EPA-460/3-73-001
DEVELOPMENT
OF LOW EMISSION
POROUS-PLATE COMBUSTOR
EOR AUTOMOTIVE GAS TURBINE
AND RANK1NE CYCLE ENGINES
U.S. ENVIRONMENTAL I'KOTKCTION \(,KNO
Office of Air ami \\ at<-r Programs
-------
EPA-460/3-73-001.
DEVELOPMENT OF LOW EMISSION
POROUS-PLATE COMBUSTOR
FOR AUTOMOTIVE GAS TURBINE
AND RANKINE CYCLE ENGINES
by
Mr. Richard J. Rossbach
General Electric Co .
Energy Systems Programs
P.O. 13ox 15132
Cincinnati, Ohio 45215
Contract No. 68-01-0461
Project Officers:
Mr. Thaddcus S. Mroz (NASA - Lewis)
Mr. William C. Cain (EPA)
Prepared lor
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Mobile Source Air Pollution Control
Alternative Automotive Power Systems Division
Ann Arbor , Michigan 48105
September
-------
This report is issued by the Office of Mobile Source Air Pollution
Control, Office of Air and Water Programs, Environmental Protection
Agency, to report technical data of interest to a limited number of
readers. Copies of this report are available free of charge to
Federal employees, current contractors and grantees, and non-profit
organizations - as supplies permit - from the Air Pollution Techni-
cal Information Center, Environmental Protection Agency, Research
Triangle Park, North Carolina 27711 or may be obtained, for a
nominal cost, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22151.
This report was furnished to the U.S. Environmental Protection Agency
by The General .Electric Company in fulfillment of Contract No.
68-01-0461 and has been reviewed and approved for publication by the
Environmental Protection Agency. Approval does not signify that the
contents necessarily reflect the views and policies of the agency.
The material presented in this report may be based on an extrapolation
of the "State-of-the-art." Each assumption must be carefully analyzed
by the reader to assure that it is acceptable for this purpose. Results
and conclusions should be viewed correspondingly. Mention of trade
names or commercial products does not constitute endorsement or
recommendation for use.
Publication No. EPA-460/3-73-001
11
-------
CONTENTS
Page
SUMMARY 1
INTRODUCTION 5
ABSTRACT 17
CONCLUSIONS AND RECOMMENDATIONS 19
GAS TURBINE COMBUSTOR 31
Discussion 31
Combustor Loading and Emission Analysis 33
Combustor Concept Feasibility Development 105
Fuel-Air Mixture Supply Development 147
Porous Plate Combustor Fabrication Development 167
Combustor Configuration and Engine Integration Design . . . 180
RANKINE CYCLE COMBUSTOR 189
/
Discussion . . . • . • 189
Preliminary Burner-Vapor-Generator Design ......... 191
Bench Tests 200
REFERENCES 232
APPENDIX 235
ACKNOWLEDGEMENT 244
iii
-------
LIST OF FIGURES
Figure No. Page
1 Schematic of Porous Plate Burner 9
2 Effect of Unburned Gas Superficial Velocity and
Equivalence Ratio Upon Flame Temperature 11
3 NO Formation Rate From Hot-Air Mechanism 13
4 Calculated W>2 Formation Rates as a Function of Porous
Plate Superficial Velocity and Equivalence Ratio. ... 14
5 Air-Cooled Porous-Plate Burner Concept 21
6 Schedule for Gas-Turbine Porous-Plate Combustor .... 22
7 Conceptual Design of Porous-Plate Burner-Vapor Generator
for Rankine System 26
8 Schedule for Rankine Porous-Plate Combustor 28
9 Low NOX Frit Plate Gas Turbine Combustor (Fuel Inside). 34
10 Fuel Scheduling of Base Line Engine 37
11 Hot-Side Burner Temperature: Primary Air from Com-
pressor Discharge 38
12 Hot-Side Burner Temperature: Primary Air from Com-
pressor Discharge 39
13 Hot-Side Burner Temperature: Primary Air from Re-
generator Discharge 41
14 Hot-Side Burner Temperature: Primary Air from Re-
generator Discharge 42
15 Hot-Side Burner Temperature 43
16 Effect of Emissivlty on Burner Temperatures 45
17 Burned Gas Temperature of Radiation Cooled Combustor
at Steady State Points 48
18 Burned Gas Temperature of Radiation Cooled Combustor
at Steady State Points 49
19 Burned Gas Temperature of Radiation Cooled Combustor
at Steady State Points 50
20 Burned Gas Temperature of Radiation Cooled Combustor at
FDC Points 51
21 Burned Gas Temperature of Radiation Cooled Combustor at
FDC Points 52
22 Burned Gas Temperature of Radiation Cooled Combustor at
FDC Points 53
23 Burned Gas Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points 54
iv
-------
Figure No. Page
24 Burned Gas Temperature of Radiation Cooled Corabustor at
Steady State Wide Open Throttle Points „ 55
25 Burned Gas Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points 56
26 Surface Temperature of Radiation Cooled Combustor at
Steady-State Points 57
27 Surface Temperature of Radiation Cooled Combustor at
Steady-State Points 58
28 Surface Temperature of Radiation Cooled Combustor at
Steady State Points 59
29 Surface Temperature of Radiation Cooled Combustor at
FDC Points 60
30 Surface Temperature of Radiation Cooled Combustor at
FDC Points. . . i 61
31 Surface Temperature of Radiation Cooled Combustor at
FDC Points. 62
32 Surface Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points 63
33 Surface Temperature of Radiation Cooled Corabustor at
Steady State Wide Open Throttle Points 64
34 Surface Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points „ . . . 65
35 Surface Temperature of Radiation Cooled Corabustor at
Steady State Points 66
36 Surface Temperature of Radiation Cooled Combustor at
Steady State Points 67
37 Surface Temperature of Radiation Cooled Combustor at
Steady State Points 68
38 Surface Area of Radiation Cooled Combustor at FDC
Points 69
39 Surface Temperature of Radiation Cooled Corabustor at
FDC Points 70
40 Surface Temperature of Radiation Cooled Combustor at
FDC Points 71
41 Surface Temperature of Radiation Cooled Corabustor at
Steady State Wide Open-Throttle Points 72
42 Surface Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points 73
43 Surface Temperature of Radiation Cooled Combustor at
Steady State Wide Open Throttle Points 74
44 Superficial Velocity Along Steady State Power Match
Line 76
-------
Figure No. Page
45 Combustor Temperature Along Steady State Power Match
Line 77
46 NOX Emission Index Along Steady State Power Match
Line 78
47 CO Emission Index Along Steady State Power Match
Line 79
48 Nodal System for Transient Thermal Analysis 82
49 Time-Dependent Gas Generator Speed During 3 Wide-Open-
Throttle Transients (WOT) at 85°F 85
50 Fuel Scheduling of Base Line Engine During the Four
Transients 86
51 Configuration or Materials in Porous Burner (for
Transient Analysis) 88
52 Schedule of Burner Area and Equivalence Ratio During
the 4 Transients 89
53 Calculated Thermal Response of the Porous Combustor
During WOT From Idle 90
54 Undiluted Burned Gas Temperature During the WOT
Transient From Idle to 100% Gas Generator Speed of the
Base Line Engine 92
55 Conditions at the Burner Surface During WOT From Idle
to 100% Gas Generator Speed 93
56 Heat Fluxes Back to the Burner Surface During WOT
from Idle to 100% Gas Generator Speed 94
57 Thermal Response of Porous Combustor to a Hypothetical
Transient 96
58 Calculated Thermal Response of the Porous Combustor
During the WOT Transient from 40% Engine Power 97
59 Burned Gas Temperature During the WOT Transient from
40% Engine Power of the Base Line Engine 98
60 Calculated Thermal Response of the Combustor During
the WOT Transient from 60% Engine Power of Base Line
Engine. . „ 99
61 Burned Gas Temperature During the WOT from 60% Engine
Power of the Base Line Engine 100
62 Comb us tor Response to Cold Start Transient 101
63 Burned Gas Temperature During Cold Start Transient
of Base Line Engine 102
64 Predicted N02 Emissions Index Based on the Hot Air
Mechanism 103
65 Predicted CO Emissions Index 104
66 Gas Turbine Porous-Plate Combustor Test Section .... 106
67 SIC Spring Loaded Configuration 108
vi
-------
FIKure No. Page
68 Metallic Corabustor Configurations o . . . « 110
69 Kanthal-Wrapped Burner 112
70 Kanthal-Wrapped Burner with Zirconia Cloth and
Cerapaper Insulation 113
71 Emission Measurements from Burner S/N 102 117
72 Configuration With SiC Rings Over Mullite Cylinder. . . 118
73 Schematic of Preignition Test Setup 119
74 Preignition Test Rig Hardware 120
75 Upstream Face Temperature Leading to Preignition at
3 Atmospheres 124
76 Impending Preignition at P = 4 Atmospheres 125
77 Radiant Heat Flux From A Burner 128
78 Calculated Maximum Wall Temperatures for Air-Cooled
Burner for Cooling Air Initial Temperature, 400°F „ . . 129
79 Calculated Maximum Wall Temperatures for Air-Cooled
Burner for Cooling Air Initial Temperature, 1400°F. . . 130
80 Heat Flux Back to the Porous Burner as a Function of
Pressure and of Superficial Gas Velocity (V2J 132
81 Design of Air-Cooled Porous Burner 106 133
82 Air-Cooled Burner #106 During Fabrication 136
83 Schematic of Gas Turbine Corabustor Test Facility. . . . 138
84 Scott Exhaust Gas Analyzer 140
85 Emissions from Air-Cooled Burner S/N 106 at 1 Atm.
Pressure 142
86 N02 Measurements at 1 to 4 Atmospheres on Air-Cooled
Burner 143
87 Burner 107 Pressure Drop 144
88 Measurements of CO and NOV Emissions on Burner 107. . . 146
X
89 Measured CO Emissions on Burner 108M with Both Heated
and Unheated Mr 148
90 Measured NOX Emissions on Burner 108M with Both Heated
and Unheated Air 149
91 Fuel Vaporizer for Gas Turbine Combustor 151
92 "Sonicore" Fuel Vaporizer 153
93 Estimated Performance of Fuel Vaporizer for Gas Turbine
Corabustor ...>..........••••••••• 157
94 Vaporizer and Inlet and Exhaust Lines 159
95 Vaporizer and Inlet and Exhaust Lines 161
96 Vaporizer Air .and Fuel Supply System 162
97 Experimental Vaporization Quality as a Function of the
B Group Parameter 166
vii
-------
Figure No. Page
98 Weight Gain of Fe-25Cr-4Al-Y Alloy During Oxidation
Testing in Air 174
99 Poroloy Permeability Retention After Exposure at 1800
and 2000°F (982 and 1093°C) in Air. Alloys GE 1541 and
H 875 175
100 Nichrome Porous Cylinder No. 3 As Sintered at 2350°F. . 181
101 Porous Cylinder No. 3 with Tube Sheets Brazed on Each
End 182
102 Air-Cooled Porous-Plate Burner Concept 185
103 Conceptual Design of Porous-Plate Burner-Vapor
Generator for Rankine System 192
104 Fuel-Air Mixer and Fuel Vaporizer Double-Swirl
Carbureting Concept 195
105 Preliminary Installation Configurations . . 196
106 Rankine Engine Fuel-Air Mixer Cup Flow Characteristics. 198
107 Estimated Vaporizer Performance 199
108 Half-Filled Copper Shot Burner, Illustrating Burner
Construction and Coolant Tube Manifolding 201
109 Schematic Diagram of Burner Test and Gas Sampling
Arrangement 202
110 Burner Heat Flux Back to the Burner from the Flame;
Propane-Air Mixtures 213
111 Burner Heat Flux Back to the Burner from the Flame;
Gasoline (EPA)-Air Mixtures 214
112 Comparison of Gasoline (EPA) and Propane-Air Burned
Gas Temperatures 215
113 CO Analyses from Two Flames 220
114 CO Experimental Measurements on Water-Cooled Burners
with 20 Milliseconds Residence Time 222
115 Variation of NO Formation Rate from "Hot-Air" Mechanism
as a Function of Reciprocal Absolute Temperature. NO
Formation Rate Adjusted to 2% Molar 0 224
116 N02 Experimental Measurements on Water-Cooled Burners
with 20 Milliseconds Residence Time . 230
A-l GDF I Combustor Test Facility 236
A-2 Test Cell Heaters and Flow Metering Sections 237
A-3 Fuel Storage and Pumping 239
A-4 Heater Control Console 240
A-5 Facility Control Console 241
A-6 Scott Exhaust Gas Analyzer „ 243
viii
-------
LIST OF TABLES
Table No. Page
1 Base Line Engine Combustor Test Points 36
2 Summary of Calculations 47
3 Silicon Carbide Combustor Configurations 109
4 Metallic Combustor Configurations 114
5 Preignition Tests 122
6 Parts List of Air-Cooled Porous Burner Shown in
Figure 81 134
7 Air-Cooled Burners (Cylindrical) 137
8 Gas Turbine Fuel Vaporizer Test Data 165
9 Porous Combustor Materials 168
10 Candidate Alloys for Combustor Components 170
11 Tensile Properties of Candidate Alloys for Combustor
Components 173
12 Critical Thermal Properties of Selected Ceramics . . . 176.
13 Porous Rankine Combustor 193
14 Rankine Model Burner Test Results 216
15 Rankine Model Burner Emission Results 219
ix
-------
SUMMARY
The requirements of the 1970 Clean Air Act Amendments were that the
Environmental Protection Agency enforce stringent exhaust emission standards
for new automobiles on unburned hydrocarbons and carbon monoxide in the
1975 model year and on nitrogen oxides in the 1976 model year (although
these requirements may be relaxed for a time). As the standards originally
stood, the most difficult standard to meet was that relating to NO ,
A
especially since it is difficult to obtain low NO and low CO together.
X
Experiments with porous-plate combustors previous to this contract had
shown very low NO levels, so further development work was undertaken.
X
The purpose of this contract was to evaluate analytically and experi-
mentally the use of the porous-plate combustor for use in the gas-turbine
or Rankine-cycle advanced automobile engines to control exhaust emissions.
By providing a uniform mixture of fuel and air (pre-mixed) and
passing it through a porous plate, combustion is initiated and a uniform
flame can be stabilized on the downstream surface of the porous material.
Because of the closeness of the flame to the plate and short flame height
(1-4 mm), heat is transferred from the flame to the plate, cooling the
flame and preventing the adiabatic flame temperature from being attained.
The heat transferred to the plate must be removed by radiation or a
coolant in the case of the gas-turbine combustor and by the working fluid
in the case of the Rankine-cycle combustor. This combustion process
virtually eliminates unburned hydrocarbon emissions as long as there is
-------
a small excess of oxygen. The reduction of the flame temperature from
the adiabatic to some lower temperature dramatically reduces the NO
emissions. The CO emission can be traded off with the very low NO
emissions in terms of residence time to make them low also.
THE GAS TURBINE COMBUSTOR
The following conclusions were drawn concerning the porous-plate
gas turbine combustor from the work described herein:
1. Analytical work indicates that the combustor equivalence ratio
and the porous-plate burner area must be variable in order
for the engine to operate between the steady-state level-road
load fuel requirements and the wide-open-throttle acceleration
fuel requirements.
2. The operational temperature of the downstream side of the
porous plate must be limited to a specific value at the higher
power levels to prevent flashback through the porous plate
and pre-ignition of the fuel air mixture. Air cooling of the
porous plate matrix, in addition to radiation cooling, is re-
quired to avert this condition.
3. Models of air-cooled porous-plate combustors have successfully
been made from sintered nichrome and tested.
4. Emission measurements with propane-air mixtures at inlet tem-
peratures of approximately 400°F and one atmosphere were
favorable. The NO emission index was below the 1976 Federal
Jk
Standards and the CO emission index was below the Standards
at combustor conditions corresponding to light engine loads
but somewhat above it at heavier loads. The unburned hydro-
carbon values were at about the detection level which is much
below the Standard.
-------
5. Tests of a fuel atomizer and vaporizer using an air assist
fuel nozzle spraying into a swirling flow resulted in all but
a few points having 93% or more of the fuel vaporized.
Based upon these results, a full size porous-plate combustor con-
cept has been developed for the Baseline Engine having a total porous
2
plate area of 2.2 ft and which extends only 2 inches higher vertically
than the conventional combustor for the engine. In this concept, the
total porous surface is made up of a number of air-cooled wedge-shaped
flat-plate segments. It is recommended that this combustor be designed,
fabricated, and tested with gasoline.
THE RANKINE CYCLE COMBUSTOR
The following conclusions were drawn concerning the porous-plate
Rankine-cycle combustor from the work described herein:
1. The Arrhenius plots of superficial velocity versus reciprocal
absolute flame temperature at constant equivalence ratio which
are used to design porous-plate combustors were confirmed for
five fuels including gasoline.
2. Both the emissions indexes for NO and CO increased with super-
X.
ficial velocity at constant equivalence ratios and for li
fuels were only below the Standard at conditions corresponding
to moderate engine loads.
3. The unburned hydrocarbon emission indexes were very low, about
at the detection level.
A. The fuel atomizer and vaporizer developed for the gas turbine
combustor are adequate for the Rankine combustor. All but a
few of the test points indicated 93% or more fuel vaporization.
-------
The original Rankine-cycle combustor concept is unchanged. Since
the sintered copper water-cooled porous-plate combustors are success-
fully being made by the General Electric Company and based upon the
above results, it is recommended that the full-size water-cooled Rankine
combustor be designed, fabricated, and tested using gasoline.
-------
INTRODUCTION
The purpose of the work under this contract was to evaluate the
use of the porous plate combustor systems for gas turbine and Rankine
cycle advanced automotive engines.
Evaluation of this unique combustion system included emission levels,
burning limits, air and fuel requirements, response to transients, general
operating characteristics, start up, durability, sensitivity to degree
of inlet fuel vaporization, and pressure drop. This work was intended
to provide proof of feasibility of use of porous plate burners by demon-
strating adequate performance in the following areas:
• Emission characteristics of the model or segments as developed
• Definition of start up characteristics
• A complete combustor operating map sufficient to define full
scale combustor designs. The information to be generated in-
cludes mixture quality control requirements, mixture distribu-
tion requirements, corabustor active area control, means of con-
trolling equivalence ratio, etc.
The approach to the gas turbine and Rankine cycle combustor develop-
ment included the following tasks:
Gas Turbine Combustor Development
IB. Combustor loading and emission analysis
IIB. Combustor configuration and engine integration design
TUB. Combustor concept feasibility development
-------
IVB. Fuel-air mixture supply development
VB. Porous plate combustor fabrication development
VIB. Prepare recommendations
Rankine Cycle Combustor Development
IR. Preliminary burner vapor generator design
IIR. Bench tests
IIIR. Engineering test unit
IVR. Prepare recommendations
The development of the Rankine Cycle Combustion system was performed
in parallel to the gas turbine combustor development. A complete inter-
change of technology was made, generated on either gas turbine or Rankine
cycle burners.
The requirements of the 1970 Clean Air Act Amendments are that the
Environmental Protection Agency enforce stringent exhaust emission standards
for new automobiles on unburned hydrocarbons and carbon monoxide in the
1975 model year and on nitrogen oxides in the 1976 model year. The
original 1976 standards are as follows:
Unburned hydrocarbons, grams/mile 0.41
Carbon monoxide, grams/mile 3.4
Nitrogen oxides, grams/mile 0.4
Although the automobile industry is attempting to meet the emission
standards with the conventional automobile reciprocating engine, there
are several other promising engine types which have a high potential of
meeting the 1976 emission Standards. Both the gas turbine and the Rankine
cycle engines have the potential for low exhaust emissions. The combustor
for each engine can be designed separately for efficient and controlled
-------
(1,2,3)
combustion while still being integrated into the engine configuration.
Currently, combustors designed for these two engine types still have
emissions problems, especially with difficulty in obtaining low NO
emissions. Many conventional combustors burn from sprayed droplets.
The result is that large variations in the local equivalence ratio can
exist in the primary combustion zone due to inefficient atomization and
mixing; for an overall lean equivalence ratio, part of the combustion
can occur at near-stoichiometric conditions, resulting in high peak flame
temperatures, with attendant high NO production rates, while the rest
X
of the combustion occurs at a very lean equivalence ratio at lower flame
temperatures with lower NO production rates. The overall result is
high NO , since the NO production rate increases logarithmically with
X X
temperature; the areas of low NO production cannot balance the areas of
X
high NO production.
X
In order to minimize NO emissions, both the flame temperature and
the residence time at temperature must be controlled. The flame tempera-
ture determines the NO production rate, while the residence time is the
time over which the NO production rate is operative. The flame tempera-
X
ture should not be excessively depressed; if the flame temperature •*?> too
low, then the CO decay rate is also low, and not enough CO will be burned
off to form CO , for a given residence time. Similarly, flame temperatures
that are too low will also result in high emissions of unburned hydro-
carbons, since the burn-off rates depend strongly on flame temperature.
Previously, experiments with a porous plate combustor had shown
very low NO levels, so further development work was undertaken. By
X
providing a uniform mixture of fuel and air and passing it through a
-------
porous plate, a uniform flame can be stabilized on the downstream surface
of the porous material. Because of the closeness of the flame to the
plate, heat is transferred from the flame to the plate, cooling the flame
and preventing the adiabatic flame temperature from being attained. The
heat transferred to the plate must be removed by radiation or a coolant
in the case of the gas-turbine combustor and by the working fluid in the
case of the Rankine-cycle combustor. The evenness of the flame virtually
eliminates unburned hydrocarbon emissions as long as there is a small
excess of oxygen. The reduction of the flame temperature from the adiabatic
to some lower temperature dramatically reduces the NO emissions. The
X
emission of CO can be traded off with the very low NO emissions in terms
J x
of residence time and flame temperature to make them low also.
