION
Monolithic catalysts have been successfully tested at high pressure in
the newest of the Westinghouse combustion facilities—the Advanced Coal
Conversion Department (ACCD) Test and Development Center. The principal
accomplishments thus far have been the following:
o Observed CATATHERMAL operation at pressures up to 190 psig in
four inch diameter on distillate oil fuel.
o Advanced catalytic combustor test technology to 15 atmos-
pheres. Whereas earlier successful testing at 3 atmospheres
was carried out with relative ease, high pressure testing has
required extensive rig and instrumentation development.
o Recorded excellent emissions of 12 ppmv CO and 1 ppmv NO at
190 psig, 2200°F outlet temperature (CATATHERMAL operation).
o Demonstrated adequate fuel preparation zone (FPZ) performance
for research purposes with the multiple-conical (NASA-type)
fuel-injector.
o Observed no apparent catalyst deterioration following liquid
impingement during startup or during brief episodes of pre-
mature ignition in the FPZ.
FACILITY DESCRIPTION
The Test and Development Center, operated by the Westinghouse Advanced
Coal Conversion Department, is located at the Corporation's Waltz Mill
Site in Madison, Pennsylvania. It was designed and constructed between
1975 and 1978 to provide flexible, modern facilities for conducting
development tests on gas turbine combustors, as well as life tests on
turbine blade materials and coatings, and for determining the
corrosion/erosion/deposition characteristics of fuels and cooled-air-
foil surfaces.
385
-------
The Waltz Mill test complex is shown on Figure 1. The 600-square-foot
main laboratory contains a 33 foot high- bay at the north end, and two
stories of laboratories and offices in the south half. The experimental
area is monitored from a second floor control room that overlooks the
high bay. A near-by Compressor Building houses a centrifugal and a
reciprocating compressor; a fuels blending facility with day tanks for
testing fuel oils and pumps for the cooling water systems; and a Fuel
Storage Area containing underground storage tanks.
The schematic in Figure 2 shows the installed facilities and the con-
nections required to serve the test passage.
Non-Vitiating Air Preheaters
Two dual-fuel fired heat exchangers heat the air from the process air
compressors before it enters the test passages. The air side operates
at full process pressure. The complete unit can raise 7.5 Ib/sec from
200 to 1200°F or the entire 12.5 Ibm/sec to a lower temperature. The
heaters are fired by natural gas or No. 2 fuel oil.
Fuel System
No. 2 fuel oil and clean coal liquids are stored underground in two
10,000 gallon tanks and in one 5,000-gallon tank. The fuel can be
transferred from these tanks to five 500 gallon day tanks located in the
Fuels Blending Building. These day tanks are connected to the Process
Laboratory combustion passage test area.
On-Line Gas Analysis Facility
On-line continuous monitoring of combustion rig emissions is accomplished
with a Beckman-designed gas analysis system. The system includes a
heated sample line, a smoke meter with detached densitometer, and the
instruments listed below:
386
-------
Analyzer Method Type
NO/NO Chemiluminescence Beckman 951 H
x\
CO Infrared Beckman 865
COp Infrared Beckman 865
^0 Infrared Beckman 865
02 Magnetic susceptibility Beckman
Hydrocarbons Flame ionization Beckman 402
Smoke SAE Beckman
The gas analysis includes provisions for continuous recording of all
parameters except smoke. Sample flow to any instrument can be inter-
rupted for on-line calibration check. General operating practices
include calibration prior to and following each test period.
DATA ACQUISITION SYSTEM
All sensors provided in the combustion laboratory at Waltz Mill are
electronic, and the site is equipped with a PDP-11/34 computer. A
Digital Equipment Co. data acquisition system serves as interface be-
tween sensors and computer. A similar computer has served the Coal
Gasification Process Development Unit at the same site for four years.
CATALYTIC COMBUSTOR TEST RIG AND INSTRUMENTATION
The catalytic combustion test passage (Figure 3) is constructed of 8 in.
O.D. Sch 80 stainless steel pipe with a 5 in. I.D. combustor section.
This internal dimension allows for cooling air and insulation around the
combustor internals. The 8 in. pipe is the pressure boundary and is
rated at 300 psig at 1000°F. The test passage consists of four sections:
o Air preparation
o Fuel vaporization
o Combustion
o Cooling
387
-------
The air preparation section takes the preheated process air and directs
it to the vaporization section. The process air must have an even
distribution for good catalytic combustion, which is accomplished with
an appropriate device upstream of the fuel injector.
The fuel vaporization section consists of a NASA-type fuel nozzle with
24 injection tubes. The vaporization section is designed to mix the
fuel with preheated air to yield a homogeneous fuel/air mixture.
The combustion section contains the catalyst where oxidation of the
fuel/air mixture takes place. The inside liner of the combustion section
has cooling fins and cooling air flow on the outside wall. This cooling
air mixes with the combustion air at the exit of the combustion section.
The cooling section is the place where the hot combustion gases are
cooled by a high pressure water spray and exhausted to the muffler
house.
Test data are recorded by the data acquisition system (DAS), for sub-
sequent retrieval and analysis. These data include the following:
o Air mass flow
o Oil flow
o Preheat temperature
o Test passage pressure
o Catalyst pressure drop
o Catalyst outlet temperature profile
o Combustion gas temperature
o Exhaust gas analysis.
The DAS has the capability of collecting all required data at the rate
of one set every ten seconds. This data rate is ideal for steady-state
operation, yet is sufficient for characterization of start-up conditions
because the start-up transient takes several minutes.
388
-------
Process air mass flow rate is measured by the pressure drop of a Venturi
and is compensated by pressure and temperature corrections. The mass
flow-rate computed from the Venturi Ap, pressure, and temperature is fed
back to the process air controller. The oil flow is measured by a full
flow turbine meter. Test rig temperatures are measured by either type K
or type S thermocouples. The gage pressure and differential pressures
of the test rig are measured by transmitters. The combustion gas products
are sampled and analyzed for the following contents:
o Unburned hydrocarbons
o Nitric oxide
o Nitrogen dioxide
o Oxygen
o Carbon monoxide
o Water vapor
o SAE smoke number
These data are continuously recorded so that combustion efficiency and
catalyst performance can be evaluated. The specific emissions analysis
instrument action available at Waltz Mill was described earlier.
For this test, Engelhard provided a catalyst designated as DXE-442,
which uses palladium as the principal precious metal component supported
by a stabilized alumina washcoat. The monolithic support is a zircon
composite having a cell density of 256 channels per square inch. The
catalyst configuration consist of two 3.8 inch diameter catalysts, each
2 inches long, with a 0.25 inch space between the two segments as shown
in Figure 4.
RESULTS
The test procedure for the subject catalyst included operation in the
CATATHERMAL Mode at 110, 150, and 190 psig. At each pressure, air flows
were varied so that the CATATHERMAL reaction was extinguished and
re-established. The results at each pressure are shown in Figures 5, 6
389
-------
and 7. The results show that the catalyst pressure drop was never
greater than 4.02 percent.
The flat temperature profiles verify the adequacy of the fuel preparation
zone. The predicted temperature, based on measured fuel and air flow
rates is shown on each figure. The temperature measured by means of
exhaust gas analysis at the center!ine is in agreement with the other
temperatures.
There appears to be a slight pattern factor developing in the catalyst
exit temperature profile at the higher pressures. This occurence is
still being investigated, but the most likely explanation is that it is
due to a change in the FPZ.
CONCLUSION
The catalyst has been shown to operate as well or better at 190 psig as
it does at lower pressures. Catalyst transition from non-CATATHERMAL to
CATATHERMAL operation at pressures up to 190 psig has been demonstrated
through testing so far.
390
-------
to
VO
•\fLIOUID PROPANE
] STORAGE TANK
•5,000 GAL
TOLUENE TANK
•MIXED GAS
STAND PIPE
(2) 1O,OOO CAL
FUEL OIL TANKS
3.OOO GAL
WASTE OIL
TANK
FUEL STORAGE AREA
COOLING TOWERS
S~\NITRQGEN\
\_JSUPPLY |
COMPRESSED a LIQUIFIED
GAS STORAGE AREA
SWITCHGEA
HOT WATER BOIL Eft
I.OOO GAL N? 6 FUEL OIL
500 GAL TOLUENE
-45) 500 GAL.
FUEL OIL
FUELS BLENDING BLDG.
WATER SYSTEM—
UTILITY RM-
FUEL/GAS ANALYSIS ROOM-
PROCESS LABORATORY
BLDO. - FIRST FLOOR
CENTRIFUGAL
CQMPRESSOB
•AIR RECEIVER
SWITCHGEAR
- I.OOO GAL
WASTE OIL TttNK
COMPRESSOR BUILDING
N
SECOND FLOOR
3028
FIGURE 1. WESTINGHOUSE WALTZ MILL COMBUSTION/PROCESS LABORATORY
-------
N« 2
FUEL OIL/1
STORAGE
ZO.OOO GAL\
RUN
TANKS,
500 6*1
£ACH
HIGH PRESS. AIR
45O PS/6
10 SCFM
PROPANE
4O TO 45 O
PS/6
455 SCFM
N*6
FUEL OIL
STORAGE
COOLING
WATER
INJECTION
SYSTEM
is a SPM
AT 450 PS/
OJ
STEAM
J Lss/sa
4OO PS/G
MISC. GAS
SUPPLY
AIR |—I
COMPRESSOR
75 LBS/SCC
15,760 SCFM) ' 1
30O PS/G
EQUIPMENT
COOLING
SYSTE
AIR
IPREHEATER
\I2OO "F
30O PS/G
COMBUSTION
SECTION
AUXILIARY -
TEST
HARDWARE
TEST PASSAGES
2,5OO°F 3OO PSI6
OOP
DO
CONTROL ROOM
PRODUCT GAS
CHARACTERIZATION
AIR
PREHEATER
IOOO°F
3OO PS/C
I,ZOO 6PM
AT 7* PSI
REVV
!020-77
AIR
COMPRESSOR
5.O LBS/SEC
(3,840 SCFM)
30O PS 16
3021
FIGURE 2. OVERALL FLOW DIAGRAM FOR PROCESS LABORATORY
-------
DILUTION
AND
COOLING
FLOW STRAIGHTENING
FIGURE 3. CATALYTIC COMBUSTION TEST PASSAGE
-------
CV^**^?'-"
FLOW
FIGURE 4. CATALYST IN HOLDER
-------
Ln
1000 1500 2000
—
—
i I
O
s
,
PREDICTED FROM MEASURED „
FUEL/AIR RATIO
I l
I
v v3.s
) p
AP/P
f/a
to
s
(CO)
(UHC)
(NOX)
P.P.'
1
1
1000 1500 2000 2500
= 706 F
= 80 FPS
= 110PSIG
= 3.45%
= 0.0257
= 1.8 PPS
- T/C, 2-3/4"
- EGA. 21-3/4"
- T/C, 20-3/4"
= 4 PPM
- OPPM
= I.3PPM
12%
OUTLET TEMP (DEC F)
FIGURE 5. TEST RESULTS: 110 psig
-------
LO
VD
1000 1500 2000
—
0
PREDICTED FROM MEASURED
FUEL/AIR RATIO
i i
1000 1500 2000
TIM
V3.8
P
V AP/P
f/a
GA
i D °
D
A
(CO)
(UHC)
(NO )
9
P.P.
2500
= 713 F
= 87 FPS
= 150 PSIG
= 4.02%
= 0.0241
=2.6 PPS
- T/C, 2-3/4"
- EGA, 21-3/4"
- T/C, 20-3/4"
= 9 ppm
= 5 ppm
= 0.9 ppm
= 3.0%
OUTLET TEMP (D?f? F)
FIGURE 6. TEST RESULTS: 150 psig
-------
0
VO
_L
1000
1500
2000
IN
= 686 F
V3.8 = 79 FPS
P = 190 PSIG
AP/P = 3.58%
f/a = 0.0249
GA =3.0 pps
O - T/C, 2-3/4"
n - EGA, 21-3/4"
A - T/C, 20-3/4"
(CO) = 12 ppm
(UHC)= 6 ppm
(NOX)= 1.3 ppm
P.F. = 12.1%
PREDICTED FROM MEASURED
FUEL/AIR RATIO
AS FIG 5
2500
OUTLET TEMP (DEC F)
FIGURE 7. TEST RESULTS: 190 psig
-------
COMBUSTION CATALYST STUDY
FOR SIMULATED AIRCRAFT IDLE MODE OPERATION
By:
I. T. Osgerby (Speaker)
R. M. Heck
R. V. Carrubba
C. C. Gleason (General Electric)
E. J. Mularz (NASA-Lewis Research Center)
Engelhard Industries Division
Engelhard Minerals and Chemicals Corporation
Menlo Park, Edison, New Jersey 08817
Presented at
Fourth EPA Workshop on Catalytic Combustion
Cincinnati, Ohio
May 14-15, 1980
Prime Contractor: G.E. from NASA-Lewis Contract NAS 3-20580
Subcontractor: Engelhard from G.E.
Contract Manager: C. C. Gleason (General Electric)
Program Monitor: E. J. Mularz (NASA-Lewis)
398
-------
ABSTRACT
One of the advanced combustor design concepts under Investigation
involves the use of a catalytic reactor downstream of a conventional gas
turbine combustor to reduce CO and unburned hydrocarbon emissions at engine
idle conditions. Laboratory screening tests were conducted to evaluate the
effect of catalyst cell geometry on clean-up performance for CO and propane
fuels.
The gas feed to the catalyst configurations tested simulated combustion
gases from a conventional combustor. Analysis of the screening test results
provided performance and design criteria for assessing clean-up conversion
and pressure losses over a wide range of operating conditions. A preferred
catalyst configuration was selected from this study based on the design goals
of:
Catalyst Pressure Loss AP < 3.%
P
Emissions Index CO $ 10 g/Kg fuel
UHC ^ 1. g/Kg fuel
NOX emissions for this design approach would be determined by the flame
preburner, and were essentially zero in all laboratory simulation tests.
(Preburner was simulated with a CATCOM* catalyst used to completely combust
the Jet A fuel.)
Parametric studies were conducted to measure the benefits of a
homogeneous reaction contribution at the higher catalyst operating
temperatures typical of the intended application. These results showed
that at operating temperatures of 960°C the CO emission requirements were
attained, but the UHC emissions were borderline. At operating conditions
above 960°C, it is anticipated that the emission requirement will be met.
* CATCOM is a trade name of Engelhard Minerals and Chemicals Corporation.
399
-------
COMBUSTION CATALYST STUDY
FOR SIMULATED AIRCRAFT IDLE MODE OPERATION
By:
I. T. Osgerby (Speaker)
R. M. Heck
R. V. Carrubba
C. C. Gleason (General Electric)
E. J. Mularz (NASA-Lewis Research Center)
Engelhard Industries Division
Engelhard Minerals and Chemicals Corporation
Menlo Park, Edison, New Jersey 08817
Presented at
Fourth EPA Workshop on Catalytic Combustion
Cincinnati, Ohio
May 14-15, 1980
Prime Contractor: G.E. from NASA-Lewis Contract NAS 3-20580
Subcontractor: Engelhard from G.E.
Contract Manager: C. C. Gleason (General Electric)
Program Monitor: E. J. Mularz (NASA-Lewis)
398
-------
ABSTRACT
One of the advanced combustor design concepts under investigation
involves the use of a catalytic reactor downstream of a conventional gas
turbine combustor to reduce CO and unburned hydrocarbon emissions at engine
idle conditions. Laboratory screening tests were conducted to evaluate the
effect of catalyst cell geometry on clean-up performance for CO and propane
fuels.
The gas feed to the catalyst configurations tested simulated combustion
gases from a conventional combustor. Analysis of the screening test results
provided performance and design criteria for assessing clean-up conversion
and pressure losses over a wide range of operating conditions. A preferred
catalyst configuration was selected from this study based on the design goals
of:
Catalyst Pressure Loss AP_ < 3.%
P
Emissions Index CO £ 10 g/Kg fuel
UHC « 1. g/Kg fuel
NOX emissions for this design approach would be determined by the flame
preburner, and were essentially zero in all laboratory simulation tests.
(Preburner was simulated with a CATCOM* catalyst used to completely combust
the Jet A fuel.)
Parametric studies were conducted to measure the benefits of a
homogeneous reaction contribution at the higher catalyst operating
temperatures typical of the intended application. These results showed
that at operating temperatures of 960°C the CO emission requirements were
attained, but the UHC emissions were borderline. At operating conditions
above 960°C, it is anticipated that the emission requirement will be met.
* CATCOM is a trade name of Engelhard Minerals and Chemicals Corporation.
399
-------
NOMENCLATURE
a Geometric total surface to volume ratio
DJJ Hydraulic diameter
Df-air Diffusivity of species in air
D Combustion efficiency
F* Apparent friction factor
L Catalyst channel length
%ec Reynolds' number in catalyst channels
Ng Sherwood number (dimensionless mass transfer parameter)
Ng Schmidt number (dimensionless mass transfer parameter)
P Pressure
A Pressure drop
p Density
CJ Percent open area of honeycomb support
T Temperature
V Catalyst face velocity
X Mole fraction of seed fuel
X Conversion
400
-------
I. INTRODUCTION
Exploratory investigations of the application of catalytically supported
combustion to gas turbine combuator design (References 1, 2, 3) have
previously confirmed the potential for achieving ultra-low emissions under
controlled, high power conditions. These studies further identified engine
idle conditions as a potential problem area, since low compressor discharge
and operating temperatures lead to high hydrocarbon emissions.
One design concept proposed to achieve low emissions at engine idle is
to employ a preburner and to utilize the catalyst in the "clean-up" mode to
reduce preburner hydrocarbon emissions. Conceptually, at higher power the
preburner operation is terminated, and the catalyst operated under
catalytically supported combustion conditions, to maintain ultra-low
emissions.
An experimental program was carried out to determine whether acceptable
hydrocarbon and CO conversion levels could be achieved with a
combustion catalyst operated as a catalytic converter at simulated aircraft
idle mode conditions. Correlations obtained in this study were utilized to
design a clean-up catalyst for testing in combustor sector hardware as
reported separately (Reference 4).
The performance goals for the catalytic converter were as follows:
A. Clean-up capability to maintain combustor emissions less than
10 g/Kg fuel of CO and 1 g/Kg fuel UHC.
B. Converter pressure losses of ^ 3%.
401
-------
An additional constraint was imposed by the hardware design in that
the space available for catalyst was restricted to a maximum length of
9.5 cm.
Combustor Design Concepts
One catalytic converter design approach is illustrated in Figure 1A
in which the catalyst is placed downstream of a conventional combustor
with the total combustion mass flow passing through the catalyst for
emission reduction. Preliminary calculations indicated that severe
pressure losses would be incurred and low conversion levels obtained at
the operation conditions listed in Table I (Mode A-l). Reducing the
catalyst mass flow by bypassing combuator air round the catalyst, and
mixing the bypass air with the catalyst exhaust products, reduces the
pressure losses and increases anticipated catalytic conversion performance.
Additional calculations indicated that using a configuration with 50 percent
bypass air (Figure IB) provides a broad range of combustor operation with
pressure losses < 3%, while maintaining peak catalyst inlet temperatures
under 1260°C. Catalytic converter inlet conditions for 50 percent bypass
air are listed in Table I under Mode A-2.
Combustor A-2 is assumed to be operated as a conventional gas turbine
combustor primary zone with the balance of the combuator catalyst air
introduced between the primary zone and catalyst to mix with and cool the
primary zone products. It was assumed that at least 98% combustion efficiency
is maintained in the primary zone. This corresponds to an emissions level
at the catalyst inlet of approximately 600 Vppm to 1200 Vppm as carbon over
402
-------
the operational range of (overall) fuel/air ratios 0.008 to 0.0155.
Catalytic conversion of 95% of the remaining hydrocarbons and 80% of the
carbon monoxide would then yield the required emissions goals. Thus, the
catalyst will be operated in a "clean-up" mode in which emissions conversion
will be predominantly catalytic with any contribution due to homogeneous
reaction dependent on the operating temperature.
The initial phase of the laboratory test program was designed to allow
evaluation of purely catalytic performance of various CATCOM catalysts
(described in Table II and shown in Figure 2) by selecting test conditions
which precluded homogeneous combustion. The basis for evaluating catalyst
performance was the trade-off between catalyst conversion-length requirements
and pressure loss-length restrictions. The catalyst configuration chosen
for this test series were all essentially identical Engelhard proprietary
preparations of palladium on zircon composite monolithic catalyst and
differed only in cell density of the honeycomb support. These were 6, 14
and 39 holes per square cm (see Table II). Catalyzed honeycomb lengths were
selected based on equal geometric surface area, with one additional length
of catalyst tested to provide additional information for prediction of
catalyst conversion. These screening test conditions (listed in Table III)
were sufficiently comprehensive to allow comparative emissions performance
and pressure drop criteria to be established and to allow a preferred
catalyst-length configuration to be selected.
* CATCOM is a trade name of Engelhard Minerals and Chemicals Corporation.
403
-------
An additional series of tests was carried out with the preferred
catalyst configuration to determine the enhancement in emissions clean-up
capability with simultaneous catalytic and homogeneous combustion
contribution.
