EPA-650/2-73-014
August 1973
Environmental Protection Technology Series
:-:•:•>;•:•:• •:•:•:•:•:
-------
EPA-650/2-73-014
INVESTIGATION OF SURFACE COMBUSTION
CONCEPTS FOR NOX CONTROL
IN UTILITY BOILERS
AND STATIONARY GAS TURBINES
by
W. U. Roessler, E. K. Weinberg,
J. A. Drake, H. M. White, andT. lura
Urban Programs Division
The Aerospace Corporation
El Segundo, California 90245
Aerospace Report No. ATR-73(7286)-2
Grant No. R-801490
Program Element No. IA20I4
EPA Project Officer: David W. Pershing
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
August 1973
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
11
-------
ACKNOWLEDGMENTS
Appreciation is acknowledged for the guidance and continued assistance
provided by Mr. D. W. Pershing of the Environmental Protection Agency,
Control Systems Laboratory, who served as EPA Project Officer for this study.
The following technical personnel of The Aerospace Corporation made
valuable contributions to the study performed under this grant.
J. A. Drake
E. K. Weinberg
H. M. White
LJ, Mr /
W. U. Roessler, Manager
Surface Combustion Study
Approved
\^<^f^i^-
Toru lura
Associate Group Director
Environmental Programs
Group Directorate
/Joseph Meltzer
^-'Group Director
Environmental Programs
Group Directorate
ill
-------
ACRONYMS, TERMS, AND ABBREVIATIONS
AEC
AFAPL
AGA
AiResearch
American Lava
APU
CFR
Chrysler
CO
Corning
DuPont
EPA
EVC
FID
Ford
General Electric
General Motors
HC
Hold en
LEL
LPG
Matthey Bishop
Atomic Energy Commission
Air Force Aero Propulsion Laboratory
American Gas Association
AiResearch Manufacturing Company
American Lava Corporation
auxiliary power unit
Cooperative Fuel Research
Chrysler Corporation, Amplex Division
carbon monoxide
Corning Glass Works
E.I. DuPont de Nemours & Co., Inc.
Control Systems Laboratory, Environmental
Protection Agency
externally vaporizing combustor
flame ionization detector
Ford Motor Company
General Electric Company, Space Division
General Motors Corporation
hydrocarbons
Holden Company
lower explosive limit
liquified petroleum gas
Matthey Bishop, Inc.
-------
ACRONYMS, TERMS, AND ABBREVIATIONS (Cont)
NASA/Lewis
NDIR
NO
x
Perfection
Pratt & Whitney
Selas
TEL
National Aeronautics and Space Administration,
Lewis Research Center
nondispersive infrared
oxides of nitrogen (NO plus NO^)
Perfection Products Company
Pratt & Whitney Aircraft Division, United
Aircraft Corporation
Selas Corporation of America
tetraethyl lead compound (one of several lead
compounds used in gasoline)
VI
-------
CONTENTS
ACKNOWLEDGMENTS iii
ACRONYMS, TERMS, AND ABBREVIATIONS
1. SUMMARY 1-1
1. 1 Study Results 1-1
1. 2 Conclusions 1-4
2. INTRODUCTION 2-1
3. STATE-OF-THE-ART REVIEW OF NONCATALYTIC
SURFACE COMBUSTION DEVICES 3-1
3. 1 Introduction 3-1
3.2 State-of-the-Art Review 3-6
3.2.1 American Gas Association 3-6
3.2.2 Bureau of Mines 3-8
3.2.3 Burnham Corporation 3-11
3.2.4 General Electric Company 3-18
3.2. 5 Selas Corporation of America 3-2Q
3.2.6 University of Wisconsin 3-34
3.2.7 Other Organizations 3-41
REFERENCES 3-45
4. STATE-OF-THE-ART REVIEW OF
CATALYTIC DEVICES 4-1
4. 1 Introduction 4-1
4.2 Catalyst Features 4-2
4.Z.I Catalytic Process 4-2
4.2. 2 Catalyst Requirements 4-2
4.2. 3 Typical Catalysts 4-4
4.2.4 Substrates 4-9
VI1
-------
CONTENTS (Cont)
4.3 Catalyst Degradation 4-16
4.3. 1 Catalyst Poisoning 4-18
4.3. 2 Alumina Phase Change 4-21
4. 3. 3 Thermal Shrinkage 4-21
4.3.4 Thermal Differential Expansion 4-22
4. 3. 5 Melting 4-22
4. 3. 6 Vibration Effects 4-22
4.4 Catalytic Converters and Combustors 4-23
4.4. 1 Automotive Oxidation Catalysts 4-23
4.4.2 Tail Gas Abatement Systems 4-25
4.4.3 Low-Temperature Catalytic Combustors 4-39
American Gas Association 4-40
EPA, Control Systems Laboratory 4-41
Matthey Bishop, Inc 4-44
4.4.4 High-Temperature Catalytic Combustors 4-45
Air Force Aero Propulsion Laboratory 4-45
Engelhard Industries 4-51
EPA, Control Systems Laboratory 4-55
NASA/Lewis 4-56
REFERENCES 4-59
5. EVALUATION OF POROUS-PLATE AND CATALYTIC SURFACE
COMBUSTORS FOR LARGE UTILITY BOILERS 5-1
5.1 Porous-Plate Surface Combustors 5-1
5. 1. 1 Introduction 5-1 '
5.1.2 NO Emission Level Target 5-3i
J\.
5. 1.3 Boiler Design Requirements 5_3
5.1.4 Burner Characteristics 5-^8
5.1.5 Conclusions 5-15
5. 2 Catalytic Combustors 5-17
REFERENCES 5-19
Vlll
-------
CONTENTS (Cont)
6. EVALUATION OF CATALYTIC AND NONCATALYTIC SURFACE
COMBUSTION CONCEPTS FOR STATIONARY GAS TURBINES. . 6-1
6. 1 General Considerations 6-1
6.2 Gas Turbine Operating Characteristics 6-2
6. 3 Gas Turbine Emissions 6-5
6.4 NO Emission Goals for Large Power Installations 6-13
3C
6. 5 Fuel Preparation and Preignition Problems 6-15
6.6 Catalytic Combustors 6-20
6.6. 1 Catalytic Combustor Design Parameters 6-20
6.6.2 Potential Problem Areas 6-23
6.7 Porous-Plate Surface Combustors 6-27
6.7. 1 Introduction 6-27
6. 7. 2 Discussion 6-27
6. 7. 3 Conclusions 6-30
REFERENCES 6-31
APPENDIXES
A. VISITS AND CONTACTS A-1
A. 1 Organizations Visited A-1
A. 2 Organizations Contacted by Telephone A-2
B. UNITS OF MEASURE--CONVERSIONS B-l
IX
-------
TABLES
3-1 NO and CO Emissions of Conventional and Infrared Burners. . . 3-8
x
3-2 Peak NOX, CO, and HC Concentrations and Flame Tempera- '
tures for Meeker Noninfrared and Ceramic Infrared
Burners --Methane -Air Mixtures ..................... 3-12
3-3 Operating Conditions and NOX Emissions for a Commercial
Boiler with Radiant Burner (Model PB-340 boiler; Burnham
B030 Burner) ................ ... ............... 3-16
3-4 Burnham Burner Cost Data ..................... ... 3-18
4-1 Catalyst Lightoff Temperature- -Decomposition of Nitrous Oxide 4-3
4-2 Relative Reaction Rates- -Catalytic Oxidation of Methane ..... 4-4
4-3 Catalytic Removal of HC and CO--Platinum Group Metal Auto-
motive Catalysts ...................... ...... 4-6
4-4 Base Metal Catalysts- -Industrial Processes .............. 4-7
4-5 Catalytic Removal of HC and CO--Base Metal Automotive
Catalysts ..................................... 4-8
4-6 High Temperature Porous Materials ............... .... 4-10
4-7 Americal Lava Monolithic Substrate Physical Properties ..... 4-12
4-8 Corning W-l Monolithic Substrate Physical Properties ....... 4-15
4-9 DuPont To rvex Monolithic Substrate Physical Properties ..... 4-17
4-10 Emissions from. Commercial Catalytic Heaters at Maximum
Heat Output ................................... 4-42
4-11 Matthey Bishop Catalytic Heater Characteristics (Cataheat
System) ...................................... 4-44
5-1 Characteristics of Existing Boiler for Units 5 and 6 of ! '
Haynes Steam Plant (Los Angeles Department of Water
and Power) .................................... 5_8
6-1 NOX and CO Emissions for Gas Turbines and Steam Power
Plants
-------
FIGURES
3-1 Surface Burner Configurations ..................... 3-2
3-2 Kinetic Data- -Natural Gas Fuel .................... 3-5
3-3 Bureau of Mines Flat-Flame Meeker Burner
Test Setup .................................. 3. 10
3-4 Burnham Dual Fuel Burner Sectional Schematic ......... 3-13
3-5 Burnham Corporation Burner Dimensions and Ratings ..... 3-15
3-6 NOx Emission Concentration vs Percent Flue Gas
Recirculation--320 MW Corner Fired Unit ............. 3-17
3-7 Schematic of Porous -Plate Burner ......... • ......... 3-19
3-8 Alternate Frit Configurations ..................... 3-21
3-9 Porous Radiant Gas Turbine Combustor ............... 3-22
3-10 Conceptual Design of Porous -Plate Burner -Vapor
Generator for Rankine System ..... ................ 3-24
3-11 Effect of Unburned Gas Superficial Velocity and
Equivalence Ratio Upon Flame Temperature ............ 3-26
3-12 Heat Flux Back to the Burner for Propane-Air
Mixtures ................................... 3-27
3-13 Heat Flux Back to the Burner for Gasoline- Air
Mixtures ................................... 3-28
3-14 "Sonicore" Fuel Atomizer ........................ 3-29
3-15 Preliminary Emission Measurements (Not Corrected
for Water Content) ............................. 3-30
3-16 Typical Selas Radiant-Cup Air-Gas Burner ............ 3-32
3-17 NO vs Gas Temperature for a Selas Inspirator
Burner .................................... 3-^4
3-18 University of Wisconsin Burner Test Section ........... 3-35
3-19 University of Wisconsin Fuel and Air Inlet System ........ 3-37
XI
-------
FIGURES (Cont)
3-20 Q/Qmax vs Equivalence Ratio Profiles for Various Mass
Flow Rates 3-39
3-21 NO vs Equivalence Ratio Profiles for Various Mass
Flow Rates 3-40
3-22 Surface Burner Geometry 3-42
4-1 Conversion of CO to Methanol 4-9
4-2 American Lava Stacked and Rolled Corrugated
Structures 4-13
4-3 Corning Substrate Geometries 4-14
4-4 Effect of Lead Additives on Catalyst Efficiency 4-20
4-5 Engelhard Monolithic Catalysts with Improved
Catalytic Thermal Stability 4-24
4-6 Effects of Thermal Aging on Matthey Bishop
AEC3A Catalyst 4-26
4-7 Conversion Characteristics of Base Metal and Noble
Metal Catalysts (General Motors Bench Test
Evaluation) 4-27
4-8 Pressure Drop of Monolithic and Pellet Catalysts 4-29
4-9 n-Heptane Fume Abatement Effectiveness
(115, 000/hr space velocity) 4-30
4-10 n-Heptane Conversion Efficiency; vs Temperature and
Space Velocity--Torvex Straight-Through Honeycomb 4-30
4-11 n-Heptane Conversion Efficiency vs Temperature and
Space Velocity--Torvex Cross-Flow Honeycomb ........ 4-31
4-12 n-Heptane Conversion Efficiency vs Temperature and '
Space Velocity--Granular Catalyst Substrate 4-31
4-13 Lightoff Characteristics--Platinum Catalyst 4-33
4-14 Lightoff Characteristics--Palladium Catalyst 4-34
XII
-------
FIGURES (Cont)
4-15 Palladium Catalyst Performance 4-35
4-16 Performance of Stabilized and Nonstabilized
Catalysts--Nitric Oxide Abatement 4-38
4-17 Performance of Stabilized Aged and Unaged Catalysts vs
Space Velocity--Nitric Oxide Abatement 4-38
4-18 Catalytic Space Heater Schematic 4-39
4-19 Hydrocarbon Emissions vs Fuel Flow Rate — Catalytic
Space Heater 4-43
4-20 Aircraft Gas Turbine Emission Goals 4-47
4-21 Catalytic Combustor Test Setup--Air Force Aero
Propulsion Laboratory 4-48
5-1 NO Formation Rate 5-2
5-2 Predicted NOX Emissions vs Mode of Operation for
Scattergood No. 3 Unit 5-4
5-3 Kinetic NO Formation for Combustion of Natural Gas
at Stoichiometric Mixture Ratio--Atmospheric
Pressure 5-5
5-4 Haynes Supercritical Steam Boiler--Units 5 and 6 5-6
5-5 Burning Velocities of Methane-Air Mixtures vs
Reciprocal Temperature 5-10
5-6 Porous Plate Burner Configuration Schematic 5-12
5-7 Computed Heat Flux to the Porous-Plate Burner as a
Function of Burning Velocity 5-13
6-1 Performance Characteristics of Simple-Cycle
Gas Turbines 6-3
6-2 Performance Characteristics of Regenerative-Cycle
Gas Turbines 6-3
6-3 Simple-Cycle Gas Turbine Combustor Inlet Temperature
vs Compressor Pressure Ratio 6-4
xui
-------
FIGURES (Cont)
6-4 Regenerative-Cycle Gas Turbine Combustor Inlet
Temperature vs Turbine Inlet Temperature--Regenerator
Effectiveness 0. 90 6-6
6-5 Regenerative-Cycle Gas Turbine Combustor Inlet
Temperature vs Turbine Inlet Temperature--Regenerator
Effectiveness 0. 70 6-7
6-6 Predicted Air-Fuel Ratio vs Turbine Inlet
Temperature (Natural Gas) 6-8
6-7 Ford Experimental Externally Vaporizing
Combustor (EVC) 6-11
6-8 Permissible Gas Turbine NOX Concentrations to Meet
Rule 67 of the Los Angeles Pollution Control District
(315 MW Power Output) 6-14
6-9 Autoignition Temperatures for Methane and
Kerosene vs Pressure 617
6-10 Effect of Temperature on Lower Limits of
Flammability of 10 Paraffin Hydrocarbons in Air
at Atmospheric Pressure . 6-18
6-11 Ignition Delay Time vs Inlet Temperature at
1 Atmosphere 6-19
6-12 General Electric Air-Cooled Burner No. 106
During Fabrication 6-29
xiv
-------
SECTION 1
SUMMARY
Based on a review of the state of the art of surface combustion concepts, an
assessment has been made of their applicability to large utility boilers and
stationary gas turbines.
The term "surface combustion, " as used in this study, refers to those con-
cepts in which combustion occurs in close proximity to a solid surface. The
interest in these concepts arises from their low-emission characteristics, in
particular NO , which result from the combustion process occurring at
Ji.
reduced temperatures. This process can be either noncatalytic or catalytic.
In the noncatalytic surface combustor, a fraction of the heat of combustion is
immediately transferred from the flame layer to the adjacent solid surface.
This heat is then removed from the burner surface by means of radiation and/
or conduction into imbedded cooling tubes. This heat transfer mechanism
results in gas temperatures below the adiabatic flame temperature. In the
catalytic combustor, reduced gas temperatures are achieved by operation
with very lean fuel-air mixtures. The catalyst serves the function of pro-
moting chemical reactions, which, under these operating conditions, would
otherwise occur too slowly for efficient low-emission burning.
1.1 STUDY RESULTS
The results of the investigation are summarized as follows:
1. Little information is available on the design and operational
characteristics of catalytic combustors, except for some
exploratory test work conducted by Engelhard Industries and by
the Control Systems Laboratory of the Environmental Protection
Agency. In these tests, extremely low NOX emission levels have
been achieved with concurrently low HC and CO emissions.
1-1
-------
2. Emission test data obtained by Engelhard Industries on small-size
catalytic combustors, operated with propane and gasoline, indi-
cate unde tec table NOX, less than 60 ppm CO, and less than
10 ppm HC. These values are for fresh catalyst materials and
do not include potential long-term performance degradation effects
due to catalyst poisoning and high-temperature operation.
3. Further reduction of the HC and CO emissions might be achieved
by increasing the operating temperature of the catalytic combustor
and the residence time of the combustion products. Current state-
of-the-art technology limits the temperature of catalysts to about
2400°F, but advanced monolithic or pellet substrates might be
developed which have a higher temperature capability. At these
temperature levels the NOX emissions are still low.
4. The available information indicates that catalytic combustors
might be applicable to both existing and new stationary gas tur-
bines. However, a number of potential problem areas require
resolution before a complete assessment of the feasibility of
these devices can be made. Systematic and carefully planned
experimental and theoretical investigations are recommended to
provide the technical information needed for a meaningful evalu-
ation of catalyst durability, specific heat release rate, emission
characteristics, mechanical integrity, catalyst materials opti-
mization, ignition characteristics, and pressure drop.
5. It is unlikely that catalytic combustors can be-incorporated
economically into existing utility boiler installations, which pper-
ate with little excess air. Since the catalyst requires a lean fuel-
air mixture, several combustor stages would be needed with fuel
addition and heat rejection between stages. A change in the boiler
heat transfer mechanism from primarily radiative to primarily
convective would be involved -which, in turn, would require a com-
plete redesign of the boiler and its air induction system. The
concept might be feasible for newly designed power plants, but
additional preliminary design studies would have to be carried out
to resolve this issue.
6. No information is presently available regarding the application of*
noncatalytic surface combustors, of the flow-through or direct-
fired types, to large utility boilers. However, theory and limited
small-scale burner experiments indicate that a surface combustor
has the potential of achieving NOX emission levels appreciably i
lower than those obtained in the best current state-of-the-art
large utility boilers. Due to the large surface area requirements
of these combustors, novel design configurations must be utilized
to accommodate the burner within the boiler envelope. Although
1-2
-------
the porous-plate surface combustor concept might be technically
feasible, it lacks attractiveness because of limited fuel com-
patibility and complex problems of packaging. The porous-plate
combustor may not be the optimum concept for removing heat
from the flame zone.
7. Porous-plate surface combustors require large burner surface
areas and volumes. For example, for an equivalence ratio of
unity, a system pressure of 1 atmosphere, and a gas temperature
of 3000°F, the cooling capacity of porous-plate burners is about
45, 000 Btu/hr-ft2, resulting in a surface area requirement of
about 20, 000 ft2 for a typical utility plant (300 MW output).
8. NOx emissions below 40 ppm are predicted for noncatalytic sur-
face combustors operated at typical steam boiler conditions and
maximum gas temperatures of 3000°F. Preliminary test data by
General Electric are in reasonable agreement with the NOX pre-
diction. The CO emissions from, those tests are of the order of
500 ppm.
9. Sintered porous metal surface combustors incorporating imbed-
ded cooling tubes represent a design approach which merits
further evaluation for low-pressure gas turbines of the type
projected for automotive and highly regenerated stationary gas
turbines, and possibly combined gas turbine/steam turbine
cycles. However, because of the high heat loads and surface
temperatures resulting from the high system pressures of typ-
ical simple-cycle gas turbines, porous-plate surface combustors
are not considered feasible for this particular engine type.
10. Porous ceramic surface combustors are not considered feasible
for use in either simple-cycle or regenerative-cycle gas tur-
bines, because acceptable burner surface and gas temperatures
require the application of imbedded cooling tubes. However,
because of the large differences in thermal coefficients of expan-
sion of ceramics and metals and the consequent excessive thermal
stress, incorporation of cooling tubes is not currently possible.
11. Before a complete assessment of porous-plate surface combustors
can be made, substantial research and development efforts are
required to demonstrate satisfactory burner operation for long
periods of time in an environment typical of large utility boilers
and gas turbines, without clogging of the micron-size pores.
12. To prevent fuel coking and local overheating of the burner mate-
rials, catalytic and porous-plate surface combustors are limited
to premixed gaseous and prevaporized, premixed distillate fuels.
The sulfur and metal content of the fuel must be maintained at
1-3
-------
very low, but as yet undetermined, levels to minimize poisoning
of the active catalyst and clogging of the micron-size porous
burner pores.
13. In prernixed/prevaporized systems, preignition of the combus-
tible mixture presents a potential problem area, if the mixture
temperature is higher than the autoignition temperature of the
fuel. With methane, preignition could occur in highly regenerated
gas turbines (with pressure ratios of the order of 4 to 6) and high
pressure ratio simple-cycle gas turbines, unless special care is
exercised in the design of the fuel prevaporization and mixing
chamber. For kerosene-air mixtures, the autoignition temper-
atures and the ignition delay times are low enough that preignition
could occur in all gas turbines. In steam boilers, the auto-
ignition temperatures of methane and kerosene are above the typ-
ical air preheat temperatures of about 650°F. Therefore, pre-
ignition does not pose a problem in well designed induction
systems.
14. Either monolithic or pellet type substrates are considered appli-
cable to catalytic combustors. The catalyst and substrate com-
position and manufacturing processes are considered proprietary
by the manufacturers. Currently, platinum group metals are
favored for automotive oxidation catalysts, but base metal or
promoted base metal formulations might be developed for use in
second-gene ration catalysts. The noble metal content of the pro-
moted catalyst formulation is expected to be sufficiently low so
as to have no significant impact on the world platinum group
metal supply and demand balance.
1.2 CONCLUSIONS
In summary, it is concluded that the catalytic combustor concept looks prom-
ising for application in gas turbines. However, comprehensive experimental
and theoretical investigations are required to provide the information needed
for a meaningful assessment of this concept. Catalytic combustors are not
considered feasible for use in existing steam boilers, but might be applicable
to newly designed installations.
The sintered metal porous-plate combustor concept with imbedded cooling
tubes, while technically feasible for steam boilers and possibly for low pres-
sure ratio gas turbines, lacks attractiveness because of packaging problems
and the requirement of gaseous and distillate fuels.
1-4
-------
SECTION 2
INTRODUCTION
The purpose of this report is to present a review of the state of the art of
surface combustion concepts and an assessment of their applicability to large
utility boilers and stationary gas turbines.
The term "surface combustion, " as used in this study, refers to those con-
cepts in which combustion occurs in close proximity to a solid surface. The
interest in these concepts arises from their low emission characteristics, in
particular NO , which result from the combustion process occurring at re-
j£
duced temperatures. This process can be either noncatalytic or catalytic. In
the noncatalytic surface combustor, a fraction of the heat of combustion is
immediately transferred from the flame layer to the adjacent solid surface.
This heat is then removed from the burner surface by means of radiation and/
or conduction into imbedded cooling tubes. This heat transfer mechanism
results in gas temperatures below the adiabatic flame temperature. In the
catalytic combustor, reduced gas temperatures are achieved by operation
with very lean fuel-air mixtures. The catalyst serves the function of pro-
moting chemical reactions, which, under these operating conditions, would
otherwise occur too slowly for efficient low-emission burning.
Surface combustion devices have been in wide use for industrial heating and
chemical process applications, and low-temperature catalytic combustors
have long been used for space heaters. Because of their potential for achiev-
ing very low NOX emissions, these devices have recently received consider-
able attention with respect to application in gas turbines as well as Rankine
cycle engines. This study will review these recent activities to determine
whether a potential exists for applications in utility boilers and stationary gas
turbines, and, if so, to identify potential problem areas and to delineate
further research efforts necessary for their resolution.
2-1
-------
The technical data presented in this report are based on discussions with a
number of organizations that are active in the area of surface combustion
research, development, and manufacture. The information is supplemented
by material from contractor reports and other literature related to the gen-
eral subject.
The results of this study are presented in the following manner: Sections 3
and 4 include state-of-the-art reviews of noncatalytic and catalytic surface
combustion devices, respectively; Section 5 treats the evaluation of these
devices for large utility boilers; and Section 6 is concerned with an evalua-
tion of these concepts as applied to stationary gas turbines. Appendix A lists
the organizations that were contacted during the course of this investigation.
Appendix B presents a table of metric conversion factors.
2-2
-------
SECTION 3
STATEl-OF-THE-ART REVIEW OF NONCATALYTIC
SURFACE COMBUSTION DEVICES
3. 1 INTRODUCTION
The technical information presented in this section is based on material
extracted from the literature and from the discussions -with a number of
organizations known to be involved in the development and manufacture of
noncatalytic surface combustion devices. These include primarily the
American Gas Association, U. S. Bureau of Mines, Burnham Corporation,
General Electric Company, Selas Corporation, and University of Wisconsin.
In the broadest sense, the term "surface combustion" involves the promotion
of gas-phase oxidation and reduction reactions between fuel and air in close
proximity with a solid surface. Surface combustors, which are also known by
several other designations such as nonadiabatic, flat-flame, radiant, refrac-
tory, and infrared burners, have shown promise as low-pollution devices. In
these burners, a fraction of the heat of combustion is immediately trans-
ferred from the flame layer to the solid surface, which generally consists of
a porous metallic or refractory ceramic material. The energy transfer is
adjusted such that the combustion temperature is reduced sufficiently to limit
the formation of NO emissions without significantly compromising the CO and
HC oxidation reactions.
Basically, there are two surface combustor concepts, classified as (1) flow-
through or transpiration burners, and (2) direct-contact radiant burners. The
design features of these two configurations, shown in Figure 3-1, are briefly
discussed as follows:
The flow-through burner consists of a porous plate which is comprised of
either an open-celled matrix of closely spaced interstices or ports which
3-1
-------
oAo AIK
POROUS
REFRACTORY
s
A^
r
i
i
I
\v
' \\X\V
m
' ' ' ' 't$
••••;'$
!:j%j||
1
•*":.••.•£
•;/:;'. ;.|
.'.'"' • : •*•
• - '• '. • >i5i
• . . . • :;•!
iVV\V\N
RADIATING
SURFACE
/
/
/
\
\
X
/
GAS-AIR
|NTAKE
SUPPORT
BRACKET
REFRACTORY
BLOCK
RADIATING
»- SURFACE
FLAME
Flow-Through (Transpiration) Porous
Burner
Direct-Contact Radiant Burner
Figure 3-1. Surface Burner Configurations
-------
allow the passage of the fuel-air mixture. These burners, which require
a prevaporized and premixed fuel-air charge, are limited to gaseous and
distillate liquid fuels. The pore structure depends upon the fabrication methods
which include laminating, foaming, blowing, and powder or granule bonding.
Metallic or refractory ceramic materials are used in the manufacture of the
burner plates, which can be made in the form of flat plates, cylinders, and
hemis phe r e s.
In the direct-contact burners, the fuel-air mixture is discharged through a set
of nozzles in such a manner that combustion occurs along a refractory surface.
Gaseous fuel-air mixtures are preferred for these burners, but they can be
modified to handle certain liquid fuels.
Porous refractory surface combustors are used commercially in various
applications including dryers, space heaters, and metal treating and brazing
systems. Also, experimental combustors of the metallic and ceramic types
are currently being developed for potential use in automotive gas turbine and
Rankine cycle engines.