The porous burner is shown schematically in Figure 1. Premixed
fuel and air pass through a porous matrix. A substantial amount of heat
is conducted from the flame back to the cooled porous matrix, since the
flat laminar flame is maintained close to the metal surface. The burned
gas (flame) temperature obtained with the porous burner varies from the
adiabatic flame temperature down to about 1500°F (2340°F) as the super-
ficial velocity is decreased from the adiabatic flame velocity. NO pro-
A
duction rates decrease logarithmically with burned gas temperature;
hence, since burned gas temperatures with the porous combustor are lower
than the adiabatic flame temperature, NO production rates are advantageously
lower than in conventional combustors at the same combustion rates.
The characteristics of flames burning on cooled porous plugs have
been discussed from various points of view. *• » » ' The operating principle
of the porous combustor relative to NO production rates is that a very
A
-------
SINTERED
POROUS
METAL
BURNED GAS
T 0
Figure 1. Schematic of Porous Plate Burner
-------
uniform, flat, laminar flame, about 1 mm high, is formed on the surface
of the porous combustor. This flame is non-adiabatic, with a predictable
and appreciable fraction of its heat release being transferred from the
unburned side of the flame back to the porous plate. Since the flame is
non-adiabatic, the burned gas temperature can be selected by design at
a value below the adiabatic temperature so as to control NO production
rates; NO production rates are extremely temperature dependent, and re-
X
ductions in burned gas temperatures result in significant reductions in
NO production rates. The heat absorbed by the porous plate can be re-
X
moved either by allowing the plate to operate at a high temperature such
that heat removal is accomplished by thermal radiation and preheating of
inlet unburned gas mixture or by incorporating cooling tubes in the
porous matrix for compressor discharge air in the case of the gas turbine
or for the working fluid in the Rankine application.
Measurements of burned gas temperature (T) with the porous burner
are related to superficial unburned gas velocity through the burner (V_,-)
and equivalence ratio (R) by straight lines on a log V.^ versus 1/T plot,
yielding a single line for each value of R; data from Kaskan is shown
in Figure 2. The four porous burner combustion lines are independent of
gas inlet temperature, to a first approximation. The lines at which
adiabatic combustion occurs are also shown for two different unburned
gas inlet temperatures. The inlet temperature of the fuel-air mixture
has a significant effect on the maximum velocity above which the flame
lifts off the porous surface. The points of intersection of the adiabatic
and porous burner lines determine the unburned gas velocities at which
the flat flame will begin to lift off the porous plate, the upper limit
10
-------
Temperature, °F
4000 3800 3600 3400 3200
—i 1 1 1 r
3000
—r
2800
100
10
Estimated Liftoff Limit for
Inlet Fuel-Air of 1200°F
Liftoff Limit
for Inlet of 77°F
Porous Burner
Combustion Lines
Equivalence Ratio
Note:
I
V_5 is at P - 1 atm.
I
I
4 x 10"4 5 x 10~4
Reciprocal Absolute Temperature, (1/°K)
6 x 10
.-*
Figure 2. Effect of Unburned Gas Superficial Velocity and Equivalence Ratio
Upon Flame Temperature.
11
-------
of operation. The lower limit of operation (or flame extinguishment) is
the gas velocity at which the burned gas temperature drops below that
required to sustain the flame reaction, which occurs at about 4 cm/sec.
As seen in Figure 2, the flame temperature obtained with the porous
burner can be considerably below the adiabatic flame temperature by varying
the unburned gas velocity, independent of equivalence ratio.
Shown in Figure 3 are NO formation rates based on the "hot-air"
mechanism. These rates are derived from the basic kinetic data used by
Fenimore . It can be seen that significant reductions in NO formation
rates can be achieved by lowering the actual flame temperatures by only
several hundred degrees. For example, a temperature reduction of 360°F
(200°C) reduces the NO production rate by a factor of 20.
The porous burner produces a very uniform, stable flame; hydrocarbon
and particulate emissions from lean flames on this burner have been
measured, and are approximately zero. CO emissions can be kept at a low
level by providing sufficient burned gas residence time at temperature
to allow decay to CO . However, the gas temperature after dilution must
be low enough so that the NO production rates are low. The emissions
X
design of the porous plate burner therefore simplifies at first approxi-
mation to selection of a flame temperature low enough to limit NO forma-
tion. The NO production rates of Figure 3 can then be used to compute
NO formation rates as a function of superficial velocity V-_ from the
flame temperature data of Figure 2; Figure A presents the NO formation
rates thereby estimated as a function of V... The extreme reductions of
NO that can be obtained with the porous combustor are apparent.
12
-------
10.0
3200 3000
2800
800 3600
(for N - 0.1)
NO Formation Rate by
Hot Air Mechanism
r = .; Normalized to 2200K
dT
At 1 Atn Total Pressure and No -0.1
2
J. Experimental Data from Various Flames
I _ I _ I
0.001
4 4.5 5 5.5
Reciprocal Absolute Temperature (1/T) x 10*, (1/°K)
Figure 3. NO Formation Rate From Hot-Air Mechanism.
13 '
-------
Equivalence Ratio R - 1.0
0.7
Full Load
10
1.0
Adlabatic Plane
1200°F Inlet
Adiabatlc Flame
77 F Inlet
Typical Mean
Load
Note: V2 Is at P-l atm
10 20 50 100
Superficial Gas Velocity, V (cm/sec)
Figure 4. Calculated N0£ Formation Rates as a Function of Porous Plate
Superficial Velocity and Equivalence Ratio.
-------
Until analytical ami experimental techniques were employed to evaluate
tin.' feasibility of the porous-plate combustor for both gas turbine and
Kankino atitomohM c- engines. As regards the gas turbine application, this
i,-purl contains ana.l yt leal results on the burner area requirements for
the various operating conditions of the-Baseline Engine as well as exhaust
eni.ssi.on predictions. The design concept of an air-cooled, variable-area
combustor for this engine is presented. Operational and emissions data on
several experimental combustors are presented along with the fabrication
development leading to these combustors. Finally the demonstration results
for a full-scale fuel-air mixture system are presented.
With regard to the automotive Rankine engine application, heat load
and emissions data are presented for propane and four liquid fuels. Although
the fuel-air mixture system developed for the gas turbine combustor is
directly applicable, alternate systems were.investigated. Finally
recommendations for the development of both gas turbine and Rankine
combustors are presented.
15
-------
ABSTRACT
Experiments with porous-plate combustors previous to this contract
had shown very low NO levels, so further development work was under-
X
taken. The purpose of this contract was to evaluate analytically and
experimentally the use of the porous-plate combustor for use in the gas-
turbine or Rankine-cycle advanced automobile engines to control exhaust
emissions.
For the porous-plate combustor for the gas turbine, the following
conclusions were drawn: (1) Analytical work indicates that the burner
area must be variable in order to meet the fuel flow requirements for
operation between steady-state load and wide-open-throttle acceleration;
(2) Air-cooling of the porous burner (in addition to radiation cooling)
is required to avoid flashback; (3) Models of air-cooled porous-plate
combustors have been successfully built from nichrome and tested; (4)
Emission measurements of NO , CO and unburned hydrocarbons with propane-
air mixtures at inlet temperatures of approximately 400°F and one atmosphere
were favorable; (5) Tests of a fuel atomizer and vaporizer resulted in
all but a few points having 93% or more of the fuel vaporized.
For the porous plate combustor for the Rankine-cycle expander, the
following conclusions were drawn: (1) The Arrhenius plots of superficial
velocity versus reciprocal absolute flame temperature at constant equiv-
alence ratio were confirmed for five fuels including gasoline; (2) Both
the emission indexes for NO and CO increased with superficial velocity
A
17
-------
at constant equivalence ratios and for liquid fuels (including gasoline)
were only below the Standard at conditions corresponding to moderate
engine loads; (3) The emission indexes for unburned hydrocarbons were at
about the detection level, well below the Standards; (4) The fuel atomizer
and vaporizer developed for the gas turbine combustor are adequate for
the Rankine combustor.
18
-------
CONCLUSIONS AND RECOMMENDATIONS
THE GAS TURBINE COMBUSTOR
The following conclusions were drawn concerning the porous-plate
gas turbine combustor from the work described herein:
1. Analytical work indicates that the combustor equivalence ratio
and the porous-plate burner area must be variable in order for
the engine to operate between the steady-state level-road load
fuel requirements and the wide-open-throttle acceleration fuel
requirements.
2. The operational temperature of the downstream side of the porous
plate must be limited to a specific value at the higher power
levels to prevent flashback through the porous plate and pre-
ignition of the fuel air mixture. Air-cooling of the porous
plate matrix, in addition to radiation cooling, is required to
avert this condition.
3. Models of air-cooled porous-plate combustors have successfully
been made from sintered nichrome and tested.
4. Emission measurements with propane-air mixtures at inlet tem-
peratures of approximately 400°F and one atmosphere were favorable.
The NO emission index was below the 1976 Federal Standards and
the CO emission index was below the Standards at combustor con-
ditions corresponding to light engine loads but somewhat above
it at heavier loads. The unburned hydrocarbon values were at
about the detection level which is much below the Standard.
19
-------
5. Tests of a fuel atomizer and vaporizer using an air-assist fuel
nozzle spraying into a swirling flow resulted in all but a few
points having 93% or more of the fuel vaporized.
As a result of these accomplishments, it is recommended that the
full scale air-cooled porous-plate combustor shown in Figure 5 for the
Base Line Engine be designed, fabricated and tested with gasoline in the
General Electric Combustor Test Facility described in Appendix A. A
schedule for the recommended work is shown in Figure 6.
The full size porous-plate combustor concept shown in Figure 5 and
2
developed for the Baseline engine has a total porous plate area of 2.2 ft
and extends only 2 inches higher vertically than the conventional com-
bustor for the engine. In this concept, the total porous surface is
made up of a number of air cooled wedge-shaped flat-plate segments.
Development Plans
The air-cooled porous-plate combustor for the automotive gas turbine
engine has been developed thus far by means of 1/5 scale models. Tests
to date (at lower values of combustor inlet temperature than would be
the case in the engine) have shown NO emission levels considerably be-
X
low the Federal Standard and UHC levels at or below the detection level.
CO emissions are either below or above the Standard depending upon operat-
ing conditions. By adjusting residence time and/or dilution staging,
the CO values can be lowered with the expectation that the NO emissions
will increase somewhat but still be below the Standard. It is now ap-
propriate to test a module of the full-scale porous plate combustor at
combustor inlet conditions commensurate with engine operating conditions,
followed by testing of the full-scale combustor. The recommended work
would be carried out in the following four tasks:
20
-------
• TO
C
re
D.
o
ft
o
C
o>
fa
3
a
ft
n
o
n
n
•v
-------
Master Schedule Sheet
BASE LINE ENGINE FULL SIZE POROUS COMBUSTOR DEMONSTRATION
fMn, Months
Design
Analysis
Configuration
Shop Drawings
Fabrication
Sintering Trials
Sinter Wedges
Sheet Metal Parts
Assembly
Installation
Design Adapters
Fabricate Adapters
Prepare Facility
Check Out
Propane
Gasoline
Test
Plan
Map Burner
Demonstrate
Evaluate Data
Reduce and Plot
Compare to Prediction
Summary Report
1
—
2
3
4
5
—
6
Figure 6. Schedule for Gas-Turbine Porous-Plate Combustor
-------
Task I - Design and Fabrication of the Full Scale Combustor - Making use
of the design analysis and the test results on the air-cooled porous
plate combustor described above, a full scale combustor will be designed
to the operating conditions of the Base Line Engine. More than one com-
bustor will be fabricated having cooling tubes embedded in the sintered
nichrome porous surface. Modules of the full-scale porous plate com-
bustor will also be designed for the purpose of confirming emission
measurements and module integrity under cycle conditions. In addition,
all adapters will be made for installation in the General Electric GDF-I
Combustor Facility described in the Appendix. In the fabrication of the
combustor, care will be taken to properly control particle size distribu-
tion, sintering temperature, braze proportions and mixing and dimensions.
Completion of the corabustor will include the integrated fuel-air mixture
supply, Joining of the porous plates to the remainder of the corabustor
structure and insulation upstream of the porous plate to prevent pre-
ignition of the fuel-air mixture.
Task II - Installation and Checkout - After assembly of the full scale
combustor it will be installed in the ESP GDF-I corabustor test facility.
Adapters and complete instrumentation will be provided to the faciiicy.
Provisions for measurement will include inlet flow rates, temperatures
and pressures, outlet temperatures and emissions. The emissions will be
measured using the Scott Dilute Exhaust Gas Emission Analyzer which is
capable of accurately measuring, on a continuous basis, NO, NO, CO, CO
and unburned hydrocarbons. The facility will provide 2.5 Ibm/sec of air
flow in two streams, one being capable of being heated and controlled to
temperatures up to 1400°F and the other to 500°F. The facility provides
air at the test section at pressures up to 4 atmospheres.
23
-------
Following installation of the combustor and calibration of the in-
strumentation, exploratory tests of the combustor will be made to assure
that all required conditions of the test can be met and that all facility
controls are functioning properly. The necessary adjustments to the
facility will be made that are required for the test program.
Task III - Test and Evaluation - The operating performance of the com-
bustor, including pressure drop, exit temperature, and emissions will be
mapped over the appropriate ranges of burned gas velocity (V__)t equivalence
ratio, temperature and pressure level. Testing of full-scale modules of
the combustor will precede testing of the full-scale combustor. After
this, the test points delineated in EPA Test procedure for Low NO Com-
A
bustor Final Evaluation will be established and full measurements in-
cluding emissions will be made. These data will be obtained at the air-
flow and fuel-flow rates, inlet temperature and pressure values specified
in the EPA Test Procedure for Low NO Combustor Final Evaluation. The
X
appropriate equivalence ratio established for each test point will also
be set.
The reduced test data will be evaluated in terms of the Base Line
Engine requirements. The exhaust emissions data will be interpreted in
terms of emission index for the combustor performance map. However, for
the points specified by the EPA Test Procedure for Low NO Combustor
X
Final Evaluation, the data will be presented in terms of grams of pollutant
per mile, using the mutually agreed upon schedule of fuel economy.
Task IV - Summary Report - The analysis, design and combustor test data
will be documented at the end of the program along with appropriate draw-
ings and specifications.
24
-------
THE RANKINE CYCLE COMBUSTOR
The following conclusions were drawn concerning the porous-plate
Rankine-cycle combustor from the work described herein:
1. The Arrhenius plots of superficial velocity versus reciprocal
absolute flame temperature at constant equivalence ratio which
are used to design porous-plate combustors were confirmed for
five fuels including gasoline.
2. Both the emissions indexes for NO and CO increased with super-
ficial velocity at constant equivalence ratios for all fuels
tested and for liquid fuels were only below the Standard at
conditions corresponding to moderate engine loads.
3. The unburned hydrocarbon emission indexes were very low, about
at the detection level.
4. The fuel atomizer and vaporizer developed for the gas turbine
combustor are adequate for the Rankine combustor. All but a
few of the test points indicated 93% or more fuel vaporization.
The tin-coated copper porous-plate burners tested on this program
were available from previous General Electric research programs. It
was not necessary to develop the sintering techniques because they are
established and available for use. The liquid-cooled porous-plate burner
operates at temperatures below 700°F and as a result the thermal stresses
are moderate. The Rankine cycle porous-plate combustor operates at at-
mospheric pressure which limits the heat flux conducted back to the porous
plate from the flame, limiting the heat load.
Based on the above accomplishments, it is recommended that the full-
scale water-cooled porous-plate combustor shown in Figure 7 for the
25
-------
Cxluuit
0»t»» Out
Air rroa bp*Dd«:
Driven Co>pr«»o>
to
o\
-."..'X '.>, T
',. • , "••/ > • •-,;' ' '•"' \, ":l
Cbabujtloo
Volua*
ru«l/Alr Htalfold
Figure 7. Conceptual Design of Porous-Plate Burner-Vapor
Generator for Rankine System
A-A
-------
Rankine-cycle engine be designed, fabricated, and tested with gasoline.
The schedule for the recommended work is shown in Figure 8.
Development Plans
The water-cooled porous-plate combustor for the automotive Rankine-
cycle engine has been developed thus far by means of bench tests on flat
burner models. Tests to date suggest NO emission levels below the
A
Federal Standard at most points and unburned hydrocarbon levels at or
below the detection level. CO emissions may be either below or above
the Standard depending upon operating conditions. By adjusting residence
time and/or dilution staging, the CO values can be lowered with the ex-
pectation that the overall average of NO emissions over the cycle will
X
increase somewhat but still be below the Standard. It is now appropriate
to test a full-scale porous-plate combustor. The recommended work would
be carried out in the following four tasks:
Task I - Design and Fabrication of the Full Scale Combustor - Making use
of the design analysis and the test results on the water-cooled porous-
plate combustor described above, a full-scale combustor will be designed
to the operating conditions of a specified full-size Rankine engine.
The combustor will be fabricated having cooling tubes embedded in the
sintered copper porous surface. In the fabrication of the combustor,
care will be taken to properly control particle size distribution,
sintering temperature, braze proportions and mixing and dimensions.
Completion of the combustor will include the integrated fuel-air mixture
supply including the air preheater, joining of the porous plates to the
remainder of the combustor structure, and a tube bundle.
Task II - Installation and Checkout - After assembly of the full scale
combustor, it will be installed in a test facility which can supply air,
27
-------
Master Schedule Sheet
Event
Design
BASE LINE ENGINE FULL SIZE POROUS COMBUSTOR DEMONSTRATION
Months
Analysis
Configuration
Shop Drawings
Fabrication
Sinter Modules
Sheet Metal Parts
Assembly
Installation
Prepare Facility
Check Out
Propane
Gasoline
Test
Plan
Map Burner
Demonstrate
Evaluate Data
Reduce and Plot
Compare to Prediction
Summary Report
1
—
2
3
4
—
5
u
6
Figure 8. Schedule for Rankine Porous-Plate ComLustor
-------
water, and fuel at the appropriate conditions. Provisions for measure-
ment will include inlet flow rates, temperatures and pressures, outlet
temperatures and emissions. The emissions will be measured using the
Scott Dilute Exhaust Gas Emission Analyzer which is capable of accurately
measuring, on a continuous basis, NO, NO , CO, CO and unburned hydro-
carbons.
Following installation of the combustor and calibration of the in-
strumentation, exploratory tests of the combustor will be made to assure
that all required conditions of the test can be met and that all facility
controls are functioning properly. The necessary adjustments to the
facility will be made that are required for the test program.
Task III - Test Evaluation - The operating performance of the combustor,
including pressure drop, exit temperature, and emissions will be mapped
over the appropriate ranges of burned gas velocity (V_,.), equivalence
ratio, temperature and pressure level. After this, mutually acceptable
test points covering the expected range of operation of a specified
Rankine cycle engine will be established and full measurements including
emissions will be made. The appropriate equivalence ratio established
for each test point will also be set.
The reduced test data will be evaluated in terms of the specified
Rankine cycle engine requirements. The exhaust emissions data will be
interpreted in terras of emission index for the combustor performance map.
Task IV - Summary Report - The analysis, design and combustor test data
will be documented at the end of the program along with appropriate
drawings and specifications.
29
-------
GAS TURBINE COMBUSTOR
DISCUSSION
Conventional combustors for the gas turbine engine often have low
emissions of CO and unturned hydrocarbons at high power levels; but the
NO emissions are often high. In conventional gas turbine combustors,
fuel from atomizing nozzles is sprayed into the primary combustion zone,
and combustion occurs as the fuel vaporizes off the dropletSo The local
equivalence ratio of this combustion is uncontrollable, and the NO pro-
X
duction rates are high because of the variable local equivalence ratio.
The equivalence ratio has a string effect on flame temperature. Since NO
X
production rates have a logarithmic temperature dependence, the NO rates
X
at low temperatures can not offset the NO rates at high temperatures.
X
Furthermore, for regenerated cycles, the combustor inlet temperature is
usually the regenerator exit temperature; this high temperature also re-
sults in high NO production rates, since the resulting flame temperature
X
is high.
With conventional combustors, efforts to reduce the NO emissions
X
often result in increased emissions of CO and unburned hydrocarbons. At
the same equivalence ratio and inlet air temperature, the flame temperature
with the porous-plate combustor can be independently controlled and main-
tained below the adiabatic flame temperature. This independent control
of flame temperature is not possible with conventional combustors.
With a porous plate combustor, the lean uniform flame eliminates
unburned hydrocarbons; the controllable reduction in flame temperature
31
-------
below the adiabatic flame temperature results in reduced emissions of
NO and allows design freedom in controlling the CO. The flame temperature
X
can be controlled so that the NO production rates are very low, while
the CO decay rates are high enough with the available residence time to
reduce the CO in the exhaust gas.