II. APPARATUS AND PROCEDURES
Configuration A-2 was simulated using a two-stage catalytic reactor
as shown in Figure 3. This reactor is a 2.54 cm Inconel 601 pipe and
contains the following sections:
A. A primary stage in which preheated air and Jet A fuel are
combusted using a catalyst operated adiabatically under
catalytically supported thermal combustion conditions to provide
hot, vitiated air with negligible emissions.
B. A mixing section (2.54 cm diameter, 2.54 cm long) in which
additional preheated air is added to obtain a desired total
mass flow and mixture temperature. Seed fuels CO and C^^,
premixed with this additional (secondary) air, are added to
simulate incomplete combustion in the primary zone of the
combustor.
C. A secondary stage containing the test catalysts for which
emissions clean-up capability is to be determined.
The primary flows to the reactor were metered using a calibrated
rotameter for the air and a calibrated Crane double-acting piston pump
for the fuel. The primary and secondary air were separately preheated
using two electric furnaces. The secondary flows (air, CO and C-jHg) were
404
-------
metered using calibrated rotameters. Mixing of all feed gases was
accomplished by passing each feed through three stages of static mixers.
The actual concentration of the seed gases (CO and C^R.^) in the secondary
feed was also monitored by passing samples through the carbon monoxide and
hydrocarbon analyzers.
The catalysts (primary and secondary) were mounted in thin-walled
Inconel 601 holders which were inserted into the vertical tubular reactor.
The reactor was operated downflow.
Thermocouple arrays and pressure taps were located upstream of the
primary catalyst, in the mixing section and downstream of the test
(secondary) catalyst. Pressure tap installations conformed to ASME codes.
A 0.64 cm diameter water-cooled sampling probe was mounted 10.1 cm
downstream of the test catalyst. Probe placement was set at this distance
to prevent severe cooling of the test catalyst.
The instrumentation system used for emissions measurement included
a Beckman Model 402 heated oven flame ionization detector (F.I.D.) for
measurement of unburned hydrocarbons (UHC), a Beckman Model 315 BL
non-dispersive infrared analyzer for measurement of carbon monoxide (CO)
and a Beckman Model 951 chemi-luminescence analyzer for measurement of
nitrogen oxides (NO/NOX).
405
-------
Preliminary Checkout Tests
A. Pressure Losses
Pressure losses were measured via the static taps and manometers.
Preliminary tests were carried out which confirmed that pressure
drop measurements accurately reflected test catalyst performance
and were not affected by the method of secondary feed addition.
B. Mixing Tests
Tests were also carried out which confirmed that the secondary air
and seed fuel mixtures were well mixed with the hot, vitiated
exhaust from the primary stage. In these tests a blank (uncatalyzed)
honeycomb test piece was installed in the secondary catalyst holder
to form the downstream boundary of the mixing region, and the
unreacted species downstream of the blank were sampled. Test
measurements were uniform and non-fluctuating.
C. Preburning Tests
Preliminary tests were carried out which defined operating
mixture temperature limits below which no homogeneous combustion
occurred in the mixing section or secondary (test) catalyst
section. These tests were carried out with the honeycomb blank
in the secondary catalyst holder. Up to 930°C, no homogeneous
reaction was observed.
406
-------
III. RESULTS AND DISCUSSION
The objective of this experimental program was to obtain pressure
loss and emissions clean-up performance data on the alternative test
catalysts configurations, and provide a basis for determining the preferred
catalyst design. The selection of a preferred catalyst configuration
involves determining the trade-off between having sufficient catalyst
length for emission clean-up, and yet not exceeding the allowable pressure
loss limits. Because of the very broad range of operating temperatures of
the clean-up, catalysts (700-1300°C), it was anticipated that catalytic
reactions would predominate at the lower temperatures, while gas phase
homogeneous reactions would be significant at the higher temperature. In
order to uncouple these two effects and evaluate the catalytic clean-up
performance of each configuration, a two part experimental program was
devised.
A. A screening test program conducted with the four catalyst
configurations over a sufficiently broad range of operating
conditions to allow correlations to be developed to predict both
catalytic clean-up performance and pressure losses as a function
of catalyst length. The operating temperatures and seed gas
concentrations for these runs were selected to minimize the
contribution of homogeneous gas phase reactions.
B. A brief parametric study conducted with the preferred catalyst
configuration, selected from the performance correlations derived
from the screen test program, at higher operating temperatures
(see Table V) to measure the effects (and benefits) of homogeneous
gas phase reaction on emissions clean-up.
407
-------
Screen test data are shown in Tables V through VIII.
Catalytic Clean-Up
Previously reported results (Reference 3) have indicated that at the
operating temperature range of the screening test program (540-760°C), it
is highly probable that the catalytic oxidation reaction rate is mass
transfer controlled. Dimensionless mass transfer numbers were derived
from the screen test data on CO and C3H6 conversion to test this hypotheses
and derive design correlations.
Since the temperature rise across the catalyst configurations is
negligible for these conditions, a simplistic mass transfer model was
derived.
Conversion (X) - 1. - Xf>out/Xf in (1)
1. - expi_21±I±±l_ (2)
'LvDR
where Xf is the seed fuel concentration and subscripts
in and out denote catalyst inlet and exit. All
other terms are defined in the nomenclature
section.
Physical properties are evaluated at the average temperature
in the catalyst. Diffusivities of CO and C3Hg were estimated
using standard formulas (Reference 5).
408
-------
Calculated values of the dimensionless mass transfer and
flow parameters can be used to develop a plot of
Log
NSh
-56
i
vs. log; NRe H
where Ngc is the dimensionless Schmidt number, and
is the dimensionless Reynold's number.
This correlating technique was developed by Votruba (Reference 6) for
mass transfer in honeycomb catalysts. Data that is in the kinetically
controlled regime will not be correlated by this technique because of
the strong effect of operating temperature in the Arrhenius rate expression.
Figure 4 shows the experimental results that gave a linear correlation on
a log-log plot. Additional proof of mass transfer control is the excellent
agreement on this plot for all four different catalyzed honeycombs.
Mass transfer control performance allows extrapolation to other
configurations using the following mass transfer scaling criteria derived
from Equation 2:
L! In (1 - X1)V1(NshDf_a± )2
- .
L2 In (1 - X2)V2(NShDf_air)1
Pressure Drop
Since the tests were essentially isothermal, the pressure losses (AP)
at each test condition were reduced to dimensionless isothermal (apparent)
friction factors following standard friction factor correlation procedures.
The free stream face velocity is utilized to calculate the channel velocity
from Vch'(a) = V, and therefore
409
-------
F* _ H. _ , .
F 2 PV o L 2 VV7
ch
where p is the density and a is the percent open area of the test catalyst.
The apparent friction factor includes entrance and exit flow momentum
losses, as well as frictional losses, for both developing and fully developed
flow, and hence is higher than the conventional pipe flow friction factor.
The data for the four catalysts are plotted in Figure 5, with the pipe flow
equation provided for reference.
•\
Note that the friction factor for the 6 holes/cnr catalyst C is far in
excess of the reference value. The explanation for this anamolous value has
not been resolved.
Catalyst Comparison
The pressure drop and mass transfer correlations provided as Figures 4
and 5 were utilized to estimate the performance of the catalyst configurations
at the idle mode design condition of
T = 890°C
P = 300 kPa
V = 35 m/sec
The additional constraint of a maximum length of 9.5 cm narrowed the choice
to the selection of the channel size that most closely achieved the pressure
drop and conversion requirements.
The result of this analysis is shown as Figure 6, where for 9.5 cm
of each catalyst, the pressure drop and mass transfer conversion are
calculated. These curves indicate that at 35. m/sec catalyst A (the
39 hole/cm2 monolith) will yield
410
-------
CO = 80% conversion
UHC - 72% conversion
%AP/P -3.4% conversion
Thus, at this or lower flow velocities, the catalyst only slightly exceeds
pressure drop constraints. CO emissions goals can be met, but UHC objectives
cannot be attained by mass transfer reaction alone. Since the design
temperature is close to a level where thermal reactions become significant,
gas phase reactions are anticipated to increase hydrocarbon conversion to the
design objective.
Lower cell density catalysts were well within the pressure drop
constraint but did not approach the mass transfer conversion goals for
the program. Based on these results, the 39 hole/cm2 catalyst cell density
was selected.
Design Point Tests
A series of runs was carried out with the selected catalyst design at
conditions simulating the idle mode design point. Note that the operating
pressure was reduced from 300 kPa to 200 kPa due to rig limitations.
Test results are given in Table IX. Conversion performance with two seed
fuel concentration levels, 600 Vppm and 1200 Vppm carbon, were measured at
each test condition.
£Q The data indicate that the CO performance goal was achieved
at 910°C and higher. Comparison with calculations indicates
little or no gas phase reaction contribution.
411
-------
The performance goal of 95% conversion was approached at a
temperature of 910°C. Comparison with calculations shows that
homogeneous reaction provided a significant contribution to the
conversion at 910°C. Some conversion to CO was observed,
attributable to the gas phase reactions.
% Pressure Drop
Measured values were slightly above predicted values, due perhaps
to the lower operating pressure in the laboratory runs.
Further runs made in the test series evidenced a falling off in catalyst
activity. Subsequent examination indicates that some catalyst damage had
incurred, attributed to an overtemperature of the primary stage catalyst
during fuel/air ratio adjustment. Inspection of Table IX indicates the
malfunction occurred during run 5. The conclusions listed are primarily
drawn based on the screen test results and the first four runs.
IV. CONCLUSIONS
Laboratory tests were carried out in simulation of the use of a
catalyst reactor for jet engine emissions clean-up at idle mode conditions.
This design concept implies the use of a flame preburner, with 50% air bypass
around the catalyst. The test series was carried out with three different
honeycomb supports having 6, 14 and 39 holes per square cm and an identical
catalyst preparation. The tests were performed to develop design criteria
for use of a combustion catalyst in the "clean-up" mode. The following
summarizes the test results:
412
-------
The preferred catalyzed honeycomb configuration is 39 holes per
crn^ in the configuration 5.08 cm + .64 cm space + 3.8 cm
of Engelhard DXD-222, to be installed downstream of a 98%
efficient preburner.
CO emissions were below the emission index design goal of
10 g/Kg fuel at temperatures above 910°C in parametric tests
with the preferred catalyst.
emissions approached the emissions index design goal of
1 g/Kg fuel at a temperature of 910°C in parametric tests with
the preferred catalyst configuration.
- Estimated pressure losses are 3% at 30.5 m/sec at the design
point pressure of 3 atmospheres and 890°C temperature: design
goal was 35 m/sec.
- Homogeneous reaction contribution is necessary to meet Cj&fr
emissions goal.
- Yield of CO from homogeneous combustion of C-jHg was significant
and may affect emissions performance if C-jH, is truly
representative of UHC clean-up.
V. ACKNOWLEDGEMENT
This program was performed at Engelhard Industries R&D under NASA
Contract NAS3-20580 and subcontract to the General Electric Aircraft
Engine Group.
413
-------
VI. REFERENCES
1. Pfefferle, W. C., et al. "CATATHERMAL Combustion: A New Process
for Low-Emissions Fuel Conversion", ASME Paper Number 75-WA/FU-l,
Nov. 1975.
2. Anderson, D. N., Tacina, R. R. and Mroz, T. S., "Performance of a
Catalytic Reactor at Simulated Gas Turbine Operating Conditions",
NASA-Lewis Research Center Technical Memorandum NASA TMX-71747
(June 1975).
3. Blazowski, W. S. and Walsh, D. E., "Catalytic Combustion: An
Important Consideration for Future Applications", Combustion Science
and Technology. 10, pp. 233-244 (1975).
4. Mularz, E. J., Gleason, C. C. and Dodds, W. J., "Combustor Concepts
for Aircraft Gas Turbine Low-Power Emissions Reduction", Fourteenth
Propulsion Conference co-sponsored by the American Institute of
Aeronautics and Astronautics and the Society of Automotive Engineers,
NASA TM-78875, July 1978.
5. Bird, R. B., Stewart, W. E. and Lightfoot, E. N., "Transport
Phenomena", John Wiley & Sons, Int., 1960.
6. Votruba, J., "Heat and Mass Transfer in Honeycomb Catalysts - II",
Chem. Eng. Sci., 30_, 1974.
7. Hawthorn, R. D., "Afterburner Catalysts - Effects of Heat and Mass
Transfer Between Gas and Catalyst Surface". A.I. Chem. E. Dallas
Meeting, February 1973.
414
-------
TABLE I. CLEANUP CATALYST OPERATING CONDITIONS
FOR TOO OPERATING MODES
Operating Range Fuel Concentration
Operating P V T CO UHC
Mode (kPa) (m/s) (OG) Vppm Vppm C
Full Flow
Combustor 200-400 46-122 430-820 Up to 2000 Up to 1000
(A-l)
50% Bypass
Combustor 200-400 18-61 700-1290 Up to 1200 Up to 1200
(A-2)
^ ^Assumes conventional combustor operating at 98% efficiency
upstream of catalytic converter
415
-------
TABLE II. CATALYST DESCRIPTION
CELL DENSITY LENGTH HYDRAULIC
CODE NO. HOLES/CM2 CELL SHAPE CM DlAflETER CM
A. DXD-222 39 SINE WAVE 7.62 .0975
B. DXD-221 14 TRAPEZOID 11,43 .1722
C. DXD-225 6 SINE WAVE 15,24 .2997
D. DXD-221 14 TRAPEZOID 15.24 .1722
SUPPORT: ZIRCON COMPOSITE
CATALYST: PALLADIUM ON STABILIZED ALUMINA
PER CENT
OPEN AREA
.655
.542
,733
.542
416
-------
TABLE III. SCREENING TESTS FOR SELECTION OF
PREFERRED CATALYST CORE: CLEAN-UP MODE
BASIS:
Low pressure drop
High conversion of first stage emissions
TEST CONDITIONS:
Pressure:
Temperature:
Velocity:
Fuels added:
200 kPa
540, 760eC
24, 30, 49, 60 m/sec
200/400 Vppm CO
200/400 Vppm
TABLE IV. PARAMETRIC TESTS
RUN PRESSURE TEMPERATURE VELOCITY SEED 1
# kPa °C m/sec. Vppm CO
SEED @
Vppm
1
2
3
200
980
870
760
36.6 600 & 1200 200 & 400
417
-------
TABLE V. SCREENING TEST RESULTS: CATALYST A
Conditions
Results
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
P
atm
2.
"
"
••
11
"
M
"
"
»
"
II
'•
II
II
I
T
°F
617
1427
"
501
10CO
"
612.
K36.
"
501
107G
"
619
1414
U
U
fWsec
53.
89.
"
92.
126.
"
62.
98.
"
99
129
"
57.
93.
"
T
1.2
2.52
"
3.1
4.24
n
1.61
3.33
"
3.67
5.55
"
1.46
2.07
"
xco
" (1}
.75f2\
.77
_
.19
.21
_
.59
.65
-
.17
.22
.
.67
.71
XUHC
—
.79,,v
,77Z1
.
.56
.56
.
•
-
*»
.46
.51
-
.75
.73
(1) Fuel feed concentration 200 Vppm C
(2) Fuel feed concentration 400 Vppm C
418
-------
TABLE VI. SCREENING TEST RESULTS: CATALYST B
Conditions
Results
Run
1
2
3
4
5
6
7
8
9
10
11
12
P
atm
2.
11
"
»
"
"
11
11
11
T
«p
635
14)8
"
572
1022
"
617
1436.
ii
572
1126.
"
U
ft./sec
48.
82.
"
81.
114
"
54.
S3.
"
74
113.
n
V
.95
1 .82
"
2.20
3.6
"
1,11
2.21
"
1.92
3.40
"
XCO
.66,}
.72m
*»
,20
.31
m
,71
.75
^
.33
.4
*UHC
.68, ?
,74m
—
.55
.53
„
.75
.77
—
.52
.53
(1) Fuel feed concentration 200 Vppm C
(2) Fuel feed concentration 400 Vppm C
419
-------
TABLE VII. SCREENING TEST RESULTS: CATALYST C
Conditions
lesults
Run
1
2
3
4
5
6
7
8
9
10
11
12
P
atm
2.
n
,.
"
"
.,
"
11
.,
it
"
T
°F
617.
1400
581.
1031
M
635
1418
599
1162
U
ft./^ec
50
C5.
65
119.
56.
96.
79.
119.
11
% AP
— P
.98
2.06
2.46
3.75
"
1.15
2.28
"
2.06
3.5
XCO
• m
.68
.
.21
.24
-
.72
.73
-
.43
.43
XUHC
- m
(2\
,66f Z1
-
.51
.51
-
.78
.76
-
.48
.48
Fuel feed concentration 200 Vppm C
Fuel feed concentration 400 Vppm C
420
-------
TABLE VIII. SCREENING TEST RESULTS: CATALYST D
Conditions
Results
Run
t
1
2
3
4
5
6
7
C
9
10
11
12
P
JttTl
2.
n
"
..
"
"
»
"
"
..
"
"
T
op
626
1391
M
599
1112
M
626
1396.
n
617.
1238*
it
U
ft. /sec
47
81
"
84
120.
"
53.
89.
"
76.
117.
II
"T
1.16
2.33
4.03
11
1.39
2.89
it
2.45
4.31
*co
~ m
.7
.12
.21
.
.66
.74
.36
.44
XUHC
x«
.69
ff>
.56 '
.54
—
.74
.73 '
v>
.5
.55
rn Fuel feed concentration 200 Vppm C
Fuel feed concentration 400 Vppm C
421
-------
TABLE IX. PARAMETRIC TEST RESULTS: 3.5 INCHES OF DXB-222
Conditions
Results
Run
#
1
2
3
4
5
6
7
8
10
11
12
13
14t
p
& *
"
n
»
"
"
it
"
n
it
T
°F
1670
"
1760
«
1598
1463
"
1603
ii
1612
"
1769
U
ft. /sec.
112.
il
118
n
120
M
125
"
124
M
95
"
89.
n
X AP
4.66
3.76
..
3.19
"
3.13
"
4.56
U
2.93
n
2.66
ii
XCO
rn
.89m
.77
.80
.45
.50
.31
.44
.47
„
.41
.49
.54
.65
XUHC
M
^
.88 rn
.95^
.64
.74
.48
.54
.46
.56
.74
"
~t Catalyst repacked due to low % AP/P on tests 5 and 7.
r{r Post-test inspection showed catalyst inlet damaged.
Fuel feed concentration 600 Vppm C
" 1200 Vppm C
" "
422
-------
ALTERNATIVE OPERATING MODES
A-l FULL CATALYST FLOW
HlGh
AIR
"!
FUEL
A-2 A
AIR
If
— fe
K
''- ^= HlGF
\x o- — — — ^— •
VA\ ==, Low
?v BY
Low
PASS
f
CV/, Z^fT 1
i\ -=-~T= r
l\ rz^r 1
ELOCITY U
HIGH PRESSURE Loss AP
Low TEMPERATURE T
FICIENCY n
FUEL
MEDIUM
MEDIUM
MID TO HIGH
HIGH
U
A
T
n
Figure 1. Idle Mode Combustor Configurations
423
-------
Catalyst
o
4 holes/c/n.
Catalyst "B"
holes/cm.2
Catalyst "A"
holes/c/n.
Figure 2. Catalyzed Honeycomb Types Tested
424
-------
AIR
JET A FUEL
A!R
PREHEATER
T/C
M_ nnriirATrn
r *^ PRLHLAILK
TO GAS --
ALYSIS TRAIN "*
PRODUCT
rnni TMH -«
LUULIWU ^
SECTION
/
s:-^-:
•^
?wwv
,-. \\i[
PRIMARY
. STAGE
/ CATALYSi'
MIXING
^ SECTION
~ T/C
"" SECONDARY
STAGE
CATALYST
—TWO-STAGE
REACTOR
T/C
THERMOCOUPLE
Figure 3. High Temperature Two-Stage Reactor
425
-------
100
NSc.56
10
CO
o
u
o
A
C3H6
CATALYST A
CATALYST B
CATALYST C
CATALYST D
HIGH SURFACE CATALYST
HAWTHORN
(JJEFERENCE 7)
VOTRU
(REFERENCE 6)
MONOLITH
10
DH
100
300
Figure'4. Mass Transfer Conversion vs. Reduced Reynolds' Number
-------
to
0,1 r-
0,01
0,001
102
O CATALYST A
O CATALYST B
O CATALYST C
A CATALYST D
.051
f
NRE
APPARENT FRICTION FACTOR
FRICTION FACTOR (SINUSOID CHANNELS)
MONOLITH CHANNEL REYNOLDS' NUMBER
103
Figure 5. Friction Factor Correlation
-------
90
80
270
GO
o
CJ
60
50
CO UHC
0 + 39 HOLES/CM2
A A 14 HOLES/CM2
El • 6 HOLES/CM2
01234
% PRESSURE LOSS
Figure 6. Pressure Drop vs. Mass Transfer Conversion
428
-------
EXPERIMENTAL EVALUATION OF CATALYTIC COMBUSTION WITH
HEAT REMOVAL AT NEAR STOICHIOMETRIC CONDITIONS
By:
Daniel Bulzan
National Aeronautics and Space Administration
Lewis Research Center
Cleveland, Ohio 44135
ABSTRACT
Two concentric tube configurations were tested to evaluate
catalytic combustion with heat removal at near stoichiometric
conditions. Tests were conducted at one pressure, 1.5 x
10 Pa, inlet fuel-air mixture temperatures from 780 to 960K,
combustion air flowrates from .78 to 1.5g/sec, equivalence
ratios up to .87, and a range of cooling air flowrates.