Exploratory work on surface combustion was conducted as early as 1820,
with experiments to induce chemical reactions in air-gas mixtures below
their ignition points on the surface of solid materials. In 1902, W. A. Bone
began systematic investigations on the effect of hot surfaces on combustion
(Ref. 3-1). By 1907, his experiments had shown that (1) hot surfaces have
the capability of accelerating combustion of gaseous fuel-air mixtures, (2) the
chemical reactions are confined to the boundary layer region between the
gaseous and solid phases, (3) the accelerating effect of a hot surface on com-
bustion increases with increasing temperature, and (4) surface combustion
depends upon adsorption of the combustible mixture on the surface whereby
it becomes "activated" by association with the surface. Based on these
considerations, Bone concluded that if a gaseous fuel-air mixture is passed
through a porous refactory material under suitable conditions, combustion
would take place very rapidly in a thin layer adjacent to the surface of the
plate. The heat released during combustion maintains the surface in. a
3-3
-------
state of incandescence without a visible flame, thus realizing the idea of a
flameless, incandescent surface combustion. According to Bone, the
principal advantages of this concept include the capability of efficient
combustion with a minimum of excess air and uniform heat transmission
from the incandescent surface by means of radiation.
The surface combustion concept has since been further developed by Bone
and by others and is currently applied in various industrial heater designs.
More recently, the concept has generated considerable interest because of
its low NO emission potential.
The theory supporting the contention that surface combustion can diminish
pollution emissions stems from the present knowledge of post-flame combus-
tion gas behavior, taking into account the combustion energy release and heat
transfer mechanisms. The detailed steps in the chemical transformations
are not fully understood, but are founded in classical thermodynamic and
kinetic theories. Many of the arguments focus attention on the formation and
accumulation of nitrogen oxides, particularly NO, in burner gases. The
earliest theories presumed the establishment of equilibrium between molecu-
lar nitrogen and oxygen within the post-flame, high-temperature region of a
combustor. Difficulty quickly became apparent with this simplified approach
because practical combustors rarely provide sufficient residence time for
equilibrium concentrations to be attained. Efforts were then directed to
nitric oxide kinetic models. NO concentrations computed by means of the
Zeldovich model, which describes the rates of thermal fixation acquired
through atomic oxygen-nitrogen interactions in the post-flame gases, are
presented in Figure 3-2 as a function of stoichiometry, selected combustion
temperature, and residence time (Ref. 3-2).
The simple Zeldovich mechanism is adequate to predict the NO formation
rates at high temperatures. However, at temperatures below approximately
1500 K, the actual NO formation rates are substantially higher than those
predicted by the Zeldovich model (Refs. 3-3 and 3-4). Under these
3-4
-------
RPCllK-IDRCOMCfltlOf
III
; S S
*
•in 100
IVli ci-.l ill thi'or.-tK dl combustion .til
Figure 3-2. Kinetic Data--Natural Gas
Fuel (Ref. 3-2)
conditions, the mechanism involving N_O can no longer be neglected. As
indicated in Figure 3-2, the formation of NO , which is exponentially related
IX
to the combustion temperature, can be reduced substantially by lowering the
flame temperature. In surface combustion, this can be accomplished by
dissipating a fraction of the total heat of combustion by means of radiation
from the incandescent surface or conduction into imbedded cooling tubes.
In either of these approaches, the adiabatic combustion temperature is not
attained and, for this reason, these burners are also known as nonadiabatic
burners.
3-5
-------
3.2 STATE-OF-THE-ART REVIEW
Following is a review of the state of the art, presented by organization:
3.2.1 American Gas Association
The combustion-related R&D work currently being conducted by the American
Gas Association (AGA) laboratories is directed toward the development and
evaluation of refractory porous-plate-type radiant burners for potential use in
household applications (Ref. 3-5).
3.2. 1. 1 Combustor Design Features
The refractory burners tested by AGA to date are of the porous-plate type,
the drilled-port plate type, and the direct-fired type. In the porous-plate and
drilled-port configurations, the air-gas mixture flows through the plate at
low velocities. Conversely, the direct-fired burner utilizes a blower to supply
the fuel-air mixture, which is burned so that the flame impinges on the refrac-
tory burner face (Ref. 3-6). The burners have a temperature capability of
about 2000 F. Although AGA is not interested in higher temperatures for its
applications, at least one of these burners has been tested up to 2300 F. AGA
feels that the maximum temperature of the burner is limited only by the
capability of the refractory material. However, cracking of burner plates
may pose serious problems at these higher temperatures.
3.2.1.2 Operational Characteristics
The refractory burner test work at AGA is limited to gaseous fuels, including
natural gas, liquefied petroleum gas, and manufactured natural gas. The man-
ufactured gas is obtained by cracking natural gas and consists of CO and H?
and some CH, which is added after the cracking process to adjust the heating
3
value to 530 Btu/std ft . According to AGA, this heating value is required
to satisfy the appliance code, which was written many years ago when the
gas supplied by the producers was of that type.
3-6
-------
The burner tests conducted by AGA are based on fuel-air equivalence ratios
between 0. 9 and 1. 1. For efficiency reasons, they are not interested in fuel-
air ratios outside this range. The pressure drop of the burners is dependent
upon the porosity of the refractory plate material and the flow rate used in the
particular application, and is of the order of 4 to 5 inches of water. With
respect to turndown ratio of these burners, AGA states that a factor of 3:1
can be easily realized by means of inlet flow throttling. Further turndown
could be achieved by shutting off individual sections of the burner (sectional
burner design). The burner is lighted by means of a pilot light or electric
spark or glow plug. Since the AGA programs are limited to gaseous fuels,
there are no problems related to fuel vaporization and air-gas mixture non-
uniformity. After lightoff, the burner operates with a blue flame, approxi-
mately 3/8 inch high.
Apparently, under normal operating conditions (<1850 F), thermal shocks
have never been a problem with the type of refractory materials used in the
AGA test programs. Cyclic durability tests (15-minute on/off cycles) extend-
ing over several years have been conducted on some of these burners without
incurring a failure. However, some cracking of the refractory material has
been observed within several hours of cyclic operation at temperatures above
2100°F.
3.2.1.3 Heat Transfer Characteristics
Radiation test data from a number of gas-fired infrared burners show heat
release rates between 14,000 aWd 63,000 Btu/hr-ft , depending upon the
burner design, surface temperature, and combined emissivity of the burner
surface and the "flame" layer adjacent to that surface.
In 1966, development work on a compact radiant burner projected for use in a
refrigerant boiler was initiated by the AGA (Ref. 3-7). This burner-boiler
design consisted of two rectangular refractory burners and a "coolant wall"
located between the two burner surfaces. Each burner had a surface area of
3-7
-------
0. 85 ft . The unit was operated on natural gas with about 8 percent excess
air, at a surface temperature of 1800°F and an estimated flue gas tempera-
ture of about 2200 F. Based on an emissivity of 0. 8, the design radiant heat
2
flux of this unit was 36, 000 Btu/hr-ft . The total heat flux of the unit was
2
54, 000 Btu/hr-ft , indicating that about 1/3 of the total heat of combustion
was transferred to the coolant wall by means of radiation.
3.2.1.4 Emission Characteristics
Average NO and CO emission test data for refractory and conventional burners
tested by the AGA are listed in Table 3-1. (Ref. 3-5)
The data in the table indicate that the NO emitted from the AGA infrared burner
x
is only about 10 percent of that obtained with conventional burners and about
1/3 of that of commercially available infrared burners. Conversely, the CO
emissions of the commercial infrared burners are approximately three times
higher than from conventional burners. However, AGA feels that the higher
CO emission levels of the radiant burners would probably be acceptable from
an air quality and health effects point of view.
3.2.2 Bureau of Mines
The Bureau of Mines has conducted an investigation of the emission charac-
teristics of three different burner concepts designed to simulate the operation
Table 3-1. NOX and CO Emissions of Conventional and
Infrared Burners
Burner Type
Conventional
AGA infrared
Commercially available
infrared
Emissions, ppm
(corrected to stoichionietric)
NO
78
9
12
N02
16
0
12
NOX
94
9
24
CO
91
N.A.
293
3-8
-------
of gas appliances such as space heaters and water heaters (Ref. 3-8).
These burners, including a flat-flame Meeker burner, a ceramic infrared
plate burner, and a multiport ring burner, were operated at lean,
stoichiometric, and rich fuel-air ratios using methane as test fuel. The
Meeker burner was also operated on propane and a mixture of propane and
recycle gas to simulate the effects of flue gas recirculation on the emissions.
3.2.2.1 Burner Design Features and Operation
The Meeker burner test setup is presented in Figure 3-3. As indicated, the
burner consists of a 3 /8-inch-high stainless steel grid with 1/8 -inch- square
openings. The port is surrounded by a water-cooled steel tube arrangement.
The flame and burner tubes were enclosed in a cylindrical steel duct equipped
with a viewing window on one side and a traversing slit for probes on the
3
other side. The primary flow rate through the burner was about 4.5 ft /min,
resulting in a mixture velocity of 20 in/ sec at the burner port. Secondary
3
air was supplied through a calming bed at a rate of 3. 7 ft /min, corre-
sponding to an axial velocity of 5 in/ sec. In operation, the small conical
flames emerging from the square openings of the burner approximate a
flat flame, about 0. 10 inch high.
The ceramic infrared burner used in the program consists of a 5 X 7 inch
plate composed of 0. 05-inch holes uniformly arranged at a rate of 210 holes/
2
in . The burner was installed in a 9. 5 X 12-inch duct. The premixed flow
3
rate through the burner was 4. 5 ft /min, corresponding to a velocity of
3. 7 in/sec at the port. Secondary air was admitted coaxially through a
•a'
calming bed at a rate of 3. 9 ft /min and a velocity of 1. 4 in/ sec.
Both burners were operated at primary flow fuel-air equivalence ratios of
0. 85, 1. 0, and 1.2. The gas samples extracted from the burner at various
axial and radial positions were analyzed for NO , CO, HC, and CO^. The
NO concentrations were measured by the phenoldisulfuric acid method, HC
was determined in a flame ionization detector gas chromatograph, and
by means of an activated charcoal column.
3-9
-------
h 2j"diamH
6" id pipe
Secondary air
Figure 3-3. Bureau of Mines Flat-Flame Meeker Burner
Test Setup (Ref. 2-8)
3-10
-------
3.2.2.2 Emission Characteristics
The peak concentrations of NO and CO of the two-burner types are presented
in Table 3-2. As indicated, the emissions of the infrared ceramic burner are
always lower than the emissions from the Meeker burner. In the infrared
burner, the measured NO concentrations are essentially equal at the burner
X.
surface and at a distance of 1 inch above the surface, reflecting the lower
temperature level of the infrared burner. Conversely, the NO concentration
3£
of the Meeker burner increases substantially with increasing distance from the
burner surface. This is due to the fact that the NO concentration is roughly
proportional to the residence time. In both burners, the gas temperatures are
sufficiently high to achieve substantial reduction of CO and HC within 1 inch of
the burner surface. For fuel-rich mixtures, the CO remains very high because
of a lack of sufficient oxygen at this condition.
3.2.3 Burnham Corporation
The Burnham Corporation manufactures commercial water boilers for home
and industrial usage, and markets a line of patented dual-fuel (liquid-gas)-
powered burners for various applications (Ref. 3-9). According to Burnham,
these burners achieve low pollution combustion and better fuel economy than
cqmparable conventional burners.
3.2.3.1 Combustor Design Features
Figure 3-4 is a sectional view of the Burnham dual-fuel burner concept. The
burner is comprised of an air inlet nozzle, a startup manifold, a main fuel
inlet manifold, and an igniter spark plug mounted to an attachment plate.
The discharge side of the burner is a cube-shaped cast silica-alumina refrac-
tory tile containing the diverging combustion section and a hot gas recircula-
tion tube.
3-11
-------
Table 3-2. Peak NOX, CO, and HC Concentrations and Flame Temperatures for Meeker
Noninfrared and Ceramic Infrared Burners--Methane-Air Mixtures
Vertical
Distance
from
Burner
Surface,
inches
Temperature, °F
Non-
infrared
Infrared
NO Concentration,
ppm
Non-
infrared
Infrared
CO Concentration,
ppm
Non-
infrared
Infrared
Total Hydrocarbons,
ppm
Non-
infrared
Infrared
LEAN FLAMESa (equivalence ratio 0. 85)
0
1.0
2,950b
2,940
2,650
2,410
10
30
10
10
3,300b
400
2,600b
0
72,900b
0
750b
0
STOICHIOMETRIC FLAMESa
0
1.0
3,380b
3,210
2,930
2, 580
40
100
230b
10
20
23,500b
3, 100
6,700
700
22,660b
0
370C
0
RICH FLAMES0 (equivalence ratio 1.20)
0
1.0
3,320b
3,150
2, 790
2, 720
50
80b
30
30
41, 100b
39,100
36, 500
34, 000
42,415b
0
310
0
-------
IGN1TOR
FUEL GAS INLET
START UP
OIL
INLET
LIQUID FUEL
OR
LIQUID WASTE
INLET
t
(B)
Figure 3-4. Burnham Dual Fuel Burner Sectional Schematic
3.2.3.2
Operational Characteristics
According to Burnham, several inherent advantages accrue to this type of
burner configuration, which can operate on either gaseous or liquid fuels
independently or simultaneously. Burnham states that combustion takes
place mainly along the hot refractory walls. The burner operates on an
induction principle designed to obtain good mixing of the fuel and air ahead
of the flame zone. Prevaporization of liquid fuel is accomplished through
t
heat exchange with (about 10 percent) recycled products of combustion which
are drawn through the return tube by differential pressure forces. This
eliminates the need for a sophisticated spray atomization system. The
principal features of the recycle gas vaporization, turbulent mixing, incan-
descent refractory wall combustion concept are (l) efficient and uniform
transparent flame combustion for both gaseous and liquid fuels, and (2) low
HC, CO, NO, and particulate emissions.
3-13
-------
Fuels generally used in these burners include natural gas and other gases,
jet fuel, kerosene, and No. 2 fuel oil. Fuels such as Nos. 4 and 5 oils are
more difficult to vaporize, but have been successfully burned in these com-
bustors. Some exploratory tests with powdered coal have been conducted to
show feasibility of clean combustion to white ash. Burner turndown ratios
are nominally 3:1.
During initial development and early production, cracking of the cast refrac-
tory materials was encountered. The curing methods have since been revised
and this problem has been eliminated. With liquid fuels, carbon deposition
has occurred in the return tube. However, the carbon can be burned out simply
by switching to gaseous fuels for some time. There have been no maintenance
problems with these burners other than infrequent spark plug replacement.
3.2.3.3 Heat Transfer Characteristics
The presence of an incandescent refractory surface allows a portion of the
total heat of combustion to be transferred by radiation. Estimates by
Burnham vary between 10 and 15 percent of the total heat release of the
burner.
As shown in Figure 3-5, various sizes and operating ranges of Burnham
industrial burners are available, ranging from 0.3 X 10 Btu/hr to 15 X 10
Btu/hr. Based on the overall dimensions of the refractory burner cube, the
specific heat release rates are of the order of 3 to 5 X 10 Btu/hr-ft .
3. 2. 3. 4 Emission Characteristics
Tests have been performed on a Burnham boiler rated at 80 hp, using a
model B030 burner and No. 2 fuel oil. These tests, which were conducted in
1971 by the Scott Research Laboratories, verified previous effluent results
obtained on individual burners by another laboratory. In the Scott tests, flue
gas samples were collected and were later subjected to analysis. A summary
of the results is contained in Table 3-3.
3-14
-------
TO END OF BURNER
DIMENSIONS and RATINGS
BURNER
MODEL
B002
BOOB
B010
B030
B090
MAX. FIRING
CAPACITY
OIL-GPH
2.25
6.5
17
50
110
GAS • MBH
315
910
2,380
" 7.000
15.400
A
5"
7"
9"
12-3/4"
22-1/2"
e
51/2"
9"
9"
121/8"
16-1/2"
c
9-1/2"
15"
19"
16"
27"
D
6"
8-1/2"
10"
11-1/4"
•
E
1/2" D
1/2" D
9/16" D
5/8" D
3/4" D
' See certified drawings for details and exact dimensions of air and oil inlets, and mounting details.
Figure 3-5. Burxiharn Corporation Burner Dimensions and Ratings
3-15
-------
Table 3-3. Operating Conditions and NOX Emissions for a
Commercial Boiler with Radiant Burner
(Model PB-340 Boiler; Burnham B030
Burner)
Fuel Type No. 2 Fuel Oil
Heating Value l9,530Btu/lb ' ••'•
Flow Rate 25. 7 gal/hr
Heat Flux 2. 75 X 1Q6 Btu/hr-ft^
Flue Gas Composition (Avg. )
CO2, % 13.5
O *5n ^
S0x, ppm 81
NOx, ppm 126
CH., ppm. 5
Calculated NO9 Emission,
L*
lb/106 Btu 0.2
The NO emissions listed for the Burnham burner are substantially lower
than the NO levels of about 300 to 700 ppm emitted from conventional uncon-
x rr
trolled utility boilers. However, when comparing the NO emissions of the
various burner-boiler designs, consideration must be given to the fact that
the Burnham burner utilizes about 10 percent flue gas rec-irculation. Tests
conducted by Southern California Edison Company (Ref. 3-10) and Combustion
Engineering Company (Ref. 3-11) on gas-fired utility boilers indicate that
flue gas recirculation is an effective means of reducing the NO emissions
.X
from conventional boiler units. This is illustrated in Figure 3-6, showing
measured NO concentrations as a function of gas recirculation rate for
corner-fired installations (Ref. 3-10). At 10 percent recirculation, the NO
emissions of these utility boilers are of the order of 150 ppm. This indicates
that the relatively low NO level achieved in the Burnham combustor might
3x
be primarily the result of flue gas recirculation, and not due to the refractory
burner block design.
3-16
-------
350
300
NOTE: 1. Data from different
units of same type
2. Natural gas firings
10 20 30 40
FLUE GAS RECIRCULATION, percent
50
Figure 3-6. NOX Emission Concentration vs Percent Flue Gas
Recirculation--320 MW Corner Fired Unit
(Ref. 3-10)
3-17
-------
3.2.3.5
Burner Costs
The Burnham burner models are the only commercial devices where cost
data were made available. The breakdown summary of unit costs for several
burner models purchased in quantities is given in Table 3-4.
3.2.4 General Electric Company
The Space Division of General Electric is currently involved in the develop-
ment of low-emission porous burners under both in-house and EPA-funded
projects. The funded work is aimed at automotive gas turbine and Rankine
cycle engines, and the in-house work involves porous flow-through burners
for potential use in steam power plant boilers (Refs. 3-12 and 3-13).
3.2.4. 1
Combustor Design Features
Conceptually, the General Electric burner consists of a porous, open-celled
matrix usually configured as a right circular cylinder in system application
design studies. Flat porous plates also have been used as test components.
As illustrated in Figure 3-7, a mixture of air and fuel, premixed to; the
Table 3-4. Burnham Burner Cost Data
Model
Number
B002
B005
B010
B030
B090
Burner
Rating ,
(Btu/hr) X 10
0.3
0.9
2.4
7.0
' 15.5
Basica
Burner Price
$
200
260
310
730
1210
Auxiliary"
Components
Price
$
530
680
710
900
1250
Total
Assembled
Cost
$
730
940
1020
1630
2460
aQuantity purchase price/unit
"Valves, fittings, igniter, electrical controls, etc., fitted for multifuel
operation
3-18
-------
BURNED GAS
TBURNED
-------
materials. Metallic burners have been fabricated, from various steel,
chromium, and nickel powder compositions. Ceramic materials considered
by General Electric include silicon carbide and alumina-silica compositions.
Composite laminates, comprised of a mullite base and a thin silicon carbide
overlay, are commercially available.
Four possible porous burner configurations for gas turbines are shown in
Figure 3-8 (Ref. 3-13), as follows:
1. A simple cylinder.
2. A composite structure consisting of a high-temperature, high-
conductivity (silicon carbide) outer material and a low-
conductivity inner material that is shown to be circumferentially
segmented with the segments joined at wedging surfaces loaded
by expansion springs.
3. A segmented structure, the pieces of which are made from an
integrally bonded combination of a high-temperature, high-
conductivity outer material and a low-conductivity inner material.
The radial joints are pressurized at the ends by conically cut
rings to •which axial compression is applied by a spring.
4. An alternate segmented structure made from composite material
pieces, as in the third configuration. The segment joints are
pressurized by containment, at the ends, of alternate segments
between outer cylinders and inner expansion springs. Experience
has indicated that a small leakage flow between the pressurized
joints is, under normal (sub-liftoff) operation, indistinguishable
from the normal flow through the porous burner.
Figure 3-9 shows a conceptual combustor design for potential use in a gas
turbine. Principal parts of the combustor include the outer containment shell;
the radiation heat sink screen and support baffle; the segmented porous burner
structure; the forward end burner support structure with its axial loading
spring; the aft end burner support structure, with the integrally attached
primary fuel-air mixing tube; the fore and aft radial expansion springs for
pressurizing the burner segment joints; the vaporizer, which incorporates
the air atomizing nozzle; and the primary air-flow control valve with actuator.
Heated atomizing-vaporizing air and fuel are introduced into the air atomizing
3-20
-------
to)
I
! 1 1 1 1 1 1 1 1 1 1 1 II 1 1 1 |
1 II 1 1 1 1 1 II [ 1 1 1 1 1 1 1
o
o
0
0
o
o
o
o
o
T
i 1 1 1 1 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 MI
1' III II
~~^\ loo o
, I il
i i
I A-1
too
^3J
00 01
O O
EL
o nU
o o o o
Figure 3-8. Alternate Frit Configurations (Ref. 3-13)
-------
00
i
IV
TERTIARY AIR
TRANSPIRATION
COOLED SURFACE
COOLING 4. SEALING
AIR
FLOW TO TURBINE
PRIMARY AIR FLOW
CONTROL VALVE
TANGENTIAL SLOTS-^>'
AlR ATOMIZING^"
NOZZLE
-HEATED ATOMIZING, VAPORIZ
Figure 3-9. Porous Radiant Gas Turbine Combustor (Ref. 2-13)
-------
nozzle. Spray from this nozzle is mixed with heated air issuing from
tangential slots. This heated, rich fuel-air mixture passes down the central
tube of the vaporizer and back up the surrounding annulus. It is then injected
into the main primary air stream. To facilitate mixing, the latter stream
issues in jets from orifices immediately upstream of the injection outlets of
the vaporized rich mixture. The burner is sealed at its ends on axial faces
•which bear on the end support pieces. Cooling and sealing air is injected
into the leakage paths across these surfaces in order to ensure that no fuel-
air mixture bypasses the porous burner. On the downstream end, a trans-
piration-cooled surface on the burner support structure is employed where
the effluent combustion gas passes over this structure. Tertiary air baffles
are employed downstream of the primary combustion gas and secondary
dilution air mixing zone. The separation of secondary and tertiary air delays
the reduction of the combustor outlet gas temperature to the final turbine inlet
temperature value until sufficient time for CO combustion has elapsed. Heat
is transferred from the flame to the porous burner, causing it to become an
incandescent radiation source. From here, heat is radiated to the heat sink
screen-baffle structure, through which is passed the secondary air from the
regnerator. Upstream of the combustor is the primary air-flow control
valve.
Combustor concepts have been also devised for use with Rankine cycles.
In this approach, shown in Figure 3-10, combustion heat feedback from the
flame is dissipated in part to a tubular heat exchanger imbedded in the porous
matrix. Superheated vapor is produced in a second heat exchanger by means
of heat exchange with the partially cooled combustion gases. Experimental
burners for an automotive Rankine power plant have been fabricated from
conductive copper and copper alloys.
3-23
-------
Air From Expander
Driven Compressor
Air Controlled
Pressure Reservoir
•nd Manifold
Conbustlon
Volume
fuel/Air Manifold
Figure 3-10. Conceptual Design of Porous-Plate Burner-Vapor
Generator for Rankine System (Ref. 3-13)
Section A-A
-------
3.2.4.2 Operational Characteristics
Limits characterizing burner behavior are expressed by General Electric
in terms of the gas superficial (entrance) velocity into the burner, u7c; (at
LJ D
25°C) and the reciprocal absolute temperature of the products leaving the
porous plate. At high superficial velocities, the flame zone becomes detached
and insufficient heat feedback occurs; consequently, combustion can no longer
be supported. As shown in Figure 3-11, this point represents the liftoff limit.
A minimum velocity is required for stability at low heat release rates.
For the experimental work, superficial gas velocities have ranged between
4 and 50 cm/sec. Fuel-air equivalence ratios have ranged between 0.7 and
1. 0, with some more recent data at approximately 1. 1 using gasoline. As
shown in Figures 3-12 and 3-13, heat feedback into the burner varies with
superficial velocity, fuel type, and operating pressure. Maximum turndown
ratios of about 7:1 have been achieved by General Electric with single burners
operated with propane, and higher overall turndown ratios might be feasible
with combustor staging.
The most difficult task has been to find a matrix material that will maintain
its integrity and give satisfactory operating life. Cracking has been a prob-
lem with ceramic materials, and overheating and formation of local hot spots
have occurred in the metallic configurations. Operation at elevated pressures
has resulted in excessive surface temperatures. Surface oxidation also
appears to be a materials problfem which is being carefully addressed.
To assure complete fuel vaporization and mixing, which is deemed essential
for proper combustor operation, General Electric has developed an air-
driven acoustic fuel atomizer, shown in Figure 3-14, combined •with an
advanced small-volume tangential mixing chamber. With this design, Gen-
eral Electric feels that flashback, which represents a potential problem in
all premixed and prevaporized systems, can be avoided. To date, flashback
has not been encountered in any of General Electric's tests.
3-25
-------
o
ttt
100
E
o
in
CM
=1
CB
O
o
u_
oc.
LU
D_
in
NOTE: u25 is at P = 1 atm
and T = 25° C
ESTIMATED LIFTOFF
LIMIT FOR INLET
AIR-FUEL OF 1200°F
LIFTOFF LIMIT
FOR INLET
OF 77° F
0.7
POROUS PLUG
COMBUSTION CURVES
R = 1.0
EQUIVALENCE
RATIO
4x10"4 5x10'4
RECIPROCAL ABSOLUTE TEMPERATURE, (1/°K)
6 x 10'
4000 3800 3600 3400 3200 3000 2800
TEMPERATURE, °F
Figure 3-11. Effect of Unburned Gas Superficial Velocity and Equivalence
Ratio Upon Flame Temperature (Ref. 3-13)
3-26
-------
oo
I
tv)
-J
40,000
35,000
30,000
I 25,000
f,
4J
to
(9
V
20,000
M 15,000
I
10,000
5,000
Propane-Air Mixtures
(^ 25°C)
10 15 20 25 30
Superficial Gas Velocity u c, cm/sec
Equivalence
Ratio
1.0
0.9
0.8
35
40
Figure 3-12. Heat Flux Back to the Burner for Propane-Air Mixtures (Ref. 3-13)
-------
45,000
CD
_ 40,000 —
y. ^ 25, ooo
5*
£ 15,000
S£
CD
10 20 30
SUPERFICIAL GAS VELOCITY u25, cm/sec
Figure 3-13. Heat Flux Back to the Burner for Gasoline-Air Mixtures (Ref. 3-13)
-------
RESONATOR CHAMBER
LIQUID
CENTER OF
IMPLOSION
AIR OR GAS
Figure 3-14. "Sonicore" Fuel Atomizer (Ref. 3-13)
3.2.4.3 Emission Characteristics
The porous plate combustors developed by General Electric for use in gas
turbine and Rankine cycle engines are designed to meet the 1976 Federal
Emission Standards for light-duty vehicles. As indicated in Figure 3-15, the
NO emission index equivalent to the 1976 standards (assuming a fuel economy
n
of 10 miles per gallon) has been achieved when operating the Rankine engine
burner at 1 atm with propane as test fuel (equivalence ratio 0.9). In this
particular test series, the CO standard was met for superficial gas velocities
below about 25 cm/sec. No data were provided by General Electric for higher
operating pressures. With prevaporized liquid fuels, General Electric
obtained significantly higher NO emissions, even after accounting for the
conversion to NO of the fuel bound nitrogen.