There are several design problems with the porous-plate combustor
for the gas turbine application. At pressure levels of four atmospheres,
the heat flux back to the burner which must be removed by the heat sink
can be large, depending on the operating point. Possible cooling methods
to remove this heat flux include independent air cooling with tubes em-
bedded in the matrix and also radiation to a sink. The cooling can be
done with cold-side regenerator discharge air, which is cooler than the
flame temperature. A value of equivalence ratio approaching unity without
violating the temperature limitations of the burner materials is desirable
from the standpoint of minimum flow bypass around the regenerator. As the
equivalence ratio increases towards unity, the amount of compressor discharge
air required for combustion is favorably reduced, but the hot-side burner
temperatures increase toward material limitations. In addition to ma-
terials limitations at the higher equivalence ratios, the problem of pre-
ignition upstream of the burner also increases. Insulation on the up-
stream side of the burner is employed to prevent pre-ignition. Materials
for both the porous burner matrix and the insulation must be suitable for
high temperature operation. The porous-plate combustor also requires a
fuel vaporizer, since a pre-mixed fuel-air charge is necessary for opera-
tion.
Another problem that must be met for combustors for the automotive
gas turbine is that there is a wide range of fuel flow required to meet
32
-------
all of the transient flow conditions. For the porous plate combustor,
this means a wide range in superficial velocity (V_-) of the fuel-air
mixture, and hence a wide range in heat flux back to the burner which
must be removed by the cooling sink. Because of the fact that the flame
can lift off the porous-plate burner if the superficial velocity becomes
too high, sufficient porous-plate surface must be provided for the com-
plete operating range of the engine.
Illustrated in Figure 9 is the general design concept of a cylindrical
porous combustor for the gas turbine. The pre-mixed fuel-air mixture
provided by a device which atomizes and vaporizes the flow is introduced
to the inside of the combustor. The laminar flat flame burns on the out-
side of the cylinder where it is immediately quenched by the diluent air.
The porous plate made either of metal or ceramic material radiates the
heat removed from the flame to the diluent baffle, which is cooled by
the diluent air. Also shown in Figure 9 are tubes embedded in the matrix
of the porous combustor for air cooling in the event that radiation heat
transfer can not provide enough cooling.
COMBUSTOR LOADING AND EMISSION ANALYSIS
In order to investigate the burner temperatures during operation
and the limits imposed by the allowable range of superficial velocity,
a steady-state combustor performance analysis was done. Estimates were
made of the exhaust emissions of NO and CO. Then, the thermal response
Ji
of the combustor to four engine transients was calculated.
Steady-State Combustor Performance Analysis
The results of a combustor performance analysis in parametric form
are based upon the Test Procedure for Base Line Engine Combustor Final
Evaluation on the Test Rig provided by EPA which specified the combustor
33
-------
Figure 9. Low NO Frit Plate Gas Turbine Combustor (Fuel Inside)
-------
test points shown in Table 1. Later, the transient fuel-flow map of the
Base Line Engine, shown in Figure 10 became available. Steady state calcu-
lations for the points shown on Figure 10 were then done.
A computer code for the thermal analysis of radiation-cooled porous
combustors was developed and used to perform parametric analyses. The
assumptions made to calculate temperatures of the porous burner include:
• One dimensional, steady-state heat transfer
• Experimental relationship between superficial velocity, V2_, burned-
gas temperature, and equivalence ratio according to Kaskan
• Lift off at the adiabatic flame velocity
• Configuration: Hollow cylinder
• Engine data from the Base Line Engine.
The following values were assigned for the parametric study of burner
temperatures with the computer code:
Porous Plate Emissivity 0.9
Radiation Receptor Emissivity 0.9
Gas Emissivity 0.0
Minimum Stable Flame 4.0
Superficial Velocity, cm/sec
Burned Gas Specific Heat, Btu/lb-°F 0.38
Porous Material SiC
Porous Plate Thickness, in. 0.125
Porous Plate Effective Conductivity, 13.8
Btu/(hr-ft-°F)
Pore Size, microns 100
Porosity, % 20
Shown in Figures 11 and 12 is the hot-side burner temperature as a function
of burner area for equivalence ratios (R) of 0.7, 0.8, and 0.9 with primary
35
-------
Table 1. BASELINE ENGINE COMBUSTOR TEST POINTS
BE.l Simulated Federal Driving Cycle
Fuel
Flow
Wf
(pph)
Heat
Exchanger Turbine
Comb us tor Exit Inlet
Pressure Temperature Temperature
PI Tl T2
(psig) (°F) (°F)
6 27 1380 1450
10 8 900 1100
12 8 1000 1250
13 8 1100 1370
16 10 1380 1680
65 16 1100 2100
BE. 2 Steady Speed Mode
Wf
(pph)
12
PI Tl T2
(pslg) (°F) (°F)
8 1100 1350
lU 8 1200 1490
16
25
31
39
48
59
69
82
10
16
20
25
29
35
40
47 i
1500
1590
1620
1680
1730
1780
1810
f 1850
Total
Air
Flow
Wa
(Ib/sec)
1.50
0.85
0.85
0.85
0.95
1.15
Wa
(Ib/sec)
0.85
0.85
0.95
1.15
1.30
1.45
1.60
1.80
2.00
2.30
Veh. Speed
(MPH)
30
40
50
60
70
80
90
100
108
119
36
-------
160
140
120
100
5 80
60
40-
20-
85°F Inlet
Preferred Metering
at Upper Limit
Points to
Simulate Wide
Throttle
Points to
Simulate
Steady
State Speed
Steady State
Match Power
Minimum Steady Fuel Flow During Engine
Braking
10
20 30 40 50 60
Gas Generator Speed, %
Figure 10. Fuel Scheduling of Base Line Engine
70
80
90
100
-------
0.7
oo
4000
Equivalence Ratio
0.8
3000
0)
M
0)
§•
3
m
co
I
2000
1000
0
Typical Design Line
Flame Extinguishment
(No. 1 only)
(119 MPH)
Design Point
Typical Design Line
Design Point
(119 MPH)
—Flame Extinguishment
FDC Point
Figure 11. Hot-Side Burner Temperature:
0 1
2
Burner Area, ft
Primary Air from Compressor Discharge
-------
4000
3000
0)
M
3
0)
§•
-------
air to the combustor coming from the compressor discharge. Shown are
lines for the six Federal Driving Cycle (FDC) points as well as the De-
sign Point (119 raph) taken from Table 1. The maximum area above which
flame extinguishment is assumed to occur (V_,. = 4 cm/sec) is shown; the
problem of flame extinguishment occurs only with the first FDC point.
Also shown is the minimum area to avoid lift-off. The velocity, V^ , is
the superficial velocity of the unburned gas based on the porous plate
upstream area, and referred to the density of the fuel-air mixture at
25°C (76°F) and at pressure. A typical design line corresponding to a
fixed area is shown on the figures. The conclusions from Figures 11 and
12 with the primary air being compressor discharge are that:
• Decreasing equivalence ratio from 0.9 to 0.7 generally increases
required burner area and decreases the hot-side surface tempera-
ture level.
• A given design (fixed area) usually cannot be accommodated by a
constant equivalence ratio.
Similarly, the variation of hot-side burner temperature is shown in
Figures 13 and 14 as a function of burner area for equivalence ratios of
0.5, 0.7, 0.8 and 0.9 when the primary air is from the regenerator dis-
charge; an equivalence ratio of less than 0.7 is required to keep the
burner temperatures below 2000°F.
Shown in Figure 15 is a comparison of the case of primary air at
R = 0.7 being the compressor discharge from Figure 11 and of the case of
the primary air at R = 0.5 being the regenerator discharge from Figure 13.
From Figures 13 and 15, it can be concluded that:
• At an equivalence ratio of 0.5 using regenerator air, the hot-
side surface temperatures are generally higher than for an
40
-------
4000
0.5
Equivalence Ratio
0.7
0)
3
4-1
2
0)
I
H
M
I
O
3000
2000
1000
— Typical Design Line
I
(119 MPH) Design
Point
Flame Extingui;
(No. 1 Only)
6
1
Lift Off
I
hment
— Typical Design Line
(119 MPH) Design
Point
FDC Point
6
Plane
Extinguishment
(No. 1 Only)
Lift Off
I
12012
2
Burner Area, Ft
Figure 13. Hot-Side Burner Temperature:Primary Air From Regenerator Discharge
-------
Equivalence Ratio
4000
3000
NJ
s
hi
0)
E
•3
•H
CO
2000
1000
R - 0.8
R - 0.9
Typical Design Line
Design Point
(119 mph)
Flame Extinguishment
(No. 1 only)
"Liftoff
1
*— Typical Design Line
Design Point
(119 mph)
'Liftoff
Burner Area, ft'
Figure 14& Hot-Side Burner Tenperature: Primary Air From Regenerator Discharge
-------
CO
o>
M
4J
03
t-i
-------
equivalence ratio of 0.7 using compressor discharge air.
• Decreasing equivalence ratio generally increases burner area
and decreases surface temperature level.
• A variable equivalence ratio permits limiting surface tempera-
tures, while staying within (V«5) velocity limits.
2
Calculated burner temperatures at one burner area (1.15 ft ) are
shown in Figure 16 for the FDC points and the steady-state vehicle speed
points for the Base Line Engine for an SIC burner operated at an equiva-
lence ratio of 0.8. The peak hot-side burner temperature occurs at 119
2
mph at this burner area. With the burner area fixed at 1.15 ft , the
first FDC point has a V at 4.2 cm/sec, which is near flame extinguish-
ment; the sixth FDC point has a V above the adiabatic lift-off data
/ o \
of Dugger. '
The effect of emissivity on hot-side temperature is shown in Figure
16. If the emissivity is only 0.75 instead of 0.9, then the hot-side
2
burner temperature is about 2544°F at this burner area of 1.15 ft (or
almost 300°F higher than the case of 0.9 emissivity).
Combustor Analysis Using Base Line Engine Fuel Schedule - The code for
steady-state analysis of burner temperatures mentioned previously was
used to perform parametric analyses of 18 engine cycle points. These
points are: a) six steady-state operating points from idle to full
power, b) the six Federal Driving Cycle (FDC) points, and c) six engine
points simulating wide-open-throttle (WOT). Figure 10 is the fuel flow
map for the Chrysler Base Line Engine which shows the six steady-state
operating points selected and the six engine points simulating wide-open
throttle; the six FDC points are shown in Table 1. The corresponding air
44
-------
3000
w.
s
V
kl
01
*
H
2000
1000 -
>
Actual Flame
Temperature (R*v-0.8)
Hot-aide Burner
BLE Conditions
ft
Bum *'
Enlaalvlty G_ • 0.0
Gas
'Burner ' eSink " °'9 '
'Burner " €Sink "
Cold-side Burner
SIC Burner
Particle size, 100 alcrons
Porosity, 20Z
1/8" thick
Effective conductivity, 13.8 BTU/hr ft F
I I
I 1 J 1 I
*
I i
I I
30
50
70
90
110
Simulated FDC Modes
Steady-State Vehicle Speed oph
Figure 16. Effect of Baisslvlty on Burner Temperatures
45
-------
flow rates and temperatures shown for each engine point are used in the
analysis. Compressor discharge air is used as the primary air for com-
bustion, and the diluent air is cold-side regenerator discharge air. The
analysis for each of the three sets of six operating points was done for
emissivities of the burner and radiation receptor of 0.9 and 0.7, and for
equivalence ratios of 0.9, 0.8 and 0.7. At each of the 18 operating points,
the burner area was varied from areas that correspond to flame extinguish-
ment (at superficial velocity, V-,- = 5 cm/sec) to areas that correspond
to the lower of either 50 cm/sec or the lift-off velocity for each condi-
tion (as determined from the combustor inlet temperature and the equivalence
ratio). These calculations are for mixtures of propane-air, since data
for V versus 1/T with gasoline were not available. The specific heat
of the burned gas in the flame was taken as 0.38 Btu/lbm-°F.
Table 2 summarizes the calculations of burned gas temperatures which
are given in Figures 17 to 25 and of hot-side burner temperatures which
are given in Figures 26 to 43. The burned gas temperature is determined
solely by the superficial velocity V._ and the equivalence ratio, R,
according to available data; hence, the emissivity of the burner has no
effect on the burned gas temperatures.
The values of V at the operating points are indicated by the ap-
propriate symbol on each of the figures.
On the figures of hot-side burner temperatures, operating envelopes
corresponding to a hot-side temperature limit of 2300°F and of burner
2 2
areas between 1 ft and 2.5 ft are shown. The 2300°F limit for steady-
state operation on the hot-side is selected based on results from flash-
back tests at Energy Systems Programs and Corporate Research and Develop-
ment which indicated that hot-side burner temperatures in the range of
2
2000°F to 2732°F (1500°C) can cause flashback. Burner areas up to 2.5 ft
are considered reasonable for packaging considerations.
46
-------
Table 2. SUMMARY OF CALCULATIONS
Steady-State
Operating Points
Equivalence Burned Gas
Ratio Emissivity Temperatures
0.9 0
0.8
0.7
0.9 0
0.8
0.7
9 Fig. 17
18
19
7 17
18
19
Hot Side
Burner
Temperatures
Fig. 26
27
28
35
36
37
FDC
Burned Gas
Temperatures
Fig. 20
21
22
20
21
22
Points
Hot Side
Burner
Temperatures
Fig. 29
30
31
38
39
40
WOT
Burned Gas
Temperatures
Fig. 23
24
25
23
24
25
Points
Hot bide
Burner
Temperatures
Fig. 32
33
34
41
42
43
-------
4000
3000
u,
o
3
J-)
(3
Id
01
Q.
01
H
05
a
o
•a
v
M
33
2000
1000
Superficial Velocity
25
cm/
sec
50 6U 70 80
Gas Generator Speed,
9U 100
Radiation Cooled Conbustor
Steady State Points
Emissivity of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.9
Burner Area, ft
Figure 17. Burned Gas Temperature of Radiation Cooled Combustor at
Steady State Points
48
-------
3000
c*
o
§
4J
2
01
a
2000
vo
o>
1000
Superficial Velocity
V , cm/sec
50 toO 70 80
Gas Generator Speed, %
90
Radiation Cooled Combustor
Steady State Points
Emissivity of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.8
01 23 456 78 9
Burner Area, ft2
Figure 18-. Burned Gas Temperature of Radiation Cooled Combustor at Steady State Points
-------
3000
§
4J
a
M
0)
u
•o
0)
§
a)
2000
1000
Superficial Velocity
, cm/sec
50
60 70 80
Gas Generator Speed, %
90
Radiation Cooled Combustor
Steady State Points
Emissivity of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.7
5 6
Burner Area, ft2
8
10
11
Figure 19. Burned Gas Temperature of Radiation Cooled Combustor at Steady State Points
-------
4000
3000
-------
4000
Oi
K)
3000
2
4J
0)
C3
•O
I
aq
2000
1000
i ' T
Superficial Velocity
, cm/sec
2345
FDC Points
Radiation Cooled Combustor
FDC Points
Emissivlty of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.8
0
0.1
X I
1.0
10
20
Figure 21*
Burner Area, ft
Burned Gas Temperature of Radiation Cooled Combustor at FDC Points
-------
Co
4000
3000
0)
M
4J
-------
4000
3000
0)
0)
2000
o
0>
I
1000
I 1 I
Superficial Velocity
, cm/sec
50
100 90,80,70,60
Gas Generator Speed, %
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emlsslvity of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.9
j L
1 I I
0.1
1.0
10
30
Figure 23.
Burner Area, ft
Burned Gas Temperature of Radiation Cooled Combustor at Steady State
Wide Open Throttle Points
-------
4000
I '
3000-
P*
o
2
20001-
Superficial Velocity
, cm/sec
30
25'
25
20
15
10
100 90,80,70,60
Gas Generator Speed, %
Oi
1000-
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emissivlty of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.8
J i I
0.1
1.0
10
Burner Area,
Figure 24, Burned Gas Temperature of Radiation Cooled Combustor at Steady State
Wide Open Throttle Points
30
-------
4000
3000
0)
9-2000
s
o
1000
i I I I r
i i i
1 ' '' I
Superficial Velocity
, cm/sec
50 100 90,80,70,60
Gas Generator Speed, %
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emlssivlty of Burner and Sink, 0.9 and 0.7
Equivalence Ratio, 0.7
I
J I
I I I
l i
0.4
1.0
10
2
Burner Area, ft
100
Figure 25, Burned Gas Temperature of Radiation Cooled rombustor at Steady State
Wide Open Throttle Points
-------
T i r i
3000
Gas Generator
Speed, Z
°0
100
2000
u.
e
0)
a
§
H
01
u
(0
IM
t-t
3
1000
Radiation Cooled Combustor
Steady-State Points
Emissivity of Burner and Sink, 0.9
Equivalence Ratio, 0.9
Superficial Velocity
on/sec
O 5
A 10
D 15
O 20
V 25
e 30
A 35
• 40
• 45
V 50
Lift Off
i i i
0.3 1.0 2 10
Burner Area, ft
Figure 26. Surface Temperature of Radiation-Cooled Cotnbustor
at Steady-State Points
57
-------
3000
Gas Generator
Speed, Z
2000
Ch
e
»
HI
3
-------
3000
I I I I I
1 I
I I I I I
2000
Operating
Limit
01
t-i
3
i-
01
H
0)
u
10
D
t/3
50
60 70 80
1000
Gas Generator
Speed, %
Radiation Cooled Combustor
Steady-State Points
Emissivity of Burner and Sink, 0.9
Equivalence Ratio, 0.7
Superficial Velocity
Vyet cm/sec
O 5
A 10
D is
O 20
V 25
• 30
A 35
• 40
• 45
V 50
* Lift Off
i i i i i i i
0.3 1.0 10
2
Burner Area, ft
Figure 28- Surface Temperature of Radiation Cooled Combustor
at Steady State Points
59
-------
2800
T
l '
2000
1
4J
2
0)
0)
o
1000
1 i '
Radiation Cooled Comluustor
FDC Points
hmissivity of Burner and Sink, 0.9
Equivalence Ratio, 0.9
Operating
Limits \
T
l i
Superficial Velocity
V25, cm/sec
O 5
D15
O20
V25
• 30
A 35
040
• 45
50
Of I
i ill
0.07 0.1 1.0 2 10
Burner Area, ft
Figure 29, Surface Temperature of Radiation Cooled Combuator at FDC Points
20
-------
2500
2000
I
4J
CJ
H
0)
1000
T I I I
Radiation Cooled Combustor
FDC Points
Emlsslvlty of Burner and Sink, 0.9
Equivalence Ratio, 0.8
I T
Operating
Limits
I I | T
Superficial Velocity
V-e, cm/sec
O 5
A10
015
Q20
V25
A 35
• 40
50
( iii
j i
Off
j I
I I I I
0.1
1.0 10
2
Burner Area, ft
Figure 30. Surface Temperature of Radiation Cooled Combustor at FDC Points
20
-------
25UO
2000
ft,
o
§
tsi
H
« 1000
VI
0.1
Radiation Cooled Combustor
FDC Points
Emlssivlty of Burner and Sink, 0.9
Equivalence Ratio, 0.7
I I I
, , , , I
1.0
^Operating
Limits
10
Burner Area, ft
Figure 310 Surface Temperature of Radiation Cooled Combustor at FDC Points
20
-------
4000
Pu
o
I
g 2000
0)
8
0
CO
1000
Superficial Velocity
V, cm/sec
P 5
Aio
Dl5
3000»- 020
V25
• 30
A 35
• 40
• 45
¥50
100 Z Gas Generator Speed
Off
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emissivlty of Burner and Sink, 0.9
Equivalence Ratio, 0.9
I I i
0.1
1.0
Burner Area,
10
40
Figure 32o
Surface Temperature of Radiation Cooled Combustor at Steady State
Wide Open Throttle Points
-------
4000
3000
Q>
M
2 2000
0)
u
4
W
1000
I I I
Superficial Velocity
'25'
cm/sec
O 5
A 10
D is
O 20
V 25
• 30
A 35
• 40
• 45
V 50
•fr Lift Off
Operating
Limits
I I I
1 I
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emissivity of Burner and Sink, 0.9
Equivalence Ratio, 0.8
90 70
I I i
Gas
Generator
Speed, Z
90
0.1
1.0
10
Burner Area, ft'
Figure 33. Surface Temperature of Radiation Cooled Combustor at Steady State
Wide Onen Throttle Points
40
-------
4000
Ul
3000
0)
p
3
V
8-
0)
H
0)
O
09
en
2000
1000
Superficial Velocity
25'
cm/sec
O 5
A 10
D15
O 20
V 25
• 30
A 35
B 40
• 45
V 50
. Lift Off
Operating
Limits-
T
I I
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emissivity of Burner and Sink, 0.9
Equivalence Ratio, 0.7
Gas
Generator
%
0.2
1.0
Burner Area, ft
10
40
Figure 34, Surface Temperature of Radiation Cooled Combustor at Steady State
Wide Onen Throttle Points
-------
I
Gas Generator
Speed, Z
3000
100
2000
V
(J
3
4J
18
hi
O
o.
I
s
-------
i i r
3000 -
2000
[b
O
tO
M
01
01
H
0)
O
d
3
w;
1000
Gas Generator
Speed, Z
Radiation Cooled Combustor
Steady State Points
Emissivlty of Burner and Sink, 0.7
Equivalence Ratio, 0.8
Superficial Velocity
V, en/sec
O 5
A 10
D 15
O 20
V 25
• 30
A 35
• 40
• A5
T 50
-- Lift Off
i i
i i
0.3
1.0
10
Burner Area, ft
Figure 36. Surface Temperature of Radiation Cooled Combustor
at Steady State Points
67
-------
3000
2000
w
cfl
i-
01
o
rt
CO
1000
0
0.4
Operating
Limits 1
i i i
Gas Generator
Speed, Z
Radiation Cooled Combustor
Steady State Points
Emissivity of Burner and Sink,0.7 • *°
Equivalence Ratio, 0.7 • AS
Superficial Velocity
V25, cm/sec
O 5
A 10
D 15
O 20
V 25
• 30
A 35
T 50
•# Lift Off
j I
I i
1.0
10
Burner Area, ft
Figure 37. Surface Temperature of Radiation Cooled Combustor
at Steady State Points
68
-------
3000
Pb
o
8 2000
8
«0
M
1000
I I
I i r
I I
Radiation Cooled Combustor
FDC Points
Emissivity of Burner and Sink, 0.7
Equivalence Ratio, 0.9
. .