Propane and Propylene fuels were used. Both configurations
utilized air flowing through the center tube for cooling and
combustion occuring in the annulus on the catalytic surface.
One configuration had the catalyst applied to the outside diam-
eter of the inner tube for a length of 56 cm. The outside
diameters of the inside and outside tubes were 1.27 and 2.54 cm
respectively. At an inlet mixture temperature of 960K, a com-
bustion air flowrate of 1.5 g/sec, and a fuel flowrate of 6.0 x
_o
10 g/sec, 34% of the heat available from the propylene fuel
was transferred to the cooling air using a Pt catalytic tube.
Conversion of the fuel was very low for this configuration.
The second configuration had the catalyst applied to the inside
diameter of the outer tube for a length of 56 cm. The outside
429
-------
diameters of the inside and outside tubes were .95 and 1.59 cm
respectively. At an inlet mixture temperature of 925K, a com-
bustion airflow of 1.0 g/sec, and a fuel flowrate of 6.0 x
10 g/sec, 58% of the heat available from the propylene fuel
was transferred to the cooling air with a Pt/Rh catalyst. At
equivalence ratios as high as .87, NO emissions less than
A
.2g NO /kg fuel were obtained. Combustion efficiencies
A
greater than 99.5% were measured with this configuration for
some operating conditions.
430
-------
SESSION IV
ALTERNATE FUELS AND THE
CATALYTIC COMBUSTOR
431
-------
FUEL NITROGEN CONVERSION - THE IMPACT OF
CATALYST TYPE
By:
B. A. Folsom, W. D. Clark, C. W. Courtney, M. P. Heap, W. R. Seeker
Energy & Environmental Research Corporation
8001 Irvine Blvd., Santa Ana, CA 92705
ABSTRACT
Data has been presented comparing the performance of two bench scale
catalytic reactors burning simulated LEG doped with ammonia. The performance
criterion of interest was the conversion, or retention of, NH» to other nitrogen
compounds under both fuel-rich and fuel-lean conditions. The two reactors
tested were an alumina-supported platinum catalyst and a zirconia-supported
nickel oxide catalyst. The major difference in their performance was the
concentration of total fixed nitrogen (TFN equals NH- + HCN + NO) measured in
fuel-rich products as a function of percent theoretical air.
432
-------
LIST OF FIGURES
Figure 1 Experimental Apparatus
Figure 2 Catalytic Reactor and Housing
Figure 3 Catalytic Reactor Test Apparatus
Figure 4 NH., Processing in LEG Platinum Catalytic Reactor
Figure 5 NH3 Processing in LBG-Fired Nickel Oxide Catalytic Reactor
Figure 6 Unstaged Nickel Oxide Catalytic Reactor with 500 ppm NH_
Figure 7 Staged Nickel Oxide Catalytic Reactor with 500 ppm NH»
Figure 8 Summary: Nickel Oxide Catalytic Reactor with 500 ppm NH3
Figure 9 Summary: Nickel Oxide Catalytic Reactor with 2400 ppm NH
433
-------
LIST OF TABLES
Table I Nickel Oxide Catalytic Reactor Constant Test Conditions
434
-------
ACKNOWLEDGMENT
The investigations described in this paper were carried out under EPA
contract 68-02-2196. The authors wish to express their appreciation to
Mr. G. B. Martin, the EPA Project Officer, and to Dr. J. Kesselring of the
Acurex Corporation for his help in obtaining the catalysts.
435
-------
SECTION 1
INTRODUCTION
The integrated gasifier LEG - fired gas turbine - steam turbine cycle is
one of several advanced coal based power cycles currently being developed. The
primary advantage of this cycle is the potential for low emission of sulfur
products without the economic and efficiency penalties associated with stack
gas sulfur removal. Since the coal gasification process operates fuel-rich,
the sulfur in the coal is converted mainly to hydrogen sulfide (H_S) in the
low Btu offgas which can be removed (potentially) more easily than sulfur
dioxide (S0_). During the gasification process, some of the nitrogen in the
coal is converted into ammonia (NH_) in the LEG. Under typical gas turbine
combustor operating conditions a large fraction of the NH_ may be converted to
NO resulting in significant emission.
The concentration of NH~ in the LEG depends upon gasifier design and
operating parameters and may be as high as 0.38 percent (1). The gas cleanup
system, which functions primarily to remove sulfur products and particulates
from the LEG, may remove a portion of NH.,, but the remainder will enter the
gas turbine combustor and may be oxidized to nitrogen oxides (NO ). A concen-
X
tration of 0.38 percent NH_ in the LEG would produce approximately 1370 ng/J
f j
(3.2 Ib N02/10 Btu) if all NH~ is converted to N0_. At present there are no
New Source Performance Standards (NSPS) for LEG fired combined cycle power
systems. However, it is reasonable to expect that when NSPS are promulgated
they will be at least as stringent as current NSPS for other power systems such
as gas turbines or gas fired steam generators. The current NSPS for gaseous
fossil fuel-fired steam generators with greater than 73.3 MW (250 x 10 Btu/hr)
heat input is 86 ng/J (0.2 Ib N02/10 Btu) (2). For a combined cycle firing
an LEG with 0.38 percent NH, to meet this emission level, the overall conversion
436
-------
of NH» to NO would have to be less than 6.3 percent.
This paper describes some of the work being carried out under EPA Contract
68-02-2196 to develop low NO emission combustor concepts for LEG containing
ammonia. Comparative tests have been carried out with two catalysts to assess
their suitability for the primary stage of a two stage combustor to minimize
fuel nitrogen conversion.
437
-------
SECTION 2
EXPERIMENTAL SYSTEMS
The experimental systems used in this investigation can be conveniently
divided into the following subsystems:
• Reactant Metering and Supply
• Modular Reactors
• Analytical Train
A schematic of the total system is shown in Figure 1. Although several different
reactor systems have been tested, only those involving catalyst beds will be
described below.
REACTANT METERING AND SUPPLY
The LEG was synthesized by blending together high purity gases from
cylinders. All gases were high purity grade (99.97 percent or better) with the
exception of CO (99.0 percent). NH», NO and HCN were supplied as custom grade
mixtures in nitrogen (± 2 percent accuracy) to facilitate metering. The
oxidant was dry air. All gases were metered with sapphire jewel orifices
operated in the critical flow (sonic) regime. The pressures upstream of the
orifices were measured with high accuracy variable capacitance pressure trans-
ducers which were calibrated periodically against a laboratory reference. The
pressures downstream of the orifices were maintained constant and the flow rate
through each orifice was calibrated by filling an evacuated tank. The estimated
total inaccuracy in each gas flow rate was 0.5 percent.
Variations in water vapor content of the LBG were achieved by metering
distilled water with a calibrated rotameter and prevaporizing before mixing
with the other reactant gases. The mixture was maintained well above the dewpoint
438
-------
temperature to prevent subsequent condensation. Other design details of the
reactant flow system are discussed in Reference 3.
Catalytic Reactors
Two catalytic reactor systems were used during the investigation. The
initial investigations were carried out with the system shown in Figure 2.
Later work was carried out with the series staged configuration shown in
Figure 4.
The premixed catalytically supported reactor is shown in Figure 2. The
two catalysts used in these tests were an alumina-supported platinum and a
zirconia-supported nickel oxide catalyst with some platinum on the upstream
cells to promote light-off; both were supplied by the Acurex Corporation. The
desirable features of the graded cell catalyst have been discussed previously
(4). The maximum recommended temperature for this catalyst is 1588 K (2400 F).
Type K thermocouples were cemented in two of the cells in the downstream
segment to monitor maximum monolith temperature. To minimize heat losses and
approximate adiabatic conditions, the catalyst was mounted in a refractory tube
and electrically backheated. The reactants passed through a sintered stainless
steel disc immediately upstream of the catalyst and combustion products were
sampled with a stainless steel water cooled probe. Additional details of this
reactor's construction and operation are documented in Reference 5.
The series staged configuration is shown in Figure 3; it is similar to the
earlier reactor except that:
the sintered stainless steel flashback arrester was moved about two
inches upstream,
a reactant sample tap was installed to permit the NH,, concentration
in the reactants to be measured immediately upstream of the catalyst,
a small bead (% 0.007 in dia) type K thermocouple was installed
immediately upstream of the catalyst to measure reactant preheat
and sense flashback,
a shorter (4.0 in.) refractory-lined tube was used to couple the
reactor housing to the mixer,
a secondary air injection system was added to allow burnout of the
rich primary zone products.
439
-------
Analytical Train
Both fuel-rich and lean products were analyzed for 02> CO, C02, NO, N0x>
NH and HCN. The sample train components were constructed entirely of stainless
steel, glass and Teflon. Ammonia and HCN were trapped by bubbling known volumes
of combustion products through three water baths in series. The absorbed NH3
and HCN concentrations were then measured by ion specific electrodes. The
other product species were monitored continuously. The sample was dried to a
dewpoint of 273 K in a cyclone water trap located close to the sample probe to
minimize residence time between probe and trap. Nitric oxide and NO were
3s.
measured with a Thermo Electron Corporation Model 10A chemiluminescent analyzer.
NO concentrations were measured after the sample had passed through a stainless
X
steel converter operated at 1073 K. This converter could not be operated under
oxygen-deficient conditions because of the well-known reduction of NO.
440
-------
SECTION 3
RESULTS
Previous papers have described the results obtained with the alumina-
supported platinum catalyst. Typical results illustrating the following
qualitative behavior are presented in Figure 4.
• For fuel-lean conditions the majority of the NH_ in the reactants
was converted to NO.
• For fuel-rich conditions NO, NH» and HCN were measured in the
combustion products in amounts depending on stoichiometry.
- NO decreased as the stoichiometry was reduced (more fuel-rich)
- NH~ and HCN increased as the stoichiometry was reduced
• A distinct minimum in total bound nitrogen species (ZXN = NO + NH»
+ HCN) occurred under fuel-rich conditions.
• For stoichiometries richer than that corresponding to the minimum,
ZXN increased sharply.
Typical results obtained with nickel oxide catalyst are shown in Figure 5.
Under fuel-lean conditions the results were similar to those obtained with the
platinum catalyst. However, under fuel-rich conditions the concentrations of
NO, NH.j and HCN were much lower than those measured with the platinum catalyst
and there was no distinct minimum ZXN. Under very fuel-rich conditions (44
percent T.A.), the ZXN for the nickel oxide catalyst was 2.5 percent of the
original NH3 in the reactants, while ZXN was essentially 100 percent for the
platinum catalyst. Tests of the nickel oxide catalyst, over a range of
adiabatic flame temperatures (1000-1400°C) and with CH, as the fuel yielded
similar results.
441
-------
The performance of the nickel oxide catalyst suggests that a simple staged
combustor could be constructed to burn nitrogen containing fuels. However,
because such a behavior had not been reported by others, further tests were
carried out to verify the repeatability of the results with the nickel oxide
catalyst. The series staged reactor was operated under the test conditions
listed in Table 1. It was fired with 500 and 2400 ppm of NH3 in the low Btu
gas and operated with and without secondary air injection (staged and unstaged).
During all of these tests the combustion products were sampled for Q^* co and
C00 to verify operating conditions, and for NO and NO or NO, NH» and HCN.
£. X J
For most tests the NH~ concentration in the reactants was also measured.
The test results involving 500 ppm in the low Btu gas are presented in
Figures 6, 7 and 8. The unstaged results in Figure 6 are similar to those
previously obtained with this catalyst (Figure 4) except that the ZXN levels
under fuel-rich conditions are slightly higher (7 to 18 percent versus 3 to 6
percent). There also appears to be a slight minimum at about 80 percent
theoretical air. The staged results are shown in Figure 7 and are consistent
with essentially full conversion of the EXN in the rich combustion products
to NO in the lean secondary combustion zone. Figure 8 compares the ZXN
X
measured in the reactants, and in the unstaged and staged products. The
reactant ZXN was all NH, and should have corresponded to 100 percent "conversion."
The differences are attributable to inaccuracies in reactant preparation and/or
NH.j concentration measurements. In any case, correction of the results to
account for the measured NH- concentration in the reactants would only slightly
lower the total ZXN (expressed as a fraction of the reactant NH_ concentration).
Test results involving 2,400 ppm of NH- in the LBG are summarized in
Figure 9 and are essentially similar to those obtained at lower ammonia
concentrations.
442
-------
SECTION 4
CONCLUDING REMARKS
Data has been presented comparing the performance of two bench scale
catalytic reactors burning simulated LEG doped with ammonia. The performance
criterion of interest was the conversion of NH» to NO or HCN or the retention
of NH,. under both fuel-rich and fuel-lean conditions. The two reactors tested
were an alumina-supported platinum catalyst and a zirconia-supported nickel oxide
catalyst. The major difference in their performance was the concentration of
total fixed nitrogen (TFN equals NH» + HCN + NO) measured in fuel-rich products
as a function of percent theoretical air. The platinum-based catalyst gave
almost total conversion of NH» to XN under fuel-lean conditions. A minimum
conversion of 35 percent was achieved at 70 percent theoretical air. The TFN
concentration increased as the primary air percentage was decreased further
giving almost complete conversion at 40 percent theoretical air. For fuel-lean
nixtures, the results with the nickel oxide catalyst were similar to those
obtained with the platinum catalyst. However, for fuel-rich mixtures, TFN
:oncentration drops to less than 10 percent conversion of the inlet ammonia as
;he reactant mixture approaches 80 percent theoretical air. This conversion
ippears to be independent of the stoichiometry of the primary zone and remains
.ow even for reactant mixtures approaching 40 percent theoretical air. Also,
:urther tests indicate that these conclusions are also valid for reactant mixtures
7ith higher inlet ammonia concentrations.
These preliminary results are particularly significant to the design of
staged catalytic combustors for nitrogen-containing fuels. If the primary zone
.s to be operated fuel-rich, two design objectives must be satisfied:
1) The inlet reactants should have a composition which will minimize
the TFN concentration of the first-stage reactants to reduce exhaust
NO concent rat ions.
x
443
-------
2) The temperature of the first-stage catalyst should be less than the
critical level for catalyst degeneration. If the heat-release in
the primary zone is too high, the combustor must be designed to allow
heat extraction from the primary zone. The need for heat extraction
will depend upon the stoichiometry of the primary reactants.
Based upon these investigations, the nickel oxide catalyst would allow the
first stage to operate very fuel-rich and therefore without heat extraction, and
yet minimize exhaust TFN concentrations. In addition, combustor operation is
made easier because there is no need for precise control of the inlet reactant
stoichiometry.
A scaled-up two-stage reactor using a nickel-based catalyst is currently
being tested at pressures up to ten atmospheres. Additional bench scale tests
are planned to investigate the influence of fuel nitrogen speciation.
444
-------
REFERENCES
1. Robson, F. L., Blecher, W. A. and Giramonti, A. J. Combined-Cycle Power
Systems. EPA-600/2-76-149, U. S. Environmental Protection Agency,
Washington, D. C., 1976. p. 359
2. Environmental Protection Agency Title 40, Chapter 1, Subcharter C,
Part 60-Standards of Performance for New Stationary Sources, Federal
Register, Vol. 36, No. 247, December 23, 1971.
3. Folsom, B. A. and Courtney, C. W. Chemiluminescent Measurement of
Nitric Oxide in Combustion Products. In Proceedings of the Third
Stationary Source Combustion Symposium, March 1977.
4. Kesselring, J. P., Krill, W. V. and Kendall, R. M. Design Criteria for
Stationary Source Catalytic Combustors. EPA-600/7-77-073c, U. S.
Environmental Protection Agency, Washington, D. C., July 1977. p. 193
5. Folsom, B. A., Courtney, C. W. and Heap, M. P. Environmental Aspects
of Low Btu Gas-Fired Catalytic Combustion. Proceedings of the Third
Workshop on Catalytic Combustion, October 3-4, 1978.
445
-------
TABLE 1. NICKEL OXIDE CATALYTIC REACTOR
CONSTANT TEST CONDITIONS.
LBG Composition (percent by volume)
H2
CO
CH,
20
20
5
55
Total Reactant Flow Rate
Heat Release Rate
Preheat Temperature
Adi aba tic Flame Temperature
(adjusted by N2 dilution)
Measured Reactor Temperatures
Space Velocity
Actual Reactant Velocity
93.5 scfh
3730 Btu/hr
200°C
1200°C
1000-1200°C
1.44 x 105/hr
10.0 ft/sec
446
-------
Exhaust
Stack
High Purity Gas
Cylinders
Cylinders
9999
Vent
Gas Bubbler
Train for Wet
Sample
8 Channel
Critical Orifice
Flow Metering
System
Zero and Span Gases
O
Preheater
500°C
Max.
Rotameters
u g
NO/NO..
Chem1lum1ntscent
Analyzer
*
CO
NDIR
y
co2
NDIR
"r
°2
Paramagnetic
* * *
^
\
1 Bypass and
4 Wet Sample
Flowrate
Measurement
i
Figure 1. Experimental apparatus.
-------
Exhaust
Stack
Monolith
T.C.
Steel
Shell"
ck Heat
Preheat
i.cr\
100 n
Sintered
S.S. Disc
T
44
mm
75
nm
25
mm
\
""""""1"""
Water-Cooled
S.S. Sample Probe
Graded Cell
Catalyst
Electrical
.Heating
Element
1500 w.
Ceramic Wool
Insulation
^Castable
Refractory Cylinder
Figure 2. Catalytic reactor and housing.
448
-------
.p-
.p-
Water-Cooled Mixer
Gas Metering System
Secondary Air Injection
In
Refractory
Insulation
Small Bead
Thermocouple
Electrica,!
Preheat
Reactor
Housing
Refractory
Lined Tube
Sintered
Stainless Steel
Disc
Reactant
Sample
Graded Cell
Catalytic
Reactor
Electrical
Back Heat
Combustion
Product
Sample
Figure 3. Catalytic reactor test apparatus.
-------
o
VI
u
-------
40
300
200
§.
a.
100
O NO
D NH3
A HCN
Owu «*•» un
LAN Or IWj(
NH3 In LBG • 500 ppm
Approximate Full
Conversion
pHmary
pr1mary
Figure 5. NH~ processing in LBG-fired nickel oxide catalytic reactor.
-------
Ul
N3
100
X
t
en
80
(U
3
60
40
20
40
100J! Conversion
r
pi
60 80 100 120
300
• \
200
X
OL
100
O NO
D NH3
A HCN
O «N or NOX
NH3 1n LOG • 500 ppni
Test Conditions: See Tablet
Approximate Full
Conversion
Figure 6. Unstaged nickel oxide catalytic reactor with 500 ppm NH_,
-------
Cn
100
80
CO
o 60
c
o>
I
w 40
20
100 % Conversion
40 60 80
*pr1mary
100
0.
I
I
I
120
300
200
X
g.
100
O NO
O NOX
NH3 1n LOG = 500 ppm
Test Conditions: See Table 1
40
60
Full Conversion
10
•a
is
80
*pr1mary
100
120
Figure 7. Staged nickel oxide catalytic reactor with 500 ppm NH .
-------
140
120
100
o 80
« 60
o
o
40
20
0
40
o
o
o
o
O
o
60 80 100 120
O Unstaged
O Staged
\/ Reactant
NHa In LOG = 500 ppm
Test Conditions: See Table 1
Figure 8. Summary: Nickel oxide catalytic reactor with 500 ppm NH».
-------
in
in
120
o
100
80-
*•»
o
40
20-
0
40
o
60
80
100 120
O Unstaged
O Staged
\/ Reactant
NHa in LBG • 2400 ppm
Data Taken: 12/79
Test Conditions: See Table I
Figure 9. Summary: Nickel oxide catalytic reactor with 2400 ppm NH,.
-------
CATALYTIC COMBUSTION CHARACTERISTICS
OF LOW AND MEDIUM BTU GAS FUELS
By:
H. C. Lee
I. T. Osgerby
Engelhard Industries Division
Engelhard Minerals and Chemicals Corporation
Menlo Park, Edison, New Jersey 08817
Presented at
Fourth EPA Workshop of Catalytic Combustion
Cincinnati, Ohio
May 14-15, 1980
Sponsor: U.S. Department of Energy
Power Systems Department
Contract No. EF-77-C-01-2683
W. W. Bunker, Contract Monitor
456
-------
ABSTRACT
A test program was carried out to determine the feasibility of
catalytically-supported thermal combustion (CATATHERMAL*) of low and
medium BTU gas fuels.
Combustion characteristics of CATCOM* catalysts were investigated
over the range of heating value 4.5 to 12.0 MJ/m3 (120 ^-320 BTU/f t3) ,
using simulated low and medium BTU gas fuels. The laboratory test
results showed that low and medium BTU gas can be catalytically burned
to accomplish high efficiency and essentially emissions-free combustion
with acceptable pressure losses at 30 m/sec reference velocity and 300°C
inlet temperature.
Efficiency for catalytically-supported thermal combustion is
significantly affected by operating temperature and reference velocity.
The maximum reference velocity for low emissions combustion was halved
when the catalyst inlet temperature was decreased to 150°C. The turn-
down capacity of the catalyst system at 30 m/sec reference velocity and
300°C inlet temperature is estimated to be around 1.4, provided the
maximum operating temperature for practical catalyst supports is limited
to 1500°C. The effect of pressure on combustion efficiency was neglig-
ible over the range (1 to 5 atm) covered in laboratory test. Combustion
efficiency was also shown to be a function of the ratio of hydrogen to
combustible components in low and medium BTU gas. The higher the pro-
portion of hydrogen, the better the combustion performance regardless of
total fuel heating values.