3.2.5 Selas Corporation of America
The Selas Corporation manufactures a variety of industrial furnaces and
"radiant cup" burners which are used by various metal, glass, chemical, and
petroleum processors. A significant part of Selas1 technical capability
3-29
-------
LU
3
LL.
E
u>
o
o
o
_
o
a
E
en
X*
LU
a
g
VO
LU
10
1.0
0.1
0.01
0.001
1976 FEDERAL EMISSION STANDARDS
(10 miles/gal)
CO
NO,
CO MEASURED
PREDICTED CO
EQUILIBRIUM FOR THE FLAME
TEMPERATURE AND RESIDENCE
TIME
NO2 CONVERTED FROM MEASURED NOX
EQUIVALENCE RATIO: 0.9
PRESSURE: 1 atm
A MEASURED CO
D MEASURED.NOW
I
5 10 15 20 25
SUPERFICIAL GAS VELOCITY, u25, cm/sec
30
Figure 3-15. Preliminary Emission Measurements (Not Corrected for
Water Content) (Ref. 3-12)
3-30
-------
involves manufacture and assembly of industrial pyrolysis furnaces wherein
hydrocarbon feedstocks are catalytically or thermally converted (cracked)
to commercial by-products. Typical process fluid outlet temperatures vary
between 1400 F and 1800°F, with pressures up to 500 psi. These units are
horizontal or vertical tube furnaces containing a number of radiant-cup
burners which are designed to provide precise and uniform control of the
combustion gas composition and heat flux density patterns along the furnace
profile (Ref. 3-14).
3.2.5.1 Combustor Design Features
Selas burners are procured and operated either as nozzle mix, premix, or
aspirator configurations, depending upon the fuel type and the application.
The feature which is common to Selas burners is that the combustible fuel-
air mixture is introduced at a centrally located burner tip and forced to flow
radially across the surface of a parabolic-shaped refractory cup where igni-
tion and combustion occur. The general outline of a typical radiant-cup, gas-
fired burner is shown in Figure 3-16.
Selas manufactures a line of gas or naptha (kerosene)-fired radiant burners,
the largest being rated at about 2x10 Btu/hr. This particular unit has a
diameter of 12 inches.
3.2.5.2 Operational Characteristics
In steady-state operation, as the mixture flows along the contour of the cup,
rapid heat exchange occurs along the surface, enabling the cup to act as a
radiative source. This is coupled with an internal recirculation flow pattern.
Theoretically, the combustion is completed within the cup envelope. Thus,
the need for additional or excess secondary air to complete the combustion
process, as common with more conventional burners, is claimed to be
unnecessary.
3-31
-------
ji > 4p fe r; - ^
FLAME
REFLECTOR
(REFRACTORY)
HEAT RESISTANT
ALLOY CASING
THREADED REFRACTORY
ORIFICE
Figure 3-16. Typical Selas Radiant-Cup Air-Gas Burner
3-32
-------
These burners, which are designed for a variety of gaseous fuels, can operate
continuously with little or no maintenance. Special vortex-type configurations
are available,which enable safe operation with high hydrogen-content gases
often available to refineries. The nominal turndown ratio for most Selas
production burners is between 5:1 and 10:1. Variable area burners also are
available for larger turndown ranges. Oil-fired burners are not in production.
Occasional breakage of ceramic burner tips (not cups) necessitates replace-
ment. Burners are designed to permit such replacement from outside the
furnace wall after the individual fuel supply is shut off. However, the furnace
remains operational during the replacement procedure.
t
3.2.5.3 Heat Transfer Characteristics
The radiant cup is designed to offer controlled heat transmission by radiating
energy in a symmetrical beam. This is considered important in industrial
furnace units operating at high internal tubular heater temperatures and pres-
sures. Selas states that the high degree of symmetrical heating obtained with
these cup burners tends to minimize distortion and bowing of heater tubes.
The furnaces are designed and rated for continuous operation over several
years.
3.2.5.4 Emission Characteristics
Test data on commercial furnace units are not available. However, tests
conducted by Selas on laboratory configurations indicate that radiant-cup
burners produce lower NO emissions than conventional burners. According
to Selas, this is due to the intimate contact between the burning gases and the
cup surface, which permits a portion of the heat of combustion to be trans-
ferred into the cup--resulting in lower gas temperatures. Moreover, aero-
dynamic forces generated within the cup are thought to induce recirculation
of cooler furnace gases and the combination of these effects rapidly quenches
the NO formation process. The NO emissions of an aspirator-type lab-
Jt Jt
oratory furnace operated on natural gas near stoichiometric are shown in
3-33
-------
Figure 3-17. As indicated, the NO varies between 30 and 80 ppm, depending
ji
upon the selected furnace gas temperature. According to Selas, comparable
NO emissions were obtained with No. 2 fuel oil.
^C
3.2.6 University of Wisconsin
An experimental program of graduate research on low nitric oxide burners
was undertaken by Peters at the University of Wisconsin (Ref. 3-15). The
investigation consisted of determining the degree of nitric oxide reduction
that can be achieved by means of nonadiabatic combustion of propane-air
mixtures on a cooled porous metal plate.
3.2.6.1 Combustor Design Features
A cutaway drawing of the combustor arrangement is shown in Figure 3-18.
It is comprised of a gas mixer, a porous metal disc burner, and a main
burner housing. The steel top plate which holds the replaceable porous metal
disc includes a water-cooling system for the porous disc. The main burner
100
80
§: 60
40
Z
Ul
o
20
10
o-
o
o
I
2000
2200 2400
GAS TEMPERATURE, °F
2600
Figure 3-17. NOX vs Gas Temperature for a Selas Inspirator
Burner
3-34
-------
GAS MIXER
THERMOCOUPLES
WATER
OUTLET
TOP PLATE
WINDOW .'HOLDER
POROUS METAL DISK
BURNER HOUSING
FIREBRICK
VERMICULITE
STAINLESS STEEL
EXHAUST
SCALE: 3/8
Figure 3-18. University of Wisconsin Burner Test Section
(Ref. 3-15)
3-35
-------
housing is insulated and includes windows, igniters, gas sample probes, and
thermocouples used for flame observation and monitoring of the combustion
products.
Three types of porous disc materials were utilized in the program, including
grades 340 and 460 of oilite filter material manufactured by the Amplex Div-
ision of Chrysler Corporation, and 70-micron rated P/N S-2121 sintered
316 stainless steel manufactured by the Sintered Specialties Division of Parker
Pen Corporation. All the discs were 2. 0 inches in diameter and either 0. 25,
0. 38, or 0. 50 inch thick.
A schematic of the combustor test setup is presented in Figure 3-19, showing
the main components of the air and fuel supply systems. The .air is provided
by a compressed air source, and the air flow rate is measured with an orifice
meter. The "natural grade" propane fuel used in the program is supplied
from a tank and controlled by a single-stage regulator. From the regulator,
the fuel is routed into a filter with a liquid separator and then through a
rotameter to measure the flow rate. Following the rotameter, the fuel passes
through a needle valve and into the gas mixer located upstream of the burner
inlet section. Ideally, the mixer delivers a homogeneous, premixed propane-
air mixture to the burner. The combustible mixture then flows through a
water-cooled, sintered.porous metal disc and is ignited on the back side of
the disc. The combustion products pass through an insulated tube, where gas
samples are taken, and finally into an exhaust system.
In the test program, the combustor was operated at fuel-air equivalence ratios
between about 0.7 and 1.2, pressure levels of 19.3, 39.9, and 54.3 psia, and
a mixture inlet temperature of 530°R. The mass flow rate through the burner
was varied. For example, at 39.9 psia pressure, the flow rates covered the
range between 3.3 Ib/hr and 30.6 Ib/hr. Smaller flow variations were con-
sidered for the other operating pressures.
3-36
-------
OJ
I
Co
REGULATOR
MAIN
BURNER
SECTION
AIR SUPPLY
FILTER
Figure 3-19. University of Wisconsin Fuel and Air Inlet System (Ref. 3-15)
-------
3.2.6.2 Heat Transfer Characteristics
In the test program, the fraction of the heat of combustion returned to the
porous plate was varied by controlling the mixture equivalence ratio and the
burner mass flow rate, as shown in Figure 3-20, for an operating pressure
of 39.9 psia. The heat flux reaches a maximum at equivalence ratios slightly
higher than stoichiometric and generally increases with decreasing flow rate.
These trends are in agreement with the results obtained by General Electric
and others. As indicated in the figure, as much as 30 percent of the com-
bustion heat can be transferred to the cooled porous disc. Similar results
were obtained for the other operating pressures.
3.2.6.3 Emission Characteristics
The nitric oxide emissions measured at an operating pressure of 39.9 psia
and a residence time of 40 msec are plotted in Figure 3-21 as a function of
fuel-air equivalence ratio and mixture flow rate. As expected, the NO con-
centrations are very low at low equivalence ratios, especially at low flow
rates, reflecting the lower temperatures of the combustion products at these
operating conditions. The NO emissions increase with increasing equiv-
a
alence ratio, reaching a maximum near stoichiometric. Also shown in Fig-
ure 3-21 are the computed adiabatic equilibrium NO concentrations. In the
^t
lean regime, the NO emissions with flame cooling are substantially lower
2t '
than the corresponding equilibrium values.
Small variations in the measured NO emission levels were observed for the
2t
three different porous-plate configurations evaluated in the program. These
differences are attributed to the variations in plate thickness, thermal con-
ductivity of the material, and contact resistance around the circumference of
the discs.
3-38
-------
0.25
i
w
sD
£
o
I-
ac
x
D
la
x
0.20
0.15
0.10
0.05
0.00
I
I
0.7 0.8 0.9 1.0 1.1
FUEL/AIR EQUIVALENCE RATIO
1.2
1.3
4600ILITE BRONZE PLATE
PRESSURE: 39.9PSIA
INLET TEMPERATURE: 530°R
Q - HEAT FLUX TO PLATE
UMAX ~
FLOW RATE
D
A
O
•
AUIADA 1 It, 1
Ibm/hr
3.3
= 6.2
= 11.3
= 16.6
Figure 3-20. Q/Q vs Equivalence Ratio Profiles for Various Mass
Flow^ates (Ref. 3-15}
-------
CL
Q.
«
o
111
O
O
u
TIME: 40 msec
PRESSURE: 39. 9 psia
INLET TEMPERATURE: 530°R
O ADIABATIC EQUILIBRIUM
VALUES
_ ADIABATIC ZELDOVICH
* MECHANISM EQUATIONS
FLOW RATE Ibm/hr
• = 30.6
= 22.9
= 16.6
= 11.3
= 6.2
= 3.3
20 r—
10
0
0.8 0.9 1.0 1.1
FUEL/AIR EQUIVALENCE RATIO
1.2
Figure 3-21,
NO vs Equivalence- Ratio Profiles for Various
Mass Flow Rates (Ref. 3-15)
3-40
-------
The CO emissions from these burners were very low and in reasonable
agreement with the equilibrium values computed for the particular fuel-air
equivalence ratios, indicating good combustion efficiency under all operating
conditions.
Except for a few isolated cases, the observed hydrocarbon emissions were
less than 10 ppm hexane. According to Peters (Ref. 3-15), the experimental
accuracy was probably no better than ±10 ppm.
3.2.7 Other Organizations
Several other organizations are involved in the development and/or manufac-
ture of heterogeneous (radiant) combustion devices, including AiResearch,
Holden, Institute of Gas Technology, and Perfection. The information pro-
vided by these companies is briefly discussed in the following sections.
3.2.7.1 AiResearch Manufacturing Company
Porous-plate combustors/heaters are currently being studied for potential
use in closed Brayton cycle (gas turbine) engines. Emission test data with
propane fuel indicate HC, CO, and NO emission levels of about 10 ppm or
j£
less, at an equivalence ratio of 0. 8. Diesel fuel data gave slightly higher
emissions (Ref. 3-16).
3.2.7.2 A. F. Holden Company
This company manufactures porous refractory blocks and "luminous wall" fur-
naces for various industrial heating and metal treatment applications. Allow-
able maximum wall temperatures of 2000°F and above have been maintained
in certain applications. The refractory brick can be operated over a range of
fuel-air equivalence ratios, but operation near stoichiome'tric is preferred.
Maximum heat input is about 150, 000 Btu/hr-ft . Currently, these burners
are used only with gaseous fuels, but certain liquid fuels might be acceptable
o O
as well. A minimum gas flow rate of 200 ft /hr-ft is required to cool the
brick and prevent flashback. Emission data are not available for these
burners (Ref. 3-17).
3-41
-------
3.2.7.3 Institute of Gas Technology
This organization has recently completed a study for NASA/Lewis relating to
emissions from porous-plate combustors designed for use in closed Brayton
cycle (gas turbine) systems. The surface combustor considered for this
application is schematically shown in Figure 3-22 (Ref. 3-18). Based on this
study the Institute of Gas Technology concludes that porous-plate burners
-12 in.
OOOOOOOO
GAS-AIR
MIXTURE
•vv
«w*
^^
vwt
«w
.1 » ^4 ft f * *
1/4 in.
O.D. TUBES
POROUS PLATE
BURNER
PROCESS FLUID
IN
t>J fc»
I OUT
12 in.
Figure 3-22. Surface Burner Geometry (Ref. 3-18)
3-42
-------
fabricated from high-conductivity materials having a fine pore structure
(consistent with allowable pressure drops) provide a practical approach to a
low-emission combustion system, provided the plate temperature is kept low
and the flame is quenched rapidly. Potential problem areas cited by the
institute include overheating of the plate at high heat loadings, and flame front
instability.
3.2.7.4 Perfection Products Company
Perfection manufactures a line of flow- through gas-flame radiant burners for
use in industrial space heaters, food processing, drying ovens, and other
uses. Emphasis is on operating safety and burner lifetime, which is mea-
sured in years. High radiation efficiency is achieved with its newly devel-
oped waffle-face silicon carbide burner tile which has a maximum temperature
capability of 1850°F. Total design heat release rates of these burners are
below 100, 000 Btu/hr-ft . Operation of the burners at near-stoichiometric
conditions results in CO emissions of about 300 to 400 ppm. Data on NO
emissions are not available.
3-43
-------
REFERENCES
3-1. W. A. Bone, "Surface Combustion with. Special Reference to Recent
Developments in Radiophragm Heating, " Gas Journal, 423-428
(16 May 1923).
3-2. W. H. Barr and D. E. James, "Nitric Oxide Control--A Program
of Significant Accomplishments, " ASME Winter Annual Meeting,
New York, N. Y. , 26-30 November 1972, Paper 72-WA/Pwr-l 3.
3-3. Li. S. Carretto, et al. , "The Role of Kinetics in Engine Emission
of Nitric Oxide, " Combustion Science and Technology, ^3, 53-61
(1971).
3-4. D. T. Pratt and P. C. Malte, "Formation of Thermal and Prompt
NO in a Jet Stirred Combustor, " 75th National AIChE Meeting,
Detroit, Mich. , 3-6 June 1973, Paper 34B.
3-5. Personal communication with the American Gas Association,
Cleveland, O. , 5 April 1973.
3-6. D. W. DeWerth, A Study of Infrared Energy Generated by Radiant
Gas Burners, Research Bulletin 92, American Gas Association
Laboratories, Cleveland, O. (November 1962).
3-7. W. O. Specht and E. J. Weber, Development of a Compact Refrigerant
Boiler for an Air Conditioning Application, Research Report 1447,
American Gas Association Laboratories, Cleveland, O. (August 1967).
3-8. M. E. Harris, et al. , Reduction of Air Pollutants from Gas Burner
Flames, Bulletin 653, U.S. Department of the Interior, Bureau of
Mines (1970). ,
3-9. Personal communication with Burnham Corporation, Lancaster, Pa. ,
2 April 1973.
3-10. F. A. Bagwell, et al. , "Utility Boiler Operating Modes for Reduced
Nitric Oxide Emissions, " Journal of Air Pollution Control
Association, 21 (H) 702-708 (November 1971).
3-11. C. E. Blakeslee and H. E. Burbach, "Controlling NOX Emissions from
Steam Generators, " 65th Annual Air Pollution Control Association
Meeting, Miami, Florida, June 1972, Paper 72-75.
3-45
-------
3-12. Personal communication with General Electric Company, Cincinnati, O. ,
6 April 1973.
3-13. R. J. Rossbach, Development of Low-Emission Porous-Plate Com-
bustors for Automotive Gas Turbine and Rankine Cycle Engines.
Quarterly Progress Report GESP-736, 18 September 1972 to
31 January 1973, General Electric Company, Cincinnati, O.
3-14. Personal communication with Selas Corporation, Dresner, Pa. ,
2 April 1973.
3-15. B. D. Peters, "Nitric Oxide Reduction by Heat Transfer in a Porous
Disk Burner, " Ph.D. Thesis, University of Wisconsin (October 1972).
3-16. Personal communication with AiResearch Manufacturing Company,
Phoenix, Ariz.
3-17. Personal communication with A. F. Holden Company, Milford, Mich.,
8 December 1972.
3-18. A. Kardas and R. B. Rosenberg, Determination of Anticipated
Emission Levels from a Surface Type Combustor Module, Institute of
Gas Technology (October 1972).
3-46
-------
SECTION 4
STATE-OF-THE-ART REVIEW OF CATALYTIC DEVICES
4. 1 INTRODUCTION
The discussion of catalytic combustion devices presented in the following
sections is primarily based upon the technical information acquired from a
number of organizations, including Engelhard Industries, NASA/Lewis, Air
Force Aero Propulsion Laboratory (AFAPL), and the Environmental Pro-
tection Agency (EPA), and on supplementary information extracted from the
open literature.
With the exception of a small number of exploratory tests conducted by
Engelhard and the Combustion Research Section, Control Systems Laboratory
of the EPA, there is no information available at this time regarding the per-
formance characteristics of high-temperature catalytic combustors of poten-
tial interest to stationary gas turbine and power plant applications. However,
a considerable amount of experience has been gained from the development
and operation of catalytic converters used in many industrial processes, as
well as the exhaust treatment of automotive engines under a variety of
operating conditions. In particular, the effects of catalyst deactivation due
to poisoning and high-temperature operation on the lightoff characteristics
and the service life of these systems have been evaluated by the automotive
and chemical process industries. Since catalytic combustors are likely to
have design features similar to those of industrial and automotive catalytic
devices, and since they are expected to be influenced by the same factors
affecting performance and durability, a brief discussion of these parameters
is considered appropriate.
The design features of base metal, platinum-group metal, and nonmetallic
catalysts--supported on monolithic, pellet, and fibrous pad substrates--are
examined in Section 4.2, followed by a section on catalyst performance
degradation due to poisoning, overtemperature conditipns, and vibratory
4-1
-------
loads. Section 4.4 discusses the operational characteristics of four different
catalytic system configurations. These include automotive oxidation catalysts,
industrial tail gas abatement catalysts, low-temperature catalytic heaters,
and, finally, the high-temperature catalytic combustors currently considered
by several organizations for potential application in mobile and stationary
gas turbines.
4.2 CATALYST FEATURES
4.2.1 Catalytic Process
In chemical reactions, the molecules of the reactants are required to over-
come specific potential energy barriers (Ref. 4-1). The rate of a reaction,
which is determined by the number of molecules having sufficient kinetic
energy, can be increased by (1) increasing the temperature of the reactants,
and (2) incorporating a suitable catalyst. In the case of the catalyst, the
chemical reaction proceeds in the following manner. The reactants are
initially brought into contact with the catalyst surface by diffusion. Upon
contact, the catalyst reacts with the molecules of the reactants to create an
unstable intermediate product,which effectively lowers the energy barrier
between the molecules involved in the reaction. The intermediate product
reacts further at the catalyst surface and forms the final product,which is
then removed from the surface of the catalyst by means of desorption and
diffusion. Under ideal operating conditions, the catalyst remains unchanged
in the reaction process. Its only function is to promote the chemical reac-
tions which would otherwise not occur or occur very slowly.
4.2.2 Catalyst Requirements
Necessary attributes for catalytic devices include adequate chemical activity,
selectivity, and service life. In addition, the catalyst substrate must be
capable of withstanding the mechanical and thermal loads generated in tliie
particular duty cycles.
The activity of the catalyst under normal operating conditions has a bearing
on the size of the catalytic unit required for each application. Also, high
activity at low temperatures is desirable in order to minim.ize the amount of
4-2
-------
reactant preheating needed for catalyst lightoff. Table 4-1 presents the
lightoff temperature for a number of catalysts with respect to nitrous oxide
decomposition (Ref. 4-2).
Selectivity represents another important parameter. Frequently, different
reaction products can be obtained from a given set of reactants, depending
upon the composition of the catalyst used in the process. For example, in
ammonia oxidation, either nitric oxide and water or nitrogen and water can
be formed. The yield of nitric oxide can be increased simply by increasing
the rhodium content in the platinum base catalyst normally utilized in the
ammonia oxidation process (Ref. 4-3).
Catalyst life is a most important aspect, particularly in the case of costly
noble metal designs. Short life requires frequent replacement of the catalyst
and this results in higher overall process cost. For these reasons, indus-
trial catalysts are generally designed for several thousand hours of
maintenance-free operation.
Table 4-1. Catalyst Lightoff Temperature--
Decomposition of Nitrous Oxide
Catalyst
Cu-O
CoO
NiO
CuO
MgO
CaO
Ce02
ZuO
CdO
Cr2°3
Fe2°3
Lightoff Temperature, C
215
225
300
380
390
420
450
480
600
615
660
700
4-3
-------
4.2.3 Typical Catalysts
Literally thousands of different catalyst materials and formulations have
been evaluated over the years for potential use in a variety of chemical pro-
cesses and automotive exhaust treatment systems. These configurations
fall into three basic categories: platinum, group metal catalysts, base metal
and promoted base metal catalysts, and nonmetallic catalysts. In general,
the metallic catalyst materials are deposited on a suitable substrate. A
number of different deposition methods have been developed, including precip-
itation, electroplating, coating with a paste, deposition of colloidal mate-
rials, and impregnation of the substrate with a salt solution.
4.2.3.1
Platinum Group Metal Catalysts
The platinum group metals, especially platinum, palladium, and rhodium,
are being used extensively in industrial catalysts, either by themselves or
in the form of alloys. As indicated in Ref. 4-4, palladium is the most
effective noble metal catalyst for many processes, followed by platinum and
rhodium. This is illustrated in Table 4-2, showing the relative reaction
rates for catalytic oxidation of methane for a platinum and a palladium
catalyst supported on Y-alumina (Ref. 4-5). As indicated, the reaction rates
obtained with palladium are substantially higher than those for platinum.
Noble metal catalysts are used primarily in oxidation processes such as
nitric acid manufacture, tail gas abatement, and automotive exhaust emis-
sion control.
Table 4-2. Relative Reaction Rates--Catalytic
Oxidation of Methane
Catalyst /Substrate
P alladium / Y -Alumina
Platinum / Y- Alumina
Activation
Energy
kcal/mole
19.6
9.2
Reaction Rates Relative to that
of Palladium/ Y-Alumina at
400°C
400°C
1.0
0.44
450°C
2.76
0.71
600°C
28.7
2. 1
800°C
235.9
5.7
4-4
-------
A number of the platinum group catalysts tested by the automotive industry
are listed in Table 4-3. The HC and CO conversion efficiency of these cata-
lysts and the test conditions are also listed in this table (Ref. 4-6).
4.2.3.2 Base Metal Catalysts
Base metal catalysts employ metals or oxides of metals from the transitional
group of the periodic table of elements, including vanadium, chromium,
manganese, iron, cobalt, nickel, copper, and zinc. These catalysts are used
primarily by the chemical and petroleum industries in a variety of processes
such as hydrogenation, ammonia synthesis, methanol synthesis, and oxida-
tion reactions. A number of typical catalyst configurations used in these
applications are listed in Table 4-4 (Ref. 4-2).
Generally, several metals and their oxides are combined to increase the
effectiveness of the catalyst compared with single component configurations.
For instance, in the synthesis of methanol from carbon monoxide and hy-
drogen, mixtures of zinc oxide and chromium oxide are used extensively. As
illustrated in Figure 4-1, neither of these oxides is very effective by itself.
However, by adding about 20 to 40 percent chromium oxide to the zinc oxide,
the efficiency of the process increases substantially (Ref. 4-2). In this case,
chromic oxide acts as a promoter whose principal function is to minimize
the crystallite size of the catalyst. There is ample evidence that the effec-
tiveness of catalysts increases as the size of the crystallites is reduced.
i
A number of automobile and catalyst manufacturers are experimenting with
promoted base metal catalyst formulations containing small amounts of
platinum group metals for potential use in automotive emission control sys-
tems . These efforts are aimed primarily at the development of a less
expensive catalyst that approaches the performance of noble metal designs.
Platinum and palladium appear to be likely promoter choices (Ref. 4-7).
Table 4-5 presents a small fraction of the base metal catalyst formulations
tested by the automotive industry. Also listed in this table are the pollutant
concentrations at the inlet of the catalyst, catalyst efficiency, and operating
temperature (Ref. 4-8).
4-5
-------
Table 4-3. Catalytic Removal of HC and CO--Platinum
Group Metal Automotive Catalysts
Catalyst Composition,
% weight
!Pd/lPt/Al2O3
0.19Pt/{Ba)/Al2O3
0.19Pt/Al203
0.375Pt/0.5F/0.25Cl/Al203
0.1Pt/0.5F/Al203
0.4PtAl2O3
3.2Pd/Al203
Catalytic
Conversion, a
%
HC
83
69
92
61
93
55
70
CO
76
81
81
-
80
-
95
Test Conditions
Average
, Engine Catalyst
TEL, Type, Temperature,
ml /gal Duration cyl °C
2.7 188 hr 1
3 40 hr 8 435
3 -
present 12,000 mi 8 -
4.8 8
3 40 hr -
0.2 10,000 mi 8
Notes
Air added to exhaust
Air added to exhaust
Nonuniform distribution
of Pt
Glass -fiber thread and
fiber support; air
added to exhaust
aAt end of test period
bTEL: tetraethyl lead
-------
Table 4-4. Base Metal Catalysts--Industrial Processes
Type of Reaction
Examples
Catalysts
Ammonia synthesis
Iron, molybdenum,
osmium, uranium , etc .