Superficial Velocity
V2^, cm/sec
O 5
A 1°
D "
O 20
V "
• 30
A 35
B 40
• «5
V 50
-j. Lift Off
I I i i i I
0.05
o.i
1.0
10
20
Burner Area, ft
Figure 3&. Surface Area of Radiation Cooled Combustor at FDC Points
-------
3000
fe 2000
e
I
4J
cd
o»
CO
1000
0.1
I i i
i I '
T
Operating
/Limits
Radiation Cooled Combustor
FDC Points
Emissivity of Burner and Sink, 0.7
Equivalence Ratio 0.8
Superficial Velocity
cm/ sec
-e
O 5
A 10
D 15
O 20
V 25
• 30
A 35
40
50
Lift Off
1.0
10
Burner Area, ft .
Figure 39. Surface Temperature of Radiation Cooled Combustor at FDC Points
20
-------
2000
4)
2
S 1000
OT
I I
Radiation Cooled Conbustor
FDC Points
Emissivlty of Burner and Sink, 0.7
Equivalence Ratio, 0.7
I
I I I I
Superficial Velocity
V2^, cm/sec
O 5
A10
Dl5
O 20
V25
• 30
A 35
• 40
• 45
T50
4|frLift Off
I i i i l
0.1
1.0
10
Burner Area, ft
Figure 40. Surface Temperature of Radiation Cooled Combustor at FDC Points
30
-------
4000
3000
i
4J
2
0)
2000
8
a
<4-4
14
CO
1000
Superficial Velocity
V25» cm/sec
O 5
A10
_D15
Q20
• 30
A 35
• 40
T50
* Lift Off
i i i i
Gas Generator
Speed, Z
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emlssivlty of Burner and Sink, 0.7
Equivalence Ratio, 0.9
. ...... I
0.1
1.0
10
Burner Area, ft
Figure 41. Surface Temperature of Radiation Cooled Combustor at Steady State
Wide Open Throttle Points
30
-------
4000
3000
s
4J
2
2000
1000
Superficial Velocity
V.-, cm/sec
O 5
A 10
D15
O 20
V 25
• 30
A 35
B 40
e 45
V 50
^ Lift Off
Gas Generator
Speed, %
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emlsslvity of Burner and Sink, 0.7
Equivalence Ratio, 0.8
, , . .1
0.1
1.0 10 40
Burner Area, ft
Figure 42C Surface Temperature of Radiation Cooled Combuetor at Steady State Wide Open Throttle
Points
-------
4000
3000
g
4J
cd
H
2000
8
CO
1000
i I
Superficial Velocity
'25'
cm/sec
O 5
A 10
D 15
O 20
V 25
• 30
A 35
• 40
• 45
V 50
Iff Lift Off
Operating
Limits —i
Gas Generator
Speed, Z
90
Radiation Cooled Combustor
Steady State Wide Open Throttle
Emissivity of Burner and Sink, 0.7
Equivalence Ratio, 0.7
I I
I 1
0.1
1.0
10
Burner Area, ft
Figure 43. Surface Temperature of Radiation Cooled Combustor at Steady State Wide Open
Throttle Points
40
-------
Examination of the operating envelopes shows that all 18 operating points
fall within the envelopes over a range of equivalence ratios from 0.7 to
2
0.9 and over a range of burner areas from 1.0 to 2.5 ft .
For the plots of hot-side burner temperature for the WOT, the hot-
side burner temperature shown in the plots would be reached only if the
burner were to remain in the steady-state at these pressures and fuel
flow rates.
2
An analysis was made at an area of 1.15 ft and an equivalence ratio
of 0.8 along the steady state power matching line of Figure 10. Figure 44
shows the range of V_,. over the range of power from idle to full power,
on the line of steady-state match points. The operating velocity of 38.3
cm/sec at 100% gas generator speed is below the lift-off velocity of 48.3
cm/sec. Shown in Figure 45 are the burned gas temperatures, the hot-
side burner temperatures, and the inside burner temperatures, all at
steady state. The burner has a high-conductivity outer layer of 1/8"
thick silicon-carbide (K about: 10 Btu/hr-ft-°F) and an Inside layer of
1/8" thickness and low thermal conductivity (K about 1.5 Btu/hr-ft-°F).
Figure 46 shows the emission index for NO. emission; the calculation
uses the hot-air model for N02 production rates, as shown earlier in
Figure 3. Prompt N02 is not included in the calculation. A stratified-
plug flow model is used to estimate residence time, with a six-inch
2
residence length and an exit flow area of 34 in . This simplified model
with stratified plug-flow assumes that the velocity profile is uniform,
with the primary air from the flame and the secondary air assumed to be
unmixed while in the combustor. In reality, the flows would be mixed,
with the mixture at a lower temperature; the lower mixture temperature
would have a lower N02 production rate. Figure 47 shows the 00 emission
index, calculated with a 00 decay rate based on the flame temperature
75
-------
u
V
0)
o
•
>>
u
o
30
Equivalence Ratio, 0.8
Burner Area, 1.15 ft2
Emissivity of Burner and Sink, 0.9
o.
20
Pull Power
Idle Power
10
I
_L
50
60
70 80
Gas Generator Speed, Z
90
100
Figure 44, Superficial Velocity Along Steady State Power
Match Line
76
-------
3000
2000
u,
o
0)
^
•s
u
-------
1.40
EPA Standard
(Based on 10 mi/gal.)
0.20
I
1
x
0)
•o
c
M
C
o
Tt
tn
w
•H
W
X
o
0.10
s
Equivalence Ratio, 0.8
Burner Area, 1.15 ft2
Emlssivlty of Burner and Sink, 0.9
Idle Power
Full Power-
I
50
60
70 80
Gas Generator Speed, %
90
100
Figure 46. NOx Emission Index Along Steady State Power
Match Line
78
-------
30
T
T
T
Equivalence Ratio, 0.8
Burner Area, 1.15 ft2
Emlssivlty of Burner and Sink, 0.9
20
Pull Power
Idle Power
C
o
fl
05
CO
EPA Standard (Based on 10 mi/gal.)
10
I
I
I
I
50
60
100
70 80 90
Gaa Generator Speed, %
Figure 47. CO Emission Index Along Steady State Power Match Line
79
-------
and with the stratified plug-flow model for residence time. For both
Figures 46 and 47 the Federal Standard is indicated for the case of an
average fuel consumption of 10 mpg. Although all of the steady-state
NO points are lower than the Standard, the 00 values are higher. By
X
increasing the residence time and/or staging the diluent air the low
NO values can be traded off with the high 00 values. It is characteristic
X
of porous-plate flames that only traces of unburned, hydrocarbon are pro-
duced as long as there is some excess air.
Transient Combustor Performance Analysis
A computer code for the transient thermal response of radiation-
cooled burners was developed and used to analyze three Wide-Open-Throttle
(WOT) accelerations and a cold start transient.
Analytical Methods - The approach was to use a finite difference marching-
time (or explicit) formulation. The burner slab was divided into nodes,
and a heat balance was written on each node to include conduction through
the matrix, the enthalpy fluxes, and the heat from the flame for the case
of the node on the hot-side surface. For the node on the cold-side sur-
face, the heat out was balanced with the temperature increase in the gas
just upstream of the burner by conduction in the gas. The code allows
for 2 layers of porous materials with different properties to be used in
( 9 )
a composite burner. Following the approach of Schneider , the local
gas temperature within the porous material was taken as equal to the local
metal temperature, because of the good heat transfer coefficient and the
high volumetric density of heat transfer area in the matrix. Furthermore,
the initial steady-state temperature profile in the solid was calculated
( 9 )
using the steady-state results of Schneider for a cooled porous plate.
80
-------
The initial condition for each of the transients is assumed to be the
steady state condition when the transient begins; for example, the initial
condition of the WOT acceleration is the gas generator at 50% speed and
at a fuel flow of 10 Ibm/hr.
Looking at the nodal system as shown in Figure 48, the transient
heat going into the hot-side node at an instant of time is:
QIn - 'Air + 'PI - QRAD
The first term on the right-hand side of Eq. 1 is:
CPg (TB(1> ~ TAJ
= WPri Pg
The transient heat rate, Q , is the heat that is required to heat the
Air
primary air flow, W , from its combustor inlet temperature, T ,
L n A _
in
(which is the compressor exit temperature) to the hot-side burner exit
temperature which is taken as identical to the hot-side burner temperature,
T'(l). The prime refers to the temperature of the node at the previous
instant of time, since very small time increments with a typical size of
0.01 seconds are used in the marching time solution. (L, is the specific
heat of the unburned fuel-air mixture.
The second term is:
QF1 = WPri (3)
This transient heat rate, Q , represents the net cooling of the flame
by the sink and is calculated from Kaskan's data as a function of the
superficial gas velocity, V.,., and the equivalence ratio, R. The enthalpy
difference of the fuel-air mixture in the parentheses is the enthalpy h*
at the adiabatic flame temperature less the enthalpy h at the actual flame
81
-------
-'RAD
-
QF1 - WPri
'RAD
iurn
a F
- T
Sink
Figure 48. Nodal System for Transient Thermal Analysis
82
-------
temperature which is a function of V as shown earlier in Figure 2.
The third term is the transient radiation flux from the hot-side
burner temperature, T'(l), to the sink, I , , and is calculated with:
(4)
a is the Stefan-Boltzraann constant, and F is the overall view factor
(including emissivities) from the burner to the sink. A,. is the frontal
Burn
area of the burner.
The conduction flux out of node 1 into node 2 is approximated in the
finite difference approach with Fourier's conduction law as:
Tl(l) - T'(2)
Q, 2 = K_ A_ -5 j=-S (5)
1-z B Burn AX
where K^ is the effective thermal conductivity of the burner, and AX is
the node thickness and also the thickness over which the gradient T'(l)
D
:T'(2) is obtainedo
o
The enthalpy flux of the air flow out of node 1 is:
QW)out ~ WPri Sg (TB(1) - TRef)
while the enthalpy flux of the air flow into node 1 is taken as:
\
Ref )
The temperature, T , is an arbitrary reference temperature for the
enthalpy which drops out of the calculation method.
Then, since the "heat in" less the "heat out" represents the heat
left to increase the temperature of the node, the following heat balance
holds:
83
-------
'in + Qw)m - Qi-2 *
where At is the increment in time. For this half-node, the burner mass
is M., = pn TT^ A,, . The effective density and specific heat of the burner
D B 2. Durn
material are p_ and Cnr>, respectively. The burner temperature of the node,
B rti
Tn(l) is the new temperature of the node at the end of the time increment,
At; T'(l) is the temperature of the node at the beginning of the time
o
increment. Hence, the transient heat balance results in the following
algebraic equation for the node temperature at the current instant of time
in terms of the known temperatures from the previous instant of time:
'-a,
(9)
{lj(l) - 1
WPri CPB At
Vrn ** PB CPB
For the interior nodes, similar heat balances are applied which also
result in algebraic equations for temperatures of the nodes; the heat
fluxes in and out of the interior nodes are calculated with Fourier's
conduction equation.
Input Data - Time-dependent values of fuel flow, total air flow, and tem-
peratures were taken from the supplied Chrysler Base Line Engine data.
For the WOT acceleration from Idle, the gas generator speed is shown in
Figure 49 as a function of time as supplied for the Chrysler Base Line
Engine; this plot in conjunction with the plot of fuel flow as a function
of gas generator speed shown in Figure 50 was used to obtain the time
dependent fuel flow. Similar plots were utilized to obtain the time
dependent values of pressure level, total gas flow, compressor exit tern-
84
-------
45
WOT from
60% Power (Assumed)
CO
o
K
S
g>
ij
O
35
WOT from
40% Power (Assumed)
u)
«
a
30
25
•WOT from Idle
6th Generation Engine
100% GG Speed = 44,610 RPM
20
_J
0.4
_J
0.6
Jj
0.8
I
I
"TTT
0.2
1.0 1.2
Elapsed Time, seconds
1.4
2 0
Figure 49. Time-Dependent Gas Generator Speed During 3 Wide-Open-Throttle
Transients (WOT) at 85°F
85
-------
160
140
120
fi
.0
oo
fa
iH
0)
100 h
* 80 U
Gas Generator
Acceleration Schedule
60% Power
(Assumed)
WOT from
Idle
WOT from
40% Power
(Assumed)
Cold Idle
old Start
Steady State
Match Power
Minimum Steady State. Fuel Flow
During Engine Braking
40 50 60
Percent Gas Generator Speed
100
Figure 50, Fuel Scheduling of Base Line Engine During the Four Transients
-------
perature, and regenerator exit temperature. The gas generator speed during
the cold start transient was also supplied.
For the WOT transients from 40% engine power and 60% engine power,
the transient gas generator speeds were approximated as shown in Figures
49 and 50 since they were not supplied.
For the transient calculations discussed herein, the configuration
of materials and their thermophysical properties are shown in Figure 51.
On the flame side of the burner is a slab of Silicon Carbide (SiC) of
1/8" thickness; it is selected because of its high-temperature capabilities.
On the air inlet side is a slab of low-conductivity material which acts
as an insulator to prevent auto-ignition upstream of the burner; the thick-
ness of this slab was arbitrarily taken as 1/8", but the thickness may
need to be increased to prevent auto-ignition of gasoline which has a
lower ignition temperature than propane.
The schedules of burner area and of equivalence ratio during the
four transients are shown in Figure 520 These arbitrary schedules were
selected to avoid both lift-off and flame extinguishment during the
transients; the steady-state temperatures at both the beginning and the
end of the transient with these schedules are sufficiently within ma-
terials limitations. At the start of the transient, the equivalence
ratio undergoes a step increase to 1.0 from 0.8, and the burner area has
2 2
a simultaneous step increase to 2.20 ft from 1.15 ft . At the end of
the transients, the equivalence ratio undergoes a step decrease to 0.8,
2
as does the burner area to 1.15 ft .
Results - Shown in Figure 53 is the calculated thermal response of the
porous combustor at three locations during the WOT fuel transient from
87
-------
1/8"
1/8" Thick
Thick Low-Conductivity
SiC Material
Flame
Thermal Conductivity 10
Specific Heat 0.3
Density 162.
Cold Air Inlet
1.5 BTU/hr-ft-°F
0.3 BTU/lbm-°F
112. lbm/ft3
Figure 51. Configuration of Materials in Porous Burner
(for Transient Analysis)
88
-------
2.0 _
M
cu
c
Time, sec.
10
Legend
Cold Start
WOT from Idle
WOT from 40% Engine Power
WOT from 60% Engine Power
1.0 _
—
i\
v <
1
^_]
I1
D
F-
L^
l
1
l
l
i
i
1
1
-\
\
'j /—Transients over
I)./ at 1.53 seconds
\\t-
WOT from 60% Engine Power Begins at 0.67 sec
WOT from 40% Engine Power Begins at 0.48 sec
WOT from Idle and Cold Start Begin at 0 0 s«
1
5
Time, sec.
/Cold Start over
at 8.75 sec.
•
ap
1
10
Figure 52. Schedule of Burner Area and Equivalence Ratio During the
<•» Transients
89
-------
T
2500 _
2000
1500-
100
Transient over at 1.53 seconds
WOT From Idle
Hot-Side Surface Temp'erature
Interface Temperature
T
Cold Side Surface Temperature
10
20
40
50
30
Time, seconds
Figure 53. Calculated Thermal Response of the Porous Combustor During WOT From Idle
-------
idle of the Base Line Engine to full gas generator speed. The fuel transient
ends after 1.53 seconds. Figure 54 shows the undiluted burned gas tem-
perature during this transient. From the theory of the porous combustor,
only the superficial gas velocity, V_s, and the equivalence ratio, R,
determine the flame temperature; hence, the burned gas temperature re-
mains constant after the transient is over at 1.53 seconds because both
V and R remain constant.
To examine the time-dependent behavior of the hot-side surface tem-
perature during the WOT, Figure 55 is shown in which the time rate of
change of temperature of the hot-side surface node as well as the hot-side
surface temperature for the first 8 seconds of the WOT transient are
plotted. Similarly, Figure 56 shows the time-dependent behavior of the
three surface fluxes as defined earlier: P^/A^n* QFl/ABurn' QRAD/
A^ . When the burner finally reaches steady state after approximately
80 seconds for this transient, the equality holds that:
"RAD - 'FI (10)
For this transient, the value of Q. /A- is about 10 times as large
as QBAT/A^ for most of the transient. The overshoot in temperature
above the final steady-state value for this transient occurs because the
dT/dt of the surface node is still a large positive number when the hot-
side burner temperature first goes through the value that corresponds to
the final steady-state value. Because of the selected schedule of burner
area, the flux that heats the incoming air, Q . /A_ , undergoes a step
increase at 1.53 seconds, which is a large disturbance in the flux boundary
condition at the surface; the burner is unable to instantaneously conduct
91
-------
4000
3500
3000
Burned Gas Temperature, °F
(Undiluted)
0)
M
4J
rt
M
0)
2500
10
20
30
Time, seconds
Figure 54,. Undiluted Burned Gas Temperature During the
WOT Transient From Idle to 100% Gas Generator
Speed of the Base Line Engine
92
-------
u
00
c
m
x
u
01
V-i
3
0)
D-
E
0>
H
01
4-1
,3
0>
e
500
400 -
300 -
Time Rate of Change of Hot-Side
Surface Temperature During WOT from
Idle to 100% Gas Generator Speed
-100
Flow Transient Over at 1.53 Seconds
1
1 1 1 1 1 1
0)
M
3
o>
a,
-------
3
u
CO
TO
0)
— Flow Transient Over at 1.53 Seconds
QAlr/AB
urn
QRAD/AB
urn
urn
10
Time, seconds
Figure 56, Heat Fluxes Back to the Burner Surface During WOT from
Idle to 100% Gas Generator Speed
-------
away this step increase in heat flux with the current gradient, so the
surface node begins to heat up in order to provide the necessary tempera-
ture gradient to conduct away the heat flux.
A hypothetical transient was run to investigate the calculated
thermal response further. For this case, the initial temperature profile
was selected such that the temperature gradient in the solid was com-
parable to the final steady-state value of the WOT transient; the initial
temperatures were selected at about 200°F lower than the final steady-
state values. The results of this transient calculation are shown in
Figure 57« There is no overshoot of the hot-side surface temperature,
and the time-rate of change of the hot-side surface temperature is an
order of magnitude lower than that of the real WOT shown above in Figure
55 because the initial temperature profile is steep enough to conduct
away the incoming heat flux. The transient temperature is asymptotically
approaching the steady-state value without overshoot because the time
rate of change of the hot-side surface temperature is asymptotically
approaching zero.
The transient burner temperatures and the undiluted burned gas tem-
perature for the WOT transient from 40% engine power are shown in Figures
58 and 59. The corresponding information for the WOT transient from 60%
engine power is given in Figures 60 and 61, while the corresponding in-
formation for the Cold Start transient is given in Figures 62 and 63.
Estimates of NO™ produced by the hot-air mechanism during the 4
transients are shown in Figure 64, as a function of gas generator speed.
Estimates of CO decay are shown in Figure 65, The calculations are quasi-
steady-state; the NO^ calculations are made using the NO- production rate
95
-------
o
-------
3000
2000
1000
WOT from 40% Engine Power
Hot-Side Surface Temperature
Interface Temperature
Cold-Side Surface Temperature
-Transient Begins at 0.48 seconds
-Flow Transient Over at 1.53 seconds
J_
_L
J.
_L
_L
_L
_L
-L
_L
10
20
40
50
30
Time, seconds
Figure 58. Calculated Thermal Response of the Porous Combustor During the WOT Transient from 40% Engine Power
-------
3500
-------
3000
2000
PL,
e
M
0)
vO
1000
WOT from 60% Engine Power
Hot-Side Surface Temperature
Interface Temperature
Cold-Side Surface Temperature
0
-Transient Begins at 0.67 seconds
"Flow Transient over at 1.53 seconds
_L
JL
J_
0
10
20 30
Time, seconds
40
50
Figure 60. Calculated Thermal Response of the Combustor During the WOT Transient from 60% Engine
Power of Base Line Engine
-------
4000
3000
t! 2000
0)
H
1000
Burned Gas Temperature (Undiluted)
WOT from 60% Engine Power
-Transient Begins at 0.67 seconds
•Flow Transient Over at 1.53 seconds
20
60
Time, seconds
Figure 61. Burned Gas Temperature During the WOT from 60% Engine Power
of the Base Line Engine
100
-------
400
300
0)
M
3
0)
a
e
0)
H
200
100
Cold Start
-Flow Transient Over at 8.75 seconds
I
I
I
I
Figure 62.
10 20
Time, seconds
Combustor Response to Cold Start Transient
30
101
-------
4000
3500 -
3000 -
2500
fe.
o
0)
u
% 2000
1-1
0)
a
E
v
H
1500
1000
500
Burned Gas Temperature (Undiluted)
Cold Start
-Flow Transient Over at 8.75 seconds
I I I
10 20
Time, seconds
30
Figure 63. Burned Gas Temperature During Cold Start Transient
of Baseline Engine
102
-------
-J
CJ
O
O
O
O
e
rH
^x
x"
•z.