The upper limit of reference velocity in the region of operability
for the combustion of low and medium BTU gas was bounded by the pressure
loss requirement (< 5 percent) above an inlet temperature of 235°C and
by combustion efficiency below 235°C. The low limit boundary of
reference velocity was set by the requirement of flashback-free operation.
The catalyst performance at a given pressure loss was also shown to
be a function of physical properties of substrates, i.e., open area,
hydraulic diameter and cell shape.
* CATCOM and CATATHERMAL are tradenames of Engelhard Minerals & Chemicals
Corporation.
457
-------
NOMENCLATURE
a Numerical constant
DH Channel hydraulic diameter
J* Apparent friction factor
L Length of catalyst
p Pressure
AP Pressure drop
T Temperature
U Velocity
a Open fraction of substrate
^c Combustion efficiency based on carbon balance
Subscripts
I Isothermal
in Inlet
Ref. Reference (catalyst inlet-face)
T Combustion
458
-------
INTRODUCTION
Low and medium BTU gas obtained from gasification of coal is one
of the principle new fuel sources being investigated by industry and
the U.S. government. The commercial utilization of low and medium BTU
gaseous fuels in a combined cycle power plant requires the solution of
critical engineering problems in both the gasifier and turbine systems.
The laboratory-scale investigation into the catalytically-supported
combustion of low and medium BTU gas was carried out under contract to
the Department of Energy (Ref. 1) to evaluate the performance character-
istics of this combustion process. Potential applications are to
pressurized gasifiers, where advantage may be taken of the excellent
combustion stability of this process in the presence of fluctuations in
gasifier heating value.
Incentives for utilizing catalytically supported combustion include:
: high combustion efficiency
: excellent combustor stability
: negligible thermal NOX emissions
: combustion both inside and beyond flammability limits
: uniform temperature profiles
Carrubba et al (Ref. 2) reported the low emissions combustion of
H2/CO/N2 mixtures with heating values from 2 MJ/nr*, well below the con-
ventional flammability limits. Additional confirmation of these perfor-
mance characteristics was provided by DeCorso et al (Ref. 3). This
latter evaluation of catalytic reactor performance utilized a 5 inch dia-
meter pressurized static test rig. Exit gas temperature profiles were
459
-------
measured and reported to be as narrow as + 30*C at 1300°C catalyst exit
temperature.
The present study was directed toward establishing a technical data
base with respect to low and medium BTU gas combustion, flashback potential,
catalyst design, and low emissions operating ranges. All tests were
carried out utilizing proprietary Engelhard precious metal CATCOM'
catalysts evaluated in a one inch diameter test rig.
The program was divided into the following steps:
: Select five nominal fuel compositions (Table I) for the test
to represent typical low and medium BTU coal gas fuels and
define the test conditions to simulate gas turbine combustor
operation.
: Determine the boundary of flashback free operation at the maximum
preheat levels and select operating conditions without flash-
back for subsequent tests.
: Evaluate five different configurations varying reference
velocity, inlet temperature, pressure and fuel heating value,
and determine a preferred catalyst configuration to meet the
objectives of >99% combustion efficiency and <5% pressure loss
at 30 m/sec reference velocity and 300°C inlet temperature.
: Define the region of operability for constraints of combustion
efficiency, pressure loss and flashback-free operation with
the preferred configuration, and generate performance data
applicable to the design of a prototype combustor for combustion
of low and medium BTU gaseous fuels.
Engelhard Minerals and Chemicals Tradename
460
-------
EXPERIMENTAL FACILITIES
Experimental facilities used in the test program are described in
Reference 4, including fuel-air supply system, preheating system,
mixing system, reactor system, control instrumentation and analytical
system.
Figure 1 gives a schematic drawing of the flashback reactor. A
one inch diameter (nominal) test catalyst is loaded into a catalyst
holder with a flashback arrestor to prevent fuel-air mixture from pre-
igniting before entering the catalyst. The flashback arrestor consisted
of a sintered blank honeycomb and a refractory ceramic liner insert.
Partially open channels of catalyst core are completely blocked by high
temperature cemment and channeling of the preheated gas in the reactor
is prevented by Flberfrax 'packing around the catalyst core.
Air preheat temperature was maintained by indirect electric preheat
furnaces and reactor pressure was adjusted with a pressure regulator on
the exhaust line, downstream of a cooling jacket. Throughout the tests,
gas samples were taken with a water-cooled probe located 0.1016 m down-
stream of the catalyst core for measurement of emissions (CO, unburned
hydrocarbon, NO/NOX, 02 and C02 concentrations in the reactor effluent
stream). The sample probe location was evaluated during previous tests
to keep the distance as close as possible to the catalyst without un-
desirable cooling effects.
RESULTS AND DISCUSSION
Flashback Tests
Flashback tests were conducted to establish the region of operability
without flashback for configuration selection and parametric tests. The
flashback arrestor developed for low BTU gaseous fuel (<20% H2 concentration)
did not prevent flashback in fuel/air mixtures containing high hydrogen
concentration.
f Fiberfrax is a registered trademark of the Carborundum Company.
461
-------
2
Accordingly, various types of arrester were tried with 7.5 MJ/m
fuel (containing 40% H2) at the maximum preheating level of 300°C.
As shown in Figure 1, an uncatalyzed honeycomb (0.0127 m long, located
0.00635 m in front of catalyst with 15° slant angle to the flow) to-
gether with a refractory ceramic insert (placed just upstream of an
arrestor) prevented flashback at the following limiting conditions:
Pressure < 4 atmospheres
Inlet Temperature < 300°C
Reference Velocity > 15 m/sec
Heating Value » 4.5-12.0 MJ/m3
Preferred Catalyst Configuration Selection
A series of tests were carried out to select a preferred catalyst
configuration which could meet the objectives of low emissions and high
combustion efficiency (> 99%) at acceptable pressure losses (< 5%).
Previous tests (Ref. 4) demonstrated that combustion efficiency is
strongly affected by reference velocity and inlet temperature, but the
effect of pressure on combustion efficiency and pressure loss in turbulent
flow is negligible. Thus, configuration selection tests were performed
at a constant adiabatic flame temperature of 1260°C and varying reference
velocity by changing pressure at a constant mass flow rate. The fuel
o
with a medium BTU heating value of 12.0 MJ/m was selected for the tests
due to its lower hydrogen concentration to minimize flashback problems.
Five different catalyst configurations (A.2, C.2, E.2, G.2 and H.2
in Table II) with palladium on stabilized alumina on a Zircon composite
support were evaluated, using a test procedure developed in Reference 4,
Test results for combustion efficiency and pressure losses versus
reference velocity at 150 and 300°C inlet temperatures are presented in
Figures 2 through 4 and summarized in Table III.
462
-------
Combustion Efficiency
Figure 2 shows a linear relationship of the negative logarithm of
complement combustion efficiency vs. reference velocity at 300°C inlet
temperature. All catalyst configurations except catalyst A.2 yielded
combustion efficiencies greater than 99% at 30 m/s reference velocity.
A combustion efficiency >98% was obtained only from catalyst configura-
tions C.2 and E.2 when the inlet temperature was reduced to 150°C as shown
in Figure 3. A significant drop in combustion efficiency was noticed
especially with catalyst configuration H.2.
Catalyst configurations A.2 and H.2 were expected to have large entrance
region effects due to low open area (A.2) and small length to cell size
ratio (H.2) in each segment. Poor performances were attributed to high mass
transfer rates near the entrance of the catalyst.
Catalyst configurations C and E were selected on the basis of
combustion efficiency for low emissions operation over a wide range of
velocities and inlet temperatures.
NOX emissions were less than 2.5ppm for all catalysts tested under
all operating conditions.
Pressure Losses
Figure 4 shows isothermal pressure losses for configurations A.2,
C.2, E.2, G.2 and H.2, indicating a linear relationship of isothermal
2
pressure loss vs. (U Ref/T). The apparent friction factor is essentially
a constant in turbulent flow and the isothermal pressure loss is expressed
as follows:
463
-------
(APT/P) - a • (£>-z-f* (1)
(UzRef/T)
Table III gives the parameters of interest for each configuration,
2 L
indicating a wider variation of a than (p^)f* for each configuration.
The combustion pressure loss for each configuration is also presented
in Table III. Low isothermal pressure losses with a high ratio of
combustion to isothermal pressure loss represents good catalyst perfor-
mance (in a single-segment catalyst configuration) to attain low emissions
operation within acceptable pressure loss. Results for configurations
C.2 and E.2 show a high ratio of combustion to isothermal pressure loss,
however, configuration C.2 does not meet the pressure loss requirement
Note that the flame arrester accounted for 20-30% of the total
pressure drop.
The superiority of comfiguration E.2 is apparent over the entire
range of operating conditions as shown in Figures 2, 3, and 4. Its
superior performance is attributed to an optimum combination of geometric
properties, i.e., open area, hydraulic diameter and cell shape.
Configuration E.2 was also selected as the preferred catalyst
configuration in combustion of low BTU gaseous fuel (Ref. 4). The
preferred configuration, Configuration E.2 consists of:
0.1619 m refractory ceramic liner + 0.0127 m
uncatalyzed 14 channels/cm^ honeycomb (flashback arrester)
+ 0.00635 m space + 0.0762 m catalyzed 6 channels/cm^
honeycomb catalyst: palladium on stabilized
alumina on zircon composite honeycomb support.
464
-------
Catalyst Performance Tests
A parametric test program was formulated to determine the region
of operability for high combustion efficiency at acceptable pressure
losses with the preferred catalyst Configuration E.2.
Emissions and pressure losses were measured, varying reference
velocity, inlet temperature and fuel heating value. Total flow rates
were set to maintain high Reynolds numbers (> 5000) throughout the
characterization tests to preserve scalability to industrial gas turbine
operating conditions. A pressure of three atmospheres was selected
for the test but some runs were carried out to investigate the effect
of pressure on emissions.
Additional tests were carried out to examine effects of adiabatic
flame temperature and fuel components on catalyst performance.
Effect of Operational Variables
Reference Velocity; The effect of reference velocity on combustion
efficiency is pronounced and the plot of the negative logarithm of the
complement combustion efficiency versus reference velocity is linear at
high combustion efficiencies >95% over the velocity range 15 to 35 m/s.
The linear relationship is invalid at combustion efficiencies <95%, as
shown in Figures 5 and 6.
Inlet Temperature; Combustion efficiency was significantly reduced at
3
low inlet temperature, except for combustion of 7.5 MJ/m fuel. Combus-
tion efficiencies >99% at 150°C were obtained for a reference velocity
<20 m/s. At inlet temperatures below 150 C, low and medium BTU gas fuels
did not light-off near the entrance of the catalyst bed due to low kinetic
activity and the selected catalyst length was not sufficient for complete
465
-------
CATATHERMAL reactions. Minimum inlet temperature to obtain combustion
efficiency greater than 99% at a reference velocity of 30 m/sec is inter-
polated to be 200 and 240°C, in Figures 5 and 6, for combustion of low
and medium BTU fuels, respectively.
In order to examine the effect of inlet temperature, combustion
efficiency was measured at a constant space velocity - corresponding to
a reference velocity of 30 m/s at 150 C - and plotted versus the
3
reciprocal inlet temperature for 7.5 and 12.0 MJ/m fuels in Figure 7.
Combustion efficiency is strongly affected by both inlet temperature
and adiabatic flame temperature due to the exponential dependence of the
reaction rate constant on temperature. However, a weak temperature de-
o
pendence was observed with the 7.5 MJ/m heating value fuel as shown in
Figure 7.
A mechanism for catalytically supported thermal combustion is
postulated to comprise regions of kinetic control, mass transfer control
and catalytically-supported thermal reactions, and the temperature
dependency for reaction in each zone has a distinct characteristic
(Ref. 5) . The activation energy for kinetic control and catalytically-
supported thermal reaction regions are greater than 15 Kcal/gmol for H2
and CO fuels, but the temperature dependency is negligible in the mass
transfer region (Ref. 6, 7).
Thus a weak overall temperature dependency for 7.5 MJ/m3 fuel indicates
that kinetic effect is not important for the combustion with 7.5 MJ/m
fuel, while a kinetic reaction is controlling for 12.0 MJ/m3 in a range
of 150°C to 300°C inlet temperature. Since the 7.5 MJ/m3 fuel has
a higher ratio of hydrogen to carbon monoxide and the rate constant for
hydrogen is 15-20 times higher than for carbon monoxide in the kinetic
zone (Ref. 7), the fuel mixture easily lights-off even at 150°C inlet
temperature and combustion efficiency is not less than 99.4 percent over
the entire range of operating conditions covered in the characterization
tests.
466
-------
Pressure: The effect of pressure on combustion efficiency was not
discernible in the pressure range covered in this test program.
Fuel Composition/Heating Value
Hydrogen; Tests were conducted with fuel composition varied at constant
heating value (7.5 MJ/m ) and reference velocity (30 m/s.). Measured
combustion efficiencies declined as hydrogen concentration decreases.
A high combustion efficiency was obtained with a fuel having a high ratio
of hydrogen to mix gas (CO + CH/) as shown in Table IV and Figure 8.
3
Comparing test results obtained with 7.5 and 12.0 MJ/m fuel, no apparent
effect of fuel heating value on combustion efficiency was observed. A
similar result was obtained with low BTU fuels (Ref. 4).
CO + CH/,; From the test results shown in Figure 8, CO and City concentra-
tions in the fuel appears to have an adverse effect on the catalyst
performance.
In order to determine the reactivity of CO and City, the combustion
efficiency derived from test data obtained at a low inlet temperature
of 150°C was plotted in Figure 9. The combustion efficiency of City is
much higher than CO on the basis of complete combustion of City. This
may be due to:
: CO may be more difficult to oxidize due to the strong
adsorption of CO on active sites of the catalyst at low
temperatures (inlet zone).
: City is partially converted to CO and hence gives a misleading
result.
467
-------
Non^Cgmbustible Components (C09, N7 and Steam) : Additions of nitrogen
and steam to the fuel-air mixture resulted in lower combustion efficien-
cies. No significant difference in effect of nitrogen or steam on
combustion efficiency was observed except for lowering the fuel/air ratio
(dilution effect). Similar result was observed with low BTU fuels
(Ref. 4).
Effect of Adiabatic Flame Temperature
The effect of adiabatic flame temperature on catalyst performance
was investigated by changing the fuel/air ratio, to determine the turn-
down temperature for 99% combustion efficiency. Two series of tests
were carried out at inlet temperatures of 250 and 300°C with 12.0 MJ/m
fuel. Test results are shown in Figures 10 and 11.
Minimum temperature required for low emissions operation is inter-
polated to be around 1160°C at 30 m/sec reference velocity and 300°C inlet
temperature. This corresponds to a turndown ratio of =1.4 provided the
maximum operating temperature is limited to 1500 C. The minimum temper-
ature was 1220°C at an inlet temperature of 250°C as shown in Figure 11.
Comparison of Low and Medium BTU Test Results
Performance test results in Figures 5 and 6 indicate that the
o
combustion efficiency with low BTU fuel (4.5 - 6.75 MJ/m ) is slightly
o
higher than results obtained with medium fuel (12.0 MJ/m ). An excellent
o
combustion efficiency was always observed with 7.5 MJ/m fuel as shown
in Table V-A.
Even though the catalytic combustion of gas fuels are complicated
by various fractions of each component in the gas mixture, overall
combustion performance is generally affected by the ratio of hydrogen to
other combustible components, regardless of fuel heating value(Table V-B).
Note that the actual combustion efficiency of coal gas fuel is higher
than results reported in this test program on the basis CO and CH^ as
reactants (Hydrogen was not included).
468
-------
Region of Operability for Low and Medium BTU Fuels
The region of operability was defined for low and medium BTU
fuels as shown in Figure 12, considering the constraints of combustion
efficiency > 99%, pressure loss < 5% and flashback-free reactor
operation. The boundary for combustion efficiency was derived from the
extrapolation of parametric test results assuming a linear relationship
in the plot of the negative logarithm of the complement combustion
efficiency versus reference velocity.
As shown in Figure 12 for low BTU fuel, pressure loss is the
limiting factor on allowable reference velocities above 210 C inlet
temperature and combustion efficiency controls the region of operability
below 210°C.
The transition temperature for pressure loss and combustion
efficiency increases to 250°C for medium BTU fuel. Thus, the region
of operability for combustion of medium BTU fuel is slightly narrower
than the region for low BTU fuel. The low limit boundary of reference
velocity is set by the requirement of flashback-free reactor operation ,
an artifact of the Engelhard reactors and will be different for a
practical combustor.
CONCLUSIONS
From the results of laboratory tests for combustion of low and
medium BTU coal gas fuels, the following conclusions have been drawn:
1. Based on the laboratory test results, catalytically supported
combustion has been shown to be capable of combusting a wide
range of low and medium BTU fuel compositions with high com-
bustion efficiency, low emissions and acceptable pressure drop
at conditions simulating gas turbine combuster operation.
469
-------
2. Combustion efficiency of 99.8 percent and pressure losses
below 4.2 percent were achieved with the preferred CATCOM*
catalyst. Configuration E, i.e., palladium catalyst on
0.0762 m zircon composite honeycomb, at 30.5 m/sec reference
velocity and 300 C inlet temperature. Emissions were:
Carbon Monoxide 80-190 Vppm
Unburned Hydrocarbon (as C.j) 1 Vppm
Nitrogen Oxides 2.3 Vppm
3. NOX emissions were always < 2.5 Vppm. Thus, NOX formation
from thermal nitrogen fixation is negligible at an adiabatic
flame temperature of 1260°C.
4. Catalyst Configuration E.2 with the highest open area was
selected as the preferred configuration for high combustion
efficiency at acceptable pressure loss in the combustion of
low and medium BTU gaseous fuel.
5. Combustion efficiency was affected by reference velocity,
inlet temperature and adiabatic flame temperature, but not
affected by pressure or heating value.
o
6. Combustion performance with low BTU fuel (4.5-6.75 MJ/m , H2/
CO = 0.68 - 0.86) better than results obtained with medium
BTU (12.0 MJ/m3, H2/CO = 0.52) fuel but an excellent combustion
efficiency was always obtained with 7.5 MJ/m (H2/CO = 1.92)
fuel attributed to the much higher ratio of H£ to CO in the
fuel.
* CATCOM is a trade name of Engelhard Minerals & Chemicals Corp.
470
-------
3
7. Since combustion efficiency with 7.5 MJ/m fuel containing
a high ratio of t^/CO (1.92) was not sensitive to inlet
temperature, the surface reaction rate might not be critical,
however, the kinetic rate controlled the overall rate of
combustion of low and medium fuels with a low ratio of H2/CO
(<0.86).
8. Pressure losses were correlated with the dynamic momentum
divided by the absolute inlet temperature with a slope
approximately equal to unity on a logarithmic plot in the
range of channel Reynolds number greater than 5000. It
indicates that the apparent friction factor is constant in
turbulent flow.
9. Flashback was more severe with medium BTU fuel and was attributed
to the increase in H2 concentration.
471
-------
REFERENCES
1. Catalytically Supported Thermal Combustion of Coal Derived Low BTU
Gas -- D.O.E. Contract EF - 77 - C - 10 - 2683.
2. Carrubba, R. V., Chang, M., Pfefferle, W. C. and Polinski, L. M.,
"Catalytically Supported Thermal Combustion for Emissions Control" -
Electric Power Research Institute NOX Control Seminar, February 6, 1976.
3. Decorso, S. M., Mumford, S., Carrubba, R. V., and Heck, R., "Catalysts
for Gas Turbine Combustors — Experimental Test Results' - ASME
gas Turbine Division Conference, New Orleans, LA (March, 1976).
4. Osgerby, I. T., Carrubba, R. V., Heck, R. M., and Bunker, W. W.,
"Investigation of Process and System Design Variables for CATATHERMAL
Combustion of Low BTU Gas", ASME 79-GT-66, presented at the ASME
Gas Turbine Conference, San Diego, California, March 12-15, 1979.
5. Pfefferle, W. C., "Catalytically-Supported Thermal Combustion",
U.S. Patent 3,928,961 (1975).
6. Hlavacek, V., and Votruba, J., "Experimental Study of Multiple
Steady States in Adiabatic Catalytic Systems", Chemical Reaction
Engineering II, American Chemical Society, Washington, 545-558 (1974).
7. Hawthorn, R. D., "Afterburner Catalysts-Effects of Heat and Mass
Transfer Between Gas and Catalyst Surface". Shell Development
Co., Rept. p. 2121, 1972.