Promoted iron (K-O, Al O.)
Methanol synthesis
CO
Copper, zinc oxide
Zinc oxide with chromia
Oxo-reaction
CO + H + RCH = CH-
Cobalt + thoria
' CHO
Hydrosulphurization
RSH
RH
Cobtal-molybdia
Iron and tungsten sulphides
Dehydration +
dehydrogenation
CH2 = CH -CH
Copper, Cu + Ni +
Oxidation
Vanadium pentoxide
-------
Table 4-5. Catalytic Removal of HC and CO--Base Metal Automotive Catalysts
Catalyst
Composition,
Percent Weight
MnO /CuO/NiO/CrO,
x
NiOjBa)/Al2O3
NiO!Ba)/Al2O3
7CuO/0.09SiO2/Al2O3
3CuO/7Si02
6CuO/6Cr 0 /Al O
CuO/Cr,O,
C, J
62CuO/5Co 0 /33A1 O
8CuO/4Co203/.V205/Al203
20CuO/0.1Ag20/Al203
6CuO/0. !Pd/6SiO2/Al2O3
O.ZS-15CuO/0.05-0.3Pd/SiO2/
A1203
5-10V205/Al203
50Co304/CaAl204
4-15 MnO /2-5Ti/Al,O,
X £* 3
4U308/A1203
^{oO /A1_O_
Exhaust Gas
Composition Before
Converter, ppm
HC
.
1,400
-
325
-
418
20,000
2,000
325
1,400
140
_
2,900
12,000
375
4,650
2,800
CO
_
29,000
10, 000
-
_
60,000
60,000
40,000
3,000
1,750
_
„
60,000
40,000
-
-
NOX
310-630
155
-
- •
1,000
„
4,000
1,500
-
-
-
_
-
-
-
-
Catalytic
Conversion,
Efficiency,
Percent3
HC
80
85
39
83
-
54
88
54
>80
76
69
72
77
62
70
71
CO
>80
77
45
95
-
.
95
_
72
>50
58
90
_
63
68
-
-
NOX
87-99
96
98
-
90
,
.
90
-
-
-
_
_
-
-
-
-
Test Conditions
TEL,"
ml/gal
present
0
3
3
present
present
present
1.6
12
-
3
12
0
present
12
3
2
Duration,
hr
341
3
120
100
-
350
.
238
50
-
-
60
_
600
75
-
-
Engine
Type
8 cyl
CFRC
CFR
CFR
2 cyl
8 cyl
8 cyl
8 cyl
CFR
-
CFR
CFR
1 cyl
1 cyl
CFR
8 cyl
-
Average
Catalyst
Temperature,
'C
425-650
485
-
-
380
_
285
480
-
-
510
„
„
-
-
-
-
Remarks
Air added for HC and
CO conversion
Federal Test Cycle
Odorous exhaust subse-
quently removed by
catalytic oxidation over
03 (7.3 vol %) added to
exhaust
High thermal stability,
high attrition resistance
Federal Test Cycle
Air added to exhaust
0. 12% S in fuel, multi-
layer catalyst
In presence of 3 TEL
HC conv. = 30 vol To
Air added to exhaust
I
00
At end of test period
tetraethyl lead
°Cooperative Fuel Research
-------
300 *C
325*C
350*0
' 375-C
*
-------
their maximum safe operating temperature, and the probable failure mode
are listed in Table 4-6. The substrates are manufactured from these mate-
rials and can be either pellets, monolithic structures, or fibrous materials.
In some special applications, metal wires or screens are used which are
coated with the catalyst material.
4.2.4. 1
Pellet Substrates
Aluminum oxide appears to be the material most frequently used in the manu-
facture of pellet type catalyst supports. These pellets, which can be of a
variety of shapes ranging from small spheres to elongated cylinders, are
manufactured by means of extrusion or drop tower processes. Spherical
pellets are frequently preferred because of their lower attrition rate. Some
heat treatment is required before the extrusion process to remove excess
moisture. Depending upon the selected temperature-versus-time profile
during dehydration of the pellets, a- or "Y-alumina is obtained. Below 1750 F,
the ~Y-pha.se of Al^O., predominates. This material has a very high surface
area, compared with the a -phase,which is formed at temperatures above
1750°F. Substrate surface area is a very important parameter because the
activity of a catalyst is dependent upon the available surface area.
Table 4-6. High Temperature Porous Materials
Material
Safe Operating
Temperature, °F
Probable Failure Mode
Al2O3-SiO2 fiber
fiber
SiC
3
-2 SiO
(alpha)
ZrO.
1900
2000
2700
2700
3000
3200
Sintering and crystallization
Sintering and crystallization
Oxidation
Sintering
Sintering
Sintering
4-10
-------
Pellet substrates are used primarily for chemical and petrochemical
applications. Some automobile manufactures are considering pellet catalysts
for use in emission control systems.
4.2.4.2 Monolithic Substrates
The monolithic type of substrate refers to a single unit structure, generally
of honeycomb design to provide the necessary surface area. While referred
to as monolithic, the structure may in fact consist of many layers of cor-
rugated sheets stacked together to comprise the total unit. Currently, mono-
lithic substrates coated with platinum group metal catalyst formulations are
favored for industrial tail gas purification systems and catalytic converters
for forklift truck applications. Also, monolithic platinum group metal cata-
lysts have been selected by most automobile manufacturers for use in their
1975 model year automobiles (Ref. 4-7).
Two ceramic corrugated monolithic configurations, a -alumina (Al Si Mag ©
614 and 775) and cordierite (Al Si Mag ^ 795) are being produced by American
Lava. Al Si Mag is the trade name for alumina-silica-magnesia compounds.
Cordierite is the mineralogical name of ternary oxide (2 MgO • 2 A19O^ : 5 SiO_)
(R)
and this compound is the primary constituent in Al Si Mag ^ 795. As indicated
in Table 4-7, the two types differ primarily in the porosity level. Al Si Mag ^
614 has a higher temperature capability but shows lower thermal shock resis-
tance (Ref. 4-9). Both rolled and stacked configurations are available (Fig-
ure 4-2). Currently, the stacked structures are preferred for automotive
applications because of their superior durability under the very severe thermal
conditions occurring in automotive installations.
Corning manufactures a truly monolithic, multi-cellular ceramic substrate
by means of a proprietary extrusion-type process. This substrate configura-
tion, which is designated W-l, has a square cell matrix and can be manu-
factured in a wide variety of shapes and sizes, as shown in Figure 4-3.
Physical properties of the W-l material are listed in Table 4-8 (Ref. 4-10).
Torvex ceramic honeycomb material is commercially available from DuPont
in two compositions, alumina and mullite (Ref. 4-11). The alumina composi-
tion contains about 96 percent cr-alumina, 3 percent magnesium aluminate,
and 1 percent mullite (3 A12O3 ' 2 SiO2). The mullite formulation is com-
posed of 55 percent mullite, 30 percent silica, and 15 percent a-alumina.
4-11
-------
Table 4-7. American Lava Monolithic Substrate Physical Properties
Materials
Property
Water absorption
Safe operating
temperature
Specific gravity of
material web
Specific heat
Coefficient of
thermal expansion
Thermal shock
resistance
Compressive strength
(parallel to passages)
Modulus of rupture
Thermal
c onductivity
Unit
%
°C
OF
Btu/lb °F
in/in/°F
70-1400°F
psi (5c/in SC)
0. 016 thick web
psi (4 in. centers,
1X1 in. beam,
5c/in SC)
Solid ceramic at
570°F ,
Btu in/hr-ft F
Alsimag 614
Dense 96%
Alumina
Highest mechanical
strength. Good
corrosion
resistance
0
1,538
2,800
3.65
0.21
4.4X 10"6
Fair
15,500
2,800
119.0
Alsimag 776
Porous 96%
Alumina
For catalyst
carriers
and special
applications
17
1,200
2, 192
2.5
0.21
3.9 X 10"6
Good
8,500
1,500
85.0
Alsimag 795
Cordierite
Good thermal shock
resistance.
Excellent as
catalyst carrier
25-30
1,200
2,192
1.7
0. 19
2. 1 X 10"6
Excellent
2,750
1,800
10
I
I-*
to
-------
ROLLED STRUCTURES .') ,
Figure 4-2. American Lava Stacked and Rolled
Corrugated Structures
4-13
-------
Figure 4-3. Corning Substrate Geometries
4-14
-------
Table 4-8. Corning W-l Monolithic Substrate
Physical Properties
MECHANICAL, STRENGTH
Crushing resistance
Axial 5,000 psi
Radial 500 psi
45° 50 psi
Impact, cumulative ft-lb
5 ft dropa
Radial, 140 ft-lb 8% wt loss
Axial, 140 ft-lb 3% wt loss
Vibration, 280 cpm, 11% wt loss
amp 1. 062 in, 80 min
Thermal expansion, 1.25 X 10~6 in/in°F
75°F to 1830°F
STRUCTURAL UNIFORMITY
Bulk density 28 ± 2 lb/ft3
Cell uniformity 0.060 to 0.061 in
(0.06 X 0.06 in)
Wall uniformity 0. 010 to 0. 0095 in
(0.010-in thick wall)
Cells/in2 (0.06 X 0.06 in 203 ± 5
cell, 0. 01-in wall)
Outside diameter ± 1/32 in
(Nominal 4-5/8 in)
Open frontal area 74% ± 2%
Porosity, by mercury 34%
porosimeter
Median pore size, by 7.5 microns
mercury porosimeter
aMounted in converter assembly
4-15
-------
The important physical properties of these materials are listed in Table 4-9.
Three geometrically different configurations are being manufactured by
DuPont for use as catalyst supports in tail gas abatement systems. These
include a simple honeycomb structure, a slanted honeycomb design, and a
crossflow-type structure that enhances flow turbulence and mixing.
According to DuPont the Torvex materials have a number of desirable fea-
tures, including good thermal shock resistance, chemical inertness, high
porosity, low bulk density, and high temperature capability. The alumina
type has a safe operating temperature limit of about 2730°F and mullite is
applicable up to 2460°F. In many cases, a wash coat is applied to the mono-
lith before deposition of the catalyst material. The wash coat consists prim-
arily of high porosity Y-alumina, which is added to increase the effective
surface area of the substrate material for the purpose of maximizing cata-
lyst activity. However, since Y-alumina is converted back to a -alumina at
about 1750 F, the operating temperature must remain below that value to
conserve the original effectiveness of the catalyst.
4.2.4.3 Fibrous Substrates
Fibrous substrate materials are manufactured from glass-ceramics or
asbestos. This type of substrate is utilized primarily in low-temperature
catalytic combustors manufactured by a number of companies for use in
domestic, agricultural, and industrial heaters.
4.3 CATALYST DEGRADATION
Catalytic converter^combustor configurations pose fundamental durability
problems related to the active catalyst constituents and the substrate mate-
rials. The service life of the catalyst is determined by two separate but
interrelated aspects: performance durability and physical durability. The
performance durability is most strongly impacted by decremental changes
in catalytic activity caused by:
1. Contamination of the catalyst surface by poisonous compounds
contained in the fuel and engine lubricating oil (e.g. , lead,
phosphorus, sulfur)
4-16
-------
Table 4-9. DuPont Torvex Monolithic Substrate Physical Properties
Composition
Maximum operating
temperature, °F
2 3
Surface area, ft /ft
Cell size, in. -
Bulk density, lb/ft3
Porosity, percent
Wall thickness, in.
m, i i -j. Btu-in
Thermal conductivity, • T
hr-ft -F
Coefficient of thermal expansion,
cm/cm °C
Ceramic Type
Alumina
96*-Al203/3 MgAl204/
1(3 Al-CX ' 2 SiCL)
M «J w
2732
60 - 384
1/8 - 3/4
10 - 34
60 - 80
0,03 - 0.05
6 (at 1470°F)
0.5 x 10"5
Mullite
55(3 AL.O, ' 2 SiO,)/
£* 3 £
30Si02/15*-AL,03
2462
60 - 384
1/8-3/4
20 - 35
60 - 80
0.03 - 0.05
6 (at 1470°F)
0.5 x 10"5
I
l-k
~J
-------
2. Reduction in the porosity of the substrate surface as a result
of alumina phase change occurring at excessive temperature
3. Alumina thermal shrinkage due to excessive temperatures
The physical durability of the catalytic unit is most strongly impacted by:
1. Thermal expansion differences between monolithic ceramic
substrates and their supporting container
2. Local melting of monolithic ceramic substrates due to local
overtemperature conditions
3. Failure of pellet retaining screens due to overtemperature
4. Cracking of monolithic ceramic substrates and breakup of pellet
substrates due to vibratory loads
These factors are analyzed in the following sections.
4.3. 1 Catalyst Poisoning
There is universal agreement that the effectiveness of a catalyst is reduced
substantially when exposed to certain metallic and nonmetallic substances,
such as lead, sulfur, and phosphorus. Among the possible poisoning mech-
anisms are the following:
1. Chemisorption of the poisonous materials on active surface sites,
thereby impeding the reactions between the reactants and the
catalyst
2. Deposition of a poison or coating on the catalyst surface, render-
ing the surface inaccessible to the reactants
Chemisorptive poisons are substances that are capable of forming stronger
bonds with the catalyst surface than the bonds normally formed by reactants
and products. According to Maxted (Ref. 4-12) the most toxic substances for
metal catalysts are the elements of groups Vb and VIb of the Periodic Table
of Elements, certain compounds containing metal ions, and molecules and
ions with multiple bonds such as carbon monoxide.
Lead, sulfur, and phosphorus have been identified as particularly severe
poisons for most metal catalysts. There are indications that base metal
catalysts are less susceptible to poisoning by these elements than platinum
group metal catalysts (Ref. 4-13). According to Wheeler (Ref. 4-14), catalyst
4-18
-------
poisoning occurs either nonselectively or selectively. In the nonselective
mode, the active surface of the catalyst is reduced uniformly by slow
adsorption of the poison, which results in a gradual reduction in the catalyst
efficiency. However, if the toxic substance is rapidly adsorbed, the mouths
of the catalyst pores are then selectively poisoned. This results in a very
rapid decline of the catalyst performance because the internal pore surfaces
are then no longer available to participate in the chemical reactions.
4.3.1.1 Lead Effects
The effect of lead contaminant level in the fuel on the performance of an
Engelhard PTX 3 automotive oxidation catalyst is illustrated in Figure 4-4
(Ref. 4-15). As indicated, the conversion efficiency decreases substantially
with increasing lead content in the fuel. Since these data are based on con-
stant speed operation of a spark ignition reciprocating engine, the data may
not be directly applicable to catalytic combustors of the type projected for use
in gas turbines. Also, the operating temperature of a catalytic combustor
for gas turbines would generally be higher than the average temperature of
an automotive catalyst, and this might alleviate potential lead poisoning
problems to some degree. In any case, a detailed test program would be
required to fully assess the effects of lead contaminant level in the fuel
on the performance of catalytic combustors.
4.3.1.2 Phosphorus and Sulfur Effects
Very little information is available regarding the deleterious effects of sulfur
and phosphorus on catalyst activity. Tests conducted by General Motors on
automotive catalysts indicate no "significant" differences in the effects of
these contaminants on base metal catalysts as opposed to noble metal cata-
lysts (Ref. 4-16). According to General Motors, the sulfur problem
could be alleviated if the catalyst temperature would be maintained above
1300°F at all times. Since most catalytic gas turbine combustors would be
operated at temperatures above that level, except perhaps at idle, sulfur may
not pose a serious problem in these applications. In General Motors' opinion,
phosphorus is bad, regardless of the operating temperature.
4-19
-------
I
ro
o
CO
u.
o
u
o
ce
60
40
20
PTX3 0.2% PT
DATE COMPLETED 3-28-72
EVALUATION CONDITION
INL. CAT. TEMP. 800° F
ENGINE SPEED 1000 rpm
WITHOUT AIR
1
V
FUEL:
Pb FREE
0.035 gm/gal Pb
0.07 gm/gal Pb
OIL: ASHLESS
DURABILITY CYCLE
MODE
1
2
3
4
TEMP.
1000° F
1200° F
1250° F
1200° F
TIME
14 min
15 min
6 min
7 min
200 400
TIME ON TEST, hr
600
Figure 4-4. Effect of Lead Additive on Catalyst Efficiency (Ref. 3-15)
-------
4. 3. 1. 3 Other Contaminants
Other potentially toxic substances contained in liquid fuels and lubricating
oil include zinc, barium, halides and other trace elements.
Coking of catalyst surfaces has been observed in tail gas abatement systems
under fuel-rich operating conditions. Although the effectiveness of the cata-
lyst is reduced by the coking layer, the original effectiveness can be restored
by operating the catalyst with excess air for some time.
Obviously, zero fuel and oil contaminant levels would be desirable, but
practical considerations such as lead contamination of the fuel during ship-
ment and the sulfur and trace metals normally found in kerosene-type fuels
dictate that trace levels of these contaminants will have to be tolerated by
catalytic combustors. The exact contribution of lubricating oil constituents
to catalyst deactivation is not evident.
4.3.2 Alumina Phase Change
Pellet substrates used in automotive catalysts are generally composed of
activated alumina material. Frequently, the monolithic substrates have a
wash coat of v-alumina which provides the high porosity required for high
catalyst activity. Although the alumina material does not melt until a temp-
erature of 3600°F is reached, it does undergo a phase change from Y-alumina
to the low porosity a -alumina at 1750°F, which results in a substantial
i
reduction in catalyst effectiveness.
4. 3. 3 Thermal Shrinkage
Pellet substrates are also subject to shrinkage in physical volume at elevated
temperatures. The effect of the thermal shrinkage is to reduce catalyst
effectiveness via reduced surface area and to cause "loosening" of the pellets
In the catalyst container unless special design features are added to com-
pensate for any loss in volume.
4-21
-------
According to General Motors (Ref. 4-17). excessive shrinkage has occurred
in early pellet designs at temperatures of about 1400 F. However, current
pellets have satisfactory shrinkage properties up to 1800 F, and further
improvements are likely.
No thermal shrinkage problems have been observed on monolithic substrates,
4.3.4 Thermal Differential Expansion
Both pellet and monolithic substrates have thermal expansion coefficients
different from the catalyst canisters housing them. Upon bed warmup, the
pellets can become looser in the bed than originally packed. Monolithic
catalyst elements also can become "loosened" with respect to the container.
Both may then be subject to mechanical attrition effects, as discussed in
Section 4. 3. 6.
4.3.5 Melting
The cordierite material used for monolithic substrates has a melting point of
approximately 2500 F to 2600 F. Even though the overall bed temperature
might be below this level, it is conceivable that local zones of the catalyst
might be subjected to overtemperature unless a homogeneous fuel-air mix-
ture is provided.
4.3.6 Vibration Effects
Although vibration levels in gas turbines are generally very low, it is con-
sidered appropriate to dwell briefly upon the deleterious effects that excessive
vibratory loads can have on the durability of catalysts. In automotive installa-
tions, breakup of pellet and monolithic catalysts has been observed by a num-
ber of small-car manufacturers (Ref. 4-7). Apparently, this problem is
related to high second order vibrational loads occurring in 4-cylinder engines.
Although this problem has not yet been resolved, it appears that catalyst/
substrate attrition and cracking could be alleviated by utilizing properly
designed containers.
4-22
-------
4.4 CATALYTIC CONVERTERS AND COMBUSTORS
Very little work has been conducted to date by industry and governmental
agencies on high-temperature catalytic combustors of the type applicable to
gas turbines and stationary power plants. The only known data in this
particular area were taken by Engelhard (Ref. 4-18) and by the Control Sys-
tems Laboratory of EPA (Ref. 4-19). In both cases, small catalytic com-
bustor sections were tested over a limited range of operating conditions.
To provide the information required for a meaningful assessment of high-
temperature catalytic combustors, it is considered appropriate to briefly
discuss the operational characteristics and the problem areas related to
the following catalytic converter/combustor categories:
1. Automotive oxidation catalysts
2. Tail gas abatement systems
3. Low-temperature catalytic burners
4. High-temperature catalytic burners
4. 4. 1 Automotive Oxidation Catalysts
Oxidation catalysts are being considered by all domestic automobile manu-
facturers as principal emission control system components projected for
their 1975 and later model year cars. The catalytic converter type that
appears most frequently in the first-choice systems is the platinum group
metal-monolithic configuration, exemplified by the Engelhard PTX design.
These catalysts are installed downstream of the engine exhaust manifold
in order to convert the unburned hydrocarbon and carbon monoxide species
leaving the engine. Under normal operating conditions of the engine, the
catalyst temperature is of the order of 1200°F to 1400°F. However, at high
engine loads, the temperature increases to about 1700 F to 1800 F, and even
higher temperatures have been observed at excessively fuel-rich operating
conditions of the engine. The monolithic catalysts are designed for space
velocities of the order of 100,000 hr~ .
4-23
-------
450
CM
CO
400
o
I °
I S
I £
350
«J, 30°
5 °
00
< Q_
250
CO
200
o
o
150
J_
T
95% CONFIDENCE
LIMITS (±2a)
STANDARD PTX
I
IMPROVED PTX
(July 1971)
IMPROVED PTX
(January 1972)
VIRGIN 1400 1600 1800
TEMPERATURE |°F] OF CUMULATIVE AGING IN AIR
(24 hr at each temp.)
Figure 4-5. Engelhard Monolithic Catalysts with Improved
Catalytic Thermal Stability (Ref. 4-20)
4-24
-------
The most pervasive problem relative to the 1975 emission control systems
appears to be the lack of adequate catalyst durability. When fresh, the per-
formance of many catalysts is adequate to meet the Federal emission stand-
ards for light-duty vehicles. However, with increasing car mileage, the
effectiveness of the catalyst deteriorates steadily due to the combined effects
of poisoning by lead, sulfur, and phosphorus contained in trace quantities in
the fuel, and to catalyst overheating which can occur under extreme engine
operating conditions. Catalyst failures have occurred also as a result of
cracking and local melting of the substrate material, attributed to excessive
vibratory loads and oveirtemperature conditions.
The catalyst durability problem is treated in several ways. One of these is
the use of improved substrate materials. Another technique involves improve-
ments in the control of the catalyst operating environment. In addition, efforts
are under way to improve the lightoff and thermal stability characteristics of
the catalysts. The progress made by Engelhard in this regard is illustrated
in Figure 4-5 (Ref. 4-20). These data show the greatly increased retention
of activity of the improved catalyst, even after severe thermal aging. Similar
results were reported by Matthey Bishop (Ref. 4-21). As shown in Fig-
ure 4-6, the lightoff temperature of the Matthey Bishop improved catalyst
increases with aging temperature. However, the steady-state efficiency of
the catalyst is essentially unaffected by aging.
I
There are some indications that certain base metal catalysts might have
lower lightoff temperatures than noble metal catalysts. This is illustrated
in Figure 4-7 which presents HC and CO conversion efficiencies determined
by General Motors in bench tests of noble metal and base metal catalysts
(Ref. 4-17).
4. 4. 2 Tail Gas Abatement Systems
Catalytic units are employed in a number of tail gas abatement systems used
by the chemical process industry to reduce the hydrocarbon and nitric oxide
4-25
-------
100
£ 50
o
o
0
200
INITIAL TEST
650-750° C
840-970° C
- 1100 C
300
INLET TEMPERATURE, °C
400
STATIC ENGINE DURABILITY TEST:
EFFECT OF TEMPERATURE OVER
A 24-hr TEST ON AEC 3A
Figure 4-6.
Effects of Thermal Aging on Matthey Bishop
AEC3A Catalyst (Ref. 4-21)
4-26
-------
I
t\)
A HC
O CO
- NOBLE METAL
BASE METAL
400
500
TEMPERATURE, °F
600
700
Figure 4-7.
Conversion Characteristics of Base Metal and
Noble Metal Catalysts (General Motors Bench
Test Evaluation) (Ref. 4-17)
-------
emissions from these processes. The operational characteristics of these
systems are briefly discussed in the following sections.
4.4,2.1 Hydrocarbon Fume Abatement
In the past, direct flame incineration was the favored hydrocarbon fume
abatement method. More recently, catalytic combustion has become attrac-
tive, primarily for economic reasons. According to a study conducted by
Hein (Ref. 4-22), the cost of fume abatement by the lower temperature
catalytic process is only about 20 to 70 percent of the cost of direct flame
burners, depending upon the particular application. Based on the encourag-
ing results obtained by a number of investigators at UCLA (Refs. 4-23 to
4-25) in the area of catalytic oxidation of various hydrocarbon species,
DuPont carried out an experimental program to establish the feasibility of
catalytic combustors in fume abatement systems. In these tests, platinum
catalysts were used in conjunction with pellet and Torvex mullite monolithic
substrate configurations (Ref. 3-26 and 3-27).
With regard to catalyst manufacture, the pellet substrate materials were
first coated with a Y-alumina wash coat, which has a specific active surface
2
area of 180 to 200 m /gm. The monolithic substrates, which have a super-
2
ficial surface area between 1. 17 and 1.26 m /liter of substrate, were coated
also with Y-alumina to provide a catalytic surface area of about 2. 5 X
4 2
10 m /liter. The platinum was deposited on the substrate from an aqueous
solution by chemical reduction, or by solution impregnation and subsequent
hydrogen reduction. The platinum loadings used were 2 and 3 percent of the
Y-alumina mass. The catalysts were operated at space velocities between
30, 000 hr"1 and 175,000 hr'1.
The pressure drop characteristics of these catalytic combustor units are
presented in Figure 4-8. As indicated, the straight cell honeycomb has the
lowest pressure drop, followed by the cross-flow honeycomb and the
spherical pellet configuration.
4-28
-------
0 10 I 00
PRESSURE DROP. IN H,0/IN 9£D
Figure 4-8. Pressure Drop of Monolithic and
Pellet Catalysts (Ref. 4-26)
Fume abatement effectiveness data obtained by DuPont are shown in
Figures 4-9 through 4-12 (Ref. 3-26). Figure 4-9 presents the conversion
efficiency of n-heptane as a function of reactor inlet temperature for a cross-
flow honeycomb catalyst, a spherical catalyst, and, for comparison, a
direct-flame combustor. As indicated, the catalytic burners are substantially
more effective at the lower temperatures. These characteristics clearly
show the advantage of catalytic combustors,which can operate efficiently at
rather low flow inlet temperatures. This results in substantial fuel cost
savings.
The effects of space velocity on the lightoff temperature of the straight-flow
honeycomb, cross-flow honeycomb, and spherical pellet catalysts are
depicted in Figures 4-10 through 4-12. In each case, the lightoff tempera-
ture increases with increasing space velocity. The cross-flow honeycomb
design has the lowest lightoff temperatures of the three configurations.