O
CO
CO
CSI
O
2
CN
O
2.0
1.0
FEDERAL STANDARD
(0.4 gm/mi)
(Based on 10 mi/gal)
FUEL
0.1
0.01
0.001
0.0001
I
BASED ON STEADY-STATE
CALCULATION AT THE
TRANSIENT POINTS
I
I
20
40 60 80
PERCENT GAS GENERATOR SPEED
100
Figure 640 Predicted N02 Emissions Index Based on the Hot Air Mechanism
103
-------
100
u:
e
.a
c
c
c
8
X
w
c
2
10
8
i
FEDERAL STANDARD (3.4 gra/mi)
(Based on 10 mi/gal) -
BASED ON STEADY-STATE
CALCULATION AT THE
TRANSIENT POINTS
2.0
I
I
20
40 60 80
PERCENT GAS GENERATOR SPEED
100
Figure 65. Predicted CO .Emissions Index
104
-------
with the "hot-air" mechanism shown in Figure 3 and are made with flame
temperatures based on the instantaneous values of superficial gas velocity,
V , and of equivalence ratio, R. The stratified plug-flow model was
used to estimate residence times in the combustors. Also shown for com-
parison in Figures 64 and 65 are the Federal Standards.
COMBUSTOR CONCEPT FEASIBILITY DEVELOPMENT
The object of this task was to fabricate a porous plate burner model
(approximately 1/5 size) which could be run over the combustor operating
conditions of the Base Line Engine and to fully test this burner so as
to obtain emissions and operational data.
First, initial screening work on radiation-cooled geometries was
done. However, cracking and preignition occurred a number of times, and
tests to examine this phenomenon were conducted. Because of the occurrence
of preignition, attention was turned to air-cooled combustors. A design
analysis of air-cooled combustors was done, and several were built and
tested.
Initial Screening of Radiation-Cooled Combustors
Screening tests of various porous plate combustor configurations
were carried out with models one-fifth the size of the combustor for the
Base Line Engine application. The combustor configurations tested in-
cluded porous ceramics, porous metals, and burners made up of layers of
screens and insulating materials. These configurations are radiation-
cooled geometries.
The combustor test rig used for these screening tests is shown in
Figure 660 The primary air-fuel mixture enters the test rig through the
pipe at the left. The secondary or diluent air enters through the fitting
105
-------
Figure 66. Gas Turbine Porous-Plate Combustor Test Section
-------
at the top left of the drawing. The cylindrical burner (about two inches
in diameter and six inches long) is surrounded by a conical screen which
is a radiation sink for the burner and which is cooled by the secondary
flow. The hexagonal shape is a view port which permits visual observa-
tion of the burner during testing. Another view port downstream of the
flange is used to detect the presence or absence of flames or plumes
from the burner. Farther downstream is the water injection nozzle which
is used to cool the gas stream ahead of the valve which controls the
pressure level in the system.
Ceramic Combustors - The first screening tests were done with porous
silicon carbide cylinders in a typical configuration shown in Figure 67.
Since it was recognized prior to the present contract that high service
temperatures were required for the porous plate, SiC cylinders were pro-
cured with the largest porosity (15 to 20%) commercially available. In
a combination of bench and rig testing, all 4 of the cylinders tried
were fractured with each cylinder having a different test history. It
was tentatively concluded that the fractures were related to non-uniform
density (porosity). The silicon carbide tests are summarized in Table 3.
Metallic Combustors - A porous stainless steel burner is shown in Figure
68. During the first test with a metal burner, the surface of a bare
porous stainless steel burner became too hot at high heat fluxes and was
damaged. To increase the temperature capability, the burner was wrapped
with a layer of Kanthal wire screen, but preignition was still a problem.
Ignition inside the burner did not occur until the temperature of the
Kanthal screen on the outside of the burner was in excess of 2100°F, as
indicated by an optical pyrometer. The results indicate that a higher
temperature differential between the burner surface and the inside surface
107
-------
o
oo
SIC CYLINDER
WIRE SCREEN
I I I
® -
-------
Table 3. SILICON CARBIDE COMBUSTOR CONFIGURATIONS
Description
Lavite End Caps
High Temp. Gaskets
Spring Loaded
• Bolt Loaded
• No Diluent Screen
• Bolt Loaded
• Flat Steel End Caps
• High Temp. Gaskets
• Attached TC's Inside Burner
• Flat Inconel End Plates
• Bolt Loaded
• Flat Inconel End Caps
• High Temp. Gaskets
Results
• Two hours of operation
• Atmospheric lift-off data
• Shattered next day after 30 rain.
operation
• Burner shattered as soon as
ignited
• 8 cycles in shop vise 1200-1400°F
to room temp.
• 6 cycles in test rig
• Burner shattered after 15 min.
operation
• Temperature gradients observed
during vise tests
• Stress relieved at 1800°C, 1 hour.
• Burner shattered while being fired up
-------
DESCRIPTION
• POROUS STAINLESS STEEL BURNER
RESULTS
• AT PEAK HEAT FLUX AND AT PRESSURE
POROUS PLATE EXCEEDS ALLOWABLE
TEMPERATURE (2000°F)
if.-...'. -
Figure 68. Metallic Combustor Configurations
-------
i;; needed to avo.id internal ignition.
Hence, a layer of ceramic high-temperature porous insulation was
pj.u-eii hiMwtu-ii HH- Knnthal screen and the porous steel burner in an
attempt to provide this temperature differential. A sketch of this con-
figuration is shown in Figure 69. Tests were run with exit gas velocities
up t:o JOO ft/sec. In most cases, the reason for termination of the test
was tlu1. presence of a hot spot on the outer stainless steel cover pipe.
Post-test inspections revealed that the burners had melted opposite the
hot spot.
Several muJ i i-.Layer burner configurations were constructed of porous
riptal with various types of insulation and Kanthal screening. In most
cases, the burner would function satisfactorily for about 10 to 15 minutes
with a uniform red glow and no plume, and then the burner would fail due
to hot spots, and/or holes in the insulation.
A burner shown in Figure 70 was assembled of Kanthal screen, four
layers of Cerapaper and two layers of zirconia cloth insulation and
Kantlial screen. The burner ran successfully for about two hours at 8
psig and V around 10-15 cm/sec. Three days later, the burner was re-
started to measure NO , but the burner did not operate properly and a
hot spot was observed. Disassembly revealed that the burner end cap on
the inlet side had failed.
This burner was rebuilt and tested in the rig. At a typical FDC
condition, the NO was measured at 15 ppm with low diluent flow. After further
testing,the burner deteriorated, and disassembly revealed that the in-
sulaiLor: had shrunk away from the end plates.
The metallic combustor screening tests are summarized in Table 4.
Ill
-------
DESCRIPTION
9 POROUS STAINLESS STEEL BURNER
• TWO LAYERS OF CERAPAPER INSULATION
• ONE LAYER OF KANTHAL SCREEN
RLSULTS
NO PREIGNITION AT OUTSIDE TEMPERATURE
OF 2150°F (OPTICAL PYROMETER)
FLAME NOT EXTINGUISHED AT AS HIGH AS
130 FPS AXIAL VELOCITY
POROUS STAINLESS STEEL
.CYLINDER CO.62 WALL)
H-
oo
e
H
vo
01
•o
•a
(6
a.
(D
l-t
2 LAYERS OF CERAPAPER
CO.62 THK)
KANTHAL
SCREEN
2.OO D!A
6.0 O
-------
KANTIIAL
SCREEN __, ; 4 LAYliRS 2 LAYERS
CERAPAPKK / ZIRCONIA CLOTK
\ttti,
tttttt
1J *• * f/fff f
-* * * Fs* *^-
RESULTS: Ran well for about two hours at 8 psig, and superficial velocity of
15 on/sec. Burner was restarted 3 days later to measure NOX emissions
but hot spot appeared. Burner end cap at inlet had failed.
Figure 70. Kanthal-Wrapped Burner with Zirconia Cloth and Cerapaper Insulation
-------
Table 4. METALLIC COMBUSTOR CONFIGURATIONS
Description
• Porous Stainless Steel Burner
Porous Stainless Steel Burner
One Layer Kanthai Screen Around Burner
Porous Stainless Steel Burner
Two Layers of Kanthai Screen Around Burner
Porous Stainless Steel Burner
Two Layers of Cerofelt Insulation
One Layer of Kanthal Screen
• Porous Stainless Steel Burner
• Two Layers Cerofelt Insulation
• One Quarter Inch High Fences; Spacing:
0.25 and 0.5 Inches
• Porous Stainless Steel Burner
• Two Layers of Cerofelt Insulation
• Two Layers Ranthai Screen 1/8 Inch Apart
Results
At peak heat flux and at pressure
porous plate exceeds allowable
temperature (2000°F)
Porous plate runs cooler
Preignition at high heat flux
Kanthal raises allowable temperature (2400*F)
Operation unchanged
• No preignition at outside temperature
of 2150"F (Optical Pyrometer)
• Flame not extinguished at 130 fps axial
velocity
• Fences shield flame from axial velocity
• Cerofelt fails due to required pressure
drop
Flame not extinguished at 200 fps axial
velocity
Inner Kanthal screen operating temperature
raised
-------
Table 4. METALLIC COMBUSTOR CONFIGURATIONS
(Cont'd.)
Description
• Porous Inconel Burner
• Two Layers Cerapaper
• One Layer Kan thai Screen
• Porous Inconel Burner
• Two Layers Cerapaper
• One Layer Kanthal Screen
• Better End Cap Design
• Porous Inconel Burner
• Two Layers Irish Refrasil
• One Layer Kanthal Screen
• Kanthal Screen
• Three Layers Cerapaper
• Kanthal Screen
• Same as above
• Kanthal Screen
• Four Layers Cerapaper
• Two Layers Zirconia Cloth
• Kanthal Screen
• Same as above
Results
Ran well for 10-15 min., then "clink"
sound, large plume
Downstream end of burner failed
Ran well for 30 min.
Burner operation changed, plume appeared
Inconel was intact, Kanthal and stainless
steel diluent screen had few holes
Developed hot spot after 10-15 min.
Insulation had split and there were large
holes in Inconel and Kanthal screen, small
hole in diluent screen
Ran well for 10-15 min., then chugging sound
and plume
Outside of burner like new
Inside Kanthal screen melted
Five or six small holes in insulation
Ran veil for 15 min.
Developed hot spot
Inner Kanthal screen melted
Multiple holes in Insulation
Several holes in diluent screen
Ran well for 2 hours
Burner was restarted after weekend but hot
spot appeared
Burner end cap at inlet had failed
Ran well for about 30 min.
Measured NO « 15 ppra at typical FDC condition
with low diluent flow. Then observed high NO
and plume.
Insulation had moved away from end cap
-------
Emission Measurements on Burner 102 - Burner S/N 102, which consisted
of a porous inconel matrix with ZrO cloth insulation on the inside di-
ameter, was tested on the bench and in the test rig. Tests on the bench
were run at equivalence ratios of 0.7 and 0.8 over the full range of un-
burned gas velocities. A few test points were obtained at an equivalence
ratio of 0.9. The burner was then installed in the test rig. Emission
measurements (CO and NO ) were made at atmospheric pressure for an equiva-
X
lence ratio of 0.7 over the full range of unburned gas velocities. The
Emission Index (El) as calculated from the ppm measurements is shown in
Figure 71. The data is not corrected for water content. Also shown are
estimates of NO production from the "hot-air" mechanism and CO decay
which use a residence time calculated from a stratified plug-flow model.
Composite Ceramic Burner - A ceramic burner of SiC rings over a mullite
cylinder shown in Figure 72 was run on the bench for about thirty minutes
over a range of V . Then the burner became cooler on the downstream
ond and the cooler area gradually increased as the pressure drop through
the burner decreased from 15 to 7 psi. Disassembly revealed that the
SiC rings were intact but the mullite cylinder had multiple cracks.
1'reignition Tests
During the initial screening tests of the cylindrical radiation-
cooled burners described previously, a number of occurrences of pre-
i(-;nition were observed. In these instances, the fuel-air mixture was
ignited upstream of the burner by the hot inner surface of the burner.
In order to determine the operating conditions which lead to preignition
in this type of burner, a special test fixture was designed and built.
Figure 73 is a sketch of the fixture and Figure 74 shows the completed
116
-------
'76 Fed. St'ds. (10 MPG)
CO
CO Measured
10
1.0
0)
£
o
o
o
Predicted CO
Equilibrium for the Flame
Temperature and Residence
Time
o
C-
6C
X
o
-c
c
c
o
0.1
NO
Converted from Measured NO
Residence Times Calculated
with Stratified Plug Flow
Model, and 6 inch Residence
Length
0.01
Predicted NO
0.001
2/20/73 Data
Measured CO
Measured NO
x
I I
D
Figure 71.
10 15 20
Superficial Gas Velocity, V25» cm/sec
Emission Measurements from Burner S/N 102
25
30
117
-------
SiC Rings
Mullite Cylinder
7711 I I I
I I I I—T~ ~\U
-e—0—o — o —e—e—e—e—e— o— o —©—9—e-
TTT—(17
\ \\ T JLJ—I—I—I—I—I—I—I i '' ' i
Figure 72. Configuration With SiC Rings Over Mullite Cylinder
-------
Porous
Incone1
Plate
Rex Radiant
Burner
Water Jacket
Fuel-Air Mixture Inlet
Spring
Figure 73. Schematic of Preignition Test Setup
119
-------
K>
o
-
Figure 74. Preignition Test Rig Hardware
-------
parts. The test burner was a commercial product of Rex Radiant Company
and consisted of a flat, circular 3-inch-diameter disc with an overall
thickness of 0.6 inches. The disc was a composite in which the layer
at the burning surface was composed of 0.1 inch thickness of porous SiC,
backed by raullite which performed the function of a thermal insulation in
order to keep the upstream surface cool while the burning surface was
running at high temperature. A porous Inconel plate was installed at
the upstream face of the ceramic burner and four thermocouples were at-
tached to the plate in order to measure the upstream surface temperature.
These thermocouples were used as a measure of impending preignition.
The fuel-air mixture was fed into the plenum upstream of the burner as
shown. After passing through the porous burner, the mixture was burned
near the downstream surface and the combustion products flowed out through
a water-cooled passage to the exhaust. A water cooled valve in the ex-
haust was used to control the burner operating pressure. Table 5 sum-
marizes the tests performed in the preignition rig.
Tests were completed on the bench at one atmosphere pressure and
an equivalence ratio of 0.9 over a range of unburned (superficial) gas
velocities (V?c.) from 15 cm/sec to 40 cm/sec. In these tests, the up-
stream surface temperature remained essentially at the temperature of the
incoming fuel/air mixture and no preignition occurred. The test section
was then installed in the pressurized rig and the same tests were con-
ducted at one and at two atmospheres pressure. No preignition was ob-
served.
The system pressure was next increased to three atmospheres with R = 0.9
and V«_ ^ 10 cm/sec. The measured upstream surface temperature of the
25 "\>
121
-------
Table 5. PREIGNITION TESTS
N>
Test
FB-1
FB-2
FBr-3
FB-4
Section R
Rex Radiant burner 0.9
0.9
0.9
Rex Radiant burner 0.9
backed with two
layers of ZrO?
cloth
Rex Radiant burner 0.9
backed with two
layers of ZrO^
cloth plus two
layers fiber-
chrome
Rex Radiant burner 0.7
Inlet
Temp.
°F Press, Atmos
55 1
55 2
46 3
60 3
45 3
65 1
65 2
65 3
65 4
V , cm/sec
15 - 40
15 - 40
40-32
40 - 32
40 - 32
15 - 40
15 - 40
15 - 40
40 - >15
Result
No preignition
No preignition
Preignition
Preignition
Preignition
No preignition
No preignition
No preignition
Preignition
-------
burner began to rise rapidly and when the temperature reached about 1400°F,
preignition occurred and the fuel was immediately shut off. Another
attempt was made to establish this point with the same result. It was
then decided to set three atmospheres pressure and 40 cm/sec. At this
high velocity, the flame was near liftoff and little heat was transferred
from the flame to the burner. The objective was to reduce the unburned
gas velocity (and increase the heat flux to the burner) in steps until
the preignition condition was established. This plan was successful and
a preignition point was established. Figure 75 is a plot showing the
temperature recorder behavior as the preignition condition was established.
In order to determine the effect of insulation on the preignition condi-
tion, two layers of ZrO_ cloth were installed between the upstream face
of the burner and the inconel plate. This had no effect on the pre-
ignition results. An additional two layers of 1/U in. thick Fiber-
chrome insulation were installed with no effect on the preignition re-
sults. Examination of the ceramic burner used in the above tests re-
vealed that it had hairline cracks extending from the outer edge along
generally radial lines.
Based on the above results it was concluded that operation at four
atmospheres pressure would be possible only if the equivalence ratio
were reduced. A new SiC/Mullite burner was installed and tests were
conducted at R = 0.7. No preignition occurred at a pressure of three
atmospheres over the full velocity range. At a pressure of four atmos-
pheres, the upstream face temperature increased to about 950°F and
stabilized for V ^ 15 cm/sec. Figure 76 is a recorder chart trace
showing this condition. At this condition, very slight changes in fuel-
123
-------
1600
1400 f-
Data Taken March 2, 1973
Pressure,? = 3 atm.
Equivalence Ratio.R % 0.9
Fuel-air Inlet Temperature, 46°F
1200
1000
800
01
Id
3
i-l
n)
U
0)
a
I
600
400
200
Preignition
1550°F
u
Ol
U
01
n
0»
•
00
U
01
o>
«M
co
B
to
CM
Q)
CO
(U
(A
41
0}
_L
12
Time, Minutes
16
20
Figure 750 Upstream Face Temperature Leading to Preignition at
3 Atmospheres
124
-------
iOOO
800
Pu
e
0)
h
3
4J
a
\->
a»
600
400
0
0)
a
200
0)
a
o
s
r
IT)
u
o>
00
u
m
01
CO
Data Taken March 3, 1973
Pressure, P » 4 atm
Equivalence Ratio, R % 0.7
Fuel Air Inlet Temperature, 65°F
_L
1
J_
12 16 20
Time, minutes
Figure 76. Impending Preignition at P
24 28
4 Atmospheres
32
36
-------
air ratio caused large effects on the upstream face temperature. Post-
test examination of this second burner revealed the same type of hairline
cracks as those observed in the first burner.
The temperature of the downstream surface (combustion side) of the
burner was not measured during the preignition tests. However, using the
theoretical heat flux from the flame to the burner, a surface temperature
required to radiate this heat away can be calculated. For a pressure
of 3 atmospheres and an equivalence ratio of 0.9, the calculated surface
temperature is 2120°F. For a pressure of 4 atmospheres and an equivalence
ratio of 0.7, the calculated temperature was about 2050°F. The emissivity
was taken as 0.9. Within the accuracy of the measurements and the un-
certainties in the calculations, these two temperatures can be considered
equal. A tentative conclusion is that at a certain hot-side surface tem-
perature, the flame propagates against the flow of fuel-air mixture in
the porous plate and eventually reaches the upstream surface igniting the
fuel air mixture prematurely. The phenomenon labeled "flashback" is
postulated to occur at a fixed hot-side surface temperature which probably
depends on such things as porosity and pore size of the porous material.
Air-Cooled Combustor
Design Analysis - In order to solve the problem of preignition or flash-
back experienced in radiation-cooled burners, some preliminary calcula-
tions were performed to determine the effect of air cooling on the tem-
perature of a metallic burner. The configuration examined had cooling
tubes embedded in the porous plate matrix, using air as the cooling medium.
With the temperature limitations of a metallic burner, neither radiation
cooling alone nor air cooling alone could reject the heat load from the
126
-------
flame at all operating conditions. The calculations were made to determine
if the combination of radiation and air cooling could reject the heat.
Some available combustor loading calculations indicated that the maximum
steady state heat flux from the flame to the porous plate was on the
order of 70,000 Btu/hr-ft2.
Figure 77 is a plot of the radiation heat flux between two long con-
centric cylinders. The inner cylinder (surface 1) represents the burner
surface radiating to a heat sink (surface 2). At a burner surface tem-
perature of 2000°F, the radiant heat flux ranges from about 28,000 Btu/
2 2
hr-ft to 37,000 Btu/hr-ft , depending upon the sink temperature. Thus
for an air cooled combustor, the difference between the heat flux of
2
70,000 Btu/hr-ft and the radiated flux must be removed by cooling, or
the burner temperature will exceed 2000°F. An analysis was performed
to determine the temperature distribution in a burner with embedded cool-
ing tubes. It was assumed that the heat flux from the flame to the
burner surface is uniform and that there is no temperature variation
through the thickness of the burner. The resulting nonlinear differential
equation describing the burner temperature distribution was solved
numerically for a number of possible operating conditions. For each of
these cases, the solution yielded the axial temperature distribution in
the burner. The maximum calculated temperature for each case was used
as a measure of cooling effectiveness. The cooling air fractional pres-
sure drop (AP/P) was held constant at 5%. Figure 78 shows the calculated
maximum wall temperature as a function of heat flux from the flame to the
burner with burner operating pressure as the parameter. The cooling air
inlet temperature was taken as 400°F which is about the maximum compressor
discharge temperature. Figure 79 is the corresponding result for a cool-
127
-------
10'
.. 3
_l
33
- 10
c
3
-a
OS
Burner Surface
at T
1400
1000 1500 2000 2500
Burner Temp., T., F
Figure 77. Radiant Heat Flux Fran A Burner
3000
128
-------
3000
2000
0)
n
V
I
NJ
VO
1000
Di = 0.1 inch, Tube Inside Diameter
TA - 400°F, Cooling Air Initial Temperature
AP/P = 0.05, Fractional Pressure Drop
TCTMV = 1400°F, Sink Temperature
a INK
L = 6 inches, Length
10,000 20,000
30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000
Heat Flux to Combustor, Btu/hr-ft^
Figure 78. Calculated Maximum Wall Temperatures for Air-Cooled Burner for Cooling Air Initial
Temperature, 400°F
-------
3000
u, 2000
o
2
I
8
1000
D, =0.1 inch,Tube Inside Diameter
T. = 1400°F, Cooling Air Initial Temperature
AP/P - 0.05, Fractional Pressure Drop
TSINK = 1400°F, Sink Temperature
L = 6 inches, Length
10,000 20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,0
Heat Flux to Combustor, Btu/hr-ft'
Figure 79. Calculated Maximum Wall Temperatures for Air-Cooled Burner for Cooling Air Initial
Temperature, 1400°F
-------
ing air inlet temperature of 1400°F which is about the maximum regenerator
discharge temperature. In order for these plots to have meaning, the
heat flux from the flame to the combustor porous surface must be known.