472
-------
TABLE I
SYNTHETIC
(dry)
GAS COMPOSITIONS
I
15.0
14.4
1.8
6.4
62.4
LOW BTU
II
20.5
15.6
2.4
8.7
52.8
FOR LOW
III
24.5
18.7
2.9
10.4
43.5
AND MEDIUM BTU FUELS
MEDIUM BTU
IV V
18.6 51.5
40.0 30.0
2.2 6.1
39.2 12.4
-
MIX:
H2
CH4
C02
RTTJ /SCF
Heating Value 120 150 180 200 320 ( 1 Atm
4.5 5.62 6.75 7.5 12.0 MJ/m3
473
-------
TABLE II
•e-
*-j
js
Catalystt
Configuration
A
B* *
C
D
E
F*
G
H**
CATALYST CONFIGURATIONS FOR LOW AND MEDIUM BTU GAS TESTS
GEOMETRIC PROPERTIES FOR SUBSTRATES
Length
(cm)
7.62
2X5.08
3.81
5.08
7.62
2X5.08
7.62
3X2.54
Channels
per cm^
14
6
40
14
6
6
17
9
Hydraulic
Diameter (cm)
0.1722
0.300
0.098
0.1722
0.300
0.300
0.1940
0.2930
Open
Area(%)
54.2
73.3
65.5
54.2
73.3
73.3
68.8
61.0
Bulk Surface
Area (cm^/cm )
12.6
9.8
26.8
12.6
9.8
9.8
14.2
8.8
* Over-temperature test (Ref. 1)
** Single-Cell multi-segment configuration
T Pd/A^O^ on Zircon Composite Support
-------
TABLE III
CONFIGURATION SELECTION FOR CONFIGURATIONS A, C. E. G. AND H
Pressure Losses
Configuration
Isothermal Combustion
Physical Parameter
(Eq'n 1)
(L/DH)f* I/a2
-p-
Ln
A.2
C.2
E.2
G.2
H.2
Combustion
Efficiency
(50
97.0
99.6
99.8
99.4
99.5
APlm+
P
at URef
2.5
2.2
1.3
1.4
1.5
APT(%)
P
= 30m /s and
7.8
6.6
4.2
5.0
5.0
APT
**I
Tin = 300°C
1.5
2.1
2.1
1.8
1.5
0.64
0.82
0.61
0.57
0.51
3.40
2.33
1.86
2.11
2.69
Note (+): Catalyst only (pressure loss through the flashback arrestor is excluded)
-------
TABLE IV
H2
40.0
35.0
30.0
25.0
40.0
EFFECT OF FUEL COMPOSITION
Catalyst
Inlet Temperature
Pressure
Reference Velocity
Fuel Heating Value
Fuel Composition (%)
CO CH4 N2 C02
18.6 2.2 39.2
21.8 2.6 40.6
25.0 3.0 42.0
28.3 3.3 43.4
18.6 2.2 - 39.2
(H2, CO, CH4, N2 and
Configuration E.2
225°C
200 kPa
30 m/sec
7.5 MJ/m3
Fuel*
Air
.2531
.2536
.2541
.2533
.2704
C02)
H2
CO+CH4
1.92
1.43
1.07
0.79
1.92
Combustion
Efficiency
.9971
.9959
.9896
<.9222
.9965
* For an adiabatic flame temperature 1260°C
476
-------
TABLE V
COMPARISON OF LOW AND MEDIUM BTU TEST RESULTS
A. COMBUSTION EFFICIENCY
Fuel Heating
Value
Configuration "A" Configuration "C" Configuration "E1
Tin=150°C 300°C
Tin=150°C 300°C Tin=150°C 300°C
6.75
(Low BTU)
7.5
(Medium BTU)
12.0
(Medium BTU)
<.90
<.90
.9967
.987 .9941
.9700
.9947
.9720 .9969
.9964
.9939 .9984
.8971 .9974
B. FUEL COMPOSITION
Fuel
Heating Value
MJ/m3
Fuel Fraction (By Volume)
CO
H2 + CO + CHA H2 + CO + CH4 CO + CH4
1. Low BTU Gas
4.5
0.46
0.48
0.86
5.62
6.75
0.41
0.41
0.53
0.53
0.68
0.68
2. Medium BTU Gas 7.5
0.66
0.31
1.92
12.0
0.34
0.59
0.52
477
-------
AIR
AND STEAM
REACTOR WALL-
INLET PIPE
REFRACTORY
LINER INSERT
PRESSURE TAPS
T/C THERMOCOUPLE
MIXING 2ONE
BLANK HCNEYCOMB
MANOMETER CONNECTION
T/C's FLASHBACK
SENSORS
/"FLASHBACK ARRESTOR
IBLANK HONEYCOMB
CATALYST HONEYCOMB
RETAINING PIN
GAS SAMPLE PFOBE
EXIT T/C ASSEMBLY
FIGURE 1, SCHEMATIC OF FLASHBACK REACTOR
478
-------
FIGURE 2 PLOT OF COMBUSTION EFFICIENCY VS.
REFERENCE VELOCITY AT 300°C INLET TEMPERATURE
TEST CONDITIONS
INLET TEMPERATURE : 300°C
PRESSURE : 100
ADIABATIC FLAME TEMPERATURE: 1260°C
FUEL HEATING VALUE : 12,0
. 9939 ——
,9995
,9990
tf —
CD
g ,9950
8 ,9990
.9500
,9000
,5000
10
V CONFIGURATION A,2
A CONFIGURATION C,2
O CONFIGURATION E,2
P CONFIGURATION 6.2
CONFIGURATION H,2
20 30
REFERENCE VELOCITY: H?vEp
o
CD
10
479
-------
FIGURE 3, PLOT OF COMBUSTION EFFICIENCY vs.
REFERENCE VELOCITY AT 150°C INLET TEMPERATURF
TEST CONDITIONS~~"
: 300°C
: 100-
INLET TEMPERATURE
PRESSURE
ADIABATIC FLAME TEMPERATURE: 1260°C
FUEL HEATING VALUE : 12,0
9999
o
- O
UJ
9995
9990
O
GO
9950
9900
,9500
,9000
,5000
V CO?IFIGURATIO:i A,2
A CONFIGURATION C,2
© CONFIGURATION E,2
CONFIGURATION 6,2
O CONFIGURATION 11,2
E,2
1
u
CD
. I
1
10
20 30 4
REFERENCE VELOCITY, UREF (M/SEC)
480
-------
FIGURE 4, COMPARISON OF ISOTHERMAL PRESSURE LOSSES
10,0
5,0
Q_
CO
UJ
CO
CO
CD
LU
CO
CO
1,0
P 0.5
CO
0,1
E,2 (PREFERRED)
ISOTHERMAL PRESSURE LOSSES
REFERENCE VELOCITY
(BASED ON EFFECTIVE AREA OF
ARRESTOR: ,000/425 M?)
TEMPERATURE (200°C-300°C)
5 10 50
[UREF/T] , FT/SEC-2 OK-1
100
I . I .... I
I . I I 1 I I i
,465 ,93
4,65
9,30
SEC~2 OK~1
481
-------
FIGURE 5, PERFORMANCE TEST RESULTS WITH LOW BTU FUEL
TEST CONDITIONS
,9999
v>
,9990
,9900
,9000
0
INLET TEMPERATURE : 150°C, 200°C, 300°C
PRESSURE : 300 KPA
ADIABATIC FLAME TEMPERATURE: 1260°C
CATALYST : CONFIGURATION E.2
FUEL HEATING VALUE
O 4,50 MJ/M^
Q 5,62 MJ/M3
O 6,75 MJ/M3
O
150 °C
1
1
TiN=300°C
3
3
2 ,
20 30 40
REFERENCE VELOCITY (M/SEC)
50
482
-------
FIGURE 6, PERFORMANCE TEST RESULTS WITH
MEDIUM BTU FUEL
PRESSURE : 300
ADIABATIC FLAME TEMPERATURE : 1260°C
CATALYST : CONFIGURATION E,2
FUEL HEATING VALUE (PREFERRED)
A : 7,5 MJ/M3
O : 12,0
,9999
-dr.9990
,9900
CO
Z3
PQ
SI
CD
,9000
0
TIN
300 °C
225°C
150 °C
150°C 225°C
I
u
2
CD
I
20 30 10
REFERENCE VELOCITY (M/SEC)
483
-------
FIGURE 7 EFFECT OF INLET TEMPERATURE
ON COMBUSTION EFFICIENCY
TEST CONDITIONS
REFERENCE VELOCITY : 30 M/SEC (a 150°C)
ADIABATIC FLAME TEMPERATURE: 1260°C
CATALYST : CONFIGURATION E,2
FUEL HEATING VALUE (PREFERRED)
A : 7,5 IWn3
O : 12,0 MJ/M3
,995 -
.990 -
LU
CQ
£i_
CD
,950 -
,900
O
CD
I
,500
RECIPROCAL INLET TEMPERATURE, i * io3 (-K-!)
TIN
484
-------
FIGURE 8, EFFECT OF HYDROGEN IN FUEL
TEST CONDITIONS
INLET TEMP,
PRESSURE
REFERENCE VELOCITY
FUEL HEATING VALUE
CATALYST
225°C
200 KPA
30,0 M/SEC
7,5 MJ/n3
CONFIGURATION E,2
(PREFERRED)
30 35
HYDROGEN COMPOSITION (%)
- 3
- 2 e-
O
I
- 1
0
485
-------
FIGURE 9, COMBUSTION EFFICIENCIES OF CO AND
TEST CONDITIONS
INLET TEMP,
PRESSURE
FUEL HEATING VALUE
FUEL/AIR
CATALYST
150°C
300 KPA
12,0 MJ/M3
0,266 MOLES/MOLE
CONFIGURATION E,2
(PREFERRED)
,9999
,9990
5-
C_3
•^
UJ
c_>
u_
Ll_
LU
00
H3
PQ
,9900
,9000
0
o
2
CD
O
UHC
_L
0
10 20 30
REFERENCE VELOCITY (M/SEC)
486
0
-------
FIGURE 10, EFFECT OF ADIABATIC FLAME TEMPERATURE
AT 300°C INLET TEMPERATURE
TEST CONDITIONS
INLET TEMPERATURE
REFERENCE VELOCITY
PRESSURE
FUEL HEATING VALUE
CATALYST
300 °C
30 M/SEC
300 KPA
12,0 MJ/n3
CONFIGURATION E,2
(PREFERRED)
,9999
LU
,9990
CO
og
o
,9900
,9000
0
1160°C
I
I
3
o
O
2
0
1100 1200
ADIABATIC FLAME TEMPERATURE (°C)
1300
487
-------
FIGURE 11, EFFECT OF ADIABATIC FLAME TEMPERATURE
ON MEDIUM BTU FUEL AT 250°C INLET TEMPERATURE
,9999
TEST CONDITIONS
INLET TEMPERATURE
PRESSURE
FUEL HEATING VALUE
CATALYST
REFERENCE VELOCITY
O 30 M/SEC
a 23 M/SEC
A 18 M/SEC
250°C
100-400
12,0
CONFIGURATION E,2
(PREFERRED)
,9990
,9900
M
,9000
0
__H£
o
CD
O
2
1100
1200
1300
ADIABATIC FLAME TEMPERATURE (°C)
488
-------
FIGURE 12, REGION OF OPERABILITY
FOR LOW AND MEDIUM BTU GAS FUELS
t BOUNDARY FOR LOW BTU GAS : BELOW SOLID LINE
• BOUNDARY FOR MEDIUM BTU GAS: BELOW DOTTED LINE
• LOW BOUNDARY OF REFERENCE VELOCITY WILL BE
DETERMINED BY FLASHBACK CRITERIA FOR LOW AND
MEDIUM BTU FUELS,
60
50
o
30
LU
20
10
0
COMBUST
PRESSURE
LOSS4 5%
^99%
REGION OF OPERABILITY
150
200 250
INLET TEMPERATURE (°C)
489
300
-------
CATALYTIC COMBUSTION OF ALTERNATIVE FUELS
By:
Henry Tong, Edward K. Chu, and Gerry C. Snow
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
ABSTRACT
In the past 2 years, Acurex has conducted several programs on the
catalytic combustion of alternative fuels. The results from three of these
programs are described. The first program used a simulated LBtu gas to define
the problems of burning gasifier fuels at low combustion temperatures. The
second program evaluated the performance of different catalytic reactors for
the combustion of coal derived liquids, and the third program examined the
feasibility of a radiatively cooled catalytic flameholder of heavy No. 6 fuel
oil. The test results show that these alternative fuels can be burned effi-
ciently with catalytic reactors; however, problems unique to combinations of
fuel, application and operating conditions need to be resolved before signifi-
cant commercialization of catalytic combustors.
Note:
The purpose of this note is to correct an error in the presentation of this
paper by the senior author at the Fourth Catalytic Combustion Workshop. In
the study on the catalytic combustion of coal-derived liquids, the third
catalytic combustor was a reactor fabricated from a Corning proprietary mono-
lith with platinum applied by Acurex to the first element to improve the
lightoff characteristics. In the paper to follow, this reactor is referred
to as the "Corning" reactor simply to distinguish it from the second reactor,
also fabricated by Acurex, which is referred to as the "Acurex" reactor.
490
-------
INTRODUCTION
In the next 10 years we expect to see significant changes in power
generation equipment. This will be a consequence of efforts to minimize our
dependence on foreign oil, increasingly stringent emissions requirements and
an increasing awareness to be energy efficient. For gas turbine engines, this
has resulted in many programs to develop efficient, low emissions combustor
concepts that can handle high nitrogen and/or low energy content fuels. Some
of the combustor concepts that show significant benefits use catalytic com-
bustion and much of the past research and development contributions have come
from Engelhard and Westinghouse. More recently, many more groups have joined
in the development of catalytic combustion as reflected in the agenda of this
and past catalytic workshops.
In much of the current combustor research, especially as it relates to
gas turbines, catalytic combustion is compared with lean premixed prevaporized
(LPP) combustion systems. Although one can easily evoke a significant debate
on the relative benefits of catalytic and LPP combustion, the authors feel
that catalytic combustion is really a special LPP combustor in which the flame
is initiated and stabilized by a catalytic surface. Except for some studies
on the catalytic combustion of partially vaporized liquid droplets, most cata-
lytic combustion concepts have a fuel preparation section to mix and vaporize
(in the case of liquid fuels) the fuel/air mixture to present a nearly homo-
geneous gaseous mixture to the catalyst.
With regard to the viewpoint that catalytic combustion is a special case
of LPP, the concept may be applied to extend the flammability limits to very
lean conditions as well as the blowout limits. Several small gas turbine power
systems are designed around combustor temperature rises of a few hundred degrees.
Mixtures this lean would be difficult if not impossible to burn (at least as
491
-------
premixed fuel and air). Thus catalytic combustion may allow greater flexibility
in the design of gas turbine cycles.
More advanced fuel delivery systems or combustor sectoring concepts for
both LPP and catalytic combustors may be necessary to obtain the full range of
turndown required for most gas turbine applications. System complexity must
therefore increase if these combustors are used to control emissions. Addi-
tional research work will be required to minimize these complexities to achieve
an acceptable level of system reliability.
In the past two years we have completed several research and development
programs to attain a better understanding of catalytic combustion characteris-
tics with the hope of applying this knowledge to design and demonstrate low
emissions, high efficiency gas turbine combustors. The objectives of these
studies were to examine the feasibility of catalytic combustion concepts, evalu-
ate catalyst stability and lifetime, evaluate fuel nitrogen conversion and the
generation of other pollutants and to measure combustion efficiency.
In this paper we will summarize the results of the following studies:
• Catalytic combustion of simulated low Btu gas (internal R&D)
• Catalytic combustion of coal derived liquid fuels (funded by EPRI)
• Rich flameholder for heavy fuels (funded by UTRC)
CATALYTIC COMBUSTION OF SIMULATED LOW BTU GAS
The objective of this experimental program was to determine some of the
combustion characteristics of low Btu gases from typical state-of-the-art coal
gasifiers. For these evaluations it was not practical to use a "real" gasifier
product so a premixed synthetic gas was used to simulate the products from a
Wellman-Galusha gasifier. The molar composition of this test gas was 29.4% CO,
15.4% H , 2.8% CH,, 3.4% CO , and 49.0% N2> The heating value of this gas was
about 170 Btu/ft . The stoichiometric air/fuel ratio was about 1.3 by volume.
Since a small amount of sulfur will be present in typical gasifier
products (even after a water wash), the target turbine inlet temperature was
selected to be less than 1600°F to minimize sulfur corrosion. This temperature
492
-------
is also representative of turbine inlet temperatures for many small gas turbine
engines.
The catalytic reactor was a two-size graded cell configuration as shown
in Figure 1. However, since the test gas contained a modest amount of hydrogen,
grading was probably not necessary and similar results are likely with a cata-
lyst of only fine cells. Past experience with similar materials as in this
reactor has shown that about 3 inches of total length was sufficient for com-
plete combustion of methane or propane at temperatures greater than 2100 F.
However, a longer configuration was selected for these tests because of the
low design combustion temperature which may limit the occurrence of the gas
phase combustion, especially CO and methane.
Data was obtained for stoichiometries ranging between 260 and 762 per-
cent theoretical air and for air preheat temperatures between 200 and 850°F.
Because there was no model fuel bound nitrogen compounds and the combustion
temperatures were low (1000 to 1900°F), NO emissions were too small to be of
X
interest. Unburned hydrocarbons and CO emissions are shown in Figures 2 and 3.
Also shown in Figure 2 are the lines of constant bed temperature. As might be
expected, CO and hydrocarbons increase as the bed temperature decreases; however,
the CO concentrations are significantly below theoretical if none of the initial
CO was oxidized. This would imply that the governing factor for efficient com-
bustion of this synthetic fuel is the rate at which CH, is oxidized. CO that
is in the fuel apparently oxidizes quite rapidly, perhaps due to catalytic com-
bustion and the presence of water vapor. This is probably true even at low
combustion temperatures. However, the CH, oxidation reactions are much slower
than H« reactions at low preheat temperatures and therefore take place further
into the reactor bed when the bulk gas temperature is sufficiently heated as a
result of H2 and CO combustion. CO formed here apparently has insufficient
residence time to oxidize. Interestingly, at low preheat and high theoretical
air (low bed temperatures) the sum of CO and hydrocarbons is the same magnitude
as the input concentration of CH, (-12,000 ppm). These emissions could probably
be reduced by higher reactor temperatures, longer catalyst beds, finer monolith
cells, or lower frontal velocities.
493
-------
CATALYTIC COMBUSTION OF COAL DERIVED LIQUID FUELS
The objective of this study was to determine the effect of coal derived
liquid (CDL) fuels on catalytic reactors. The basic scenerio of interest was
to assume that petroleum distillate fueled catalytic combustors will be used
in future utility gas turbines. In the event of a shortage of these fuels,
the utilities may elect to use CDL fuels temporarily. If this were the case,
it would be important to know if the catalyst would be poisoned or deactivated
by the CDL, or whether the catalyst would suffer any permanent change in per-
formance characteristics when the utility returns to distillate fuels. A
longer term interest was to understand the mechanisms which lead to catalyst
deactivation, especially with CDL fuels, and possibly identify ways to improve
catalytic reactors.
CDL fuels, in many respects, are much different from petroleum distil-
lates. For example, the carbon/hydrogen ratios of most CDLs are larger than
those for petroleum fuels. The heteroatom concentrations such as nitrogen and
oxygen are also larger. These differences and the metallic contaminants in
CDL fuels may lead to deactivation of catalytic reactors.
In this experimental program, two CDL fuels, SRC II (350°F-850°F dis-
tillation range) and H-coal distillate (300°F-500°F distillation range) were
used. The SRC II was obtained from the pilot plant in Tacoma, Washington,
and is a 5:1 blend of middle and heavy distillates. The H-coal was obtained
from Hydrocarbon Research, Inc. of New Jersey and was produced from the distil-
lation of atmosphere overhead products. Chemical analyses for these fuels are
shown in Table 1. Also shown in this table is the chemical analysis for a
diesel fuel that was used as a baseline fuel. The carbon contents of these
fuels are almost identical; however, the hydrogen content for SRC II is lower
than that for H-coal which in turn is lower than for diesel. Typical physical
properties for all three fuels are shown in Table 2.
The above fuels were tested on three reactor systems. The first reactor
was a proprietary noble metal (UOP) catalyst on a washcoated zirconia spinel.
This was a high activity reactor. The second reactor was an Acurex reactor
with a 2 weight percent platinum loading on a Torvex alumina substrate. No
494
-------
washcoat was used. The third reactor was a Corning proprietary monolith with
2 weight percent platinum (applied by Acurex) to the first element. Each re-
actor was a graded cell configuration composed of five 1-inch thick elements
to form a reactor 5 inches long and 2-1/2 inches in diameter.
Combustion data were obtained under fuel-rich and fuel-lean conditions,
as shown in Table 3, for a combustion pressure of 1 atm. Additional combustion
data were also obtained at 3 atm for fuel-lean conditions using H-coal and the
UOP reactor; however, this data will not be presented.
Both lean and rich combustion tests were started by first passing 30 to
40 ft/sec preheated air through the reactor bed until the bed reached a steady
state temperature of about 700°F. Initial lightoff was achieved with propane
and used to stabilize the reactor at a nominally fuel-lean stoichiometry. The
system was then transitioned over to a liquid fuel test condition. This light-
off procedure was generally completed in about 20 minutes. Rich combustion
conditions were achieved following a lean lightoff by rapidly decreasing the
combustion air flowrate while maintaining a fixed fuel flowrate. This proce-
dure was complete quickly enough so that the system thermal inertia prevented
reactor overheating as the equivalence ratio passed momentarily through unity.
Since the combustion temperature at the desired fuel-rich condition was greater
than maximum allowable reactor temperatures, nitrogen diluent was mixed and
preheated with the combustion air.