Based on these results, it is concluded that catalyst geometry and size have
a strong effect on the performance characteristics of catalytic combustors,
especially in the low temperature regime. DuPont feels that its honey-
comb support offers excellent conversion efficiency at low pressure drop.
4-29
-------
100
S
o- 80
j CELL
- CROSS-FLOW
CERAMIC
•z • r HONEYCOMB
z
8 so h
° 40
o> 20
100
TOTAL COMBUSTION
CURVES
£CT FLAME
PRE-HEATER
CATALYST BEOS 4"DEEP
10 % L.E.L. 4.44 SCFM/INZ
ZOO 300 4OO
REACTOR INLET TEMP, «C
5OO
Figure 4-9.
n-Heptane Fume Abatement
Effectiveness (115, 000/hr
space velocity) (Ref. 4-26)
100
80
SO
o
O 40
30,000 SV
70,000
100,000
g"HEX-CELL
CERAMIC
HONEYCOMB
7'DEEP BED
100
200 300
REACTOR INLET TEMP, «C
400
Figure 4-10. n-Heptane Conversion
Efficiency vs Tempera-
ture and Space Velocity--
Torvex Straight-Through
Honeycomb (Ref. 4-26)
4-30
-------
100 -
200 300 400
REACTOR INLET TEMP, «C
Figure 4-11. n-Heptane Conversion
Efficiency vs Tempera-
ture and Space Velocity-
Torvex Cross-Flow
Honeycomb (Ref. 4-26)
100
80
8j 60
g20
100 200 300 400
REACTOR INLET TEMP. "C
Figure 4-12. n-Heptane Conversion
Efficiency vs Tempera-
ture and Space Velocity-
Granular Catalyst Sub-
strate (Ref. 4-26)
4-31
-------
The lightoff characteristics of a number of other hydrocarbon fuels oxidized
in a monolithic platinum catalyst and a monolithic palladium catalyst are
presented in Figures 4-13 and 4-14 (Ref. 4-27). In each case, the hydro-
carbon concentration was 10 percent of the corresponding lower explosive
limit (LEL) of the fuel. For methane, the LEL value in air is 5 percent,
and for the other compounds the LEL. in air varies between approximately
1 and 3 percent. Comparison between Figures 4-13 and 4-14 indicates that
platinum is more active than palladium in catalyzing n-heptane. In these
tests, methane has been the most difficult fuel to oxidize. Similar results
were obtained by Accomazzo and Nobe (Ref. 4-24) with a copper oxide pellet
catalyst. With increasing carbon number in the fuel, they observed a marked
reduction in catalyst lightoff temperature.
The effects of platinum content, chemical composition of the substrate, and
substrate surface area are illustrated in Figure 4-15. As shown, the cata-
lyst lightoff characteristics are rather insensitive to variations in these
parameters. It is conceivable, however, that catalyst aging effects might
vary for the different catalyst loadings.
4.4.2.2 Nitric Oxide Tail Gas Abatement
Catalytic combustors are widely used in nitric oxide abatement systems
employed in nitric acid manufacturing plants. In these plants, nitric oxide
is generated via catalytic oxidation of ammonia over a platinum-rhodium
catalyst. The nitric oxide then reacts with residual oxygen and forms nitro-
gen dioxide which is absorbed in water to form the nitric acid. The high
pressure gases leaving the tower are composed of approximately 0. 3 per-
cent NO + NO_, 3 percent O2, and 96 percent N2, and are expanded through
a turbine to obtain useful power (Refs. 4-28 and 4-5). To maximize the
power output of the turbine and to eliminate the NO, the tail gases are
heated before expansion by catalytic combustion of the NO,,. This is
accomplished by injecting a reducing gas, such as methane or hydrogen,
into the gas stream and oxidizing the mixture catalytically to form NO,
4-32
-------
100
I
to
OJ
80
tj
UJ
u.
u.
UJ
<
<
o
60
40
20
100
SOLVENT OXIDATION ON Pt/AI203/B-HONEYCOMB
AT 10°/=J_ELa IN AIR AND 80,OOOSVb
O ETHANOL
A N-HEPTANE
D MIBK
• M-XYLENE
• MEK
A ETHYL ACETATE
200 300
CATALYST BED INLET TEMPERATURE, °C
400
lean explosive limit
"space velocity
Figure 4-13. Lightoff Characteristics--Platinum Catalyst (Ref. 4-27)
-------
100
80
0)
Q.
U 60
z
LU
U-
u.
u
H 40
O
20
200
I ' I ' I
HYDROCARBON OXIDATION ON Pd/AU03/B-HONEYCOMB
AT 10% LELa IN AIR AND 80,OOOSV^
O N-HEPTANE
A PROPANE
D METHANE
• METHANE ON
Pd/NIO(109me/g)
B-HONEYCOMB
300 400 500
CATALYST BED INLET TEMPERATURE. °C
600
lean explosive limit
space velocity
Figure 4-14. Lightoff Characteristics--Palladium Catalyst (Ref. 4-27)
-------
100
I
w
Ul
-------
and H?O. This portion of the total abatement is the so-called decolorization
phase because all the reddish-brown NO7 is converted to invisible NO. By
L* !-
injecting additional fuel, all O_ is oxidized and the NO is then converted to
N_. Although theoretically, all NO is converted in the process, the efficiency
of actual abatement systems is somewhat lower. Chemico feels that con-
version efficiencies of about 90 percent are feasible on a long-term basis
(Ref. 4-29).
In general, platinum and palladium catalysts, supported on pellet or mono-
lithic honeycomb substrates, are utilized in NO abatement systems. Units
of this type have been operated continuously for six to eight years (Ref. 4-28).
Experience with these combustors has indicated that pellet catalysts are more
durable, while the honeycomb configurations have size and pressure drop
advantages. According to Engelhard, space velocities of the order of
100, 000 hr~ are feasible for monolithic configurations, while pellet designs
are limited to about 30, 000 hr~ . Natural gas, which contains a high per-
centage of methane, is the most popular fuel for use in NO abatement units.
As illustrated in Figure 4-14, methane has a higher lightoff temperature
than most hydrocarbons and this requires preheating of the tail gas to about
500 C in order to achieve acceptable conversion efficiency.
Typical catalyst operating temperatures are of the order of 1500 F, which
represents the limit for continuous operation for both honeycomb and spherical
substrate configurations considered by Engelhard for this application
(Ref. 4-28). For short durations, the temperature can be safely increased
to about 1650 F, without decreasing catalyst life. If synthetic gas containing
hydrogen is available the catalyst inlet temperature can be reduced to about
400 F and, as a result, the exit temperature remains sufficiently low to
assure long catalyst life (Ref. 4-28).
Initially, excessive performance degradation was observed on these catalysts
as a result of a loss in active catalyst surface area; the loss was caused by
sintering of the refractory substrate materials. Subsequently, stabilized
substrates were developed in which the active surface area was increased
4-36
-------
and sintering was minimized. As shown in Figure 4-16, the durability of the
stabilized catalyst is excellent (Ref. 4-28).
The effect of space velocity on unaged and aged stabilized catalysts is shown
in Figure 4-17. As indicated, the performance of the fresh catalyst is con-
stant over the range of space velocities shown. Conversely, the effectiveness
of the aged specimen decreases substantially as space velocity is increased
above 100, 000 hr"1.
According to Engelhard (Ref. 4-5), catalyst performance and durability are
strongly impacted by carbon deposition, catalyst poisoning, and overtemp-
erature. All these factors result in lower catalyst effectiveness, higher
lightoff temperatures, and reduced catalyst life. In general, carbon deposi-
tion is related to excess fuel and/or poor air-fuel mixing upstream of the
catalyst. The deposits can be removed by carefully operating the catalyst
in an oxidizing atmosphere for short periods of time.
Although the effects of poisonous ingredients in the tail gas have not yet been
quantified, Engelhard feels that sulfur, halogens, and iron oxide are bad for
its catalysts. Apparently, platinum catalysts are less affected by sulfur
than are palladium catalysts. However, the higher activity of palladium
usually justifies incorporation of a desulfurization step in the abatement
process.
C & I Girdler, a manufacturer of nitric acid plants, has evaluated a number
of catalytic combustor units designed for use in NO abatement systems
(Ref. 4-30). Based on Girdler's own test data, it has selected pellet cata-
lysts as first choice for use with methane. However, for economic reasons,
monolithic platinum or palladium catalysts may be more attractive for those
applications in which hydrogen-bearing purge gas is available.
4-37
-------
0
UJ
.100
(
80
60
40
20
STABILIZED CATALYST
Ul
UNSTABILIZED CATALYST
500 1000
AGING, hr
2000
Figure 4-16. Performance of Stabilized and
Nonstabilized Catalysts--Nitric
Oxide Abatement (Ref. 4-28)
UNAGED
NITRIC OXIDE
REMOVAL EFFICIENCY^
ro .u o oo c
o o o o o c
N. AGED 1000
^sxr^ HOURS
li I
1000,000 150,000
300,000
SPACE VELOCITY, hr
-1
Figure 4-17. Performance of Stabilized Aged and
Unaged Catalysts vs Space Velocity-'
Nitric Oxide Abatement (Ref. 4-28)
4-38
-------
4. 4. 3 Low-Temperature Catalytic Combustors
For the purpose of this study, low-temperature catalytic combustors are
characterized by operating temperatures below about 800 F. Although no
visible flame is produced under these conditions, the amount of heat gener-
ated by the combustor is equal to the heat produced in a conventional flame-
type burner. Typically, the specific heat release rate of a catalytic burner
2
of this type is less than 100 Btu/hr-in . This type of burner is commercially
available from a number of manufacturers for use in domestic, recreational,
agricultural, and industrial heating and drying operations. Because of its
low surface temperature, the burner offers a safety feature not available
with conventional burners.
A typical low-temperature catalytic combustor unit is shown schematically
in Figure 4-18 (Ref. 4-31). In this design, gaseous fuel enters through the
metering orifice into the metal disc, which is designed to evenly distribute
the gas to all parts of the burner. The active catalyst material is deposited
on the front surface of the ceramic or asbestos fiber pad which serves both
as catalyst support and as an insulator to minimize back radiation (Ref. 4-32).
METAL DISH
ORIFICE
CATALYTIC
PROTECTIVE SCREEN
Figure 4-18. Catalytic Space Heater
Schematic (Ref. 4-31)
4-39
-------
In general, platinum, is used as the catalytic agent, although less expensive
base metals are sometimes added as promoters to enhance the catalyst
oxidation reactions. As discussed in Section 4.4. 1, base metal catalysts
may have a lower lightoff temperature. Combustion takes place on the sur-
face of the coated fibers,which are protected by a wire screen. The air
required for combustion is drawn in through the wire screen. Generally,
because of the associated improvement in air recirculation to the burner
surface, the performance of these units improves with increasing ratio of
perimeter to cross-sectional area of the pad. Burner lightoff is accom-
plished by preheating the catalyst bed either electrically or by means of a
pilot light. The low-temperature combustor test data indicate very high HC
emissions. In conventional combustors the HC emissions are generally
much lower.
The AGA, Matthey Bishop, and the EPA have tested a number of different
low-temperature catalytic combustors. The pertinent information provided
by these organizations is briefly discussed in the following sections.
4.4. 3. 1 American Gas Association
Several years ago, the AGA conducted a research program aimed at the
development of catalytic burners for use in kitchen ranges and other related
applications. As part of that effort, a number of catalytic units were tested
and evaluated. The substrate material used in these designs consisted of
asbestos fiber pads which were procured from four different manufacturers.
The substrates •were coated in-house with a variety of catalyst materials,
including platinum, palladium, platinum-palladium mixtures, and non-noble
metals. The program was limited to gaseous fuels, primarily natural gas
and liquified petroleum gas (LPG). Manufactured natural gas was utilized
in some of the tests. The burner temperatures were held below approxi- '
mately 900 F, which was the desired limit for the projected application. In
the course of the program, a number of serious problems were encountered
including incomplete combustion of the fuel and unpredictable burner lightoff
characteristics. The program was canceled by AGA before completion
(Ref. 4-33).
4-40
-------
4.4.3.2 Control Systems Laboratory, EPA
In 1971, the Control Systems Laboratory of the EPA conducted an
experimental program to assess the potential of catalytic combustion for
pollution-free domestic heating applications (Ref. 4-34). A total of eleven
commercially available catalytic heaters were tested. Eight of these units
were operated with propane and three with lead-free gasoline.
The emission data obtained by the EPA are shown in Table 4-10. As indicated,
the NO emissions are very low on all units. The CO emissions of the gas-
.X
fired heaters varied between 20 and 1560 ppm, while the liquid-fueled units
had CO emissions between 665 and 2350 ppm. The HC emissions of the gas-
fired units were significantly lower than those of the liquid-fuel designs.
Apparently, there is no correlation between emissions and burner heat out-
2
put. The specific heat release rates of the burner (Btu/hr-ft ) are also
presented in Table 4-10. Again, there is no correlation between the emissions
and this particular parameter.
In an effort to resolve the large differences in the emissions obtained with
these burners, the EPA conducted a series of parametric tests using the old
Turner burner listed in Table 4-10 as the test unit. Variation of the thick-
ness of the catalyst pad had some effect on the HC and CO emissions. Mini-
mum HC and CO emissions were obtained with a one-inch pad. Doubling the
platinum content in the pad resulted in some reduction in HC, but increased
the CO emissions. However, this unexpected trend may well be the result
of measuring inaccuracies.
By far, fuel flow rate has been the most important parameter with regard to
HC emissions. This is because the combustion is controlled by the rate of
oxygen diffusion to the pad surface. As illustrated in Figure 4-19, the HC
emissions are minimum at a fuel flow rate of about 650 cm /min, standard.
Over the range of fuel flow rates considered, the CO emissions are almost
independent of fuel flow.
4-41
-------
Table 4-10. Emissions from Commercial Catalytic Heaters at
Maximum Heat Out puta
Heater Type
LPG MODELS
Bernzomatic
Cargo safe
Impala
McGinnis
Primus
Turner (old style)
Turner (new style)
Zebco
WHITE GAS MODELS
Coleman (old style)
Coleman (new style)
Thermos
Rated
Heating
Value,
Btu/hr
7000
6000
8000
8000
8000
7000
7000
7000
8000
5000
7000
Measured
Heating
Value,
Btu/hr
4800
5970
5970
7785
6365
3375
4200
8040
8440
4750
4915
Specific
Rated Heat
Release
Rate,
Btu/hr -in2
159
52
79
35
N.A.
93
93
93
N.A.
N.A.
N.A.
Concentration, ppm
(air-free)
NO
2
16
8
6
12
0
4
0
13
8
32
CO
46
124
174
20
1,560
27
36
205
2,350
665
1,280
HC
2,550
8,650
4,050
1,000
10,250
1, 110
1,335
370
19,000
7,500
18,000
Fuel control valves set on "High"
-------
I
o
t/>
«o
2
lii
O
m
a:
o
I
1000
900
800
700
600
500
400
300
200
100
— \ STANDARD PAD
\ V
STANDARD WITH
1.0% MANGANESE
DOUBLE-PLATINUM
LOADING
1
500 550 600 650 700 750 800
PROPANE FLOW RATE, std cm3/min
Figure 4-19. Hydrocarbon Emissions vs Fuel Flow Rate-
Catalytic Space Heater (Ref. 4-34)
850
4-43
-------
4.4.^.3 Matthey Bishop, Inc.
Matthey Bishop has developed a number of platinum group catalysts for use
in low temperature heater applications. These units, which are known as
Cataheat systems, are designed to operate on various fuels including methane,
propane, butane, natural gas, and white gasoline (Ref. 4-31). In the gaseous
fuel designs, combustion is initiated by means of a pilot light or simply by
lighting a match. In the liquid-fuel units, the fuel is brought up to the cata-
lyst pad by means of a wick arrangement. According.to Matthey Bishop, the
life of the catalyst is strongly affected by impurities contained in the fuel. If
pure hydrocarbons are used, the life should be infinite.
Emission data provided by Matthey Bishop are listed in Table 4-11 along with
other pertinent information. As indicated, the CO emissions, adjusted to air-
free operation of the burners, are between 100 and 150 ppm. However, the
HC emissions are rather high, running between 1500 and 6000 ppm of hexane
for the three burners.
Table 4-11. Matthey Bishop Catalytic Heater
Characteristics (Cataheat System)
Type
Cataheat L
(liquid)
Cataheat P
(HD5 propane)
Calaheat M
(methane-
West Texas
natural gas)
Rating
J/m~
Btu/in2
123 X 106
(75)
123 X 106
(75)
123 X 106
(75)
Air-free CO
Measured by
AGA
Procedures,
ppm
125
100
150
Approx.
Ignition
Temp. ,
°C
250-280
250
330
Approx.
Surface
Temp. ,
°C
300-420
290-500
360-470
Unburned
Hydrocarbons
as Hexane,
ppm
6000
1500
5000
Life, h
5000+
10,000+
10,000 +
4-44
-------
4.4.4 High-Temperature Catalytic Combustors
Because of their low NO emissions potential, catalytic combustors are now
jt
being considered by a number of organizations for use in automotive, air-
craft, and stationary gas turbines. The majority of the mobile engines are
designed for turbine inlet temperatures of about 1900°F and lower. The tur-
bine inlet temperatures of uncooled industrial gas turbines are somewhat
lower.
Based on current catalyst-substrate materials technology, it appears that
catalytic combustors could be designed to operate at these temperatures. At
these levels, the NO emissions which are related directly to the combustion
• 3C
temperature are expected to be very low.
Little information is currently available relative to the performance and
emission characteristics of catalytic combustors operating at typical gas
turbine temperatures. The only known exploratory catalytic combustor tests
were conducted by Engelhard and by the EPA, using a small Engelhard cata-
lyst specimen. Preliminary evaluation and test programs have been initiated
by NASA/Lewis and by the AFAPL. These programs are briefly discussed
in the following sections.
4.4.4. 1 Air Force Aero Propulsion Laboratory
4.4.4. 1.1 Program Description
Efforts are currently under way at AFAPL to develop and evaluate gas turbine
combustor modifications in an effort to reduce the CO, HC, and NO emissions
from military jet aircraft. Both current and advanced technology modifica-
tions are being considered. The current technology modifications consist of
minor as well as major design changes involving the combustor liner, fuel
injector, and fuel-air mixing devices, and incorporation of water injection.
The advanced technology modifications include variable area geometry com-
bustors, staged fuel injection, and premixed fuel-air injection schemes.
These approaches are expected to be very effective in terms of CO and HC
4-45
-------
emission reduction, but probably will be much less effective in reducing the
NO^ emissions, especially at takeoff and climb thrust levels of the engine
(Ref. 4-35).
Current and projected NO emissions from jet engines are presented in Fig-
Ji,
ure 4-20 (Ref. 4-36). Curve 1 in this figure represents current commercial
and military jet engines, curve 2 is the AFAPL projection for 1979-type
hardware, and curve 3 is based on modified conventional combustors plus
water injection. Since the projected reductions are quite modest, AFAPL
decided some time ago to search for novel combustor techniques which have
the potential of significantly lower NO emissions. Based on these con-
X
siderations, it negotiated a zero dollar contract with Engelhard covering
delivery by Engelhard of a catalytic combustor that could be used by AFAPL
in simulated engine test work. In return, Engelhard will receive all test
data acquired in the program and will have some control regarding the publi-
cation of these data. The AFAPL test program is designed to provide funda-
mental performance data on the catalytic combustor under a variety of
operating conditions. Currently, the program is considered to be a phase I
feasibility effort which may be extended if the initial test results are
favorable.
The design, fabrication, and setup of the test section illustrated in Figure 4-21
have been completed (Ref. 4-3-6). Current efforts are concentrated on the
final development of an optimum fuel injection and fuel-air mixing scheme
which is capable of providing a uniform flow velocity and mixture ratio
distribution at the inlet of the catalytic combustor under all operating condi-
tions. Compressed plant air, which is available at pressures up to 17 atm
and temperatures up to 850°F, will be used in the program to simulate the
operating conditions at the inlet of a typical jet engine combustor.
4-46
-------
u
r>
u.
o
8
x
LJ
Q
2
40
30
20
10
"UNCONTROLLED"
DATA CORRELATION
50% REDUCTION
75% REDUCTION
200
400
600
800
1000
COMBUSTOR INLET TEMPERATURE, °F
Figure 4-20. Aircraft Gas Turbine Emission Goals
(Ref. 4-35)
4-47
-------
rAIR REGULATOR-FUEL MIXER
i
£>
oc
EXHAUST DUCT
CATALYTIC UNIT /
TOTAL FLOW
VENTURI
-AIR FLOW
SEPARATOR
FLOW SEPARATOR OUTER DUCT
CATALYTIC UNIT OUTER DUCT
Figure 4-21. Catalytic Combustor Test Setup--Air Force Aero Propulsion Laboratory
-------
The program is comprised of the following six tasks, which are briefly
discussed in Sections 4. 4. 4. 1. 2 through 4. 4. 4. 1. 7.
1. Ignition limits
2. Heat release rates
3. Fuel vaporization effects
4. Catalyst degradation
5. Transient operating conditions
6. Emission characteristics
4.4.4.1.2 Ignition Limits
The ignition characteristics of the catalyst will be evaluated at the beginning
of the test program and the optimum flow rate, air-fuel ratio, air pressure,
and temperature required for lightoff will be determined. As part of this
task, the feasibility of using APU exhaust gas for catalyst preheating will be
also investigated.
In addition, flashback and blowoff limits of the combustor will be established
for various operating conditions. To prevent flashback, AFAPL has tentatively
selected a lower flow velocity limit of about 20 ft/sec. Also, screens and/or
other fixes may be installed upstream of the catalyst to prevent flashback
into the flow injection-mixing region of the setup.
4.4.4.1.3 Heat Release Rates
One of the most important factors affecting feasibility considerations of cata-
lytic combustor is the maximum specific heat release rate of the unit.
Obviously, this parameter has a direct impact on the size and weight of the
combustor. Current jet engine combustors are designed for heat release
6 3
rates of about 10 X 10 Btu/hr-ft -atm. In the test program, the heat
release rate will be varied by gradually increasing the fuel and air flow
rates until a sharp rise is observed in the emissions or until other combustor
performance parameters are significantly compromised.
4-49
-------
4. 4. 4. 1.4 Fuel Vaporization Effects
In this task, the degree of vaporization of JP-type fuels required for catalytic
combustion will be evaluated. Although the test plan initially called for test-
ing of a number of fuels, including gaseous and certain distillate fuels, the
program in its present form is limited to JP fuels.
The degree of fuel vaporization will be varied by varying (1) the distance
between the fuel injection points and the catalyst, (2) the air flow rate through
the fuel-,air mixing section, and (3) the temperature of the incoming fuel.
According to AFAPL, a small amount of nonvaporized fuel might be tolerable.
4.4.4.1.5 Catalyst Degradation
The AFAPL test plan provides for a 100-hour durability test of the catalytic
combustor. The objective of this test is to determine the degree of deteriora-
tion in catalyst lightoff temperature and other performance parameters such
as emissions and combustion efficiency. However, the scheduled test time
of 100 hours is considered insufficient to yield the information required for a
comprehensive assessment of the durability characteristics of this catalytic
combustor.
4.4.4.1.6 Transient Operating Conditions
This program task is concerned "with the determination of the thermal shock
characteristics of the catalyst/substrate materials. This will be accomplished
by exposing the catalyst to rapidly changing temperature conditions of up to
500 F/sec. Vibration tests are not considered in this task.
4.4.4.1.7 Emission Characteristics
The AFAPL test plan calls for emission measurements over a wide range of
simulated engine operating conditions. Air-fuel ratios will be varied between
approximately 25 and 120. These values cover the range between full power
and idle operation of a typical jet engine. The air pressures and inlet temp-
eratures will be varied accordingly. For instance, engine idle will be,simu-
lated by operating the combustor at an air inlet temperature of about 350 F
and a pressure of 4 atm.
4-50
-------
The instrumentation available at AFAPL for emission measurement includes
heated FID for HC, nondispersive infrared for NO and CO, and chemi-
luminescence for NO . This instrumentation is considered adequate to assure
.X
the required accuracy of the test data.
AFAPL expects low NO emissions from catalytic combustors because of the
5v
low flame temperatures,which result from operation of the combustor in the
lean fuel-air ratio regime. Conversely, in conventional gas turbine com-
bustors, the fuel is burned near stoichiometric in the primary zone and this
results in much higher temperatures and NO formation rates. AFAPL
feels that NO emission indexes (defined as Ib of NO /1000 Ib of fuel burned)
x x
below 1. 0 might be achieved with catalytic combustors under all operating
conditions. As shown in Figure 4-20, this would constitute a significant
improvement over current combustor designs.
Idle combustion efficiencies above 99. 5 percent are predicted by AFAPL for
catalytic combustors. For comparison, current aircraft gas turbine engine
combustors have idle efficiencies of the order of 90 to 98 percent, depending
upon the particular combustor-engine configuration considered. The pro-
jected increase in combustor efficiency results in substantially lower CO
and HC emissions over typical landing and takeoff duty cycles of the aircraft.
4.4.4.2 Engelhard Industries
4.4.4.2.1 Program Description
To date, most of the catalytic combustor work conducted by this company
has been performed in a simulated gas turbine test rig. It is Engelhard1 s
opinion that the data obtained from these tests are scaleable to large units.
The CO and HC emissions in larger combustors should be less than those
from this unit, because of the more favorable surface-to-volume ratio of the
larger units. This fact results in a reduction of the quenching effects of the
CO and HC oxidation reactions in the outer regions of the catalyst.
4-51
-------
Based on an in-house feasibility study of catalytic combustion concepts,
Engelhard, has identified a number of potential applications. These include
kitchen equipment, hot water heaters, stationary gas turbines, automotive
gas turbines, and aircraft gas turbines (Ref. 4-18).
Currently, Engelhard has working agreements with a number of outside
organizations, including NASA/Lewis and the AFAPL. Under the terms of
these agreements, it has provided one catalytic combustor each to NASA
and to AFAPL for testing by these organizations over a range of simulated
gas turbine operating conditions.
All catalytic combustion work performed to date has been company-sponsored.
However, negotiations regarding the development of catalytic combustors
are currently being conducted with a number of gas turbine manufacturers.
To date, Engelhard has not been involved in very detailed evaluation studies of
the applicability of catalytic combustors to steam boilers. In the past, the
Atomic Energy Commission (AEC) had shown interest in a potassium boiler
and Engelhard proposed a catalytic combustor arrangement for this applica-
tion using imbedded boiler tubes. However, this concept was rejected by
the AEC because it felt that the ceramic substrate would crack under the
specified operating conditions. Engelhard still believes that the originally
proposed concept could be developed into a viable design.