Figure 80 is an estimate of that flux as a function of superficial gas
velocity for several burner pressures and an equivalence ratio of 0.9.
2
Taking the peak heat flux for 4 atmospheres, namely 108,000 Btu/hr-ft ,
results in maximum combustor surface temperatures of 1520°F and 2200°F
for cooling air temperatures of 400°F and 1400°F, respectively. Based
on these encouraging results, the decision was made to fabricate an air-
cooled burner made up of sintered nichrome powder with 1/8 inch OD cool-
ing tubes embedded in the matrix. Design of this burner was completed
and fabrication development work was done. That development work is
described in the Section on Porous Plate Fabrication Development.
Mechanical Design - Figure 81 shows the design of the first air-cooled
burner which was designated as 106. Table 6 is a Parts List for this burner.
The basic burner is a porous cylinder with a 2.25 inch outside diameter
(O.D.) and 1.75 inch inside diameter (I.D.). The porous matrix is made
up of nichrome powders bonded together by brazing as described in the
Section on Porous Plate Fabrication Development. A total of 24, 1/8 inch
O.D. cooling tubes are embedded in the porous matrix and run parallel to
the burner centerline. Insulation held in position by a screen is installed
on the inner cylindrical surface of the burner. The end plate is insulated
with a disc of insulation. The fuel-air mixture enters the burner through
the large tube at the center, passes through the screen, the insulation,
and the porous nichrome and is ignited on the outer cylindrical surface.
Cooling air discharges from four inlet tubes into a cooling air manifold.
From this manifold, the cooling air enters the cooling tubes where it is
131
-------
110,000
100,000
90,000
80,000
I
u
^ 70,000
3
_i
13
« 60,000
50,000
40,000
30,000
20,000
10,000
Nominal Limit for
Flame Extinguishment
Equivalence Ratio, 0.9
•Lift Off Limit
with Inlet Air
at 76°F
Figure 80.
20 40
Superficial Gas Velocity, V^t cm/sec
Heat Flux Back to the Porous Burner as a
Function of Pressure and of Superficial Gas
Velocity (V25)
60
-------
..x MOD- SWA&ELOK TO
~>) BUTT IME.LD ELBOW
CAT MO. 400-Z-ZVvJ 3l(.
= 1/1
BOTH ENDS,
Figure 81. Design of Air-Cooled Porous Burner 106
-------
Table 6. PARTS LIST OF AIR-CQOLED POROUS
BURNER SHOWN IN.FIGURE 81
Item 1. Assembly
2. Fuel-Air Inlet Tube
3. Cooling Air Fitting
4. End Plate
5. Side Wall of Cooling Air Manifold
6. Upstream End Plate
7. Insulation Retainer Screen
8. Zirconium Oxide Insulation
9. Porous Burner
10. Insulation Retainer Plate
11. Zirconium Oxide Insulation
12. Downstream End Plate
13. Flow Distribution Screen
134
-------
heated by the heat transferred from the flame back to the burner. The
heated cooling air discharges from the cooling tubes at the end of the
burner and mixes with the combustion products. Figure 82 is a photograph
of burner 106 during fabrication.
Table 7 is a list of the air-cooled burners which were tested on
this program. Burner 107 was a modification of 106 in which the porous
matrix was made thicker and the tubes were embedded in the center of the
matrix instead of at the inner surface. Burner 107 has an outside diameter
of 2.5 inch and an inside diameter of 1.75 inch and contains 27 cooling
tubes. All of the air-cooled burners had a burning surface length of
6 inches.
Air-Cooled Burner Tests
Tests of air-cooled burners were made to map their operational range,
determine their emission characteristics and establish their durability.
Test Procedures and Facilities - The general procedure for testing model
burnerr, was to perform a bench checkout test followed by tests in the
pressurized rig. The purpose of the bench test was to observe the flame
uniformity and to check for defects such as leaks. During the bench
tests, the pressure drop across the porous matrix was measured as a
function of superficial velocity (V?c)«
Figure 83 is a schematic diagram of the gas turbine burner test
rig. The burner being tested is installed in the pressure cylinder as
shown. Separate streams of primary, secondary and cooling air are sup-
plied from a plant air system at 100 psig pressure. The fuel used in
the burner development tests was propane. The primary air passed through
a rotameter and then througn an electrical heater where it could be
heated to temperatures simulating the compressor discharge temperature
135
-------
Figure 82. Air-Cooled Burner #106 During Fabrication
136
-------
Table 7. AIR-COOLED BURNERS (CYLINDRICAL)
Burner
Description
Tests
106
107
108
Brazed nichrome, 2.25" O.D. x 1.75" I.D.
24 cooling tubes
Brazed nichrome, 2.50 O.D. x 1.75" I.D.
27 cooling tubes
Same as 107
Bench tests
Pressure rig tests
Bench tests
Pressure rig tests
Bench tests only
109
108M
Same as 107 except with slotted surface.
Burner 108 modified with insulation
and screen on outer surface.
Bench tests
Limited pressure rig
tests.
Bench tests
Pressure rig tests
with heated combustion
and cooling air.
-------
Air Supply
at 100 psig
00
Electrical
Heater
Burner Cooling
Air
Flowmeter
Control
Valve
Electrical
Heater
Primary
Air
i
Sampling Port
Fuel Air
Mixture
:~ Combustor j
Fuel
Tank
Exhaus t
to Atmosphere
Diluent
Secondary
Mr
Valve to
Control
Combustor
Back-pressure
Secondary
Air
Figure 83. Schematic of Gas Turbine Combustor Test Facility
-------
(about 400°F) of the gas turbine. Fuel flow was measured with a rota-
meter and was then injected into the primary air stream. The fuel-air
mixture then entered the burner where it was burned. Surrounding the
burner was a conical screen which served as a heat sink for radiation
from the burner surface. The heat sink screen was cooled by the secondary
air which flowed radially inward through the screen and mixed with the
combustion products. Cooling air passed through a rotameter and an
electrical heater, where it could be heated to about 400°F, and into the
cooling air manifold in the burner. After passing through the cooling
tubes, the cooling air mixed with the combustion products and the diluent
secondary air. The combined burner exhaust then flowed out through a
water-cooled valve used to control the burner operating pressure level.
Samples of the exhaust gases were taken through a sampling line down-
stream of the burner. Measurements of CO, CCL, NO, NO- and unburned
hydrocarbons were made with a Scott Research Laboratories, Inc., Model
108H Exhaust Gas Analysis System which was purchased by the General
Electric Company. Figure 84 is a photograph of the Scott equipment.
Burner 106 - This was the first of the air-cooled burners to be tested
in the pressurized rig. Before installing the burner in the rig,
a bench checkout was performed. The equivalence ratio was held
constant at 0.8 and the superficial gas velocity (V?(.) was varied from
10 cm/sec to about 45 cm/sec. During this operation, several small hair-
line cracks appeared on the surface. The cooling air flow rate was set
equal to the combustion air flow rate. No flashback was observed.
The burner was then installed in the pressurized rig in order to
measure emissions and to determine the effectiveness of the air cooling.
139
-------
-P-
O
Figure 84. Scott Exhaust Gas Analyzer
-------
Tests were conducted with room-temperature air at burner pressure of
one, two, and three atmospheres before a flashback occurred and the test
was stopped. These tests imposed the most severe conditions, in terms
of heat flux to the burner, of any burner tested to that time.
Figure 85 is a plot of the emission index for the measurements of
CO and NO from burner 106 as a function of superficial velocity. The
solid points are the measurements and the open points are the correspond-
ing prediction; for NO , the calculation uses the "hot-air" mechanism.
X
All of the measured NO values were well below the 1976 Federal Standard.
The CO values exceeded the standard at a V ,. of around 20 cm/sec.
Figure 86 shows the measured N0« emission index for the tests at
pressures up to four atmospheres. No definite trend with pressure is
evident.
Post-test examination of the burner revealed the existance of several
cracks. The appearance suggested that the flashback probably occurred
through one of these cracks. The pressure surge associated with flash-
back split the burner along four axial cracks located approximately 90°
apart. A new air-cooled burner (designated as burner 107) was designed
and built, This burner was 50% thicker (0.375 in. wall) and placed
more of the sintered powder behind the tubes.
Burner 107 - Air-cooled burner 107 was tested on the bench, both for
pressure drop and for uniformity of flame. Figure 87 is a plot of the
measured pressure drop as a function of unburned gas velocity, V .
Measurements were made both with and without burning. The increased
pressure drop during burning is probably due to changes in burner porosity
due to heating and reduced gas density.
141
-------
40
10
cl
_o
— i
o
0
0
£ 1.0
u
1
•—I
C
E
•— i
X*
CJ
TJ
C
i— I
co
o
CO
0)
— ' 01
E "
U
' 0.01
5
1 1 1 |i
? *
8
r '
-
w
.
& (3.4 gm/mi) I
(Based on 10 mi/gal)
i
!>
N02 Fed. Std
(0.4 gm/mi)
A
" 4 ^2 A A "
t t t &
•taMMM
•
*»
H«
Z
=• i i -
A
—
A
Solid - Measured
Open - Predicted with
Stratified Plug - -
Flow Model
i
1 II 1 l
10 15 20 25 30 3!
V_5, cm/sec
Figure 85.
Emissions from Air-Cooled Burner S/N 106 at 1 Atm. Pressure
142
-------
1.4
I
Federal Standard
(0.4 gm/mi)
o
c
CNI
g
O
•O
C
c
o
V.
•ft
0.6
0.4
" °'3
0.2
0.1
D
X
O
I
Burner 106
A R = 0.79 + 0.01, 1 atm
Q R = 0.74 + 0.05, 1.98-2.36 atm
O R • 0.75 + 0.01, 3.0-3.1 atm
X R - 0. 72 + 0.005, 4.09 atm
JO
0
0
-&
10 20 30
Superficial Velocity, V _, on/sec
40
Figure 86. N0« Measurements at 1 to 4 Atmospheres on Air-Cooled Burner
143
-------
0.5
0.4
o. 0.3
o
Q
0)
3
W
M
O
0.2
0.1
Operating Pressure, 1 atm.
A
A
A
A
A
A
A
A
A
Burner 107 Pressure Drop
A With Combustion
Without Combustion
10
20
30
40.
50
60
Superficial Velocity, V?_, cm/sec
Figure 87. Burner 107 Pressure Drop
144
-------
Kigure 88 shows the CO and NCL measurements for burner 107. The
NO values were below the Federal Standard over the full range of super-
ficial gas velocity. The CO values exceeded the standard above about
15 cm/sec. This is probably due to the relatively cool exhaust tempera-
ture associated with unheated inlet air. Subsequent tests with heated
inlet air with burner 108M described below tend to confirm this hypothesis.
Tests with these burners showed that the unburned hydrocarbons (HC)
emissions were essentially zero when the burner was intact. It was soon
learned, in fact, that indications of unburned HC was a definite indica-
tion of a burner defect. Testing with burner 107 was terminated when
there were indications of hydrocarbons. Post-test examination of the
burner revealed longitudinal cracks similar to those observed with burner
106. The cause of the cracks was hypothesized to be due to hoop stresses
associated with the radial temperature gradient through the burner.
Burner 108 - This burner was identical to burner 107. Since 107 cracked
during tests, 108 was checked out on the bench only.
Burner 109 - In order to avoid the problem of cracking, the decision was
made to machine longitudinal slots in the burner extending radially in-
ward from the outer surface of the burner to the top of the tubes (one
slot for each tube). This slotted burner was designated as 109 and was
checked out on the bench. During preliminary tests in the pressurized
rig, the burner cracked and this test was terminated before any emissions
data was obtained.
Burner 108M - Burner 108 was modified into burner 108M by wrapping the
cylindrical surface with a layer of Fiberchrome insulation held in posi-
tion with a Kanthal screen. Tests were conducted in the pressurized rig
first with unheated combustion and cooling air and later with these air
145
-------
10
T
>
o
O
CO Federal Standard
(3.4 go/mi)
CO
c
2 i.o
X
01
•o
CO
c
o
•rl
CO
CO
•H
0.1
L
Predictions with Stratified
Plug Flow Model
*
N0» Federal Standard
j,(0.4 gin/mi)
O
Air-Cooled Burner #107
1 Atm. Pressure
• R = 0.89
A R - 0.82
O R " 0.71
Slash (/) is Prediction
Inlet Fuel Air Mixture
at 70°F
0.01
1
1
1
II 20 30 40 50
Superficial Gas Velocity, V5-» on/sec
Figure M. Mtasurements of CO and NO^ Emissions on Burner 107
146
-------
streams heated to temperatures approximating compressor discharge tem-
perature (400-600°F). Figure 89 shows the CO emissions and Figure 90
shows the NO data. The CO emissions are lower with heated air (open
X
points) than with unheated air (solid points), as would be expected with
the higher exhaust temperatures associated with the heated air; with the
higher exhaust temperatures, the CO decay rates are higher which result
in reduced CO. For the same reason, it is not surprising that the NO
X
values increased with the air heating, since higher exhaust temperatures
also mean higher NO production rates. However, all the data were be-
low the Federal Standards.
FUEL-AIR MIXTURE SUPPLY DEVELOPMENT
Because the porous plate combustor requires a uniform mixture of
fuel and air upstream of the porous plate, it was necessary to design,
fabricate and test a fuel-air mixture supply system.
Design and Fabrication
The technical literature indicated that, within reasonable geometrical
restraints, complete vaporization of the required quantities of gasoline
could be accomplished only if more than one of the following vaporization
mechanisms were used:
1. Preheating the fuel prior to injection into the airstream.
2. Pressure atomization.
3. Air atoraization, including shearing and sonic atomization.
4. Evaporation of droplets by the heat of the primary airstream.
5. Evaporation of droplets centrifuged onto a plate heated by heat
losses from the combustion zone.
All of the above mechanisms could be utilized when combustion is in
progress. The cold starting problem dictated that a sonic air atomizer
147
-------
10
0)
ft,
.1
O
O
O
8
J
8
1.0
s
c
O
•H
Ul
W
•H
I
O
u
0.1
Federal Standard
(3.4 gm/mi)
(Based on 10 mi/gal)0
A A
O
A
O
A
A
O
O
O
D
D
Burner //108M
CO Data at 1 Atmosphere
Primary Inlet, 390 to 550°F
Coolant Inlet, 390 to 600°F
Q R * 0.9
A R ~ 0.8
O R ^ 0.7
Solids are with Unheated Air
10
20
30
40
Superficial Gas Velocity, V__, cm/sec
Figure 89. Measured CO Emissions on Burner 108M with Both
Heated and Unheated Air
148
-------
0)
3
(JL,
o
o
o
§
u
w
-------
be utilized to produce the fuel spray. This type of injector is also
capable of meeting the high turndown ratio required with reasonable air-
flow and air and fuel pressure requirements.
Fuel-Vaporizer Test Assembly - A gas turbine fuel vaporizer test assembly
shown in Figure 91 was fabricated, instrumented, and tested. This
assembly comprises the following elements.
Burner Model - The cylindrical surface is provided with sixteen
0.04 in. inner diameter tube fittings, and the end is provided with one
of the same. These fittings are used to pass the thermocouple wires through,
to control the flow area and thus model the porosity of the porous matrix;
and the tubes attached to the fittings are used for pressure taps. The
cylinder and end outer surfaces are wrapped with electric heating wire
to partially model the operating temperature of the porous matrix. The
end plate contains a centerbody and flow director to assist the fuel-
air mixture in making a 180° turn. Three thermocouples are staggered
within the model of the porous matrix to measure its inside wall tempera-
ture.
Vaporizer Plate - An internal heated cylinder enclosed within a
cone acts as a vaporizer plate. The cylinder is closed on one end ex-
cept for a hole through which the fuel nozzle protrudes and six flow-
directing louvers. When used with the louvers, the air flow area is
variable in order to produce a constant fuel-air ratio. A high tangential
velocity is produced, which effectively creates an airflow path much
longer than the axial length of the cylinder. The vaporization of the
fuel is enhanced by the longer exposure of the fuel droplets to the
sensible heat of the air. Also, the larger droplets are vaporized by the
heated cylinder wall when they contact it either by impaction of the
150
-------
Type of 17
Fittings
\\\\\\\x\x\\x\
\_
Primary Air
Flow Path
Air Actuated
Fuel Nozzle
Upstream Flow
Meter and Pressure
Gauges for Both
Fuel and Primary
Air
Heater Wires
Vaporizer Plate
Air Supply
Tube
\\\\ \
Simulated
Burner Surface
V Thermocouple Locations
Pressure Measurement
Figure 91. Fuel Vaporizer for Gas Turbine Corabustor
-------
spray or by centrifuging. The cone acts as a flow director and as a
radiation barrier between the model of the porous matrix and the vaporizer
cylinder. Two thermocouples measure the vaporizer cylinder wall tempera-
ture.
Fuel Nozzle - Sonicore 156K and 125K atomizer nozzles made by the
Sonic Development Corporation have been tested. The actuating air as
shown in Figure 92 enters the center of the nozzle through a 1/4 in. pipe
ell. The fuel enters a chamber surrounding the air chamber through a
1/8 in. pipe connector. It is injected into the airstream through fixed
orifices of 0.032 in. diameter. The spray Immediately enters a zone
wherein a resonator cup creates soundwaves which break the spray into a
soft mist of droplets which are expected to be about 20 microns or less
under most operating conditions.
Thus, the assembly utilizes the principles of low-pressure atomiza-
tion, air shearing and sonic atomization, evaporation of droplets by the
heat of the airstream, and evaporation of droplets centrifuged onto a
plate heated by combustion zone heat losses. Other methods which can be
incorporated into this assembly include preheating of the fuel, the in-
sertion of a device into the vaporizer cylinder which will increase the
passage length and time of exposure of the fuel to the hot air, and in-
serts around the nozzle tip to create more turbulence or change the
quantity of air passing through the spray.
Summary of Fuel Nozzle Tests - The fuel nozzles were assembled in a rig
which contained an air supply with electric heater, flowmeter and gauge
which read the pressure at entry into the nozzle. The fuel system con-
tained a pressurized tank, flowmeter, and gauge which read fuel pressure
at entry into the nozzle. The spray was exhausted into atmospheric air
and was observed visually.
152
-------
RESONATOR CHAMBER
LIQUID
CENTER OF
IMPLOSION
AIR OR GAS
Figure 92. "Sonlcore" Fuel Vaporizer
153
-------
Since the air orifices were nozzles of fixed throat diameters, the
air flow rates could be set at any given level below choked flow by
varying the air supply pressure. At low fuel supply rates, the presence
of the fuel had no appreciable effect. At fuel supply rates above 50
Ib/hr, the air flow rates decreased.
Since the fuel orifices were of fixed number and diameter, the fuel
flow rates could also be controlled by varying the fuel supply pressure.
Low fuel supply rates could be achieved by the aspiration of the air-
stream, with greater aspiration achieved at a given air flow rate by the
smaller 125K nozzle.
The intent of the nozzle calibrations was to determine whether the
spray from either nozzle was satisfactory and what combinations of air
and fuel flows were best. Since the greatest amount of energy available
for atomization was supplied by the air, smaller droplets are produced
by larger air flow rates at constant fuel flow rates, and by smaller
fuel flow rates at constant air flow rates. A "wet" spray - "dry" spray
boundary was established by varying the fuel flow at constant air flows
and temperatures. The boundary was a straight-line variation with pres-
sure, being about 7 Ib/hr for the 125K nozzle and about 6 Ib/hr for the
156K nozzle, both at 15 psi air pressure differential and essentially
invariant with air temperature.
A criteria which showed more distinction between the two nozzles was
the stable combustion limit. By setting the nozzle air flow and increasing
the fuel flow, a point was reached at which a flame would not blow out
after the ignition source was removed. At 10 psi air pressure differential,
the stable combustion limit was about 34 Ib/hr fuel flow for the 156K nozzle
and about 10 Ib/hr fuel flow for the 125K nozzle. The appearance of the
154
-------
flames also favored the 125K nozzle, since its flames could be made to
appear virtually invisible, indicating good vaporization very close to
the nozzle. The 125K nozzle appeared to have a distinct advantage with
respect to quick vaporization because it produced smaller drops.
The distribution of the 125K nozzle was also superior, since under
most operating conditions the spray angle was larger than that of the
156K nozzle. The larger spray angle and the softer spray assists in
distributing the spray into the high tangential swirl of the combustion
air and in impacting the larger drops onto the vaporizer hot plate.