The fuel-lean combustion efficiencies at 1 atm pressure are shown in
Figures 4, 5, and 6 for each of the reactor systems. Except for the Corning
reactor where diesel baseline data were not run, it can be seen that the com-
bustion efficiencies for H-coal and diesel are essentially 100 percent for face
velocities between 40 and 85 ft/sec. (This upper velocity is the maximum flow-
rate capability of the test facility.) For H-coal at the higher velocities,
a very slight decrease in combustion efficiency was observed for the Corning
reactor. Also shown on these figures are the combustion efficiencies for SRC II.
Generally this combustion efficiency is a few tenths of a percentage point lower
than that for H-coal or diesel. Except perhaps for the Corning reactor, this
reduced efficiency is probably unimportant at low face velocities; however, a
rapid decline, indicating breakthrough, is seen at the higher face velocities.
Each of the three reactor systems are less active for SRC II than for the
other two fuels.
495
-------
Fuel nitrogen conversion for each reactor with H-coal is shown in
Figures 7, 8, and 9 for equivalence ratios between 0.4 and 1.3.* Each reactor
was held at constant temperature with nitrogen diluent. The UOP and Corning
reactors show nearly 100 percent conversion to NO under fuel-lean conditions.
X
The conversion decreased to about 80 percent at = 1.0 and continues to de-
crease as becomes greater than 1.0. Although there is significant data
scatter, the Acurex reactor achieved only 80-90 percent conversion under fuel-
lean conditions and increased to about 100 percent at $ = 1.0. The remaining
data are not conclusive about the conversion for greater than 1.0. As has
been demonstrated previously with other fuels, catalytic combustion under
lean conditions is, unfortunately, very efficient for converting fuel nitrogen
to NO. Also as expected, rich combustion converts a significantly smaller
X
fraction of fuel nitrogen to NO and NHq with the lowest values obtained on
X j
the UOP reactor.
The conversion of fuel nitrogen in SRC II on the UOP reactor is shown
as a function of equivalence ratio in Figure 10. Data for the other reactors
were not taken because fuel-rich combustion was unstable and steady conditions
could not be achieved. Fuel nitrogen conversion with SRC II shows the same
trend as with H-coal except that the fuel-rich percent conversion is lower.
However, the fuel nitrogen content of SRC II is about a factor of 3 (by weight)
greater than for H-coal.
Reactor degradation (combustion efficiency) is shown as a function of
time in Figures 11, 12, and 13 for lean combustion conditions. With H-coal,
all reactors performed fairly well even though some gradual degradation was
observed. With SRC II, lean stable combustion could not be maintained for
the Acurex and Corning reactors so data for these tests are not shown.
Although the UOP reactor behaved better, it too would not maintain steady
operating conditions. This is exhibited in the combustion efficiencies shown
in Figure 12.
Under SRC II fuel-rich conditions, neither the UOP nor the Acurex
reactors behaved acceptably. Deactivation of the catalysts was very rapid.
Because the fuel sulfur emissions interfered with the specific ion electrode
technique used to measure HCN, only NH» and NO conversions are shown.
496
-------
No rich SRC II tests were conducted with the Corning reactor since its deac-
tivation time was already very short, about 20 minutes, for lean SRC II com-
bustion conditions.
With H-coal under fuel-rich conditions, significant deactivation of
the Acurex and Corning reactors was observed after less than 1 hour of com-
bustion testing. The UOP reactor was tested for 7 hours of rich H-coal com-
bustion with no significant degradation in performance.
The results of this experimental study show that CDL fuels could be
effectively combusted with the right choice of catalyst. However, long term
durability data is insufficient to evaluate the potential lifetime of cata-
lytic reactors. Of the two fuels tested, SRC II had a greater detrimental
effect on the reactors than H-coal and of the three reactors, the high dis-
persion noble metal catalyst (UOP) was the least affected. Lean catalytic
combustion resulted in essentially 100 percent conversion of fuel nitrogen to
NO and although unacceptable for normal utility operations, this may be
X
acceptable in emergency situations. Fuel-rich combustion, which in a rich-
lean configuration may control fuel nitrogen conversion, was demonstrated
only for the UOP catalyst although some degradation and reduced performance
was observed with SRC II.
RICH FLAMEHOLDER FOR HEAVY FUELS
The objective of this test was to evaluate the performance of a rich
burn catalytic flameholder which could be part of a rich burn quick quench
combustor. The intent was to operate the flameholder at the rich stoichiometry
which produced minimum NO and precursors without diluent gases. This would
X
be achieved with a highly nonadiabatic catalytic reactor with radiation as
the principal energy loss mechanism. In addition, data were also obtained
over a range of theoretical air between 70 and 170 percent. No. 6 fuel oil,
with the properties shown in Table 4, was used in all tests.
Two different catalytic flameholders were tested without noticeable
differences in performance. The first was a noble metal on a proprietary
Corning monolith and the second was a noble metal on Torvex alumina. In both
497
-------
cases, the monolith was 1 inch long and had a nominal cell size of about
1/4 inch.
Results from this test series are shown in Figure 14. The rich burn
minimum was observed to occur between 70 and 80 percent theoretical air. In
this range the flameholder temperature was about 2100 F to 2300 F and is a
practical temperature for available materials.
A major problem encountered in these tests was a buildup of ash and
coke which caused blockage of the flameholder. In most cases under fuel-rich
operation, the deposit buildup starts at the outer ceramic wall of the catalyst
holder at the front face of the reactor. To minimize these problems, it was
necessary not to have any flow discontinuities in the tube walls upstream of
the catalyst. However, the flameholder monolith physically obstructs about
15 percent of the cross sectional area, causing a natural adverse gradient and
flow interference. Even under the best of conditions some deposition of ash/
coke occurred at these leading edges. These small depositions did not cause
the termination of any tests.
From these tests we conclude that stable combustion can be achieved at
high heat release rates under fuel-rich conditions based on a radiatively
cooled catalytic flameholder. In addition, the usual minimum in NO conversion
A
was observed under fuel-rich conditions. Fuel-lean conversions were similar
to conventional flame values. Ash/coke buildup could lead to catalyst plugging
or poisoning. The severity of this problem is related to the stoichiometry,
flow patterns, and degree of vaporization.
SUMMARY
In the above discussion, we have presented a brief summary of some
studies that we have conducted to evaluate the performance of catalytic com-
bustors with alternative fuels. The results show that various alternative
fuels can be efficiently burned with catalytic reactors; however, problems
which may be unique to the combination of fuel, application, and operating
condition still exist. We believe some of the more significant problems are
not conceptual but materials and systems related. It is hoped that enough of
these will be overcome in the next few years so that the benefits of catalytic
combustion can be applied to gas turbine engines.
498
-------
TABLE 1. FUEL CHEMICAL PROPERTIES
Analysis Type SRC IX H-Coal Diesel
Elemental Analysis,
wt %
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen
Heat of Combustion,
Btu/lbm
Net
Gross
Trace Metals,
ppm wt
Calcium
Iron
Manganese
Magnesium
Nickel
Vanadium
Sodium
Zinc
Copper
Silicon
Tin
Aluminum
Boron
Lead
Molybdenum
Titanium
Silver
Cobalt
Cadmium
Potassium
Phosphorus
Chromium
ASTM E191 86.09
9.09
ASTM E258 0.86
ASTM D129 0.22
ASTM D482 0.011
By difference 3.74
16,470
17,300
Wet Chemical/
spectrographic
1.5/1.5
13.0/13.0
0.22/0.26
0.34/1.40
0.11/0.10
<1.0 /0.03
0.39/<0.10
/0.60
/0.23
/24.0
/I. 03
/13.0
/0.31
/0.17
/0.03
/0.85
/0.002
/0.02
/0.15
/
/
71.4
86.20
11.19
0.31
0.094
0.010
2.08
17,640
18,700
0.31/0.18
38.0/52.0
0.07/0.11
0.04/0.16
0.08/0.03
<0. 04/0. 003
0.15/
/14.0
/2.6
/0.18
/0.32
/0.50
/0.06
/0.27
/0.05
/0.02
/0.001
/0.005
/0.17
/
/
/
86.62
12.90
0.035
0.13
0.003
0.31
18,230
19,410
0.25/0.55
0.70/1.90
<0. 05/0. 05
0.08/0.51
<0. 05/0. 13
<1.0 /
0.14/0.02
/2.0
/0.15
75.70
/0.26
71.10
70.03
/0.66
/
70.11
/0.007
/
70.35
/
70.26
/
499
-------
TABLE 2. FUEL PHYSICAL PROPERTIES
Physical Properties SRC II H-Coal
Flash point, °F 170 11.0
Pour point, °F -60 -65
Gravity, °API 10.3 29.5
Boiling range, °F 360-700 260-552
Kinematic viscosity, CS 4.42 1.17
@ 100°F
No. 2 Diesel
175
-10
32.0
358-672
2.61
TABLE 3. CDL SCREENING TEST MATRIX
Fuel SRC II H-Coal
Diesel
Reactor Lean Rich Lean Rich Lean Rich
UOP x x x x
Acurex x x x x
x x
X X
Corning Active
Monolith
500
-------
TABLE 4. CHARACTERISTICS OF #6 FUEL OIL
#6
Fuel Oil
Gravity, °API
' Moisture and sediment, %
Gross heat of combustion,
Btu/lb
Viscosity at 100°F, SUS
Flash point, COC
Nitrogen, Dohrmann, ppm
Nitrogen, Coletnan, %
Carbon, hydrogen, %
Perkin Elmer
Sulfur, XRF, %
Ash, D 482, %
Trace elements, %
12.3
0.005
18,325
4,894
182
0.81
0.82
C - 85.57
H = 10.52
0.93
0.089
Al: 0.0012
Ca: 0.0086
Cr: 0.0006
Cu : 0 . 0002
Fe: 0.0075
Mg: 0.0033
Mn: 0.0004
Ni: 0.0085
Si: 0.0009
Ti: 0.0005
V: 0.0073
Zn: 0.0014
Pb: 0.0002
Ba: 0.0001
Mo: 0.0001
501
-------
Ul
o
FLOW
• CATALYST AND WASHCOAT
- UOP PROPRIETARY
• SUPPORT
- CORNING MCB-7 ZIRCONIA SPINEL
• GEOMETRY
- 2" DIAM, 5-1/1" LONG, GRADED CELL
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
1
MEDIUM
CELL
3/16"
SQUARES
'////////>
-+ — 1-1/2" ^
MEDIUM
CELL
3/16"
SQUARES
'////////,
1-1/2" ^
SMALL
CELL
1/16"
SQ
' / / / /
*-3/4"-»-
SMALL
CELL
1/16"
SQ
/ / / / i
-«-3/4"-*-
SMALL
CELL
1/16"
SQ
'III/
•*-3/4l!-»-
Figure 1. Catalytic Reactor for LBtu gas.
-------
o
CO
O 200°F Preheat
D 400°F
A 450°F
O 700°F
V 850°F
200 250
300
350
400 450 500 550
Theoretical air (%)
600
650
700
750
800
Figure 2. UHC as a function of theoretical air for LBtu gas.
-------
Ui
o
5000 -
•; 4000
V
(J
X
0)
»*
o
•o
-------
Oi
o
Ui
O)
100.0
99.8
99.6
99.4
99.2
99.0
98.8
98.6
98.4
98.2
98.0
O OD=O
D No. 2 diesel ( ~ 0.50, Tbed ~ 2000°F)
A SRC II (4> ~ 0.40, Tbed - 2000°F)
O H-coal (4. ~ 0.50, Tbed - 2000°F)
40
50 60 70
Face velocity, V - (ft/sec)
80
Figure 4. Fuel-lean combustion efficiency versus face velocity, UOP reactor.
-------
100.0
99.8
— 99 6
*« yy'°
c 99.4
f>
c 99.2
0)
0)
99.0
98.8
98.6 -
5 98.4
98.2
98.0
<3DK>3ra^<>~<25>-K>crOO
D No. 2 diesel ( ~ 0.50, TL . ~ 2100*F)
bed
A SRC II (4. ~ 0.60, Tbed - 2200°F)
O H-coal (4, ~ 0.45, Tbe(J - 2000°F)
1
40
50 60 70
Face velocity, Vrgf (ft/sec)
80
Figure 5. Fuel-lean combustion efficiency versus face velocity, Acurex reactor.
-------
100.0
99.8
99.6
~ 99.4
A
0s 99.2
O)
5 99.0
»*-
* 98.8
0
53 98.6
J 98.4
98.2
98.0
A
A A
A
-
_
A SRC II ( ~ 0.60, Tbe(j - 2300°F)
O H-coal ( - 0.53, Tbe(J - 2200°F)
| | 1 1 1 1 1 JL 1 -L L_
o
n
i
40
50
Face velocity, Vref (ft/sec)
Figure 6. Fuel-lean combustion efficiency versus face velocity, Corning reactor.
-------
Ul
o
co
1 UU
90
80
S 70
c
x 60
0
c 50
o
(/>
b 40
§ 30
o 20
3
U-
10
n
^Qn
8
8
—
-
8
—
0 NOX
D NH3
_
o
i i i i i i i in i r~i i ii
(X
o
\^
^^
^r
0.4 0.6 0.8 1.0
Fuel-air equivalence ratio,
1.2
1.4
r.
Figure 7. Fuel-N conversion to XN, UOP reactor, H-coal
-------
^"^
X
O
4->
c
O
•t—
l/l
s_
QJ
>
C
O
Ul <->
g
r™~
cu
3
U-
100
90
80
70
60
50
40
30
20
10
n
u — u — u
9 o
9 o
-
o
—
—
0 NOX
D NH3
—
i i i i i n n O 'n ' ' ' '
Fuel-air equivalence ratio, (tip.
Figure 8. Fuel-N coversion to XN, Acurex reactor, H-coal.
-------
IUU
90
80
~ 70
X
o 60
.1 50
in
O)
> 40
o
0
£ z 30
o ^
"oi
£ 20
10
0
\ULU
O
-
0 ° ° o
0
O NOX
D NH3
i i i i i D ,n D P . D . i
0.4 0.6 0.8 1.0 1.2 1.4
O
Vo
Fuel-air equivalence ratio,
FA
Figure 9. Fuel-N conversion to XN, Corning reactor, H-coal.
-------
c
o
in
i.
100
90
80
70
60
40
o
o
=f 30
(V
2 20
10
0
-CCHD-
O N0>
D NH.
_L
'
^
0.4
0.6 0.8 1.0 1.2
Fuel-air equivalence ratio, 4>r/\
Figure 10. Fuel-N conversion to XN, UOP reactor, SRC II.
-------
Oi
(-•
N3
100.0
99.8
^ 99.6
^ 99.4
£ 99.0
o>
§ 98.8
3 98.6
jQ
O
0 98.4
98.2
98.0
rX^ (~~~ __(-* ^V"V (*} rS ^v — ^ ^^
v-«^v^^*
O
-
-
O H-coal
— 1 — 1 — 1 — 1 I 1 1 1 1 1 1 III
•coSht -
"*
2 3 4 5 6 7 8 9 10 11 12 13 14 15
Catalyst operating time at fixed conditions (hours)
Figure 11. Combustion efficiency versus elapsed time, Acurex reactor.
-------
U1
i—•
U>
100.0
99.8
— 99.6
a*
c 99.4
g 99.2
K>
-------
IUU.U
99.8
~ 99.6
^ 99.4
% 99.2
(U
U
£ 99.0
a*
§ 98.8
•*^
3 98.6
« 98.4
98.2
98.0
^ --^^xx^^gg^^
A A
A A AA
-
A
A
D No. 2 diesel
A SRC II
O H-coal
1 « 1 1 1 1 1 1 1 1 1 l I i i
-------
25001- ?0
Ul
M
Ol
2000 -
150C*—
r— 100
100? NO yield at 15'; excess air - 1160 ppm
200
Theoretical air (%)
Figure 14. Catalytic flameholder perfonnance.
-------
GASIFICATION OF A HEAVY AND A DISTILLATE FUEL
By:
E. J. Szetela, United Technologies Research Center
R. A. Sederquist, United Technologies Power Systems Division
J. A. TeVelde, United Technologies Research Center
ABSTRACT
A catalytic gasifier which consists of an air-atomizer and a catalytic
monolith was tested with a residual oil and a No. 2 oil. Test ranges included
equivalence ratios from 6 to 8, reference velocity from 30 to 35 fps, and air
temperature upstream of the injector from 740 to 1200 F. During runs of short
duration, up to 17 min, catalyst wall temperature, catalyst discharge tempera-
ture, and effluent observation indicated that considerable fuel gasification
occurred. Moderate wall temperatures were encountered allowing the applica-
tion of available monolith materials. Further work will be needed to extend
gasifier life and demonstrate catalyst and support material durability.
516
-------
INTRODUCTION
The catalytic gasification of gasoline and other distillate fuels with
air has been examined by a number of investigators. At Siemens in Germany (Ref-
erence 1), oxidative catalytic reforming in an autothermal catalytic reactor
has been carried out at fuel-air equivalence ratios as high as 10 and catalyst
temperatures of 1400 to 1600 F without the generation of soot in the product
gas. This process, which is carried out at low residence times, produces
cracked hydrocarbons, methane, hydrogen, water, and oxides of carbon. At Jet
Propulsion Laboratory (Reference 2), distillate fuels have been partially oxi-
dized and reformed over catalyst monoliths at equivalence ratios above 3 and
catalyst temperatures above 1800 F without the formation of soot. The addition
of small amounts of steam permitted higher equivalence ratio operation. At
United Technologies (Reference 3), No. 2 heating oil containing 3300 ppm sulfur
has been partially oxidized and steam reformed over a packed bed catalyst at
equivalence ratios in excess of 4 without soot formation. This process has also
been successfully applied to a number of coal-derived distillates.
These successful programs have led to the consideration of catalytic
partial oxidation for the gasification of heavy fuel which cannot be vaporized
by conventional methods. A program was initiated at the United Technologies
Research Center to determine the feasibility of the catalytic gasification con-
cept. This paper describes the experimental approach that has been taken and
the initial results that have been obtained.
OPERATIONAL REQUIREMENTS
The objective of the catalytic gasification of residual fuel is to
achieve significant improvement in fuel preparation for combustion in stationary
517
-------
gas turbines or furnaces over the conventional atomization, vaporization, and
mixing techniques. The catalytic gasifier must operate without soot formation
while being supplied fuel-air mixture in which only a small portion of the fuel
is vaporized. The presence of a large fraction of liquid fuel droplets enter-
ing the catalyst monolith must not deactivate the catalyst or quench the cata-
lytic combustion process.
Soot or carbon formation is predicted at high temperatures in thermo-
dynamic equilibrium at an atomic oxygen-to-carbon ratio of 1 (which corresponds
to an equivalence ratio of 3). Figure 1 shows thermodynamic equilibrium carbon
boundaries for the H-C-0 system. Under equilibrium conditions, operation above
an equivalence ratio of 3 (below an 0/C ratio of 1) would produce carbon. It
is therefore necessary when operating above an equivalence ratio of 3 that
equilibration of the fuel-air mixture by reforming of the fuel be limited such
that the product gas atomic oxygen to carbon ratio for the mono-carbon species
is held above the critical value of 1 as shown in Equation 1.
Y + 2Y + Y
CO C02 ^O
Y+ Y+~Y' '
CO C02 CH^
where Y is the product gas mole fraction.
In a functioning gasifier, it is probable that all of the oxygen is
consumed in the combustion process and the remaining fuel is in the form of
cracked or intermediate hydrocarbons. It is necessary to use short residence
time in the catalyst monolith and possibly an activity which controls or selec-
tively limits the amount of reforming and prevents formation of carbon. Short
residence time can also reduce methane formation and soot production by thermal
cracking of the fuel.
The presence of large quantities of unvaporized fuel droplets can present
an operational problem. At high equivalence ratios only a small fraction of the
fuel is expected to vaporize. Fuel droplets impinging on the catalyst can
crack and form coke which can plug and poison the catalyst. Metallic impurities
and ash in the fuel may affect long term catalyst activity and operation. For
these reasons it is generally concluded that a large catalyst cell size would
518
-------
minimize surface impingement of fuel droplets which can poison the catalyst
and quench the reactions by wall cooling.
EXPERIMENTAL APPARATUS AND PROCEDURE
Tests were conducted using the steady flow facility shown in Figure 2.
The facility consisted of an electric air heater, a flow straightener, a swirl-
type, air-atomizing fuel nozzle, a catalytic reactor section, a liquid sepa-
rator and an afterburner section. The diameter of the rig was 2.4 in. upstream
of the reaction section and 2.0 in. at the catalyst test section. The rig was
insulated with fiberglass to minimize heat losses. Calibrated rotameters were
used to measure the main airflow rate, atomizing airflow rate, water coolant
flow rate and propane flow rate. The fuel flow was measured using a Fluidyne
positive displacement flowmeter. The electric heater was capable of preheating
the inlet air to a temperaure of 1200 F, The amount of nozzle airflow required
to obtain a fine fuel mist was checked by visual observation of the spray into
quiescent air.
Thermocouples were used to measure the inlet air temperature, fuel tem-
perature and the fuel-air mixture temperature prior to entering the catalyst
reactor. Thermocouples located on the substrate walls and at the exit center-
line were used to indicate the performance of the reactor. Indication of
carbon or soot formation was monitored by a continuous measurement of the pres-
sure drop across the reactor section. The cyclone separator, used for non-
vaporized liquid effluent collection, was rated by the manufacturer to have a
collection efficiency of 95 percent for droplets larger than 5.0 microns.