4.4.4.2.2 Test Fuels
A number of different gaseous and liquid fuels have been utilized in Engelhard's
catalytic combustor test work. These fuels include pure propane, commercial
propane (major constituents about 73 percent propane and 24 percent propy-
lene), methane, low-energy coal gas (heating value approximately 120 to
160 Btu/scfm), gasoline, and diesel fuel. Unleaded gasoline has been
specified by the EPA for use in its automotive gas turbine development
programs. Nondistillate fuels have not yet been evaluated and Engelhard
declines to make predictions regarding the use of these fuels in catalytic
combustors.
4-52
-------
4.4.4.2.3 Projected Advantages
In Engelhard's opinion, the principal advantage of the catalytic combustor
concept lies in its capability of operating in the fuel-lean regime, which is
not possible in the primary zones of current technology combustors. As a
result, the maximum combustion temperatures will be much lower than the
peak temperatures in conventional combustors; this will minimize the forma-
tion of NO . Furthermore, Engelhard expects high catalytic combustion
ji,
efficiencies resulting in low CO and HC emissions. The organization is
confident of achieving these goals with its design, but feels that a compre-
hensive demonstration program is mandatory before a final assessment of
catalytic combustors can be made.
4. 4. 4. 2. 4 Combustor Design Features
Engelhard considers the catalyst and substrate compositions and the manu-
facturing processes to be proprietary information. The temperature capa-
bilities of Engelhard's catalytic combustor designs have not yet been
evaluated. Current substrate materials are limited to continuous operating
temperatures of about 2400°F, but higher temperatures might eventually be
feasible with advanced substrate designs. However, as the temperature is
further increased, the rate of NO formation in the combustion zone increases
rapidly; this would have to be weighed against the potential advantages result-
ing from operation at the higher combustion temperature.
The Engelhard catalytic combustors are designed for pressure drops of the
order of 1 to 3 percent of the total pressure of the fuel-air mixture entering
the catalyst. These values are comparable to the pressure losses of current
conventional gas turbine combustors.
4. 4. 4. 2. 5 Operational Characteristics
Engelhard is currently working on the development of a number of proprietary
lightoff schemes. In its opinion, different approaches might be used for the
different fuels to account for lightoff temperature differences.
4-53
-------
Preignition and flameout are two potential problem areas impacting
conventional combustor designs. With regard to preignition, Engelhard feels
that catalytic combustors are equivalent to other prevaporizing combustor
designs currently being evaluated by the automotive and aircraft gas turbine
industries. Also, flameout is not considered to be a serious problem area
because the catalyst is self-igniting above the lightoff temperature.
Catalyst durability tests have been conducted on some samples for up to
about 1000 hours. These tests, which were run at steady-state conditions,
showed no performance degradation of the catalyst. Although these results
are certainly encouraging, they are by no means adequate to predict the
performance of the catalytic combustor in actual operation in a gas turbine.
Engelhard has conducted some preliminary cycling tests in which the fuel flow
was intermittently turned on and off. No performance degradation of the cata-
lyst was observed during these tests.
4.4.4.2.6 Emission Characteristics
To date, Engelhard has conducted some exploratory emission test work on
catalytic combustors, using gaseous fuels as well as gasoline. In these tests,
the air-fuel ratio was adjusted to about 38:1 and the pressure and temperature
of the incoming air was varied up to 5 atm and 500°C, respectively. These
values were selected to simulate typical operating conditions of regenerated
automotive gas turbines. The following average pollutant concentrations
were obtained:
undetectable
HC less than 10 ppm
CO less than 60 ppm
According to Engelhard, the emissions are independent of the pressure and
temperature levels of the incoming fuel-air mixture.
Engelhard feels that the instrumentation and test procedures used in its
experimental work are adequate to assure accurate emission data, even at the
low levels typical of catalytic combustors. The instrumentation had been
originally acquired for use in Engelhard 's automotive catalyst development
work and includes chemiluminescence for NO (1 to 10 ppm low scale), NDIR
4-54
-------
for CO and CO2, and heated FID for HC. In addition, it has a scanning
electron microscope and x-ray equipment for substrate analysis.
4.4.4.2.7 Catalytic Combustor Cost
Cost data for catalytic combustors are currently not available. However,
Engelhard states that it must be competitive with other gas turbine com-
bustor configurations. For comparison, Engelhard, indicated that the cost of
a PTX oxidation catalyst delivered to the automobile manufacturer is less
than $50, and only a small fraction of that amount is for the catalyst
material itself. To minimize cost, the platinum group metals could be
reclaimed economically if such metals would be used in catalytic combustors.
According to Engelhard the catalytic combustors could be manufactured in
its existing catalyzing facilities. It would probably not be interested in
the manufacture of complete catalytic combustor units, but would supply
catalyzed substrates to the gas turbine manufacturers.
4.4.4.3 EPA, Control Systems Laboratory
4.4.4.3.1 Program Description
An exploratory test program was conducted by the Control Systems Laboratory
of the EPA using a high temperature catalytic converter provided by
Engelhard Industries (Ref. 4-19). This unit was designed by Engelhard for
operation up to 2400°F. In the EPA tests, the combustor was operated with
premixed propane-air mixtures uniformly distributed over the inlet of the
combustor. The catalyst temperature was varied between 1800 F and
2300 F by varying the air-fuel mixture ratio. The lower temperature limit
was selected to prevent combustion instability.
4.4.4.3.2 Emission Characteristics
The test data from the program indicate NO emissions less than 10 ppm at
jt
all operating conditions. Considering the reduced accuracy of the NDIR
instrument at this low concentration level and the fact that the laboratory air
4-55
-------
contained about 4 to 8 ppm MO , the NO emissions of the catalytic burner
5C A.
are very low indeed. The HC emissions are also extremely low. Because
of instrumentation problems, CO emissions were not reported by EPA.
However, there is evidence that the CO concentrations are also low.
In a final test series, approximately 65 ppm NO was added to the combustor
inlet air. Under these conditions, the average NO concentration at the com-
X.
bustor exit was about 70 ppm, indicating essentially zero NO conversion
jC
efficiency. Considering the lean fuel-air mixture used in these tests and the
associated NO reaction kinetics, this result is not too surprising.
Ji.
4.4.4.4 NASA/Lewis
4.4.4.4.1 Program Description
NASA/Lewis is currently involved in several gas turbine programs in which
catalytic combustors are being considered. These efforts are primarily in
support of the Advanced Automotive Power Systems Program of the EPA
and include development and test work related to a catalytic combustor
designed for use in the Chrysler automotive gas turbine which is currently
being developed with EPA funding. In addition, NASA/Lewis is evaluating
catalytic combustors for a number of other potential applications, including
stationary and aircraft gas turbines (Ref. 4-37).
As part of EPA's gas turbine program, NASA/Lewis will run tests on a
catalytic combustor designed and manufactured by Engelhard Industries. The
composition of the catalyst and the substrate materials used are proprietary
to Engelhard. Initial performance and emission tests are scheduled to com-
mence in May 1973, covering the following ranges of operating conditions:
air-fuel ratio 25 to 100
air inlet temperature 400 to 1000°F
air inlet pressure 2 to 6 atm
catalyst through-flow velocity 20 to 100 ft/sec
4-56
-------
NASA recognizes that a uniform air-fuel mixture and mass distribution is
required at the catalyst inlet to prevent catalyst failure due to overtempera-
ture. If liquid fuels are used, the fuel must be prevaporized. The degree
of vaporization required for satisfactory catalyst operation is currently not
known and will be determined experimentally. To provide a high degree of
vaporization, NASA is currently running tests on a small swirl-type injector
configuration which is designed to provide a well-mixed, prevaporized fuel-
air mixture at the inlet to the catalyst. This type of injector-mixing chamber
is being also considered by NASA for use in advanced jet engine combustors.
Durability testing of catalytic combustors will be conducted by NASA upon
completion of a screening test program currently under way at Engelhard.
In the Engelhard tests, different catalyst-substrate samples will be screened
and the two best designs will then be selected for the NASA durability program.
NASA's current plans are to conduct a very extensive test program on catalytic
combustors in order to acquire all the data needed for a meaningful assess-
ment of this concept.
4.4.4.4.2 Operational Characteristics
NASA states that the catalytic combustor will be preheated electrically or by
means of waste heat from auxiliary power units or other sources. In any
case, no startup problems are foreseen.
With regard to emissions, NASA expects very low CO, HC, and NO con-
centrations in the combustor exhaust.
4.4.4.4.3 Cost Data
At this time, NASA has no cost data for catalytic combustors. Upon completion
of its feasibility test program, it will initiate a thorough economic analysis of
catalytic combustors as well as porous-plate type combustors. NASA's cur-
rent opinion is that catalytic combustors might be somewhat more costly
than equivalent porous-plate designs.
4-57
-------
REFERENCES
4-1. James Wei, "Catalysis and Reactors," Chemical Engineering Progress,
65. (6)(1969).
4-2. P. G. Ashmore, Catalysis and Inhibition of Chemical Reactions,
Butterworth and Company, Ltd. , London (1963).
4-3. I. H. Hoog, "Catalysts, " Chemical Engineering. 157-168
(December 1951).
4-4. P. H. Emmet, Ed. , Catalysis, Vol. 7, Reinhold Publishing
Company, New York (1954).
4-5. G. R. Gillespie, et al. , "Catalytic Purification of Tail Gas, " Chemical
Engineering Progress, £>8 (4) 72-77 (April 1972).
4-6. An Assessment of the Effects of Lead Additives in Gasoline on Emission
Control Systems Which Might be Used to Meet the 1975-76 Motor
Vehicle Emission Standards. TOR-0172(2787)-2, The Aerospace
Corporation, El Segundo, Calif. (15 November 1971).
4-7. Status of Industry Progress Towards Achievement of the 1975 Federal
Emission Standards for Light Duty Vehicles, ATR-73(7322)-l, The
Aerospace Corporation, El Segundo, Calif. (28 July 1972).
4-8. Study of Catalytic Control of Exhaust Emissions for Otto Cycle Engines,
Stanford Research Institute, Menlo Park, Calif. (April 1970).
4-9. Presentation to the Environmental Protection Agency, Ann Arbor,
Michigan, American Lava Corporation, Chatanooga, Tenn. (3 April 1972).
4-10. Assessment of Domestic Automotive Industry Production Lead Time
for 1975-76 Model Years, ATR-73(732 1)-1, The Aerospace Corporation,
El Segundo, Calif. (15 December 1972).
4-11. Torvex Ceramic Honeycomb. Brochure A-56299, 2M9/67,
E. I. DuPont de Nemours and Co. , Inc. , Wilmington, Dela.
(September 1967).
4-12. E. B. Maxte'd, "The Poisoning of Metallic Catalysts, " Advances in
Catalysis, Vol. Ill, Academic Press, Inc., New York, N. Y. (1951).
4-13. K. D. Werner, "Catalytic Oxidation of Industrial Waste Gases, "
Chemical Engineering, 179-184 (4 November 1968).
4-59
-------
4-14. A. Wheeler, "Reaction Rates and Selectivity in Catalyst Pores, "
Advances in Catalysis, Vol. Ill, Academic Press, Inc. , New York,
N. Y. (1951).
4-15. Transcript of Proceedings, Auto Emissions Extension Hearings,
Environmental Protection Agency, Washington, D. C. , Ford Motor
Company, Dearborn, Mich. (19 April 1972).
4-16. Transcript of Proceedings, Auto Emissions Extension. Hearings,
Environmental Protection Agency, Washington, D. C. , General Motors
Corporation, Detroit, Mich. (17 April 1972).
4-17. Request for Suspension of 1975 Federal Emissions Standards, General
Motors Corporation, Detroit, Mich. (3 April 1972).
4-18. Technical discussions with Dr. G. R. Gillespie, Engelhard Industries,
Murray Hill, N. J. , 3 April 1973.
4-19. Technical discussions with D. W. Pershing, Control Systems
Laboratory, Environmental Protection Agency.
4-20. Summary Statement for EPA Hearings on Volvo Application for
One-Year Suspension of Auto Emission Standards, Engelhard Minerals
and Chemicals Corporation, Murray Hill, N. J. (10 April 1972).
4-21. Technical Data Submitted to the EPA Suspension Request Hearing
Panel, Matthey Bishop, Inc., Malvern, Pa. (17 April 1972).
4-22. G. M. Hein, "Odor Control by Catalytic and High Temperature
Oxidation, " Annual N. Y. Academy of Science (116) 656-662
(July 1964) Article 2.
4-23. M. A. Accomazzo and K. Nobe, "Catalytic Combustion of Methane,
Ethane, and Propane with Copper Oxide, " Chemical Engineering
Progress Symposium, Ser. 59, No. 45 (1963).
4-24. M. A. Accomazzo and K. Nobe, "Catalytic Combustion of C^ and C^
Hydrocarbons, " Industrial and Engineering Chemical Process Design
and Development, 4 (4)(October 1 965).
4-25. L. A. Caretto and K. Nobe, "Catalytic Combustion of Cyclohexane,
Cyclohexene, and Benzene, " Industrial and Engineering Chemical
Process Design and Development, 5_(3)(1966).
4-26. M. R. Miller and H. J. Willhoyte, "A Study of Catalyst Support
Systems for Fume Abatement of Hydrocarbon Solvents, " APCA
Journal, 11 (12) (December 1967).
4-60
-------
4-27. M. R. Miller and D. M. Sowards, Solvent Fume Abatement by Ceramic
Honeycomb Catalyst Systems, Franklin Institute, Philadelphia, Pa.
(22 November 1968).
4-28. O. J. Adlhart, et al. , "Processing Nitric Acid Tail Gas, " Chemical
Engineering Progress, 6T_ (2)(February 1971).
4-29. D. J. Newman, "Nitric Acid Plant Pollutants, " Chemical Engineering
Progress. _67 (2)(February 1971).
\
4-30. R. M. Reed and R. L. Marvin, "Nitric Acid Plant Fume Abaters, "
Chemical Engineering Progress, 68 (4)(April 1972).
4-31. Johnson Matthey and Co. , Ltd. , London, "Low Temperature Catalytic
Heaters, " Plantinum Metals Review, j^6 (l)(January 1972).
4-32. E. R. Kweller, "The Catalytic Heater, " Appliance Engineer, ^ (5)
(1972).
4-33. Technical discussion with Dr. D. W. DeWerth, American Gas Asso-
ciation Laboratories, Cleveland, O. , 5 April 1973.
4-34. R. E. Thompson, D. W. Pershing, and E. E. Berkau, Catalytic
Combustion - A Pollution-free Means of Energy Conversion.
Environmental Protection Agency, Combustion Section, Cincinnati, O-
(August 1971).
4-35. W. S. Blazowski and R. E. Henderson, Assessment of Pollutant
Measurement and Control Technology and Development of Pollutant
Reduction Goals for Military Aircraft Engines. AFAPL-TR-72-102
(November 1972).
4-36. Technical discussion with Capt. W. S. Blazowski, Air Force Aero
Propulsion Laboratory, Dayton, O. (4 April 1973).
4-37. Technical discussion with Mr. Jack Heller and Mr. C. Mroz. , NASA/
Lewis Research Center, Cleveland, O. (6 April 1973).
4-61
-------
SECTION 5
EVALUATION OF POROUS-PLATE AND CATALYTIC
SURFACE COMBUSTORS FOR LARGE UTILITY BOILERS
5. 1 POROUS-PLATE SURFACE COMBUSTORS
5.1.1 Introduction
The basic features and operating principles of porous-plate surface com-
bustion for use in Rankine or Brayton cycle power plants can be most easily
described in terms of a schematic of a typical porous-plate burner as shown
in Figure 3-7. Premixed fuel and air passes through the porous matrix at
low velocity and burns in a flat, thin (approximately 1 mm thick) laminar
flame positioned close to the porous burner surface. Because of the close
proximity of the flame to the surface, an appreciable amount of heat is
transferred from the burner gases to the porous surface by convection and
radiation. The heat is then removed from the porous matrix, either by
cooling tubes imbedded in the plate and/or radiation from the burner
surface. The removal of this heat reduces the temperature of the burned
gases below the adiabatic flame temperature. Since the magnitude of the
NO emissions is highly dependent on the maximum flame temperature,
Jt
this approach offers the possibility of achieving low NO emissions. This
3C
is shown in Figure 5-1. As indicated, a temperature reduction from 2200 K
to 2000°K reduces the NO formation rate by a factor of about 20. Peters
Ji.
(Ref. 5-1) has demonstrated in small-scale porous-plate combustors that
the NO emissions can be reduced in this manner. However, to date, no
x
one has seriously examined the application of porous-plate surface com-
bustors to large utility boilers. While the porous plate is a convenient
way of removing heat from the combustion zone, it is possible that better
approaches might be devised.
5-1
-------
3.0
1.0
0. 1
o
O
O
(M
rvi
TEMPERATURE, °F
4000 3800 3600 3400 3200 3000
2800
r i I
r2200°K = 11 ppm/msec
(for NQ =0.1)
0.01 _ NITRIC OXIDE FORMATION
RATC BY HOT AIR
MECHANISM
d[NO]
r= dT ; NORMALIZED TO 2200 K
AT 1 atm TOTAL PRESSURE AND fc
=0.1
-I
EXPERIMENTAL DATA FROM
VARIOUS FLAMES
0.001
4.5
5.5
RECIPROCAL ABSOLUTE TEMPERATURE (1/T) X 10*. (1/°K)
Figure 5-1. NO Formation Rate
5-2
-------
The objective of this section is to (1) briefly evaluate the potential of liquid-
cooled, porous-plate surface combustors in reducing the NO emissions
from large utility boilers, and (2) determine the most critical research
work needed to permit its early application. The evaluation will be carried
out by setting up low NO emission design goals for a typical large utility
3C
boiler and determining if a porous-plate combustor can be reasonably ex-
pected to satisfy the postulated requirements .
5.1.2 NO Emission Level Target
There is no apparent reason, other than achieving lower NO emission
levels, to consider the use of porous-plate surface combustion in large
utility boilers. Hence, the target NO emission should be set appreciably
X
lower than is currently obtainable with the best existing or projected abate-
ment approaches. For comparison, the NO emissions projected for the
Scattergood No. 3 unit of the Los Angeles Department of Water and Power
are shown in Figure 5-2 (Ref. 5-2). This unit has lower projected NO
2C
emissions than any other large utility boiler in the United States. It has a
capacity of 460 MW, but will be restricted in output to 315 MW by Rule 67
of the Los Angeles Air Pollution Control District. This rule limits the NO
X
emissions of an individual point source to 140 Ib/hr in order to maintain
acceptable local NO concentrations. At 315 MW, the corresponding NO
x x
emissions are 42 ppm. To ensure NO emissions appreciably lower than
those provided by the existing technology, the NO emission target goal for
-X
this study has been set in the 10 to 20 ppm range.
5.1.3 Boiler Design Requirements
5.1.3.1 Porous-Plate Surface Combustor Operating Conditions
In order to achieve low HC and CO emissions while maintaining high boiler
efficiency, the surface combustor must operate at fuel-air equivalence
ratios of slightly less than unity. Currently, no NO emission data are
available for this equivalence ratio at sufficiently low gas temperatures to
achieve the desired target emissions. However, it is apparent from Fig-
ure 5-3 (Ref. 5-2), which presents the NO emissions for an equivalence
5-3
-------
110
100
90
BO
0)
.2 70
o
£ 60
E
Q.
°- 50
O
z
40
NITRIC OXIDE LIMIT
FOR COMPLIANCE
WITH LA/APCD RULE 67
30
20
10 —
UPPER 4 BURNERS OUT OF
SERVICE OR OVERFIRE AIR
PORTS OPEN AND UPPER 2
BURNERS OUT
(33% off-stoichiometric)
GAS
RECIRCULATION
ONLY
42ppm AT 315 MW
COMBINED 33% OFF-STOICHIOMETRIC
COMBUSTION AND GAS RECIRCULATION
100
200
300
LOAD, MW
400
500
Figure 5-2. Predicted NO Emissions vs Mode of Operation
for Scattergood No. 3 Unit (Ref. 5-2)
5-4
-------
6000
1000
100
10
_ PRIMARY ZONE - ADIABATIC
NO FORMATION
- 3800° F FOR 0.05 sec -1500ppm
NO FORMED IN 1.0 sec
RECIRCULATION ZONE
NO FORMATION
3000° F FOR
2.0 sec 50 ppm
2800
3000
3200 3400
TEMPERATURE, °F
3600
PEAK
ADIABATIC
TEMPERATURE
(650°F air
preheat)
3800
Figure 5-3. Kinetic NO Formation for Combustion of Natural Gas at
Stoichiometric Mixture Ratio--Atmospheric Pressure
5-5
-------
j iB 0 _lll_ IT1 I I Jfl IP ,fU£ Ji .Jl
PRIMARY
SUPERHEATER
-AT TEMPERA
IMDARY I REHEAT
PERHEATERJ SUPERHEATER
GAS-
TEMPERING
PORTS
POROUS PLATE
COMBUST
r-BURNERSl -
GAS-
-RECIRCULATING
FAN
Figure 5-4. Haynes Supercritical Steam Boiler--
Units 5 and 6
5-6
-------
ratio of 1.0, that NO concentrations of 10 ppm can be achieved for gas
3C
temperatures of 3000°F and residence times of a few tenths of a second.
The general trend of the NO emissions with increasing excess air is
shown in Figure 3-2 (Ref. 5-3). As indicated, at a few percent excess air
the NO emissions will be somewhat higher than 10 ppm, but still apprec-
3C
iably below the lowest projected value of 42 ppm. Therefore, a gas
temperature of 3000 F and a residence time of a few tenths of a second will
be taken as the design condition for the porous-plate surface combustor
considered in this study.
5. 1.3 . 2 Fuel
Because of the small pore size of porous-plate surface combustors, (e.g.,
the burner design in Ref. 5-4 has a pore size of 250 microns), only gaseous
fuels or distillate fuels that can be readily prevaporized are considered as
candidate fuels for porous-plate surface combustors. The evaluation has
been made using natural gas, which is by far the most popular and available
gaseous fuel for large utility boilers. Distillate fuels have not been econo-
mically competitive with other fuels for this application. Methane, which
is the main constituent of natural gas, has been selected as the fuel for
this analysis.
5.1.3.3 Utility Boiler Type and Size
Because the present shortage of natural gas may severely limit new gas-
fired utility boiler plant construction, the application selected for this
evaluation is the retrofitting of an existing boiler. More specifically,
retrofitting of units 5 and 6 of the Haynes Steam Plant of the Los Angeles
Department of Water and Power has been selected for the evaluation
example. The important boiler characteristics for these units are shown
in Table 5-1 and the boiler is illustrated in Figure 5-4.
5-7
-------
Table 5-1. Characteristics of Existing Boiler for Units 5 and 6 of
Haynes Steam Plant (Los Angeles Department of Water
and Power)
Maximum gas flow rate
Number of burners
Dimensions of boiler (approx..)
Burner arrangement
Air preheat temperature
Temperature at entrance to
secondary superheater
2,760,000, scfm
24
25 X 25 X 120 ft
4 vertical rows over 26 ft elevation, 12
burners on front and 12 on rear wall
650°F
2500°F
Porous-plate surface combustors require a large surface area, because the
laminar flame speeds are very low, "which leads to large burner volume
requirements. In general, volume constraints are more critical in the
retrofitting of existing boilers than in new boiler designs. Hence, if an
existing boiler can be retrofitted, there should be no volume constraint
problem with the new design.
5.1.4
5.1.4.1
Burner Characteristics
Burner Surface Area
The required burner surface area may be determined from the following
equation
- Q ( 1 + a/f )
(1)
5-8
-------
where
A = required burner surface area
Q = rated heat input to the boiler
P = density of the air-fuel mixture on the cold side of the burner
u = flame speed on the cold side of the burner
a/f = air-fuel ratio of the burner
h - fuel heating value
The mass velocity pu may be determined from the continuity equation
PU=P25U25 (2)
where subscript 25 denotes 25 C. Kaskan (Ref. 5-5) has determined u--
for a large number of gaseous fuels; his results for methane are shown in
Figure 5-5. As indicated in this figure and from Eq. (1), the required
burner surface area increases rapidly with decreasing gas temperature.
For Haynes units 5 and 6, the required burner surface area is approxi-
fj
mately 21, 000 ft for a maximum gas temperature of 3000 F and an
equivalence ratio of unity. This area is somewhat larger than the water
wall area of the Haynes boiler. Reducing the gas temperature to 2500 F
would increase the required burner surface area approximately fourfold.
5. 1.4.2 Burner Configuration
Many different porous-plate burner configurations might be feasible. How-
ever, the configuration chosen should satisfy the following requirements:
1. To minimize the required burner surface area, the gas tempera-
ture immediately downstream of the porous-plate surface should be
as high as possible consistent with the desired NO emission level.
j£.
2. To avoid a major boiler rebuild, the porous-plate burner must
be designed to fit the existing boiler envelope, and the gas tempera-
ture entering the secondary superheater should be about 2500 F,
matching the temperature of existing boilers.
5-9
-------
100
90
80
70
60
50
40
30
o
20
810
111
9 -
m
> 8U
i 6
5 -
4 -
3 -
2 -
EQUIVALENCE RATIO: 1.0 0.80.6
^07J. "K
Figure 5-5. Burning Velocities of Methane-Air Mixtures
vs Reciprocial Temperature (Ref. 5-5)
5-10
-------
3. To minimize plant modification cost, maximum use should be
made of the existing water wall to reduce the burner gas temperature
to the level required at the entrance to the secondary superheater.
4. To achieve the desired residence time of a few tenths of a sec-
ond, the burned gases must be quickly accelerated after leaving the
burner surface and directed to a location where rapid cooling takes
place by means of forced convection or radiation.
A configuration which attempts to satisfy these requirements is shown in
Figures 5-4 and 5-6. In this design, the burner covers the full width of the
boiler which, in the case of the Haynes boiler, is about 25 ft. The gas flow
arrangement is adapted from that shown in Figure 3-10 (Ref. 5-4).
The principal uncertainty regarding the design shown in Figure 5-6 is the
fact that a high burned gas velocity of approximately 100 ft/sec is required
parallel to the porous-plate surface. To date, the effects of a high cross -
velocity on burner performance have not been thoroughly evaluated. Gen-
eral Electric (Ref. 5-6) has successfully operated with cross-velocities up
to 200 ft/sec using a wire screen between the burner surface and the high
velocity cross-flow. However, this approach is considered impractical at
the gas temperatures used in this example, unless the screen were water-
cooled; this would cause a very sizable increase in complexity and volume
of the burner. Bone (Ref. 5-7) does not explicitly state the cross-velocity
in his porous-plate combustor boiler tests, but it can be inferred from the
stated burner dimensions and the laminar flame speed that the cross-
velocity was of the order of 80 ft/sec. Hence, it is believed that 100 ft/sec,
while not demonstrated, is probably achievable.