The 125K nozzle was chosen for further testing.
Summary of Vaporizer Fluid Flow Tests - By varying the setting of the
movable louvers and the air pressure differential across them, the ef-
fective louver flow area and the tangential swirl velocity could be cal-
culated. The effective burner flow area was variable by closing some
of the seventeen Swagelok fittings of the burner model. Pressure was
measured through one of the nozzle supply tubes and the port on the end
of the burner model. The supply air was at ambient temperature and the
conditions at the burner model exit were ambient. A fuel injector was
not installed during the tests.
with a louver air pressure differential of 1 psi (tangential velocity,
270 fps) and the louvers nearly completely closed, a choked air flow of
2
160 Ib/hr was delivered with a model burner flow area of 1 to 2 in .
With the louvers fully open, an air flow of 1080 Ib/hr was delivered
(variable with burner area).
155
-------
Calculations of Degree of Vaporization - The test conditions used for
tests of the fuel-air mixture supply were those specified for the Base
Line Engine in Table 1. Using the data summarized by Graves and Bahr
and the most severe air conditions specified in Table 1 as far as high
flow rate/low air temperature conditions are concerned (the Federal
Driving Cycle point with a fuel flow of 65 Ib/hr), estimates were made
of the degree of vaporization which could be achieved by the sensible
heat of the air alone. The contact time and length of air flow path
given in Figure 93 were based upon the vectorial path length using tan-
gential swirl velocities and axial through-flow velocities.
It was expected that droplet sizes would be produced by the sonic
injector which would be at least as good as those of a simple orifice
at 300 psi fuel pressure differential, and that 300 ft/sec tangential
velocity would be achievable with reasonable air pressure differential.
Therefore, it was expected that 93 percent vaporization could be achieved
through air sensible heat alone. All other test points except the one
with a fuel flow of 65 Ib/hr require less tangential velocity and. could
tolerate coarser sprays to achieve the same degree of vaporization.
^
A thermal analysis has shown that the vaporizer plate would operate
at about 1200°F inside an operating porous burner if it were not shielded
from radiation. With the present shielding, its operating temperature
would be closer to 200°F. Therefore, by proper design of the shielding,
the vaporizer plate could be designed to operate between these extremes.
The heating wires within the vaporizer plate allow the vaporizer plate
temperature to be controlled as a test variable.
A plot of the heat required to heat the required quantities of fuel
from 100°F to 425°F and to vaporize 100% of the fuel can have superimposed
156
-------
100
90
s-t
g 80
H
g
s
W
70
60
50
100
SIMPLE ORIFICE
EQUIVALENT SPRAYS
MOST SEVERE AIR CONDITIONS
16 PSIG
W - 1200 LB/HR
237°F
Wf - 65 LB/HR
100
200
300
400
AIR VELOCITY, FT/SEC
Figure 93. Estimated Performance of Fuel Vaporizer for Gas Turbine Combustor
-------
upon it the heat provided by convection from the plate to the combustion
air at constant equivalence ratio. The heat required rises linearly with
the amount of fuel while the heat provided decreases hyperbolically. The
net result is that 100 percent of the fuel can be vaporized by the plate
up to a fuel flow of 20 Ib/hr (provided it is placed upon the plate)* The
vaporized fraction decreases to 50 percent at a fuel flow of 30 Ib/hr,
11 percent at 65 Ib/hr, 8 percent at 85 Ib/hr, and 2 percent at 160 Ib/hr.
The 11 percent vaporized at 65 Ib/hr by the plate plus the 93 per-
cent vaporized by the sensible heat of the air exceeds 100 percent. Also
it is not expected that as much as 10 percent of the fuel will be pre-
sent on the vaporizer plate.
Fuel Vaporizer Tests - The objective of the tests was to determine the
degree of fuel vaporization for typical combustor operating conditions.
Test Apparatus - The test apparatus shown in Figure 91 is shown in-
stalled in the test facility in Figure 94« The combustion air comes
from the left into the vaporizer which is covered with insulation. The
injector (heated) air and fuel lines are shown at the location where
they enter the test device. Four manifolds can be seen leaving the
vaporizer. Each manifold collects the fuel mixture from four fittings
along an axial line on the simulated porous surface (see Figure 91).
Three of these lines simply discharge the fuel-air mixture. The fourth
discharges to a 15-inch diameter porous-plate, water-cooled combustor.
The line to this burner is water cooled to arrest further vaporization.
Thermocouples to the inside of the simulator/combustor surface, the
cylindrical vaporizer tube, and the lines used to carry pressure signals
to the manometers are also shown.
158
-------
Ul
Vaporiz(
Beneath,
Insulation
Figure 94. Vaporizer and Inlet and Exhaust Lines
-------
Figure 95 is another view of the vaporizer. Shown on the figure
are the heater wires around the simulated porous surface and on the down-
stream end of the test piece. Also shown are the mixture sampling ports
located approximately 1.75 and 5.25 in. from the closed end of the simulated
combustor surface. Samples were taken from the 1.75 in. location (open
in the figure).
Shown in Figure 96 is the apparatus upstream of the vaporizer. Shown
are the flow meters and pressure gauges for the fuel, injector air and
combustion air. Pressure lines to the manometers are shown on the right.
The injector air and combustion air heaters and insulated lines are shown
in the lower part of the figure. The fuel tank and the pressurized
ambient-temperature air line are shown at the lower left.
In order to analyze the fuel-air mixture in the vaporizer,a Scott
Model 108-H Dilute Exhaust Emission Measurement System purchased by the
General Electric Company was used. This console has provisions for
measuring total hydrocarbons, NO, NO., CO, and C0?.
Testing Technique - In order to measure the fraction of the fuel
that was vaporized, two samples were taken through the total hydrocarbons
analyzer, an isokinetic and a non-isokinetic sample. The former determines
the total hydrocarbon content by sampling all of the mixture in a given
streamline; the latter the amount of unvaporized fuel by allowing the
vapor to spill out of the streamline because the sample Is extracted at
lower velocity.
In order to provide low pressures to the Scott analyzer because it
was feared that the existing pressure regulator was not capable of with-
160
-------
' Thermocouples
Vaporizer Beneath Insulation
Cooling Water
Injector \AirJ
Mixture to 15" Diameter Burner
Mixture
: • - ' " •'.'-. •'--
Figure 95. Vaporizer and Inlet and Exhaust Lines
-------
ro
Injector Air Flowmeter and Pressure Gauge
Combustion Air Flowmeter and Pressure Gauge
Flowmeter and
Pressure Gauge
Lines to Manometers
Combustion Air Heaters
Pressurized Anbient-Tenperature
Figure 96. Vaporizer Air and Fuel Supply System
-------
standing the full compressor discharge pressures, the combustor air was
maintained at about 1 psig. It was assumed that the combustion air pres-
sure effect on vaporization was small, and anticipated that this could
be checked later. The injector air and combustion air temperatures and
flow rates and the fuel flow rate were held at the conditions specified
for each engine operating point.
Although the vaporizer outer diameter and closed end were insulated
to prevent heat loss, the vaporizer heaters were not activated. All of
the vaporization tests were run with the primary air louvers fully open,
which gave the minimum amount of tangential swirl at each combustion air
flow setting. The conditions tested, therefore, represented approximately
the most severe expected as far as providing vaporization energy to the
spray and heat from the combustion process were concerned.
Test Points - The points 1-6 for the fuel-air mixture supply simulate
the six Federal Driving Cycle points in Table 1. Test points 7 and 14-18
simulate the steady speed mode:
Test Point Vehicle Speed (MPH)
14 70
15 80
16 90
17 100
18 108
7 119
The operating conditions at steady state speeds of 30-60 MPH almost
duplicated the conditions of test points 3, 4, 5, and 14.
163
-------
Fuel flow rates of both 69 and 84 lb./hr were run under test point 7.
The lower figure represents the most air flow (at R J\» 0.8) which could
be heated to the desired temperature. The higher figure represents the
desired fuel flow rate at the maximum air temperature which could be
achieved. For each of the test points,the injector air and combustion
air temperatures were set approximately equal to the calculated compressor
discharge temperature. For any given engine condition, an equivalence
ratio of about 0.8 was set using the appropriate air and fuel flow rates.
The vaporizer pressure was set at approximately 1 psig and the air pres-
sure differential across the nozzle was varied to give a reasonable range
of calculated mean droplet sizes. The injector air flow rates were those
which resulted from the air pressure differentials.
Test Results - The test data are presented in Table 8 and plotted
in Figure 97. In the figure the vaporization quality in percent is
plotted against a group parameter similar to that of Graves and Bahr
and which relates all of the pertinent terms associated with spray drop-
let: evaporation. The group parameter is:
^0.8 / v-1.2 / vO.42
I a I la
B =
where T combustor primary air inlet temperature, °R
O
U primary air velocity in mixing region, ft/sec
3.
P. pressure of air leaving injector, psia
APf equivalent fuel pressure drop through a 0.030-inch
orifice, psi
The fuel pressure drop APf is the equivalent drop in the Graves and Bahr
apparatus. Most of the data fall between 93 and 99 percent of the fuel
vaporized, although some data were as low as 87 percent. The data seem
164
-------
Table 8. GAS.TURBINEiFUEL VAPORIZER TEST DATA
Test
Point
1
u W
Hr a
lbf/hr
1428 $ 6.043
1450 (7 ^
1500 :75.944
2
3
4
1003 I 10.01
1040 (2) ^
i
1054 A ll. 94
1117 4>12.88
1125 4>
1135 V
5
6
7
14
15
16
1234 dj 15.85
0943 C|364.99
0955 (|j >
lr
1407 &>69.25
1438 £>
1448 (>
1458 £> >
1507 fc> 84
'
1320 /d> 30.71
1334 X$> J
i
1352 OJ 38.73
1403 0) 1
/
1020 4)48.05
1037 faQ 47.75
17
18
1047 VS)
1057 if) '
i
1130 |>58.94
1154 69.25
1315 <3> 68.85
1332 <3>
1340 <5> \
i
P2
psig
1.01
1.00
1.15
1.09
1.01
1.01
1.07
V
1.22
0.84
0.85
1.00
1.15
1.10
1.15
1.16
2.18
2.18
3.68+
3.42
0.56
0.57
0.58
0.56
0.76
1.04
1.01
0.94
1.01
Tl
°F
320
310
286
176
177
175
164
i
I
V
177
237
241
411
390
401
393
376
269
267
300
312
314
323
321
316
323
354
373
367
377
371
Wa
tot
Ib /hr
d
119.4
vl
128.7
189.5
187.7
221.4
241.5
!
V
292.1
1210
1215
1234
1330
1282
1307
1366
518.3
523.2
680.4
695.6
915.2
908.3
932.0
897.5
1024
1269
1221
1189
1219
Wa/wf
19.76
i
21.66
18.93
18.76
18.54
18.75
1
1
18.43
18.62
18.70
17.81
19.20
18.52
18.88
16.26
16.88
17.04
17.98
17.96
19.05
19.02
19.52
18.80
17.38
18.31
17.73
17.27
17.71
R
0.748
^
0.683
0.781
0.789
0.798
0.789
1
1
0.802
0.794
0.791
0.831
0.770
0.799
0.784
0.910
0.876
0.868
0.823
0.824
0.777
0.778
0.758
0.787
0.851
0.808
0.834
0.856
0.835
AP .
ai
psi
1.01
1.03
9.26
3.83
2.28
3.05
3.51
0
3.51
4.72
17.7
24.4
22.9
22.6
29.5
15.4
21.2
9.60
14.9
12.3
19.9
15.2
20.5
24.1
9.65
17.4
21.8
23.3
30.8
15.1
SMDo.i
M
45
41
1
1
25
22
21
OO
21
23
97
26
80
100
23
OO
CO
23
9
29
9
30
14
10
OO
125
130
85
18
00
Vap.
_
93.5
95.3
96.5
97.9
93.9
90.8
87.7
94.8
96.8
97.0
94.6
96.3
97.3
96.3
93.7
97.1
99.0
93.8
94.5
97.1
97.1
97.2
95.8
97.1
_
90.4
87.4
88.3
1000 B
20
19
68
50
15
19
18
0
18
22
61
105
176
154
280
0
0
59
87
77
128
131
174
199
0
96
125
137
258
0
-------
ON
100
95
3
o-
§ 90
•H
4J
tO
N
•H
M
O
I
85
80
0.05
Data Key In Table 8
0.10
0.15
0.20
0.25
0.30
B Group Parameter B
4-*u °-8
0.42
Figure 97e Experimental Vaporization Quality as a Function of the B Group Parameter
-------
to fall in a band which increases in vaporization quality as the group
parameter B increases, indicating the correct influence of the individual
variables in the parameter. Except for some extraneous points it appears
that the fuel mixing device developed produces a mixture with a high
degree of vaproization.
POROUS PLATE COMBUSTOR FABRICATION DEVELOPMENT
Introduction
The basic combustor material considerations are that the material
must: (a) have structural and chemical stability at the maximum-use
conditions which includes the temperature of use and the ability to with-
stand the anticipated thermal cycling, (b) be suitable for fabrication
into the desired design for the engine, (c) be compatible with other
materials of the engine design, and (d) be economically feasible for mass
production as a combustor component.
The materials which are presently available and feasible for use in
a porous combustor for an automobile gas turbine are outlined in Table 9,
which shows the usable temperature range and the reasons for the tempera-
ture limitation. The final resolution of material selection will require
more detailed test data than are available at the present, both from the
design requirements and from material behavior measurement.
These material considerations are outlined below in more detail in
terras of the metals and the ceramics. Most of these are available com-
mercially but need more detailed crucial laboratory tests to more
accurately define the behavior characteristics for this porous combustor
application. The initial selection for detailed development was a com-
167
-------
Table 9. POROUS COMBUSTOR MATERIALS
Material
Ni Alloys
Fe-Cr-Al Alloys
A1203-S102 Glass Fibers
Si02 Glass Fibers
SiC
ZrO
Probable Use
Temp. *F
1600-1800
1800-2200
1700-1900
1800-2000
2200-2700
2400-2700
2400-3000
2400-3200
Limitations
Sintering, Oxidation
Oxidation
Sintering(a\ Devitrification
Sintering***, Devitrification(b)
(A\
Oxidationv/
Sintering
Sintering
Sintering
a. Sintering of glass structures is obviously a liquid phase phenonenun and
may be accelerated by pressure due to contact surface Increase as a result
of plastic deformation at the higher temperatures where the viscosity is
- decreasing.
b. Devitrification is simply crystallization of the glass(which is, of course,
a supercooled liquid) with the result in loss of glass fiber structure.
c. The oxidation of SIC is complex because of the presence of fuel and air
which means that SiC, SiO, SiOo, hydrocarbons, H2, H^O, (OH), 00, C02 and
02 are all present. This could mean, for example, that if the SIO species
is formed in preference to CO, there may be a high rate of oxidation at
temperatures above 2400"F. Also, the oxidation of SIC also can conceivably
cause loss of permeability due to pore blocking by the SiC^ formed. The
resolution of this problem probably will depend on measuring the oxidation
kinetics in the various temperatures of Interest.
168
-------
bustor using a silicon carbide (SiC) burner face backed with mullite
(3 Al_0_-2SiO_) to provide the necessary thermal insulation to keep the
temperature of the cold side of the burner (which faces the incoming fuel-
air mixture) below the ignition temperature. However, after limited
testing of commercially available materials in radiation-cooled burners,
it became obvious that additional cooling was required at higher loading
and that an air-cooled design was essential which would have to be based
on a metal system; the preignition tests which lead to the air-cooled
design were described earlier.
Metals
The candidate alloys for combustor components are shown in Table 10,
which lists compositions of both the Ni-base alloys and the Fe-Cr-Al
alloys. These listed alloys were selected primarily on the basis of their
excellent oxidation resistance at elevated temperatures. For the stain-
less steels, in which chromium is the only alloying element, a generalized
relationship has been established between the Cr content and the maximum
use temperature in standard combustion gases. This relationship is shown
in the following tabulation:
Max. Service Temperature
% Cr in Combustion Gases, °F
0 1050
4 1200
8 1300
12 1400
16 1450
20 1650
24 1900
28 2100
169
-------
Table 10. CANDIDATE ALLOYS FOR COMBUSTOR COMPONENTS
CRYSTAL STRUCTURE
ALLOY MATRIX
AISI 442
AISI 446
GE-1541
Hoskins 875
Amco 18SR
AISI 310
Incoloy 800
Inconel 601
Rolled Alloys
333
Hastelloy X
Kanthal A-l
BCC
BCC
BCC
BCC
BCC
FCC
FCC
FCC
FCC
FCC
Fe
Bal
Bal
Bal
Bal
Bal
Bal
Bal
14
18
19
Bal
COMPOSITION, %
Cr Ni Co Mo W Al Ti Si
20 -------
27 -------
15-___4_-
23 - - - - 6 - -
18 - - - 2 0.4 1
25 20 ----- -
20 32 - - 0.4 0.4
23 Bal - - - 1.4 -
25 Bal 3 3 3 0 1.2
22 Bal 1.5 9 0.6 - - 0.5
22 0.5 - - 5.5 -
Y
-
-
1
-
-
-
-
-
-
-
_
170
-------
On this basis, the recommended maximum service temperatures under cyclic
conditions for AISI 310 SS (24Cr-20Ni) and AISI 446 SS (27Cr) have been
reported as 1900°F and 2100°F, respectively.
Significant improvement in oxidation resistance of iron-chromium
and nickel-chromium alloys can be achieved by the addition of Al or Al
and Y (yttrium) as exemplified by the 1541 iron base alloy developed by
General Electric and the nickel-base Inconel 601 alloy (Table 10)„ In
the Fe-Cr-Al system, one of the main advantages of increased Cr content
is to reduce the amount of Al required for the selective oxidation to
form the protective A1.0_ film; 15-25% Cr has been found to be optimum.
A minimum of 2% Al is required to produce the Al?0 -layer-in the 15-25%
Cr-Fe alloy. Al content above 5% result in alloys'that are brittle
and difficult to fabricate. It was found that additions :of yttrium to
Fe-Cr-Al alloys result in improved oxidation resistance under cyclic con-
ditions because of increased oxide adherence. The mechanism by which
yttrium improves oxide adherence has been postulated that it may result
from subscale formation of YJD which either promotes sintering of the
surface oxide Al-0 or provides a locking affect for the.surface oxide.
A form of a locking mechanism of the surface oxide layer by a uniform
network of internally penetrating oxides has also been suggested as
playing a role in achieving the excellent oxidation resistance in the
Al containing nickel base alloy Inconel 601. Spalling resistance of the
oxide layer is credited to the similar coefficients of thermal expansion
of the alloy and the surface oxide.
The oxidation resistance of the Fe-Cr-Al-Y alloys is such that a
0.020 inch thick sheet can resist destructive oxidation for more than
171
-------
2000 hours at 2300°F, 1400 hours at 2400°F, and 400 hours at 2500°F. The
Fe-Cr-Al-Y alloys are as oxidation resistant at 2300°F as the nickel
base non-aluminum containing Hastelloy X is at 2000°F. The use of nickel
base alloys requires that they be resistant to hot gas corrosion due to
sulfur containing environments. For this reason, it is necessary that
the Cr content be in excess of 15%. Hastelloy X (22% Cr),for example,
shows rather good resistance to hot gas corrosion. However, high Cr plus
aluminum additions (Inconel 601) provides still further improvement to
hot gas corrosion.
The strength properties of the ferritic (BCC) stainless steels and
the more oxidation resistant ferritic iron base alloys at temperatures
approaching 2000°F are very low (Table 11). If strength requirements for
the structural components exceed the limits of the ferritic iron alloys,
it will be necessary to utilize the higher strength austenitic (FCC) iron
or nickel ba«e alloys. It is well known and documented that the face
centered cubic lattice (FCC) is more resistant to creep than the body
centered cubic lattice. Limited testing data on these Fe-Cr-Al alloys
are shown in Figure 98 for typical oxidation curves at temperatures of
2000°F to 2600°F, and loss in permeability of porous plates at 1800°F
and 2000°F shown in Figure 99.
Ceramics
The most promising ceramic materials for the porous combustor are
SiC as the burner face and mullite as an insulator immediately adjacent
to the SiC. The crucial thermal properties of these and other potential
ceramic materials are shown in Table 12* Since high thermal conductivity
and low thermal expansion improve resistance to thermal shock, the selec-
tion of SiC is obvious. This selection is further supported by the high
emittahce factor which enhances radiation heat transfer* The choice of
172
-------
Table 11. TENSILE PROPERTIES OF CANDIDATE ALLOYS
FOR COMBUSTOR COMPONENTS
TENSILE PROPERTIES
Alloy
AISI 446
GE-2541
Annco 18SR
AISI 310
Incolor 800
RA 333
Inconel 601
Hast. X
Temp. ,°F
RT
1800
2000
RT
1800
2000
2200
RT
1800
2000
RT
1800
2000
2200
RT
1800
2000
RT
1800
RT
1800
2000
RT
1800
2000
2200
0.2% Y.S., ksi
51.5
-
-
50.8
2.2
1.2
1.0
65.0
-
-
36.5
8.0
5.0
3.0
47.7
50.0
15.4
49.0
52.2
16.0
8.0
3.7
Ult, ksi
83.0
2.9
1.5
76.7
3.5
2.0
1.8
85.0
2.0
1.1
84.0
14.0
8.0
6.0
87.7
100.0
18.4
107.0
114.0
22.5
13.0
5.4
El, %
29.0
_
-
14.0
135.0
154.0
161.0
29.0
—
-
50.0
52.0
35.0
35.0
42.0
50.0
53.0
45.0
43.0
45.0
40.0
31.0
173
-------
(a) GE, 0.160 In. dia rod
static air. Isothermal
IITRI, 0.010 in. sheet
cyclic exposure
(c) GE, 0.030 in. sheet
static air. isothermal
2400°F(a)
Cracking of
Oxide
2200°F(a)
Figure 98,
10/20
Time, Hours x 10
Weight Gain of Fe-25Cr-4Al-Y Alloy During Oxidation Testing in Air
100
-------
100
1800°F
75
= 50
o 25
100
H 875
GE 1541
I
I
200 300 400
EXPOSURE TIME - HOURS
500
600
Figure 99S Poroloy Permeability Retention After Exposure at 1800 and,.