Initiation of a test began with the main airflow rate being established
and heated to the prescribed inlet temperature. After a steady state rig
temperature was reached, propane and atomizing air were introduced through the
fuel nozzle to preheat the catalyst bed to an initial temperature capable of
sustaining the chemical reactions (approximately 1200 F). In order to minimize
transients during switchover to liquid fuel, the liquid flow rate was preset
using a bypass system. Upon reaching the desired catalyst preheat temperature,
the propane flow was terminated and liquid flow through the catalyst initiated.
519
-------
Catalyst performance parameters including temperature and pressure drop measure-
ments were monitored during the test. The test was terminated when appreciable
performance deterioration was noted.
FUEL
The majority of the tests have been performed using a residual heating
oil with a viscosity which was near the lower level of the residual fuel band.
The properties of the fuel are shown in Table I. An additional test was run
TABLE I. FUEL PROPERTIES
Specific gravity 0.90
Hydrogen-carbon ratio 1.77
Percent nitrogen 0.11
Percent sulfur 0.38
Viscosity, CS 122 @ 77 F
16.6 @ 194 F
with No. 2 oil in preparation for combustor tests that will be run in the near
future.
CATALYST HARDWARE
Three ceramic monoliths were procured by catalyst manufacturers, prepared
for the catalyst by applying porous alumina on the substrate surfaces, and cata-
lyzed with noble metals. Catalyst 1 had a zirconia spinel honeycomb substrate
with square cells which had 0.10 in. sides. The substrate consisted of a single
section having a length of 3 in. The substrate was catalyzed by Oxycatalyst
with a 2 Pd:l Pt coating applied at a loading of 1.4 g/m2.
Catalyst 2 had a Torvex alumina honeycomb substrate with hexagonal cells
which had an 0.17 in. distance between parallel sides. The substrate consisted
of two butted sections, each one having a length of 1.5 in. The substrates
were catalyzed by Matthey Bishop with a Pd:Pt coating applied at a loading of
4.9 g/cm .
Catalyst 3 had a zirconia spinel honeycomb substrate with square cells
which had 0.27 in. sides. The substrate consisted of two butted sections,
520
-------
each one with a length of 3 in. The substrates were catalyzed by Oxycatalyst
with a 2 Pd:l Pt coating applied at a loading of 1.7 g/m2.
Catalyst 4 consisted of a 1.5 in. length of Catalyst 3 followed by a
1.5 in. length of Catalyst 2.
EXPERIMENTAL RESULTS WITH RESIDUAL OIL
The first series of tests were run with Catalyst 1. Three tests were
run and although good results were obtained in the first run, it was found that
performance deteriorated in subsequent runs as shown in Table II. Reference
velocity is based on the flow of air at the inlet temperature immediately up-
stream of the monolith. After Run 1, the separator exit reservoir contained
TABLE II. RESULTS WITH CATALYST 1
Run No.
Run Time, Min.
Reference Velocity, FPS
Equivalence Ratio
Air Inlet Temperature, F
Temp. Downstream of Injector, F
Catalyst Wall Temp, F
Catalyst Discharge Temp, F
1
l.A
32
6.6
1170
610
1110
1500
2
1.2
35
8.2
1180
570
1020
800
3
5.3
34
6.8
1040
550
780
620
no fuel. After Run 2, a small amount of fuel was found in the reservoir and a
considerable amount of fuel was found after Run 3.
The second series of tests were run with Catalyst 2. Good results were
obtained in the first run and performance deterioration was not as severe as
with Catalyst 1. The results with Catalyst 2 are shown in Table 3.
The third series of tests was run with Catalyst 3. Four tests were run
and in each test, it was noted that pressure drop across the catalyst increased
progressively. However, during shutdown, the obstruction seemed to be removed
because the pressure loss at the beginning of any test was the same as the
pressure loss at the beginning of the previous test. Initial conditions with
Catalyst 3 are shown in Table IV. Performance changes noted during tests with
Catalyst 3 are shown in Table V.
521
-------
TABLE III. RESULTS WITH CATALYST 2
Beginning End Beginning End
Run No.
Run Time, Min.
Reference Velocity, FPS
Equivalence Ratio
Air Inlet Temperature, F
Temp Downstream of Injector, F
Catalyst Wall Temperature, F
Catalyst Discharge Temp, F
1
-
33
6.0
1120
590
1140
1560
1
9.0
33
6.0
1120
560
990
1220
2
-
32
6.0
1175
590
1120
1355
2
5.0
32
6.0
1175
590
920
1350
TABLE IV. INITIAL RESULTS WITH CATALYST 3
Run No.
Reference Velocity, FPS
Equivalence Ratio
Air Inlet Temp, F
Temp Downstream of Injector, F
Catalyst Wall Temp, F
Catalyst Discharge Temp, F
1
33
6.2
1080
605
1280
1150
2
32
6.2
1115
660
1325
1670
3
31
6.2
1045
590
1300
1275
4
35
6.2
1215
785
1365
1415
TABLE V. PERFORMANCE CHANGES WITH CATALYST 3
Run
1
2
3
4
Time Catalyst Discharge
Min Temperature , F
Beginning
End
Beginning
End
Beginning
End
Beginning
End
0
5.8
0
3.0
0
2.5
0
3.5
1150
1300
1670
1650
1275
1040
1350
1250
Pressure Drop,
in. H20
2.2
15.0
2.4
6.0
2.5
4.5
2.6
17.5
522
-------
EXPERIMENTAL RESULTS WITH NO. 2 OIL
A single test was run using No, 2 oil with Catalyst 4. The inlet air
temperature was reduced for this test to duplicate the air temperature planned
for the actual combustor test. Although the pressure drop increased at the
beginning, it returned to its initial value and the results shown in Table VI
were encouraging. Approximately 10 percent of the total fuel flowed during
TABLE VI. RESULTS WITH NO. 2 OIL AND CATALYST 4
Run Time, Min
Reference Velocity, FPS
Equivalence Ratio
Air Inlet Temperature, F
Catalyst Wall Temp, F
Catalyst Discharge Temp, F
Pressure Drop, in. H00
L
0
30
7.5
740
1530
1035
2.5
5
30
7.5
740
1570
1035
4.5
10
30
7.5
760
1460
1055
2.9
17
30
7.5
760
1410
1100
2.6
the run with No. 2 oil was found in the liquid reservoir after the test was
completed.
CONCLUSIONS
Tests to date indicate the feasibility of the catalytic gasification
of residual and alternative fuels as a means of preparing these fuels for
subsequent combustion in stationary gas turbines and furnaces. Catalytic
gasification at high equivalence ratio results in moderate catalyst tempera-
tures allowing the application of available catalyst monolith materials.
Further optimization is required to select the operating conditions, the mono-
lith dimensions, and the catalyst formation that achieves the best level of
gasification without forming carbon. Additional work is required to extend
gasifier life, determine maximum fuel flow capability, and demonstrate cata-
lyst and support material durability under all operating conditions.
523
-------
REFERENCES
1. Henzel, H. J., H. Kostka, and A. Michel. Autothermal Gasification of
Liquid Hydrocarbons by Partial Oxidation. Paper presented at ERDA meet-
ing on Fuel Processing for Fuel Cell Power Generation, April 13, 1977.
2. Voecks, G. E. and D. J. Cerini. Application of Rich Catalytic Combustion
to Aircraft Engines. Presented at Third Catalytic Combustion Workshop,
October 4, 1978.
3. Bett, J. A. S., R. R. Lesieur, D. R. McVay, and H. J. Setzer. Adiabatic
Reforming of Distillate Fuels. National Fuel Cell Seminar, San Francisco,
CA, June 1978.
524
-------
Oi
N>
ro
c
o
-d
(O
u
I
c
(U
en
o
1.0
2.0
Oxygen-carbon ratio
4 0
Figure 1. Carbon formation boundaries in the H-C-0 system.
-------
Residual oil
Primary air
Ul
Propane
Secondary air
Water
Air
Straightener
Injector
Catalyst
Burnoff
stack
Separator
Collector
Methane
Figure 2. Catalytic gasifier test rig.
-------
CATALYTIC COMBUSTION OF HEAVY
PARTIALLY-VAPORIZED FUELS
By:
T. J. Rosfjord
United Technologies Research Center
East Hartford, Connecticut
ABSTRACT
An experimental program to demonstrate efficient catalytic combustion of
fuel-lean and fuel-rich mixtures of residual fuel and air, and to assess the
influence of incomplete fuel vaporization on the performance of a catalytic
reactor is being conducted. A 7.5-cm diameter catalytic reactor was designed
and will be tested over a matrix of conditions representative of a gas turbine
combust or inlet. For each of three test phases, two series of tests with a
uniform but poorly vaporized (less than 50 percent) mixture of No. 6 fuel oil
and air will be performed. In the first series, the non-vaporized fuel will be
contained in a spray of droplets with a Sauter Mean Diameter (SMD) less than 30
microns. In the second series, the non-vaporized fuel will be characterized by
a spray SMD approximately equal to 100 microns. The designs of the fuel injection
system and the catalytic reactor are described in this paper.
527
-------
INTRODUCTION
Current concerns regarding pollutant emissions are occurring at a time
when continued development of the gas turbine engine as an industrial prime
mover depends in part on the capability to use fuels with far less attractive
physical and chemical properties than those of the fuels currently used.
Increasingly stringent emission standards prompt even greater emissions reduc-
tions, especially of NO , than may be attainable with current combustor
design technology. Therefore, to increase the options available for low
emission, fuel flexible engine designs, novel combustion concepts ought be
investigated. Catalytic combustion is a candidate approach in this application.
Catalysts have been shown to promote stable, efficient combustion at exception-
ally lean fuel-air ratios. However, the preponderance of combustion-related
work with catalysts has involved the use of either gaseous or relatively
volatile fuels coupled with fuel-air preparation systems configured so that the
catalyst was not confronted with a fuel-air mixture in which the majority of
the fuel was in the liquid state or in droplets having a large diameter. Thus,
with the growing requirement to use low-volatility alternative and residual
fuels for stationary gas turbine engines, investigations to elucidate the
behavior of catalytic combustion systems using incompletely vaporized heavy
fuel are required. Additionally, recognizing the promise of controlling NO
formed from fuel-bound nitrogen by employing a staged combustion sequence, it
is necessary to define the operational characteristics of catalytic reactors
for both fuel-rich and fuel-lean mixtures.
United Technologies Research Center, under contract from NASA-Lewis Research
Center, is conducting an experimental program directed toward demonstrating
efficient catalytic combustion of fuel-lean and fuel-rich mixtures of residual
fuel and air, and assessing the influence of incomplete fuel vaporization on
the performance of a catalytic reactor. A 7.5-cm diameter catalytic reactor
was designed and will be tested over a matrix of conditions representative of a
gas turbine combustor inlet. For each of three test phases, two series of
tests with a uniform but poorly vaporized (less than 50 percent) mixture of No.
6 fuel oil and air will be performed. In the first series, the non-vaporized
fuel will be contained in a spray of droplets with a Sauter Mean Diameter (SMD)
less than 30 microns. In the second series, the non-vaporized fuel will be
characterized by a spray SMD approximately equal to 100 microns.
523
-------
UTRC has overall responsibility for the technical program, including the
design and fabrication of the test facility and execution of the test program.
Acurex Corporation, as a subcontractor to United Technologies, will support
UTRC in the design, fabrication and evaluation of the catalytic reactors.
Problems in obtaining the desired substrate material have delayed the initiation
of the test program. Therefore, this paper describes only the fuel preparation
system and the catalytic reactor designed to satisfy the program requirements.
Fuel Preparation System
A fuel injection system was designed to satisfy three operational requirements
at the following condition: inlet air temperature • 600 K; pressure = 600 kPa;
reference velocity = 20 m/s; fuel-air ratio = 0.020. First, the system
must provide distributions of fuel-air ratio and velocity which are uniform to
within + 10 percent of the mean value at the catalytic reactor inlet. Second,
the fuel must be less than 50 percent vaporized at the catalytic reactor inlet.
Third, the device must be capable of producing sprays characterized by a SMD of
either less than 30 microns or in the range of 70-150 microns. These require-
ments were satisfied by designing a system which employed a multiple venturi
fuel injection system followed by a straight mixer/vaporizer duct. The injector
contains 19 venturi tubes distributed over a 7.5-cm diameter circular area.
The fuel will be injected normal to the airstream through 0.76-mm inside
diameter tubes which penetrate thirty percent across each venturi throat (Fig.
1). The injector was fabricated by an electrical discharge machining technique
to permit the 7-deg half angle conical diffusers to terminate along the natural
cone intersections. This geometry eliminates regions of recirculating base
flow which would promote autoignition within the fuel preparation system.
A review of correlations developed for the atomization of heavy fuels concluded
that an expression developed for an air spray injector by Jasuja (Ref. 1) was
applicable for the injector of interest. The correlation developed was:
SMD . 0.19 (Vpl> (1+ l )°-25 +0.127 ^/ D_\°-50
'uaPa°-35 AFR \Pja|/ AFR
where: SMD - droplet diameter, m
ot - liquid surface tension, N/m
Pi. - liquid density kg/m
/Lif - liquid viscosity, Ns/m
Pa - air density, kg/m
ufl - air velocity, m/s
D - orifice diameter, m
AFR - air to fuel ratio in injector by weight
529
-------
The correlation is the sum of two terms; the first is dominated by the air
velocity while the second is responsive to the liquid (i.e., fuel) viscosity.
Proper selection of injector blockage and fuel temperature will produce the
desired air velocity and fuel viscosity, respectively, to form droplet sizes
which satisfy the requirements of the program. A 0.54-cm venturi throat
diameter was specified to produce a velocity of the 245 m/s at the fuel injec-
tion site, insuring that the first term will always have a small value.
Variation in the second term, and thus the SMD, will be achieved by varying the
fuel temperature. Analysis of the fuel for this program indicated that for a
fuel temperature of 373 K, a spray characterized by SMD • 25 microns will be
obtained, while at a temperature of 329 K an SMD « 100 micron spray will be
formed.
The rate of vaporization of the injected fuel was analyzed using the UTRC
Spray Vaporization Analysis. The analytical program is a multiple streamtube
code which uses a marching finite difference solution procedure to provide a
streamwise evolution of the two-phase flow properties. Vaporization calcula-
tions were performed using No. 6 fuel oil properties. Fuel droplets with
diameters of 30 and 35 microns at an initial temperature of 380 K and of 100
and 125 microns at 330 K were "injected" into air having a velocity of 245 m/s
as would exist in the throat of any one of the nineteen fuel injector venturi
tubes. The airflow was decelerated in the injector diffuser prior to entering
a constant area streamtube representing one-nineteenth of the total flow area.
In order to simultaneously limit the vaporization to the 50 percent maximum and
achieve a mean droplet diameter of less than 30 microns at the reactor inlet,
the maximum allowable spray SMD at the injector was determined to be approximately
35 microns. Therefore, as shown by the calculation results (Fig. 2), the
mixer/vaporizer must not be longer than 18 cm. Note that over this length, the
larger droplets vaporize to a much lesser extent; after 18 cm, a droplet
initially with a diameter of 100 microns will vaporize only 10 percent. The
length required by the large droplets to achieve 50 percent vaporization is
approximately one meter, an impractical length for either this test program or
for a conventional engine configuration. Since the gaseous fuel-air ratio
corresponding to 10 percent fuel vaporization is below the minimum at which
catalytic combustors have been operated with distillate fuel, there is a
concern that the reactors will be ineffective. Two provisions are being
considered in this instance. First, a longer mixer/vaporizer could be used to
raise the level of vaporization. It is anticipated that this approach would
provide minimal benefit however since the increase in extent of vaporization
with additional length is small. Second, in an attempt to determine a threshold
for efficient reactor operation, tests could be performed with intermediate
droplet sizes (produced by raising the fuel temperature) which would provide
more favorable levels of vaporization (but still less than 50 percent). This
approach is compatible with the fuel injection system and offers the opportunity
to define an operational limit.
The mixing of the injected fuel into the airstream was also analyzed using
the UTRC Spray Vaporization program. The initial fuel concentration was
represented as a "top hat" profile (centered on the longitudinal axis of a
530
-------
venturi) with a width equal to two fuel tube diameters. An eddy diffusivity
equal to the value determined by Longwell (Ref. 2) for the turbulent transport
of fuel droplets was employed. The results from the calculations indicated
that fuel-air mixture uniformity to within approximately + 5 percent of the
mean is achieved 12.7 cm downstream from the fuel tube. The calculations also
indicated that a uniform air velocity profile would be realized. Tacina (Ref.
3) obtained uniformity at 17.8 cm downstream from a jet fuel, multiple-venturi
injector configured to distribute 21 injection sites over a 12-cm diameter
circular area. The injector being fabricated for this program will have 19
injection sites distributed over a 7.5 -cm diameter, providing an injection
site density 2.2 times the Tacina design. Assuming a similar mixing rate, the
decreased distance between injection sites should result in uniformity at 12 cm
downstream from the fuel tube. It is recognized that, because of differences
in the fuel spray (fuel properties, droplet sizes) and the precise injector
geometry, the consistency of this prediction with the UTRC model calculations
does not guarantee the mixer performance. The agreement does, however, establish
a level of confidence in the design procedure. Additionally, based upon the
vaporization calculations presented above, the mixer/vaporizer length was
specified at 18 cm providing even greater distance to achieve uniformity.
Catalytic Reactor
Acurex Corporation, as a subcontractor to UTRC, designed the catalytic
reactors for this program (Ref. 4). A 7.5-cm diameter reactor was designed to
achieve a 99.5 percent combustion efficiency at the following condition:
inlet air temperature » 600 K; pressure = 600 kPa; reference velocity = 20 m/s;
adiabatic combustion temperature = 1400 K. The maximum allowable pressure drop
through the reactor was specified at 5 percent.
A three-phase design analysis was performed. In the initial phase, it
was assumed that the reactor was adiabatic and the fuel droplets did not
ignite. The second analysis phase superimposed the droplet ignition phenomena
on the initial results by considering the autoignition and combustion time for
the droplets within the reactor. The third phase of the analysis considered
the non-adiabatic characteristics of the reactor, particularly the de-stabilizing
influence of radiation from the upstream face of the bed.
In the initial adiabatic analysis, the governing differential equations for the
bulk gas energy and mass balances, droplet energy and mass balances, and surface
energy balance were solved using a finite difference procedure. Because of the
lack of quantitative data on kinetic rates for residual fuel and air catalytic
reactions, it was assumed that the fuel consumption rate was diffusion limited.
The initial reactor investigated was a graded-cell device configured to include
5-cm, 2.5-cm, and 15-cm lengths of 0.63-ctn, 0.32-cm, and 0.16-cm diameter cell
monolithic substrates, respectively. Typical results are shown in Fig. 3
531
-------
which illustrates the variation in the substrate temperature, bulk gas tempera-
ture, droplet radius, local gaseous fuel-air ratio and fuel conversion efficiency
with distance along the reactor length for the design condition. As indicated,
the events are guided by the droplet evaporation. As the droplet vanishes the
wall temperature increases to the adiabatic temperature and the bulk gas
temperature and fuel conversion efficiencies begin to increase rapidly. These
results indicate that in the absence of homogeneous combustion or droplet
ignition the goal of 99.5 percent combustion efficiency will require a reactor
length significantly in excess of 25 cm. A benefit of the analysis is
that it provided the temperature history required to calculate the autoignition
length of the fuel droplets. It was determined that droplet ignition would
occur at approximately 21 cm from the inlet face of the reactor, and that
complete combustion of the droplet would occur within 1 additional cm. Thus,
it appeared that a 25-cm long, graded-cell catalytic reactor can achieve high
efficiency at the design condition. The principle non-adiabatic effect analyzed
was the heat loss due to thermal radiation upstream from the inlet face of the
catalytic reactor. Results indicated that approximately a 100 K depression in
the substrate temperature could be experienced on the face; the radiation loss
would affect the initial 2-cm length of the catalytic reactor. Surface reactions
in this region would be suppressed, decreasing the initial heat release and
thus having a de-stabilizing effect on the catalytic reactor operation. Since
the length of the largest cells is more than twice the length influenced by
upstream radiation it is believed that reactor stability will be preserved.
Based upon analyses of several reactor geometries at the design condition
and near-design condition, the final reactor was designed to be a graded-cell
device constructed from Corning MCB-12, zirconia-spinel substrate. It would be
constructed from 5-cm, 2.5-cm, and 17.5-cm lengths of 0.63-cm, 0.32-cm, and
0.16-cra, diameter cell elements, respectively. The catalyst is a proprietary
^formulation applied by UOP. Analyses indicate that the device should impose
approximately a 4 percent pressure loss. The reactors will be assembled by
Acurex and instrumented with high temperature thermocouples embedded in a
dedicated substrate channel along the reactor centerline.
Status
The items described in the previous two sections are the primary components
of the test rig. The fuel injection system has been fabricated; the catalytic
reactors have not been delivered because of problems in substrate availability.