5.1.4.3 Cooling Requirements
The heat Q that must be removed from the porous-plate burner in order to
c
achieve the desired gas temperature may be calculated from the following
equation
5-11
-------
-4 ft
BURNED GASES
UNBURNED
AIR GAS MIXTURE
IMBEDDED COOLING
TUBES THROUGHOUT
Figure 5-6. Porous Plate Burner Configuration Schematic
H is the enthalpy of the burned gases for adiabatic combustion, and H is
3.
the enthalpy of the burned gases at the actual gas temperature. The cooling
load Q /A, computed for an equivalence ratio of 1. 0 and a pressure of 1 atm,
is depicted in Figure 5-7 as a function of gas temperature. As indicated,
for a gas temperature of 3000°F the cooling load is 45, 000 Btu/hr-ft ,
which is approximately one-third of the total burner heat release rate.
5.1.4.4
Materials Requirements
Different materials have been applied by various investigators for water-
cooled porous-plate surface combustors. For example, sintered copper
was used by Kaskan (Ref. 5-5) and oilite and sintered stainless steel by
Peters (Ref. 5-1). No serious material problems were encountered by
these investigators. However, in all these tests the burner was operated
without air preheat and the average cooling water temperature was some-
what lower than would be experienced in a utility boiler application. The
5-12
-------
8.0
6.0
3
m
o
X
4.0
EQUIVALENCE RATIO 1.0
PREHEAT TEMP 650 F
2400 2600 2800 3000 3200 3400
GAS TEMPERATURE, °'F
3600
3800
4000
Figure 5-1. Computed Heat Flux to the Porous Plate Burner
as a Function of Burning Velocity
-------
effect of air preheat and higher water temperatures is to increase the
material temperatures by several hundred degrees over present experience.
However, based on the test data by Peters (Ref. 5-1), no serious materials
problems are anticipated. For the oilite plate, Peters' data indicate
material temperatures below 500°F in the center of the porous plate and the
surface temperature is expected to be only slightly higher. Even if the
surface temperature were increased by several hundred degrees, the oilite
material would still perform satisfactorily. However, more detailed calcu-
lations are required before an accurate evaluation of the temperature condi-
tions of the plate can be made.
5.1.4.5 Steam Rate Control
In order to maintain maximum boiler efficiency at reduced power plant load,
it is desirable always to operate the boiler at an equivalence ratio near
unity and with preheat. Since the burner flow velocity u depends only on
equivalence ratio and burner inlet air temperature, the boiler output can
only be lowered by reducing the burner surface area. This can be accom-
plished by shutting down sections of the burner.
5.1.4.6 Other Considerations
Because of the extremely small pore size of the porous-plate burner (of the
order of 250 micron) and the long maintenance-free service life require-
ment, pore-clogging due to particulate matter contained in the combustion
air present a potential problem area for utility boiler applications.
The Los Angeles Department of Water and Power operates its utility boilers
on natural gas in the summer when photochemical smog reaches its peak,
and on low-sulfur fuel oil in the winter to reduce costs and reduce the i
demand for natural gas at a time when it is needed for home heating. The
porous-plate burner previously discussed is not capable of running on fuel
oil, although it may be possible to modify it to permit operation on either
fuel (e.g., by having separate burners for natural gas and fuel oil or by
using folding porous-plate burners).
5-14
-------
5- 1. 5 Conclusions
1. No information is currently available regarding the application
of surface combustors to large utility boilers.
2. Theory and limited small-scale burner experiments indicate
that porous-plate surface combustors have the potential of
achieving NO emission levels appreciably lower than those
obtained in tlie best current state-of-the-art large utility
boilers.
3. Porous-plate surface combustors might be feasible for existing
boilers, as well as new boiler installations.
4. Porous-plate surface combustors are limited to gaseous or
distillate fuels. Current shortages of these fuels may limit
interest in porous-plate surface combustors.
5. Before a complete assessment of porous plate surface combus-
tors can be made, a substantial amount of research and develop-
ment is required in the following areas:
a. Demonstration of satisfactory operation of the burner
with high exhaust velocities parallel to the porous plate
surface.
b. Demonstration of burner operation for long periods of
time in an environment typical of large utility boilers,
without clogging the pores.
5.2 CATALYTIC COMBUSTORS
For efficiency reasons, utility boilers are limited to near-stoichiometric
operation combined with substantial air preheat. Under these conditions,
the combustion temperatures in existing natural gas-fired boilers are of the
order of 3800°F with NO emissions varying between about 200 and 700 ppm,
j£
depending upon the particular design of the boiler. Theoretically, the NO
J£
emissions from these plants could be almost completely eliminated through
the use of catalytic combustors,which would be designed to operate at a
temperature (1) low enough to limit the formation of NO and to assure
adequate service life, and (2) high enough to provide adequate CO and HC
control.
5-15
-------
Currently, commercially available catalyst substrates are limited to a
maximum safe operating temperature of about 2400 F. Although the temp-
erature capability and formulations of the active catalyst materials are
considered proprietary by the catalyst manufacturers, there are indications
that some catalysts might be able to maintain adequate effectiveness at
these temperatures over extended periods of time. It is conceivable that
other monolithic or pellet type catalysts could be developed for use at even
higher temperature levels.
During the course of this study no concrete information was uncovered regarding
the applicability of catalytic combustors to steam boilers, although Engelhard
has briefly looked at its feasibility for a potassium boiler. The feasibility of
catalytic combustors for steam boilers has been examined as part of this
study; two conceptual designs, uncooled and internally cooled, were reviewed.
In the uncooled catalytic combustor design, a conventional monolithic or
pellet type catalyst would be operated in the lean regime to limit the com-
bustion temperature to the maximum allowable level, say 2400 F. This
could be accomplished by using all the air and only part of the fuel in an
initial catalyst stage. Upon leaving the catalytic combustor, a-portion of
the combustion heat would be transferred from the combustion gases to the
water wall before these gases enter the mixing chamber of a second
catalytic combustor stage; at this point, additional fuel would be added to
raise the temperature of the combustion products to the first-stage level.
The process is then repeated until all fuel is burned.
Although basically feasible, this particular approach is complex and has
severe limitations as far as incorporation into existing power plants is
concerned. In current boiler designs, a substantial part of the heat trans-
fer into the water wall is accomplished by means of radiation. Since the gas
temperature in catalytic combustors is limited to about 2400 F-f the radia-
tive heat transfer is rather low and thus convective heat transfer would be
required to heat the water wall. As a result, the water wall would have to
be completely redesigned, and the rise of pressure across the blower would
5-16
-------
have to be increased to compensate for the higher pressure losses
associated with convective heat transfer.
In the internally cooled approach, cooling tubes would be imbedded in the
catalyst in order to transfer a portion of the heat of combustion into the
coolant by means of conduction through the substrate. As a result, the air-
fuel mixture supplied to the catalytic converter might be enriched some-
what, compared with the uncooled design, with a consequent reduction of
the required catalyst volume. However, this design approach creates
thermal stress problems that might be beyond the capability of ceramics; a
cracking of the unit could result. Utilization of sintered metals might offer
an acceptable alternative as a substrate material for internally cooled
catalytic combustors, although currently these materials are limited to
about 2000 F. In any case, incorporation of this concept would require a
major boiler redesign effort.
Like the previously discussed porous-plate combustors, catalytic combus-
tors require gaseous or prevaporized, premixed distillate fuels to prevent
fuel coking or overtemperature conditions which might lead to catalyst
deactivation or melting of the substrate material. A discussion of catalyst
requirements, potential problem areas, and uncertainties is presented in
Section 6.
Based on these considerations, it is unlikely that catalytic combustors
could be incorporated into existing power plants, but the concept might well
be feasible for new power plant designs. However, comprehensive pre-
liminary design study efforts are required before a meaningful assessment
can be made.
5-17
-------
REFERENCES
5-1. B. D. Peters, "Nitric Oxide Reduction by Heat Transfer in a Porous
Disk Burner, " Ph. D. Thesis, University of Wisconsin (1972).
5-2. L. K. Jain, E. L. Calvin, R. L. Looper, State of the Art for Con-
trolling NO Emissions. Part 1: Utility Boilers, Catalytic, Inc.
Charlotte, ft. C. (EPA Contract 68-02-0241, Task 2) (September 1972).
5-3. W. H. Barr and D. E. James, "Nitric Oxide Control--A Program
of Significant Accomplishments, " ASME Winter Annual Meeting,
New York, N. Y. , 26-30 November 1972, Paper 72-WA/Pwr-l 3.
5-4. R. J. Rossbach, Development of Low-Emission Porous-Plate Com-
bustor for Automotive Gas Turbine and Rankine Cycle Engines,
Quarterly Report GESP 738 (EPA Contract 68-01-0461) (May 1973).
5-5. W. E. Kaskan, "The Dependence of Flame Temperature on Mass
Burning Velocity, " Sixth Symposium (International) on Combustion
publication (1956) p 134.
5-6. Personal communication with General Electric Company,
Cincinnati, O.
5-7 W. A. Bone, "Surface Combustion, with Special Reference to Recent
Developments in Radiophragm Heating, " Gas Journal, 423-428
(16 May 1923).
5-19
-------
SECTION 6
EVALUATION OF CATALYTIC AND NONCATALYTIC SURFACE
COMBUSTION CONCEPTS FOR STATIONARY GAS TURBINES
6. 1 GENERAL CONSIDERATIONS
For the past several years, gas turbines have been used extensively by the
electric power utilities to supplement the power output of steam plants
during periods of high demand. Other gas turbine applications include com-
pressor drives for the petroleum and natural gas industries, total energy
systems for institutional facilities and shopping centers, and continuous
power generation for small utilities. In the main, these units are powered by
simple-cycle (nonregenerative), dual-shaft gas turbines designed for pressure
ratios between about 12 and 18 and maximum turbine inlet temperatures
between 1400 F and 1800 F. In general, these engines consist of a high-
pressure turbine driving the compressor and a low-pressure turbine providing
power to the electric generator. The thermal efficiency of current simple -
cycle gas turbines is of the order of 25 percent, •which is rather low compared
•with the efficiencies of about 40 percent achieved in steam power plants.
Because of the low thermal efficiency capability of the current state-of-the-art
gas turbines, the utilities have shown little interest in these engines for large
base power generating installations, except when used in combined gas turbine/
steam turbine installations. In these systems, the waste heat in the gas turbine
exhaust is recovered to generate steam for use by the steam turbine.
The thermal efficiency and specific fuel consumption of gas turbine engines can
be substantially improved by increasing the turbine inlet temperature and/or by
incorporating a regenerator. In this regenerator, a portion of the thermal
energy of the turbine exhaust gases is extracted and transferred to the "cold"
side of the unit in order to raise the temperature of the compressor discharge
air. As a result, less fuel is required in the combustor to reach the desired
turbine inlet temperature.
6-1
-------
Higher turbine inlet temperatures require the application of improved
turbine blade and rotor materials such as high nickel content alloys or even
ceramics, combined with advanced turbine blade cooling techniques. Cycle
temperatures as high as 3000 F are projected by several organizations for
the 1980 time period (Refs. 6-1 and 6-2).
6.2 GAS TURBINE OPERATING CHARACTERISTICS
The predicted performance of current and advanced state-of-the-art
simple-cycle and regenerative-cycle gas turbines is presented in Figures
6-1 and 6-2 over a range of engine operating conditions. (Ref. 6-1).
In this context, the regenerator effectiveness is defined as the ratio of the
heat transferred from the "hot" turbine exhaust side to the "cold" compressor
discharge side divided by the amount of heat that could be theoretically trans-
ferred. As illustrated in Figures 6-1 and 6-2, the gas turbine thermal effi-
ciency, which relates the work output of the engine to the total energy input,
increases with increasing turbine inlet temperature.
In simple-cycle gas turbines, the optimum pressure ratio increases substan-
tially with increasing turbine inlet temperature. For instance, for a turbine
inlet temperature of 1800 F, the optimum pressure ratio is about 18:1 and
the cycle efficiency is approximately 30 percent. At 3000 F, the optimum
simple-cycle pressure ratio increases to over 30:1. The corresponding cycle
efficiency is of the order of 40 percent, which is comparable to the efficiency
of current steam power plants.
Substantially higher thermal efficiencies can be achieved at much lower cycle
pressure ratios by means of regeneration. According to Figure 6-2, the
thermal efficiency of current steam power plants can be matched with regen-
erative gas turbines operating at a turbine inlet temperature of about 2000 F
and a regenerator effectiveness of 90 percent. Under these conditions, the
optimum pressure ratio is only about 4:1.
6-2
-------
u>
1 1970.DECADE TECHNOLOGY
1 EA81Y 1980'S TfCHNOlOGY
3 I ATE I980'S TECHNOLOGY
I J PRESENT-DAY GAS TURBINES
COMPRESSOR PRESSURE RATIO
TURBINE INLET GAS TEMPERTURE-F
50r
U
z
< t-
m
a:
o
"!"
2«J_
COMPRESSOR BLEED
AIR UNCOOLED
0 50 100 ISO 200 250 300 350 400 450
SHAFT HORSEPOWER PER UNIT AIRFLOW - SHP/LB/SEC
50
z
U
£ 40
I 30
< 25
FUfL-METHANE (HHV = 1000tTU/
AMBIENT ftOF AND IOOOFT
1970 DECADE TECHNOLOGY
TUIBINE INLET OAS 1EMF>E«ATUtB=2iOOF
3400F A
AIISIDE
EFFECTIVENESS • %
COMPHSSOI MESSUIE IA1IO
EAtlY JQIO'S TECHNOLOGY
120 140 160 180 200 220 240 260 260 300 320 340
SHAFT HORSEPOWER PER UNIT AIRFIOW-SHP/IB/SEC
Figure 6-1. Performance Characteristics of
Simple-Cycle Gas Turbines
(Ref. 6-1)
Figure 6-2. Performance Characteristics of
Regenerative-Cycle Gas Turbines
(Ref. 6-1)
-------
1200
1000
LJ
.
O
u)
u
o:
a
O
u
800
LJ
0.
LU
•" 600
j-
x
LU
400
200
NOTE:
COMPRESSOR EFFICIENCY 0.86
COMPRESSOR INLET TEMPERATURE 60'
4 5 6 7 8 910
COMPRESSOR PRESSURE RATIO
20
30
40
Figure 6-3.
Simple-Cycle Gas Turbine Combustor Inlet Temperature vs
Compressor Pressure Ratio
-------
To permit a more meaningful feasibility evaluation of catalytic and
porous-plate combustors in gas turbine applications, the combustor inlet air
temperatures and air-fuel equivalence ratios were computed over wide ranges
of engine operating conditions. It is well recognized that the ignition charac-
teristics of fuel-air mixtures are strongly impacted by these two factors. A
more detailed discussion of the ignition and preignition phenomena of pre-
vaporized, premixed fuel-air mixtures is presented in Section 6. 5.
Predicted combustor inlet temperatures for simple-cycle gas turbines are pre-
sented in Figure 6-3 as a function of compressor pressure ratio. The data are
based on a compressor efficiency of 86 percent, a turbine efficiency of
88 percent, and an ambient temperature of 60 F. For the region of interest,
the combustor inlet temperature of simple-cycle gas turbines varies between
about 750°F and 1000°F.
The combustor inlet temperatures of regenerative-cycle gas turbines are
depicted in Figures 6-4 and 6-5 for regenerator effectiveness values of 90 per-
cent and 70 percent, respectively. Because of the high degree of regeneration
and the low cycle pressure ratio, the temperatures are substantially higher
than those of simple-cycle gas turbines. The curves of Figures 6-4 and 6-5
are based on the same turbomachinery efficiencies used in the simple-cycle
calculations and a combined regenerator and combustor pressure loss of 4
percent of the compressor discharge pressure.
Predicted air-fuel ratios for natural gas are presented in Figure 6-6 for both
simple-cycle and regenerative-cycle gas turbines. As expected, the required
air-fuel ratio decreases with increasing turbine inlet temperature and
decreasing degree of regeneration.
6.3 GAS TURBINE EMISSIONS
NO and CO mass emission test data from a number of aircraft
.A.
and industrial gas turbine engines are presented in Table 6-1 in terms of
pollutant concentration and pounds of pollutant per megawatt output. Also listed
in this table are emission indexes computed from initial test data published by
6-5
-------
2000
1800
u.
o
a:
a:
LJ
1400
LJ
_i
5 1200
Q£
O
3
CD
O 100°
U
800
12.0
NOTE: Compressor Inlet Temperature 60° F
Compressor Efficiency 0.86
Turbine Efficiency 0.88
Combustor Pressure Loss Factor 0.04
rc = Compressor Pressure Ratio
600. , , , , , ,.
2000 2200 2400 2600 2800 3000 4000
TURBINE INLET TEMPERATURE, °F
Figure 6-4. Regenerative-Cycle Gas Turbine Combustor
Inlet Temperature vs Turbine Inlet Temperature-
Regenerator Effectiveness 0.90
6-6
-------
2000
1800
r 1600
LJ
{£
Id
| 1400
Id
I-
Ul
2 1200
on
e
m
1000
800
600
2000
NOTE: Compressor Inlet Temperature 60° F
Compressor Efficiency 0.86
Turbine Efficiency 0.88
Combustor Pressure Loss Factor 0.04
r_ = Compressor Pressure Ratio
2200
2400
2600
2800
3000
TURBINE INLET TEMPERATURE, °F
4000
Figure 6-5. Regenerative-Cycle Gas Turbine Combustor Inlet
Temperature vs Turbine Inlet Temperature-
Regenerator Effectiveness 0. 70
6-7
-------
100
80
60
o:
_i
LU
40
20
REGENERATOR EFFECTIVENESS 0.90
REGENERATOR EFFECTIVENESS 0.70
NONREGENERATED
NOTE: Compressor Efficiency 0.86
Turbine Efficiency 0.88
Combustor/Regenerotor Pressure Drop Factor 0.04
Lower Heating Value 20,500 Btu/lb
i
2000 2200 2400 2600 2800
TURBINE INLET TEMPERATURE, ° F
3000
Figure 6-6. Predicted Air-Fuel Ratio vs Turbine Inlet
Temperature (Natural Gas)
6-8
-------
Table 6-1. NO and CO Emissions for Gas Turbines and Steam Power Plants
vD
Case
1
2
3
4
5
6
7
Unit
Curtis s- Wright gas turbine
Curtis s- Wright gas turbine
Small industrial gas turbines
Typical aircraft gas turbines
Ford EVC combustor concept
Uncontrolled steam plant
Projected steam plant*
Emissions
NOX
ppm
-
-
-
-
-
400
42
Ib/MW-hr
6.7
3.0
8.8
12.3
0.17
4.2
0.44
CO
ppm
-
-
-
-
-
<50
< 100
Ib/MW-hr
-
;
2.9
0.4
1.0
0.32
0.64
Comments
Without water injection
With water injection
JT8D engines; full power
Conceptual design
Off-stoichiometric burning
and flue gas recirculation
(1)(Ref. 6-3)
Externally vaporizing combustor (Refs. 6-4 and 6-5)
' Los Angeles Scattergood No. 3 (Ref. 6-6)
-------
Ford Motor Company for their new externally vaporizing combustor (EVC)
concept currently under development for potential use in automotive gas
turbines. For comparison, the NO and CO emissions of uncontrolled and
3C
NO -controlled steam power plants are also included in Table 6-1 (cases 6
3C
and 7).
The emissions from a Curtiss-Wright peaking power unit, powered by a
Pratt &: Whitney GG4C1-D gas turbine, are listed as cases 1 and 2. Without
water injection, the NO emissions from this unit reach the permissible limit
jC
of 140 Ib/hr, specified by the Los Angeles County Air Pollution Control District
Rule 67, at a derated power setting of about 21 MW. Incorporation of water
injection results in a 55 percent reduction of the specific NO mass emissions.
j£
Carbon monoxide data are not available for this engine.
Case 3 represents a number of small industrial gas turbines generating a net
power output of the order of 500 kW (Ref 6-7). The NO emissions from these
X
engines are comparable to those of large industrial units exemplified in case 1
by the Curtis s-Wright engine without water injection.
The full-power NO and CO emissions of typical state-of-the-art aircraft gas
turbines are shown in Table 6-1, case 4. As indicated, NO is somewhat
' x
higher than for the industrial units, reflecting the higher operating temperatures
of the aircraft engines. Conversely, CO is quite low and comparable to steam
power plants. The AFAPL projects NO emission reductions up to 75 percent,
3£
to be accomplished by means of combustor modifications and -water injection
(Ref. 6-8).
Case 5 shows the NO and CO mass emissions computed from Ford's published
jfi
EVC concept test data (Refs. 6-4 and 6-5). In this combustor design, shown
schematically in Figure 6-7, the fuel enters the prevaporization-mixing cham-
ber through a coaxial nozzle and is then atomized and vaporized by the high
velocity primary air flow. Ideally, complete vaporization of the fuel droplets
is accomplished in the chamber before the fuel-air mixture enters the primary
6-10
-------
PRIMARY DILUTION
COMBUSTION ZONE
ZONE
FUEL
PRIMARY
AIR
DILUTION AIR
Figure 6-7. Ford Experimental Externally Vaporizing
Combustor (EVC) (Ref. 6-5)
6-11
-------
combustion zone through a set of adjustable inlet ports. The primary zone
fuel-air equivalence ratio is controlled between 0. 35 and 0. 5 at all times to
limit the combustion temperature to about 2800 F to 3000 F. At these tem-
perature levels, only minute amounts of NO are formed. Secondary air is
3t
then added in the dilution zone of the combustor to obtain the desired turbine
inlet temperature. The corresponding CO mass emissions reported by Ford
for this combustor are approximately three times higher than for NO . How-
3x
ever, it is the opinion of Ford that CO could be reduced substantially by
increasing the length (volume) of the combustion chamber to permit com-
pletion of the CO oxidation reactions. Based on Ford's initial test,s, it is
concluded that the EVC concept has the potential of achieving very low NO
Ji
emissions. However, a number of potential problem areas must be evaluated
before this combustor is ready for use in gas turbines. Potential problem
areas include combustor instability, preignition, startup, and durability.
Catalytic combustors and porous-plate surface combustors also offer low
emissions potential. In these configurations, the air-fuel ratio of the mixture
entering the combustor is adjusted to limit the combustion temperature to
sufficiently low levels where the NO formation rate is low. Exploratory tests
conducted by Engelhard and the EPA indicate that very low emissions can
indeed be achieved with catalytic combustors. In the Engelhard program, the
catalyst was operated at air-fuel ratios of about 40:1, system pressures up to
5 atm and air preheat temperatures up to 500 C. Under these conditions, the
NO was undetectable, HC was less than 10 ppm, and CO was less than 60 ppm.
X
The data from the EPA program are in reasonable agreement with Engelhard1 s
results.
Porous-plate combustor tests conducted by General Electric at 1 atm indicate
NO emissions substantially below the 1976 Federal emission standards for
!X
light-duty vehicles. However, the corresponding CO standard was exceeded
in some of the tests.
6-12
-------
6.4 NO EMISSION GOALS FOR LARGE POWER INSTALLATIONS
The applicability of gas turbines to large base power generating systems is
governed by two major factors, i.e., cycle efficiency and exhaust emissions.
With regard to cycle efficiency, the specific fuel consumption of the gas turbine
must be competitive with existing steam power plants. As discussed in Sec-
tion 6-1, this requires substantial advances in the state-of-the-art technology,
including incorporation of high effectiveness regenerators or recuperators and
operation at much higher turbine inlet temperatures.
The emission data listed in Table 6-1 indicate that the NO mass emissions of
current state-of-the-art gas turbines are an order of magnitude higher than
the emissions projected by the Los Angeles Department of Water and Power for
its new Scattergood No. 3 steam plant. This particular plant is designed to
incorporate flue gas recirculation and off-stoichiometric combustion, and is
scheduled to go into service in 1974 at a derated power output level of 315
megawatts (460 MW rated power).
The permissible NO concentrations in the exhaust of a hypothetical 315 MW
!X
gas turbine power plant meeting the Los Angeles County Air Pollution Rule 67
(140 Ib NO-j/hr) are presented in Figure 6-8 for simple-cycle and regenerative-
br
cycle gas turbines, respectively. These values were computed on the basis of
the gas turbine operating parameters depicted in Figures 6-1 through 6-6. As
illustrated, the allowable NO concentrations are extremely low by current
standards, especially for highly regenerated gas turbines. The allowable NO
concentration increases somewhat with increasing turbine inlet temperature as
a result of lower engine flow rates per unit power output.
These examples illustrate that the NOx emissions from current gas turbines
require an order of magnitude reduction before the engines become feasible for
use in large base power generating units. Since conventional droplet burning
combustor concepts are severely limited with respect to NOx emission
6-13
-------
24 r—
20
<
OL
UJ
O
O
UJ
_l
m
OL
UJ
a
16
12
REGENERATOR EFFECTIVENESS 0.9
NONREGENERATED
2000 2200 2400 2600 2800
TURBINE INLET TEMPERATURE, °F
3000
Figure 6-8. Permissible Gas Turbine NOX Concentrations
to meet Rule 67 of the Los Angeles Air
Pollution Control District (315 MW Power Output)
6-14
-------
reduction, it is concluded that new combustor designs of the prevaporization/
premising type are needed to meet the low NO emission requirements with
X
gas turbines. Catalytic and possibly porous-plate surface combustors have
the potential to meet these requirements.
6. 5 FUEL PREPARATION AND PREIGNITION PROBLEMS
Catalytic combustors require a completely prevaporized and uniformly
distributed air-fuel mixture at the catalyst inlet to prevent the excessive
temperatures that might result from droplet burning on the catalyst surface.
Therefore, catalytic combustors are limited to gaseous fuels such as
methane, propane, natural gas, and coal gas and certain distillate fuels
including gasoline and kerosene.
Recognizing these requirements, NASA/Lewis and the AFAPL are currently
working on the development of a number of different fuel atomization and
prevaporization-mixing approaches. These systems are designed to provide
a uniform fuel vapor-air mixture at the inlet to the catalytic combustor section
under all operating conditions of the gas turbine. For stationary gas turbines,
which are generally operated at semisteady-state conditions, the design require-
ments for the prevaporization-mixing chamber might be less stringent.
In prevaporized-premixed systems, mixture preiginition represents a potential
problem area which requires special design consideration for several reasons.
For instance, preignition of the mixture before completion of the fuel
vaporization-mixing process could result in severe local overtemperature
conditions leading to catalyst and substrate failure. On the other hand, pre-
ignition of a completely vaporized and premixed air-fuel charge could seriously
compromise the effectiveness of the catalyst in those designs in which a high
porosity Y -alumina wash coat is applied to enhance catalyst lightoff. If these
catalyst sections are exposed to temperatures in excess of about 1750 F, the
V-alumina is converted to the low porosity a -phase, resulting in a very sub-
stantial reduction in the number of active catalyst sites.
6-15
-------
The autoignition temperatures of methane and kerosene are depicted in
Figure 6-9 as a function of operating pressure (Ref. 6-9). As illustrated,
the autoignition temperatures of these fuels decrease -with increasing pressure
level. Based on an extrapolation of the available methane data, it appears
that the autoignition temperature of methane is higher than the combustor air
inlet temperature predicted for simple-cycle gas turbines operating at com-
pressor pressure ratios up to about 25:1 (Figure 6-3). Therefore, it is unlikely
that preignition of methane-air mixtures would occur in these particular engines.
Conversely, in regenerated gas turbines, the combustor air inlet temperatures
shown in Figures 6-4 and 6-5 are sufficiently high for most designs to permit
preignition of methane-air mixtures. Because of the lower autoignition tem-
perature of kerosene, preignition could occur with this fuel at the operating
conditions of most simple-cycle and regenerative-cycle gas turbines.
The occurrence of preignition is affected also by a number of additional para-
meters, including fuel-air mixture ratio and ignition delay time. As shown
in Figure 6-10, each fuel has a minimum concentration below which auto-
ignition cannot take place (Ref. 6-10). This concentration, called the "lower
flammability limit," decreases with increasing mixture temperature. For
the various simple-cycle and regenerative-cycle gas turbines, the fuel-air
mixture entering the combustor is sufficiently rich to permit ignition of all
turbine inlet temperatures above about 2100 F.
The ignition delays for kerosene and gasoline in air at atmospheric pressure
are presented in Figure 6-11 as a function of air temperature (Ref. 6-4). At
elevated pressures, the delay times tend to be somewhat higher. In general,
the ignition delay time of the.various fuels, which is composed of physical
delay time and chemical delay time, decreases with increasing air temperature.
The physical delay accounts for the elapsed time between fuel injection and the
formation of a combustible mixture and subsequent heating of the mixture to
the autoignition temperature. The chemical delay is the time from the begin-
ning of perceptible chemical reaction to'the occurrence of autoignition. The
physical and chemical delay times of gasoline and kerosene are of comparable
6-16
-------
1400
1200
1000
H
<
2
LJ
800
O 600
z
o
6 400
=>
200
METHANE
KEROSENE
I
10 20
AIR PRESSURE, crtm
30
Figure 6-9. Autoignition Temperatures for Methane and
Kerosene vs Pressure (Ref 6-9)
6-17
-------
1 I 1
% air = 100%—% combustible
Pentane
Hexane
Heptane
Octane
Nonane
Decane
-200
200
Figure 6-10.
400 600 800
TEMPERATURE, °C
1,000 1,200
1,400
Effect of Temperature on Lower Limits of
Flarnmability of 10 Paraffin Hydrocarbons in
Air at Atmospheric Pressure (Ref. 6-10)
6-18
-------
0 1000 1200 1400 1600 1800
INLET TEMPERATURE, °F
Figure' 6-11. Ignition Delay Time vs Inlet Temperature
at 1 Atmosphere (Ref. 6-4)
2000
6-19
-------
magnitude, but overlap each other to some degree (Ref. 6-11). No
reliable data have been found for methane. However, there are indications
that the ignition delay times of gaseous methane are somewhat higher than
the combined physical and chemical delays shown in Figure 6-11 for
kerosene (Ref. 6-11).
Based on these considerations, it is concluded that the ignition delay times
for kerosene and gasoline are sufficiently short to create a potential pre-
ignition problem in gas turbine combustors utilizing prevaporized fuel-air
mixtures. With methane, preignition is probably limited to regenerated gas
turbines, in which the combustor inlet temperature is generally above the
autoignition temperature. In these cases, special care must be exercised in
the design of the prevaporization-mixing chamber. These aspects are
further discussed in the following section.
6.6 CATALYTIC COMBUSTORS
6.6.1 Catalytic Combustor Design Parameters
A number of important parameters affecting the design and optimization of
catalytic gas turbine combustors have been identified during this study.
These include various approaches to minimize emissions, housing design,
catalyst and substrate configurations, combustor sizing, lightoff procedures,
and fuel preparation. These are briefly discussed in the following
paragraphs.
6.6.1.1 Emissions
The low emission potential of catalytic combustors operating at tempera-
tures of about 2000 F has been demonstrated. Consistent with theoretical
predictions, the observed NO emissions are practically zero and the
CO and HC emissions are very low and comparable to current gas turbines.
Further reductions of the CO and HC might be possible by increasing the
operating temperature of the unit, consistent with the temperature limitations
of the catalyst and substrate materials. This could be accomplished by using
6-20
-------
only a portion of the total compressor air in the combustor and adding the
remainder in a secondary combustion zone located between the catalytic unit
and the turbine nozzle inlet. Another possible approach to reduce the HC
and CO emissions involves increasing the residence time of the combustion
gases in the high-temperature zone by means of a reduction of the design
specific heat release rate of the catalytic combustor. Although the associ-
ated increase in combustor volume might preclude the application of this
approach to automotive and aircraft gas turbines, it might well be feasible
for stationary gas turbines, which generally have less restrictive volume
limitations.
6.6.1.2 Housing Design
To minimize the outer temperature of the catalyst housing and quenching of
the chemical reactions in the wall zone, insulation of the inside surfaces of
the housing is mandatory. The mechanical and thermal integrity of the
insulation requires experimental verification under simulated gas turbine
operating conditions.
6.6.1.3 Catalyst and Substrate Configuration
Currently, very little information is available regarding the catalyst and
substrate materials composition and the important design parameters, which
are considered proprietary by the manufacturers. Basically, either pellet
or monolithic substrate configurations are considered applicable to catalytic
combustors. Both types are currently utilized in industrial and automotive
catalyst systems. In automotive applications, catalyst failures due to engine
vibrations have occurred in the past, especially on four-cylinder engines.
Under normal operating conditions, the vibration levels in gas turbines are
very low; as a result, catalyst failure due to vibrational loads is not con-
sidered to be a major problem area for catalytic gas turbine combustors.
The monoliths currently available from American Lava, Corning,and DuPont
have a temperature capability of about 2400°F, which is higher than the maxi-
mum turbine inlet temperature projected for stationary gas turbines in the 1970
decade. However, substrates manufactured from other materials such as
6-21
-------
zirconia, silicon carbide, and alumina would have to be developed for use
in the advanced high-temperature gas turbines projected for the late 1980s.
6.6.1.4 Combustor Sizing
Catalytic combustors designed and manufactured by Engelhard are currently
being tested by NASA/Lewis and by the AFAPL. Both organizations are
optimistic regarding the feasibility of low-emission catalytic combustors
for automotive and aircraft gas turbine applications. Since space velocity
and specific heat release rate have a direct impact on the required volume
of the unit, it is most important to maximize space velocity within the con-
straints of the desired emission durability performance. This might be
accomplished by means of segmentation of the catalyst substrates selected
for the various applications. For example, the upstream section of the
catalytic combustor might be designed for rapid lightoff through the use of
a low lightoff temperature catalyst formulation deposited on a high porosity
•y-alumina wash coat. Since •y-alumina is limited to operating temperatures
of about 1750 F, other catalyst segments, coated with lower specific surface
area materials such as a-alumina, would be required in the high-temperature
regions of the combustor. In order to optimize the performance and dura-
bility of the unit, the various segments might be coated with different
catalyst materials, depending upon the fuel and other operating parameters
of the particular application.
6.6.1.5 Catalyst Lightoff Procedures
With regard to catalyst lightoff, a number of organizations are currently
involved in the development of preheat systems for potential use in catalytic
combustors. NASA/Lewis favors electric preheating of the catalyst bed,
while AFAPL is concentrating its efforts on a system using waste heat from
auxiliary power systems. Engelhard is in the process of evaluating several
proprietary concepts designed for use with different fuels. Automotive
and aircraft gas turbines require very rapid catalyst lightoff in order to
minimize the start delay time of the engine and the cold start emissions.
These factors are considered to be less important for stationary gas turbines.
6-22
-------
Therefore, preheating of the catalyst could occur more gradually and this
would result in lower thermal stresses and longer catalyst life. Electric
heaters as well as waste heat and pilot flame preheat techniques are potential
candidates for stationary gas turbines.
6. 6. 1. 6 Fuel Preparation
Catalytic combustors are limited to gaseous and distillate liquid fuels. To
prevent catalyst failure due to fuel coking or overtemperature conditions, a
well designed fuel prevaporization and mixing system is required. This sys-
tem, located upstream of the catalytic combustor proper, must be large
enough to provide complete fuel vaporization and uniform fuel and air mixing,
and small enough to prevent preignition of the fuel-air charge in the pre-
vaporization/mixing chamber. The design of the chamber walls requires
careful consideration of aerodynamic principles in order to avoid zones of
flow separation which might serve as flame holders.
Cost data for catalytic combustors are currently not available. NASA/Lewis
intends to conduct a comprehensive economic analysis of catalytic combus-
tors for gas turbines upon completion of its scheduled catalytic combustor
feasibility program.
6. 6. 2 Potential Problem Areas
The low-emission potential of catalytic combustors has been demonstrated by
Engelhard and by the EPA in exploratory test programs conducted over
limited ranges of simulated gas turbine operating conditions. These initial
results are encouraging. However, a number of potential problem areas
must be resolved by carefully planned experimental and theoretical investi-
gations before a complete assessment can be conducted regarding the appli-
cability of catalytic combustors to both new and existing stationary gas
turbines. These areas, which are briefly highlighted in the following
sections, include:
1. Catalyst durability
2. Specific heat release rate
3. Emission characteristics
6-23
-------
4. Mechanical integrity
5. Catalyst materials optimization
6. Ignition characteristics
6. 6. 2. 1 Catalyst Durability
The service life of catalysts is most strongly impacted by two factors: cat-
alyst poisoning and substrate degradation. Poisoning of the catalyst by lead,
sulfur, and other contaminants contained in the fuel and lubricating oil, and
loss in catalyst activity due to overtemperature conditions, are major prob-
lem areas in automotive catalysts. In industrial catalysts, temporary deac-
tivation due to fuel coking has been observed in some cases. Poisoning by
carbon buildup and by lead and other metal compounds is considered less
likely -with gaseous and kerosene-type fuels; however, because of the long
maintenance-free service life normally demanded from industrial gas tur-
bines, the effects of trace amounts of these contaminants on catalyst durabil-
ity performance must be systematically evaluated. Conversely, sulfur is
found in some natural gas grades and in most distillate fuels considered for
use in gas turbines. Prolonged exposure of the catalyst to sulfur compounds
could seriously compromise catalyst performance, in spite of the relatively
high operating temperatures projected for catalytic gas turbine combustors.
Sulfur poisoning, which has created problems in automotive base metal cat-
alysts, has been of less concern in platinum group metal catalysts used in
many petrochemical processes. In particular, platinum has shown consider-
able resistance to poisoning by sulfur.
In stationary gas turbines, which are normally designed to operate at semi-
steady-state conditions, overtemperature is not considered a major problem
area provided a -well functioning fuel prevaporization and mixing system is !
utilized. However, the effect of temperature excursions on the durability of
the projected catalytic combustors requires experimental evaluation to deter-
mine the degree of temperature and fuel-air mixture control required for the
various projected applications. A loss in catalyst activity would result
6-24
-------
primarily in higher CO and HC emissions, and would probably have a very
small effect on the NOX emissions. Therefore, CO and HC represent key
parameters which have to be very closely monitored during the course of the
recommended catalytic combustor durability program.
Some initial durability testing will be conducted by AFAPL as part of its
catalytic combustor feasibility test program. However, the scheduled test
duration of about 100 hours is by no means adequate to provide the informa-
tion required for an assessment of the durability performance of catalytic
combustors for stationary gas turbines.
6.6.2.2 Specific Heat Release Rate
The specific heat release rate of the catalytic combustor is an important
design parameter affecting the size, weight, and cost of the unit. In general,
the CO and the HC emissions from catalytic combustors are expected to
decrease as specific heat release rate (or space velocity) is reduced while
the NO emissions remain essentially constant. To permit optimization of
3t
catalytic combustors in terms of size and emission levels, the effects of
specific heat release rate on the emissions and pressure drop must be char-
acterized over the ranges of operating conditions projected for the current
and future simple-cycle and regenerative-cycle gas turbines.
6. 6. 2. 3 Emission Characteristics
To permit a meaningful assessment of the applicability of catalytic combustors,
the lightoff and emission durability characteristics of these devices must be
determined for the .different fuels as a function of air-fuel mixture ratio, air
inlet temperature and pressure, flow rate, and operating time.
In the case of liquid fuels, the effects of fuel-bound nitrogen on the NOx emis-
sions of catalytic combustors must be evaluated experimentally. It is well
known that in conventional combustors, as much as 70 percent of the bound
nitrogen is converted to NO (Ref. 6-12).
6-25
-------
6.6.2.4 Mechanical Integrity
Vibration and thermal stress levels are expected to be relatively low in
stationary gas turbine combustors, except perhaps during engine startup.
However, experimental evaluation of these parameters is considered man-
datory in order to determine the limitations and margins of safety with
respect to substrate cracking and chipping of the various projected catalyst,
substrate, and container configurations.
6.6.2.5 Catalyst Materials and Loading Requirements
Although Engelhard has presumably selected the best currently known cat-
alyst and substrate combination, it is recommended that an independent cat-
alyst screening program be conducted. The optimum catalyst formulation and
combustor design for the various fuels, fuel contaminants, and projected
ranges of turbine operating parameters should be determined. If platinum
group metals are used in the catalyst formulation, the effects of metal load-
ing on catalyst lightoff and emission durability performance should be evalu-
ated as part of the proposed screening program. These are very important
aspects which impact the tradeoffs between catalyst cost, size, and perform-
ance characteristics, and possibly even the world platinum group metal sup-
ply and demand balance.
6. 6. 2. 6 Ignition Characteristics
The optimum catalyst preheat temperature, flow rate, air-fuel ratio, and air
pressure required for smooth lightoff are currently not known and should be
determined for the fuels of interest. Also, the operational characteristics of
the fuel atomization, prevaporization, and mixing systems must be thoroughly
evaluated for the various fuels and system applications. These devices are
required to provide a uniform fuel vapor-air mixture at the combustor inlet
/
over the full range of engine operating conditions, without the occurrence of
preignition.
6-26
-------
6-7 POROUS-PLATE SURFACE COMBUSTORS
6.7.1 Introduction
Like catalytic combustors, the sintered metal porous-plate combustor
concept offers the potential of very low NOX emissions when the combustor
is operated below the adiabatic flame temperature.
Evaluation of this concept for steam boilers has indicated that porous-plate
surface combustors might be feasible for use in low-pressure applications.
However, at the elevated system pressures typical of gas turbines, the heat
transferred back to the burner plate increases proportionately and removal
of this heat at acceptable surface temperatures creates a serious problem
area. Although a detailed analysis of these parameters is beyond the scope
of this study, it is considered appropriate to briefly discuss the important
design parameters impacting the design of porous-plate surface combustors
for simple-cycle, regenerative-cycle, and combined-cycle stationary gas
turbines.
6.7.2 Discussion
General Electric, under contract to the EPA, is currently involved in the
development of a porous-plate combustor for potential use in automotive gas
turbines (Ref. 6-13). In the course of this program, General Electric has
examined the effects of pressure and air inlet temperature on the operational
characteristics of various radiation-cooled and internally cooled burner con-
figurations. With radiation-cooled burners, flashback has occurred on sev-
eral occasions, especially at higher system pressures. At a fuel-air equiv-
alence ratio of 0.9 and a system pressure of 3 atm, the measured back face
temperature of the porous burner plate was 1400°F, reflecting the higher
heat flux into the plate. Under these conditions, preignition of the fuel-air
mixture has occurred at the inlet to the burner. Attempts to prevent flash-
back by means of various types of insulating schemes were unsuccessful.
However, by reducing the equivalence ratio to 0.7, the burner has been oper-
ated successfully at pressures up to 4 atm without the occurrence of flashback.
6-27
-------
Because of the limited capability of radiation-cooled burners, General Electric
is now concentrating its efforts on internally cooled porous sintered metal
combustors of the type illustrated in Figure 6-12. In this design, part of the
heat transferred back to the plate is removed by the coolant, and the remainder
by radiation. The higher cooling capacity of the unit results in a lower burner
surface temperature. Preliminary indications are that this concept might be
developed into a viable design for use in automotive gas turbines, operating at
design point system pressures of about 4 atm. However, because of the high
coolant air temperatures of highly regenerated gas turbines, the development
of new porous metals with a higher temperature capability might be required
for these applications.
Current simple-cycle stationary gas turbines operate at pressures of the
order of 15 to 20 atm, and even higher pressures are projected for advanced
designs. At these elevated pressure levels, the heat load of the porous-plate
combustor is much higher than in steam boilers and automotive gas turbines,
resulting in high surface temperatures and large temperature differences
across the burner plate and imbedded tube materials. The severe tempera-
ture and thermal stress conditions might be alleviated to some degree by
operating the burner at very low equivalence ratios. However, under these
conditions, the burner surface area and volume requirements increase sub-
stantially, creating a potential packaging problem. Furthermore, less cool-
ing air is available in this case, which tends to further increase the size of
the burner.
In highly regenerated stationary gas turbines, the optimum system pressure
is about 4 to 6 atm, which is comparable to automotive gas turbines. As a
result, the heat flux to the burner is reduced and the concept might be feasible
for these particular applications. However, in regenerative gas turbines, the
temperature of the coolant air is very high, especially in the case of the
elevated turbine inlet temperature levels projected for advanced gas turbines.
This would result in even higher burner surface temperatures.
6-28
-------
Figure 6-12. Gene ral Electric Air-cooled Burner No. 106
During Fabrication
6-29
-------
In combined gas turbine-steam turbine cycles, the porous-plate cooling
problems could be alleviated to some degree by using water from the steam
turbine cycle to cool the porous plate combustor.
6.7.3 Conclusions
Cooling of porous-plate surface combustors by radiation alone is not con-
sidered practical at the elevated system pressures of simple-cycle and
re generative-cycle gas turbines. However, porous metal designs utilizing
imbedded cooling tubes represent a possible design approach for low-pressure
gas turbines of the type used in automotive or highly regenerated stationary
gas turbines and for combined gas turbine - s te am turbine cycles. Considerably
more research and development work is required, especially in the sintered
metals area, before the concept can be seriously considered for application
in simple-cycle, re generative-cycle, and combined-cycle stationary gas
turbines.
6-30
-------
REFERENCES
6-1. F. R. Biancardi and G. T. Peters, "Advanced Nonpolluting Gas
Turbine for Utility Applications in Urban Environments, " Gas Turbine
and Fluids Engineering Conference and Products Show, San Francisco,
Calif. , 20-30 March 1972, Paper 72-GT-64.
6-2. D. J. Ahner, Environmental Performance, Gas Turbine Reference
Library, GER-2480, General Electric Comoany, Schenectady, N. Y.
(1971).
6-3. JL. Bogden, et al. , Analysis of Aircraft Exhaust Emission Measure-
ments, NA-5007-K-1, Cornell Aeronautical Laboratory Inc. , Buffalo,
N. Y. (15 October 1971).
6-4. W. R. Wade, et al. , "Low Emission Combustion for the Regenerative
Gas Turbine, Part 1, Theoretical and Design Considerations, " Gas
Turbine Conference and Products Show, Washington, D. C. , 8-12
April 1973, Paper 73-GT-ll.
6-5. N. A. Ax.elborn, et al. , "Low Emission Combustion for the Regenera-
tive Gas Turbine, Part 2, Experimental Techniques, Results, and
Assessment, " Gas Turbine Conference and Products Show, Washington,
D. C. , 8-12 April 1973, Paper 73-GT-12.
6-6. L. Kl Jain, E. L. Calvin, R. L. Looper, State of the Art for
Controlling NO Emissions,, Part I, Utility Boilers, Catalytic,Inc. ,
Charlotte, N. CT, (EPA Contract 68-02-0241, Task 2) (September
1972).
6-7. Hybrid Heat Engine/Electric Systems Study, TOR-0059(6769-01)-2,
The Aerospace Corporation, El Segundo, Calif. (1 June 1971).
6-8. W. S. Blazowski and R. E. Henderson, Assessment of Pollutant
Measurement and Control Technology and Development of Pollutant
Reduction Goals for^Military Aircraft Engines. AFAPL Report
TR-72-102 (November 1972).
6-9. L. S. Marks, ed. , Mechanical Engineering Handbook, McGraw Hill,
New York (1952) pp 1208-1212.
6-10. M. G. Zabetakis, Flammability Characteristics of Combustible Gases
and Vapors, Bulletin 627, U.S. Bureau of Mines (1965).
6-31
-------
6-11. Telecommunication with Mr. W. R. Wade, Ford Motor Company,
Detroit, Mich. , 29 May 1973.
6-12. G. Blair Martin and E. E. Berkau, "An Investigation of the Conversion
of Various Fuel Nitrogen Compounds to Nitrogen Oxides in Oil Com-
bustion, " AIChE National Meeting, 30 August 1971.
6-13. R. J. Rossbach, Development of Low-emission Porous-plate Com-
bustors for Automotive Gas Turbine and Rankine Cycle Engines,
Quarterly Progress Report GESP-738, General Electric Company,
Cincinnati, O. (May 1973).
6 -32
-------
APPENDIX A
VISITS AND CONTACTS
-------
APPENDIX A
VISITS AND CONTACTS
A-l
ORGANIZATIONS VISITED
Organization
Institute of Gas Technology
Chicago, Illinois
Perfection Products Company
Waynesboro, Georgia
Burnham Corporation
Lancaster, Pennsylvania
Selas Corporation of America
Dresher, Pennsylvania
Engelhard Minerals &
Chemicals Corp.
Engelhard Industries Division
Murray Hill, New Jersey
Air Force Aero Propulsion
Laboratory
Dayton, Ohio
American Gas Association
Cleveland, Ohio
NASA/Lewis Research Center
Cleveland, Ohio
General Electric Company
Energy Systems Programs
Cincinnati, Ohio
Date of Visit
November 1, 1972
November 2, 1972
April 2, 1973
April 2, 1973
April 3, 1973
April 4, 1973
April 5, 1973
April 6, 1973
April 6, 1973
Primary Contact(s)
Mr. D. Larson
Dr. R. Rosenberg
Dr. P. Goodell
Mr. R. Givler
Mr. R. Reichhelm
Dr. C. Gottschlich
Dr. G. Gillespie
Dr. R. Carrubba
Mr. R. Heck
Capt. W. Blazowski
Dr. D. DeWerth
Mr. J. Heller
Mr. T. Mroz
Mr. R. Rossbach
A-i
-------
A-2
ORGANIZATIONS CONTACTED BY TELEPHONE
Organization
A. F. Holden Company
Milford, Michigan
Oxy Catalyst Inc.
West Chester, Pennsylvania
Matthey Bishop, Inc.
Malvern, Pennsylvania
Impala Industries
"Wichita, Kansas
Zebco
Tulsa, Oklahoma
Combustion Engineering
New York, N. Y.
Coleman Company
Wichita, Kansas
Shell Oil Company
Wood River, Illinois
Argonne National Laboratories
Argonne, Illinois
Pope Evans
Alexandria, Virginia
Solar
San Diego, California
Westinghouse Electric Corp.
Pittsburgh, Pennsylvania
Esso Research &
Engineering Co.
Linden, New Jersey
Gulf Oil Corporation
Harmerville, Pennsylvania
Office of Coal Research
Washington, D. C.
Date of Visit
December 8, 1972
January 19, 1973
January 19, 1973
January 23, 1973
January 23, 1973
January 23, 1973
January 23, 1973
February 5, 1973
February 5, 1973
February 5, 1973
February 5, 1973
February 5, 1973
February 6, 1973
February 8, 1973
February 23, 1973
Primary Contact
Mr. K. Kuhn
Mr. J. Houdry
Dr. A. Khuri
Mr. J. Smith
Mr. W. Duncan
Dr. Ulmer
Mr. F. Smith
Mr. L. Graiff
Dr. A. Jonke
Mr. S. Ehrlich
Mr. W. A. Compton
Mr. D. Archer
Mr. A. Skopp
Dr. B. Taylor
Dr. G. Hill
A-2
-------
Organization
Ford Motor Company
Dearborn, Michigan
City Administration
Burbank, California
Dept. of Water and Power
City of Los Angeles
Los Angeles, California
Date of Visit
May 29, 1973
May 29, 1973
June 7, 1973
Primary Contact
Mr. W. Wade
Mr. J. Hall
Mr. N. Bassin
A-3
-------
APPENDIX B
UNITS OF MEASURE--CONVERSIONS
-------
APPENDIX B
UNITS OF MEASURE--CONVERSIONS
Environmental Protection Agency policy is to express all measurements in
Agency documents in metric units. With a few exceptions, this report uses
British units of measure. For conversion to the metric system, use the
following conversions:
To convert from
ft
ft2
ft3
in
. 2
in
BTU
BTU/ft3
BTU/hr-ft2
BTU/lb
BTU/in2
ft /min-in
lb/106 BTU
lb/in2
gal/hr
hp
to
Multiply by
°c
meters
2
meters
3
meters
cm
2
cm
kcal
kcal/m
2
kcal/hr-m
cal/gm
2
kcal /cm
3, . 2
in /mm- cm
gm/10 cal
mm. Hg
1/hr
kW
5/9 (°F-32)
0.304
0.0929
0.0283
2.54
6.45
0.252
8.90
2.71
0.556
0.039
0.00439
1.80
51.71
3.78
0.746
B-l
-------
BIBLIOGRAPHIC DATA
SHEET
1. Report No.
EPA-650/2-73-014
3. Recipient's Accession No.
4. lit It .mil Sulu it l<
nvestigation of Surface Combustion Concepts for
MOX Control in Utility Boilers and Stationary
Gas Turbines
5- Ri-port Date
August 1973.
6.
7. Author(s) W. U. Roessler, E. K. Weinberg, J. A. Drake,
H. M. White, and T. lura
8- Performing Organization Rep'
No'ATR-73(7286)-2
9. Performing Organization Name and Address
Urban Programs Division
The Aerospace Corporation
El Segundo, California 90245
10. Project/Task/Work Unit No.
11. Contract/Grant No.
R-801490
12. Sponsoring Organization Name and Address
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, North Carolina 277 11
13. Type of Report & Period
Covered
Final
14.
15. Supplementary Notes
16. Abstracts -^he report gives results of a review of the state-of-the-art of concepts of
surface combustion (that occurring near a solid surface) and an assessment of their
applicability to large utility boilers and stationary gas turbines. Catalytic combustion
looks promising for gas turbines. However, comprehensive experimental and
theoretical investigations are required for a meaningful assessment. Catalytic
combustion is not considered feasible for existing steam boilers, but might apply
to newly designed units. Sintered metal porous-plate combustors with imbedded
cooling tubes ( a non-catalytic concept), although technically feasible for steam
boilers and possibly for low-pressure ratio gas turbines, has packaging problems
and requires gaseous and distillate fuels.
17. Key Words and Document Analysis. 17a. Descriptors
Air Pollution
Boilers
Utilities
Catalysis
Gas Turbines
17b. Identifiers/Open-Ended Terms
Air Pollution Control
Stationary Sources
Surface Combustion
Sintered Metal Porous-Plate Combustion
17c. COSATl Field/Group
13R. 21B
18. Availability Statement
Unlimited
19. Security Class (This
Report)
ASS1FIED
20. Security Class (This
Page
UNCLASSIFIED
21. No. of Pages
185
22. Price
NTIS-3S (REV. 3-72)
------- |