It has become clear that the availability of high temperature substrates for
catalytic reactors can become the pacing factor in conducting future catalytic
combustion programs. The test program will be initiated when the reactors have
been obtained. The initial test phase is an evaluation of the catalytic reactor
over a range of test parameters including: inlet air temperature — 600 K, 500 K1
test section pressure — 600 kPa, 900 kPa; reference velocity — 10 to 30 m/s;
532
-------
fuel air ratio — 0.016 to 0.026. The fuel injector will be operated to
produce the two ranges of droplet sizes described above over these ranges of
conditions. The second test phase is an evaluation of two alternative reactor
designs over a subset of the previously described test conditions. The third
test phase is an evaluation of the catalytic reactor over a range of rich
fuel-air ratios. Test results from these tasks will be presented as available
in the oral presentation.
533
-------
REFERENCES
1. Jasuja, A. K.: Atomization of Crude and Residual Fuel Oils. Journal
of Engineering for Power, Vol. 101, pp 250-258, April 1979.
2. Longwell, J. P. and M. A. Weiss: Mixing and Distribution of Liquids
in High Velocity Air Streams. Industrial and Chemical Engineering,
Vol. 45, 1953.
3. Tacina, R. R.: Experimental Evaluation of Fuel Preparation Systems
for an Automotive Gas Turbine Catalytic Combustor, NASA TM-78856, 1977.
4. Chu, E.: Catalytic Reactor Design for Combustion of Partially Vaporized
No. 6 Fuel Oil. Acurex Project 7543 Report, 1979.
534
-------
0.54 cm dia-
a) VIEVV FROM UPSTREAM
AIR FLOW
FUEL SUPPLY SITE
7-deg DIFFUSbR
b) VIEW FROM DOWNSTREAM
Figure 1 Multiple Venturi Fuel Injector
80-3-87-1
535
-------
100
80
o
LU
N
a:
£
LU
O
DC
LU
Q.
60
INITIAL DROPLET
SIZE (MICRONS)
8 12
AXIAL DISTANCE — cm
16
Figure 2 No. 6 Fuel Oil Droplet Evaporation Characteristics
20
536
80-3-87-2
-------
INLET CONDITIONS
LEGEND
P = 6X 103 Pa
T = 600 K
Vf = 20 mis
50 PERCENT FUEL VAPORIZATION
100
1200 -
1000
50
I
o>
Ui
LO
800 -
600
0 L
I'.i GASI.OUS I ULI.-AIH HAIH )
TW-SU8STRATE TEMPERATURE
Tg-GAS TEMPERATURE
rd-DROPLET RADIUS
T]-FUEL CONVERSION EFFICIENCY
1 15
10 15
POSITION IN REACTOR — cm
- 0.012
- 0.008
- 0.004
CO
o
o
10 I
CO
0
0
25
O
00
I
-^
0)
u
I
Figure 3 Droplet Burning History in Catalytic Reactor
-------
EVALUATION OF A NOVEL LEAN BURN
LOW NO CATALYTIC COMBUSTION CONCEPT
/\
By:
E. K. Chu, R. Chang, and H. Tong
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
ABSTRACT
An experimental program to evaluate the feasibility of a lean burn low
NOX catalytic combustion concept is being conducted. Two catalytic reactor
configurations were designed and tested over a matrix of inlet conditions
representative of gas turbine operations. The fuel tested was a doped diesel
fuel with a nitrogen content of 0.5 weight percent of the fuel. The model
fuel nitrogen was pyridine. One of the design configurations was found to
have low NOX emissions. However, CO emissions were also found to be very high.
538
-------
INTRODUCTION
The objective of the present study is to examine a low NOY, lean
A
burn, catalytic combustion concept for fuels containing fuel bound ni-
trogen. Under lean burn conditions, fuel NO formation predominates
A
and thermal NOV is subdued. Suppression of fuel NOV formation therefore
A A
becomes a primary concern.
The present concept is based on findings by various investigators
(References 1 and 2) that NOX in combustion effluents can actually be
reduced by the injection of hydrocarbon-oxygen mixtures in one case and
NH., in another. It seems that the presence of various intermediate
species of reaction (CO, CHV, CN, NHJ simultaneously with N0¥ will some-
A A A
times reduce NOV to N« given favorable temperatures and residence times.
A t
Further, the presence of a solid surface also tends to enhance such reduc-
tion of NO as revealed by coal combustion studies (References 3 and 4)
A
and a laboratory scale study by DeSoete (Reference 5). In the present study
nitrogen doped fuel is introduced into a first stage catalyst monolith where
it is partially converted to intermediate products as well as some NO . The
A
catalytic reaction is then temporarily "frozen" by introducing a small gap
between the first and second stages. This enables the intermediate products
of reaction to interact with the NOX present, hopefully to reduce the NOX to
N2. The gap is then followed by a second set of catalyst monoliths where
fuel conversion is completed.
Since the concept is highly exploratory, the optimum condition,
if any exists, cannot be defined a priori. In the present study, two
reactor arrangements were examined for diesel fuel doped with pyridine
under various conditions by varying the fuel/air ratio and the approach
velocity and monitoring the reactor effluents. If the concept is viable,
there should be some combination favoring low NOV formation.
A
539
-------
METHOD OF APPROACH
All tests were performed in the Acurex subscale catalytic com-
bustion test facility shown schematically in Figure 1. The catalyst
was mounted in a stainless steel test section lined with a ceramic in-
sulating material. Visual observation of the catalyst reactor was made
from the aft end of the test section through a quartz window. Bed tem-
peratures were monitored with thermocouples located at various points
in the catalyst bed. Exhaust gas composition was monitored continuously
with gas analyzers. In a few studies, measurements of NH3 and HCN con-
centrations were also made by absorption in solution and then reading
with ion selective electrodes. No. 2 diesel fuel doped with pyridine
(0.5 wt % N) was used as the test fuel and was introduced as a fine
spray with an air atomized injector into a mixing zone above the com-
bustor.
The test conditions examined were:
Inlet pressure: 1 atm
Inlet temperature: 700°F
Reference velocity: 30-80 ft/s
F/A (fuel/air) ratio: 0.02-0.035 (by weight)
Two catalyst arrangements were examined. The first (Figure 2)
consisted of two large cell monoliths (2 inch diameter, 1 inch thick,
9 cells/sq inch) and one mediam cell monolith (2 inch diameter, 1 inch
thick, 16 cells/sq inch) in series as the front section followed by a
rear section made up of two small cell monoliths (2 inch diameter,
1 inch thick, 200 cells/sq inch). A gap of 1-1/2 inches separated the
two sections. In the second arrangement (Figure 3), two large cell
monoliths (2 inch diameter) made up the front section and two medium
cells (2-1/2 inch diameter) and one small cell monolith (2-1/2 inch
diameter) the rear section. A gap of 2-1/2 inches separated the two
sections.
540
-------
RESULTS AND DISCUSSION
Results are summarized in Tables la, b and 2.
Tables la and b summarize results with catalyst arrangement 1.
Table la is a baseline study using pure No. 2 diesel. As expected,
NO., emissions were low in the absence of fuel nitrogen. Table Ib
assesses the same catalyst arrangement with the pyridine doped diesel.
The first four points were preliminary studies to assess the general
stability of the system. At the fuel/air ratios and approach veloci-
ties used, combustion was stable and efficient with low unburned hydro-
carbon (UHC) and CO emissions. Fuel NOX conversion, however, also
approached 100 percent.
The next series of points summarizes the effect of changing the
approach velocity at more or less constant fuel/air ratios. Again,
combustion was complete, but NO conversions were high, averaging about
/\
75 percent and remaining rather steady as the velocity changed.
The fuel/air ratio was then changed gradually at constant approach
velocity. As the fuel/air mixture got leaner, the NO conversion also
A
decreased, but with a corresponding rise in CO and UHC emissions. This
indicated that combustion was incomplete. Unfortunately, no measure-
ments were made on the presence of other nitrogen species, in particular
ammonia and cyanide. At the temperatures encountered in the combustor,
it is expected that all the pyridine would be dissociated.
The second catalyst arrangement was then studied and results
summarized in Table 2. The major differences in this arrangement com-
pared to the first are:
• The gap is wider (2-1/2 inches vs. 1-1/2 inches)
• The front section is less reactive (two large cell monoliths
vs. two large plus one medium cell monolith)
• The back section is more reactive (larger diameter monoliths)
The results show the same general trend: NOX emmissions were
low at low fuel/air ratios. This time, however, some measurements were
also made on NH^ and HCN. For the points where NO emissions were low,
541
-------
NH3 and HCN emissions were also low. Since NH3, HCN, NOX, and N2 are
the primary nitrogenous species formed in most combustion systems, our
findings above are strong indications that the fuel nitrogen has probably
been converted to gaseous nitrogen. The combustion in the second arrange-
ment with the wider gap, however, seemed less stable than in the first
arrangement. Bed temperatures fluctuated more in this arrangement and
it was more difficult to maintain a steady test condition. In fact,
we had tried an earlier series of test runs with a catalyst arrangement
similar to Figure 2 except with a gap of 5 inches. The combustion was
so unstable that we failed to maintain a constant test condition. These
results are therefore not reported.
In both arrangements, the systems were not operating adiabatically
as evidenced by the low temperatures in the catalyst bed. At a fuel/air
ratio of 0.03 and 0.024, adiabatic bed temperatures should approach
2800°F and 2400°F respectively. The gap temperatures (measured in ar-
rangement 1 and estimated in arrangement 2 due to thermocouple breakdown)
should be taken as rough estimates only. Radiation from the bed on both
sides of the thermocouple may cause higher temperature readings for the
gas phase in the gap.
In general, however, a trend seems to be established in the tests
run so far. This is summarized in Figure 4. Fuel nitrogen to NOX de-
creased with fuel/air ratios. Catalyst arrangement 1 had a higher fuel
nitrogen conversion at the same fuel/air ratio compared to arrangement 2
as evidenced by the shift in the curves. This is probably due to the
more reactive front section in arrangement 1 (more surface area) indi-
cating that too much fuel conversion in the first section may not be
desirable. In fact, to achieve comparable NO emissions, arrangement 1
A
required a higher approach velocity (70 ft/s versus 40 ft/s) which low-
ers fuel conversion and lower fuel/air ratios which lowers bed tempera-
tures.
The high heat loss in the reactor beds also makes the temperatures
in the system generally lower than a comparable adiabatic system. For
an adiabatic system, lower front section activity (less surface area or
lower catalyst activity) may be more desirable for a given fuel/air
542
-------
to achieve low NOX emissions. On the other hand, the front section must
be sufficiently active to provide some fuel conversion. The interac-
tions of fuel/air ratio, approach velocity, catalyst activity, and bed
temperature, however, are difficult to determine from the present data.
The effect of gap width cannot be assessed since in both arrangements
studied, low NOX conditions were achieved. Although it may be argued
that the second arrangement had lower gap residence time due to lower
approach velocities and longer gap distance, the front section catalyst
activities were not the same and it is difficult to uncouple all the
effects with the current data. The back section of the catalyst bed
following the gap is essentially a cleanup section where the fuel and
intermediates are fully converted. The high CO and UHC breakthrough
seems to indicate that a more reactive second section is required for
efficient cleanup under conditions favoring low NOV.
/\
CONCLUSIONS AND RECOMMENDATIONS
Results from our series of studies indicate that we may indeed
have a viable concept for low NO production from fuel containing fuel
A
bound nitrogen. However, a large number of parameters affect system
performance. The limited amount of data we have at present do not allow
us to uncouple all the viable interactions among key variables. To
design a full scale system for gas turbine applications, we need to be
able to scale up from bench scale results as well as to build in the
flexibility of operation required in order to respond to start up,
changing loads and shut down. More information will therefore be needed
around design conditions to see how various parameters affect the results.
It is also imperative that more tests be made to confirm our present
findings.
There are a number of variables which will most likely affect
our system and need to be considered in greater detail, in particular
around design conditions.
0 Fuel/air ratios —affects system temperature and reactant
concentration
543
-------
• Approach velocity - affects residence time and therefore
conversion as well as temperature
• Gap distance - determines residence time available for gas
phase reaction in the absence of catalyst
• Catalyst activity —controls temperature and conversion,
particularly important in the front section. A sufficiently
active rear section is also needed for efficient cleanup
under all conditions
• System heat balance - how well the system approaches adia-
batic conditions may determine how the other variables should
vary since heat loss affects system temperatures
t Others - pressure, fuel type and nitrogen content may all
affect system performance.
Given the large number of parameters that have to be examined,
a more comprehensive test program will be required to provide sufficient
data for performing a full scale design.
544
-------
REFERENCES
1. Myerson, A. L., "The Reduction of Nitric Oxide in Simulated Com-
bustion Effluents by Hydrocarbon Oxygen Mixtures," Fifteenth
Symposium (International) for Combustion, The Combustion Institute,
Pittsburgh, p. 1085, 1974.
2. Banna, S. M. and Branch, M. C., "NH3-NO Reaction in the Post-Flame
Region of a Flat Flame Burner," Paper No. 78-50, Fall Meeting, Wes-
tern States Section, Combustion Institute, Lagunne Beach, CA,
October 1978.
3. Bee>, J. M., et al., "NO Reduction by Char in Fluidized Combustion."
4. Kelly, J., Personal communication, Acurex Corporation, April 1980.
5. DeSoete, G. G., "Mechanisms of Nitric Oxide Reduction on Solid
Particles," paper presented at the Fifth EPA Fundamental Combustion
Research Workshop, Newport Beach, January 1980.
545
-------
TABLE la. COMBUSTION RESULTS ON CATALYST ARRANGEMENT 1 (FIGURE 2), DIESEL
F/A TA V
(n/s)
0.034 200 30
0.035 200 30
Temperature Distribution*
Front Section Gap Back Section
T1 T2 TG T3 T4
1325 1444 1065 1565 2196
1337 1465 1090 1589 2245
Emissions
NOX (ppm) NH3 + HCN UHC CO CO,, Og
Theoretical /Actual Conversion (ppm) (ppm) (ppm) (%) (%)
N.A.V20 to
N.A.f/20
IXN1"1" -- T 50 7 11.8
2 50 7 11.6
Ul
*See Figure 2 for actual location of temperature measured
N.A. = not applicable
t £XN = NOX + HCN + NH3
F/A = fuel ratio, TA = theoretical air, V = approach velocity
-------
TABLE Ib. COMBUSTION RESULTS ON CATALYST ARRANGEMENT 1 (FIGURE 2), DIESEL + PYRIDINE.
F/A
0.029
0.029
0.029
0.03
0.034
0.033
0.033
0.034
0.035
0.03
0.025
0.023
0.021
0.024
TA
(*)
240
240
240
230
200
210
210
200
200
230
280
300
330
290
V
(ft/s)
40
36
36
53
53
55
63
72
82
67
72
72
70
69
Front
Tl
..
1471
1377
1404
1221
1227
1243
1271
1275
1284
1225
1172
1114
1163
Temperature Distribution*
Section Gap Back
T2 TG T3
1576 (1259)** 1647
1681 1560 1752
1786 1831 1838
1679 1548 1634
1695 1650 1644
1744 1769 1691
1806 1938 1765
1802 1995 1750
1827 1867
1637 1628 1681
1454 1476 1552
1376 1363 1432
1436 1407 1495
Section
T4
2266
2197
2270
2227
2224
2183
2220
2229
2232
2393
2010
1710
1555
1630
NOX (ppm)
Theoretical /Actual
294/300
294/300
294/300
310/300
350/260
334/250
334/260
350/250
350/250
310/260
252/235
235/190
215/90
244/80
Emissions
MH3 + HCN
Conversion (ppm)
102
102
102
97
74
75
78
71
71
84
93
81
42
33
UHC
(ppm)
3
--
6
5
30
20
13
7
7
4
3
31
1500
3000
CO
(ppm)
50
30
30
40
50
40
50
30
60
50
70
1540
1030
1510
co2
(*)
6.7
6.6
6.8
6.9
7.5
7.4
7.5
7.2
7.2
7.5
6.3
5.5
1.4
2.3
°2
(«
12.4
12.5
12.0
12.0
11.0
11.5
11.5
11.5
11.5
10.7
12.3
13.5
18.0
17.4
Ui
*See Figure 2 for actual location of temperature measured
**
No actual gap temperature was measured, average of T^.T- taken
F/A = fuel/air ratio, TA = theoretical air, V = approach velocity
-------
TABLE 2. COMBUSTION RESULTS ON CATALYST ARRANGEMENT 2 (FIGURE 3), DIESEL + PYRIDINE
F/A
0.032
0.031
0.027
0.026
0.024
0.022
0.027
0.023
0.025
TA
215
222
256
265
287
3H
256
300
276
V
(ft/s)
37
38
43
40
44
46
59
65
61
Temperature Distribution*
Front Section Gap Back Section
Tl T2 TG~ T3 T4
1374
1380
1338
1329
1300
1314
1428
1386
1410
1475
1475
1440
1412
1368
1392
1519
1467
1504
(1518) 1561
(1517) 1559
(1476) 1513
(1440) 1469
(1399) 1430
(1419) 1447
(1554) 1590
(1491) 1516
(1532) '1560
2097
2003
1860
1781
1687
1711
1892
1760
1846
NOX (ppm)
Theoretical /Actual
326/300
320/290
276/260
274/220
247/34
226/44
276/240
236/19
256/145
Emissions
NH, + HCN
% i
Conversion (ppm)
92
91
96
80
14
30 4.2 + 18.7
87
17 2.8 + 18.9
61 2.6 + 8
UHC
(ppm)
5
4
2
12
19
11
5
25
3
CO C02
(ppm) («)
>1200 7.1
>1200 7.0
510 6.4
5.9
-8000 3.5
8300
440 6.4
-9000 2.7
-4100 5.5
°2
(X)
••M^^^OK
11.5
11.5
12.0
13.5
15.5
15.2
12.6
16.5
13.5
Ui
*-
00
*See Figure 3 for actual location of temperature measured
No actual gap temperature was measured, average of ^'^3
F/A = fuel/air ratio, TA = theoretical air, V = approach velocity
-------
Ui
£-
vo
Control
console
Rotameter
Flow control
valve
Pressure
regulator
Air, argon,
or nitrogen
supply
Oxygen
Superheated
steam
Bypass
Oxidizer
hpnt.pr
Gaseous
fuel dopants
Preheat
temperature
Exhaust
Burst
diaphram
s
Liquid
-o fuel
oil
Chamber
heater
Pressure
tap
Exhaust
I Ml 11
III Mil III
nuiimii
Pressure tap
Tubular |
test y w./w.o. catalyst
section}
Water-cooled
probe
Heated
sample line
Quartz
viewport
Figure 1. Catalytic combustion test facility.
-------
Fuel + air
}large cells (2" dia, 1" thick, 9 cells/sq inch)
Medium cells (2" dia, 1" thick, 16 cells/sq inch)
Gap 1-1/2"
'"] Small cells (2" dia, 1" thick, 200 cells/sq inch)
•J
Sample probe
Figure 2. Schematic of catalyst monolith arrangement 1
Fuel + air
V
f
— •
Jr
L
«,
I [
11
|l
>•>
}
J
Large cells (2" dia)
Gap 2-1/2"
Medium cells (2-1/2" dia)
Small cells (2-1/2" dia)
Sample probe
Figure 3. Schematic of catalyst monolith arrangement 2.
550
-------
100
90
^•v
X
§ 80
o
^ 70
OJ
3
4-
•5 60
o
'? 50
OJ
>
c
o
c
Ol
u
30
20
10
0.02
*.i
X Test arrangement 1 (Figure 2)
Test arrangement 2 (Figure 3)
1
1
1
1
0.024 0.028 0.032
Fuel/air
0.036
Figure 4. Fuel nitrogen conversion to NOX as a function
of fuel/air ratio.
551
-------
IERL-RTP-1077
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
REPORT NO.
1PA-600/9-80- 035
3. RECI
4. TITLE ANDSUBTITLE
Proceedings: Fourth Workshop on Catalytic Combustion
(Cincinnati, OH, May 1980)
5. REPORT DATE
August 1980 Issuing date.
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
John P. Kesselring, Compiler
8. PERFORMING ORGANIZATIO!
IT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Corporation
485 Clyde Avenue
Mountain View, California 94042
10. PROGRAM ELEMENT NO.
EHE624A
11. CONTRACT/GRANT NO.
68-02-3122
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANJD PERIOD C(
Proceedings; 9/78-5/80
COVERED
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer is G. Blair Martin. Mail Drop 65, 919/
541-2235. EPA-600/7-79-038 documents the Third Workshop ana contains summaries
of the First and Second Workshops.
16. ABSTRACT
The ppQgge^jjjgg document the major presentations at the Fourth Workshop
on Catalytic Combustion, held in Concinnati, OH, May 14-15, 1980. Sponsored by
the Combustion Research Branch of EPA's Industrial Environmental Research Labor-
atory (Research Triangle Park) , the workshop served as a forum for the presenta-
tion of results of recent research in the areas of catalyst performance, components
and applications of catalytic combustion systems , and the use of alternative fuels in
catalytic combustors. The workshop provided industrial, university, and govern-
ment representatives with the current state of the art in the application of catalyst
systems for pollution control and performance improvement. Applications include
fire tube and water tube boilers, and gas turbines for utility, industrial, automotive,
and aircraft systems.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Combustion
Catalysts
Boilers
Gas Turbines
Pollution Control
Stationary Sources
Catalytic Combustion
13 B
2 IB
07D
13A
13 G
3. DISTRIBUTION STATEMEN1
Release to Public
___________^——
EPA Form 2220-1 (9-73)
Unclassified
558
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
552
U.S. GOVERNMENT PRINTING OFFICE: 1980--657-165/0161
-------
|