2000°F (982 and 1093°C) in Air. Alloys GE 1541 and H 875U '
175
-------
Table 12. CRITICAL.THERMAL.PROPERTIES.OF.SELECTED CERAMICS
THERMAL CONDUCTIVITY THERMAL EXPANSION NORMAL TOTAL
IN BTU/(hr)(ft)(°F) COEFFICIENT EMITTANCE
MATERIAL
Al.O,
MgO
Zr02
Zr02-Si02
SiC
3Al.,0,-2SiO,
200eF
18
35
1.0
3.8
90
3.5
2000°F x 10~6/°F (to 2000-2200°F) AT 2200°F
3.4 5.0 0.41
3.5 8.8 0.3
1.0 6.7 0.5
2.5 3.1
10 2.7 0.85
2.4 2.9 0.6
176
-------
mullite as an insulator for SIC is based on (a) the close match of thermal
expansion plus (b) the fact that SiC forms SiO upon oxidation and the
film of this material will bond the SiC to 3Al203-2SiO . It is to be
noted, however, that the possibilities of using mullite alone or the use
of AlpO- may become economically attractive by using a Cr 0 additive to
increase the emittance factor.
The thermal stability of SiC at very high temperatures may impose
undesirable limitations on its use as a porous combustor as compared to,
for example, Al-0 . The oxidation of SiC is not known in sufficient de-
tail to make an accurate prediction at this time. For example, the
published data on oxidation^ ' are not readily applicable to the porous
combustor because these data were measured on powders of unknown ratios
of surface area to volume. As noted in Table 9, there is need for more
detailed test data to apply to this specific application.
The readily available fabricated ceramic materials which were used
entirely or in part to construct porous combustors for testing are as
follows:
1. ZrO (Y?0 stabilized) cloth is made by Union Carbide Corpora-
£• £3
tion. This material is reported to be made by saturating an
organic fiber cloth with suitable zirconium - yttrium salts and
then thermally removing the organic accompanied by converting
the zirconium-yttrium salts to the oxides. Initial tests on
this material show that it begins to become stiff after about
one hour at 2400°F, suggesting that sintering and/or crystallite
growth is occurring.
2. SiC - mullite porous combustor plates are available from sources
such as Rex Radiant. These plates are about 1/8" of SiC bonded
177
-------
to about 1/2" of mullite. Both substances are of about 16 mesh
particle size, probably bonded with some Al 0_-SiO substance.
£ J £
Heating to 2600°F for 4 hours shows an increase in grain-to-
grain bond.
3. Porous SiC-bonded SiC cylinders are available from the Norton
Company. Tests on the cylinders have resulted in fracturing
presumably because of non-uniformity in the porosity and/or non-
homogeneity in fuel injection; however, limited testing of 1/4"
rings shows no evidence of fracture.
4. Porous Al^O.-SiO- cylinders are available from Coors Porcelain
Company. Initial tests were on a "P-6" type porosity which has
a lAl-O -ISiO composition with a thermal expansion similar to
A1203.
In addition to the above, limited studies were undertaken to develop
porous ceramic parts of SiC and Mullite in which four fabrication approach
were considered. The first technique studied was to use coarse particles
of SiC and of mullite made from fine grained powders. These particles
are composed of fine grains (i.e., generally below 10 microns); this com-
position gives much higher strength than the commercially available single
crystal particles (such as Rex Radiant uses in their porous combustor).
A second approach was to use a volatile organic particle in fine grain
compacts to produce the desired porosity. This process will result in a
fine grain structure of optimum strength for the high porosity. A third
method was to use large single crystal particles and sinter these much
the same as the Rex Radiant porous material. The reason for this latter
approach was that there may be a porosity loss in use because of sintering
at the use temperature which will be minimized by the large grains. A
178
-------
fourth possibility was to use a material with high emissivity for radia-
tion heat transfer, for example, one of the above ceramics incorporating
a high emissivity material such as chromium oxide (Cr.0_) or cerium oxide
(Ce02).
However, porous metal materials for combustors were selected as the
main effort when it became obvious heat removal by cooling air was essential.
Initially these were commercially available alloys in a sheet or plate
form. Subsequently, the most crucial work was directed to fabrication
of porous metal combustors with air cooling tubes using 80Ni-20Cr powders
and 310 stainless steel tubes.
Fabrication of Porous Air-Cooled Burner
The initial porous-plate air-cooled burner designed for development
tests was 2.25 inches in outside diameter and 1.25 inches in inside
diameter and 6 inches long with 24 coolant tubes 0.125 inch in diameter
imbedded on the surface of the inside wall of the porous surface. The
porous-plate material was 80% Ni-20% Cr alloy in the form of coarse
powder in a particle size range of 25 to 50 mesh which was procured as
the only readily available powder in this size range. By making sample
pieces, it was found that coarse particles could be sintered in hydrogen
at 2400°F to form a soft porous surface. The mixing of 5 to 10 percent
of GE-81 braze material with powder was found to improve the strength of
the sintered plate and permitted the use of a lower sintering temperature.
A total of nine porous cylinders of approximately the aforementioned
dimensions were made and tested on this program. Problems that had to
be solved included mold design to minimize the effects of differential
expansion between mold and part, mesh size, elimination of density varia-
tions in the porous structure by vibration of the mold during filling,
179
-------
the choice of the correct proportion of braze powder to be added, uni-
formly mixing the braze powder with the powdered metal. Shown in Figure
100 is the sixth porous cylinder after sintering. Some disproportionate
braze distribution areas can be seen on the surface. Shown in Figure 101
is the same porous cylinder with the tube sheets brazed on. This porous
combustor (No. 6) was fabricated with an outer diameter increased to
2.50-inches and the 27 coolant tubes located on a 2.125-inch diameter
ring so as to be imbedded in the 0.375-inch thick frit wall. This re-
quired a new mold design with the 27 coolant tubes held in position by
temporary tube sheets of GE-1541 alloy which was pre-oxidized to prevent
bonding to the porous material. The coolant tubes were coated with braze
to improve the bond with the porous material. The -35 mesh NiCr powder
was pre-treated in hydrogen at 2000°F to reduce the oxygen content and
improve sinterability and then mixed with 5 percent of GE-81 braze. This
assembly was initially sintered at 2300°F and after removal from the mold
was resintered at 2370°F for 4 hours. The dimensions of this porous-
plate unit were 2.45-2.80 inch outside diameter, 1.70-inch inside diameter,
with a wall thickness of 0.35 to 0040-inch0 Radiographic examination
of this porous cylinder revealed that considerable improvement in uni-
formity of density was achieved by vibrating the mold during loading of
the NiCr powder, although some layering effect was noted as a result of
incremental additions of the powder. This cylinder was fitted with tube
sheets and the joints were brazed at 2250°F using GE-81 braze powder.
COMBUSTOR CONFIGURATION AND ENGINE INTEGRATION DESIGN
The accomplishments on the several tasks described above have pro-
duced results which can be synthesized into a porous-plate burner design
-------
00
.'
re-
•m.
•••'••':.-
Figure 100. Nichrome Porous Cylinder No. 3 As Sintered at 2350°F
-------
Figure 101. Porous Cylinder No. 3 With Tube Sheets Brazed on Each End
182
-------
concept for the Base Line Engine. An important result stemming from the
Combustor Loading and Emission Analysis is that the porous plate area
2
must change from 1.15 ft for operation along the vehicle level road-
2
load line to 2.2 ft for wide open throttle accelerations in order for
the combustor to operate between values of superficial velocity (V«-)
which correspond to flame extinguishment and to lift-off at the two limits.
During such operation, the equivalence ratio must be varied to help limit
superficial velocity and to limit porous-plate hot side temperature with-
in the temperature capabilities of available porous materials. Because
of the excessive heat flux conducted back to the porous plate from the
flame at elevated pressure levels (4 atm.) which heat results in surface
temperatures beyond available material limits, it is necessary to cool
the burner porous plate by means of embedded air-cooling tubes. The
thermal stresses in the air-cooled porous material must be reduced for
part integrity by employing a series of flat plates instead of a hollow
cylinder. Adequate fuel atomization and vaporization can be obtained
by spraying the fuel into compressor discharge-temperature air with an
air-assisted nozzle while at the same time avoiding preignition of the
fuel. Emission testing with propane indicated that NO emission index
values were below the Federal Standards. CO emission index values were
below the Federal Standards at superficial velocity (V2,-) levels corres-
ponding to low engine load and slightly above at conditions corresponding
to high loads. The unburned hydrocarbon emissions were barely detectable
and, therefore, adequately low. Prompted by these results, a design
concept for a porous-plate combustor for the Base Line Engine was formulated
and is described below.
183
-------
Combustor Conceptual Design
Figure 102 illustrates the air-cooled porous plate burner concept
which is proposed for application to the gas turbine Base Line Engine.
Figure 102 was prepared using a drawing of the Base Line Engine. The
porous plate combustor shown extends vertically above the engine center
line only 2 inches more than the standard combustor for this engine and
2
provides the full 2.2 ft of porous-plate surface required for wide open
throttle acceleration. The flat-plate porous burner elements made of
sintered nichrome are grouped in five rings of V-shaped burners arranged
inside a circular shell which is enclosed within the combustor outer
casing. Passages for the flow of primary air,bled from the compressor
discharge, and of secondary air from the regenerator discharge are pro-
vided in the annular space between the burner support shell and the outer
casing. Primary air entering the top of the combustor is admitted through
sixteen controllable (on-off) butterfly valves to the vaporization tubes.
In each of the vaporization tubes, an air atomizing fuel nozzle is lo-
cated immediately downstream of the air valve. Vaporization and fuel-
air mixing occur during passage of the fuel and air down the length of
the 6-inch vaporization tubes. The fuel is vaporized in these tubes by
heat from the primary (compressor discharge) flow and from the secondary
flow (regenerator discharge) which surrounds the vaporization tubes. At
the end of these tubes, the fuel air mixture reverses flow direction and
passes down channels formed between the V-shaped burners and the burner
support shell. From these channels the fuel-air-mixture flow is admitted
to the upstream faces of the porous burners.
Secondary air flowing oppositely to the primary air passes between
the vaporization tubes and is admitted to the porous burner cooling tubes.
184
-------
OQ
c
1-1
a.
o
K>
g1
o
M
n>
a.
00 O
U) •-(
o
to
rt
n
a
n
o
o
0
O
•s
MCOMQMT1 MCI
-------
The V-shaped burners are axially segmented into short sections which are
scaled by end plates. This prevents severe distortion under the radial
temperature gradient existing in the burner porous surfaces. All of tho
burners in a given .line, are attached to the sides of one of the fuel-air
mixture distribution ducts.
Burning occurs on the outer surfaces of the wedge—shaped burners in
the space between opposing burners. Power modulation involves variations
in air flow rate and equivalence ratio and in addition involves shutting
burners on and off by means of the butterfly air valves. Only the most
extreme transient operating condition requires all burners in operation.
Non-operating burner surfaces continue to be cooled by secondary air
flow through the embedded tubes, and thus serve as radiation heat sinks
for the opposing operating burners.
Because of the short length and close spacing of the embedded air-
cooling tubes the cooling effectiveness will be quite high. Since this
burner concept employs both radiation and air cooling, the surface will
run cooler than the values reported in the Section on Loading and Emission
Analysis.
Combustor Concept Development
As reported above, cylindrical porous nichrome burner models with
embedded air-cooling tubes have been run, mostly at one atmosphere, and
have indicated favorable emission results. It is expected that the use
of flat plate segments instead of complete cylinders of porous material
will alleviate the cracking problem experienced at elevated pressure
levels where the heat release rates are increased over the atmospheric
case. The sintering parameters must be optimized and the joining methods
must be specified which will assure a good product.
186
-------
The design concept shown in Figure 102has the flexibility of varying
2 9
the active area of the burner in eight steps between 1.15 ft and 2.2 ft"
or in as few as one step. This flexibility will be useful in transient
operation because when more than half of the burner area is being used
at least some of the wedge-shaped segments will have no radiation sink.
By controlling the active area and the equivalence ratio, the superficial
velocity (V9r) can be made high enough so as to limit the heat flux back
to the porous plate to a value that can safely be removed by cooling air
alone. This, of course, will have to be balanced against the increase
in NO because of the smaller reduction in flame temperature.
X
It is desirable to use regenerated air instead of compressor air for
the primary flow from the standpoint of engine fuel mileage. In
previous work an analysis of an engine with and without regenerator
bypass for the combustor primary air flow was made for the case of a
constant equivalence ratio of 0.9. For the Federal Driving Cycle the
fuel mileage values with and without regenerator bypass were 10.6 and 12.8
(23)
tnpg. There is some evidence, for example ' that the atomization and
vaporization of gasoline can be carried out in a time interval smaller
than the ignition delay using regenerator air, thus eliminating the re-
generator bypass. A compromise would be mixing compressor and regenerator
discharge flow for the combustor primary flow, thus reducing the fuel
mileage penalty without using an extreme temperature that might lead to
preignition. For the porous-plate combustor, since the flame temperature
is set by the superficial velocity (V-r) independent of the burner inlet
temperature, no adverse effect on NO emission index would be expected
with the elevated primary air temperature.
187
-------
Another means for improving fuel mileage is to use advanced sintered
materials having higher operational temperatures. By so doing, the
equivalence ratio of the burner could be increased, thereby reducing the
regenerator bypass; this bypass could be zero, in the limit. To utilize
materials with higher operational temperatures than nichrome, materials
processing methods must be worked out which avoid producing the protective
oxide coating of the sintering material before sintering.
In light of the present status of the air-cooled porous plate burner
for the gas turbine automobile engine, it is important to fabricate a
full-size combustor and test it on gasoline over the expected operating
range it would have in the Base Line Engine. Because of at least partial
success with the nichrome air-cooled burners, nichrome should be utilized
in the full size burner. This will require finalizing the sintering
parameters for this material and establishing the joining techniques to
be used between the porous material and the remainder of the combustor
structure.
188
-------
RANKINE CYCLE COMBUSTOR
DISCUSSION
\-Jhen any fuel is burned with air, some nitric oxide is almost in-
evitably formed in the hot product gases; of the contributors to air
pollution from combustion processes, it may be the most difficult to
eliminate or control. The principal mechanism of its formation is that
of the hot-air reaction; and until recent years, this was thought to be
the only important process. It is now known that in some circumstances
NO can evidently be made in hydrocarbon-air flames by other processes '
and in amounts that are significant in terms of the low source emission
goals now established.
The main problem in NO reduction is still that of controlling the
A
high temperature ("thermal or hot-air") production of NO. The rate of
the process unfortunately becomes appreciable just in the range of burned
gas temperatures common to most combustion processes. It has a very
large temperature dependence; the rate is approximately doubled by a 90°F
increase in temperature so that any preheating (for example, by precom-
pression in an engine cycle) drastically increases NOX emissions. Con-
versely, any effective scheme for reduction of the maximum temperature,
as well as the residence time at that temperature, can result in sub-
stantial decrease in NOX emission, for which various approaches have been
conceived and are being pursued.
The work described here is based on one such idea: limiting the
189
-------
maximum temperature of a premised flame by heat removal from the unburned
gas, in effeet,before the mixture is burned, so that it is equivalent to
precooling the mixture. The concept is derived essentially from the
cooled porous-plug burner or flameholder, first investigated by Botha
and Spalding ' and by Kaskan . Burners of this kind have since been
used extensively in flame studies, and have also been applied in various
(18 19 20}
forms in practical combustion devices. ' * The background and
rationale of the concept were discussed above.
For the corabustor boiler in a Rankine-cycle automotive engine, the
porous plate combustor is attractive for three general reasons:
1. It appears certain to be readily controllable to yield products
low in NO and hydrocarbons and acceptable in CO content.
2. It is conceptually well matched to such a cycle, since its re-
quired coolant can simply form part of the working-fluid path,
with complete recovery of the heat necessarily absorbed by the
burner prior to its passage to the boiler.
3. Although it requires a reasonably uniform premixture, the prepara-
tion of the fuel-air mixture in this application should be fairly
straightforward.
The approach taken is a combination of analytical and detailed ex-
perimental work to determine the feasibility of the porous-plate combustor
for the Rankine Cycle engine. Furthermore, since the porous-plate com-
bustor requires the fuel and air to be premixed, concepts of a fuel vapor-
izer were examined.
It was the purpose of the work described here to examine experi-
mentally porous plate combustion particularly with respect
190
-------
to emissions, and to obtain test data that is relevant to the design of
this combustor for a Rankine Cycle engine. Preliminary designs of both
the combustor and the vaporizer suitable for the Rankine Cycle were first
made.
While the approximate distribution .of heat absorption between the
burner and the rest of the boiler system can be roughly estimated from
available data, it is desirable to determine this with some accuracy for
the anticipated range of flame conditions. Many such burner heat flux
measurements were made for this purpose, as well as for indirectly es-
timating or checking the actual burned gas temperature.
Further, measurements of oxides of nitrogen, carbon monoxide, and
unburned hydrocarbons were made over the range of operating conditions
at the atmospheric pressure of the Rankine Cycle. These measurements
were made on flames with air and five different fuels-(EPA'gasoline, AMOCO,
propane, toluene, and n-heptane).
PRELIMINARY BURNER-VAPOR-GENERATOR DESIGN
Preliminary Rankine Combustor Design
Figure 103 shows the conceptual design of the Rankine Engine burner
and vapor-generator. Combustion air from the blower passes through a
preheater where it is heated by the hot exhaust gases. The preheated
air flows through a mixing device where the fuel is injected and atomized.
Vaporization of the fuel takes place in the flow ducts leading to the
transpiration burners.
The specifications and preliminary design of a porous Rankine burner-
vapor-generator with the configuration of Figure 103 are summarized in
Table 13C This component has an overall diameter of 18 inches; there is
191
-------
Air fro* Expander
Driven
Superheated
Out
Air Controlled
Preaaure Reaervolr
•nd »• If old
= Air —
— Preheeter —
j^ (Plate-Mo Type) =
fu.l/Alr fcnlfold
Figure 103, Conceptual Design of Porous-Plate Burner-Vapor
Generator for Ranklne System
auction A-A
-------
Table 13. POROUS RANKINE COMBUSTOR
SPECIFICATIONS
Combustor Heat Load 1.87 x 10 Btu/hr
Feedwater Temperature 200°F
Steam Flow Max. 1296 Ibm/hr
Min. 40 Ibm/hr
Fuel Flow 101 Ibm/hr
Feedwater Pressure 1200 psi
DESIGN
Equivalence Ratio 0.9
Fuel/Air Flow Rate 1801 Ibm/hr
2
Burner Area 5 ft with V-_ =40 cm/sec
Minimum Area to Avoid Lift-off 3.9 ft2
at Maximum Flow
4 Concentric Burner Plates 1/2 in. Thick
Overall Diameter of Burner 18 in.
Burned Gas Exit Velocity with 1/2" 65 ft/sec
Radial Gap for Each Exit Plate
Pressure Drop at Design Condition 105 psi
(215 Microns Pore Size, 35% Porosity)
3 Stages Required to Meet Turndown Requirement of 32.4:1 with
Individual Burner Turndown of 5:1
193
-------
a 1/4 in. radial gap for each of the 4 concentric burner plates, which
gap results in a burned gas exit velocity of 65 ft/sec. Three stages are
required to supply an overall turndown ratio of 32.4:1, since the maximum
turndown of the individual burners is approximately 5:1. The package
2
shown in Figure 103 has a porous burner area of 5 ft .
Preliminary Fuel Supply Design
In conventional combustors using liquid fuel, the fuel is typically
injected into the combustion zone in the form of small liquid droplets
which are subsequently vaporized and burned. The porous-plate burner
requires that the fuel be vaporized and well mixed with air before en-
tering the porous structure of the burner.
Figure 104 shows the fuel-air mixer and fuel vaporizer concept which
is similar to that used in a current advanced aircraft engine development
program. The fuel enters the mixer through the central tube and is in-
jected radially into the primary air stream through the small orifices.
The primary vanes impart a swirling motion to the air which then passes
through a venturi section. Additional air is introduced through a radial
inflow counter-rotating secondary swirler. The intense shear region be-
tween these counter-rotating flows provide good atomization of the fuel
which is then vaporized on the downstream duct.
The turndown ratio compatibility of a single burner is in the order
of 5:1 for constant pressure and equivalence ratio. In order to achieve
the required 20-30 turndown ratio, staging is required.
Figure 105 shows two possible installation configurations for the fuel/
air mixers. Preliminary calculations were made to estimate pressure drop
as a function of cup dimensions and degree of fuel vaporization as a
function of air velocity, air preheat temperature and length of vaporizer.
194
-------
Figure 104. Fuel-Air Mixer and Fuel Vaporizer
Double-Swirl Carbureting Concept
195
-------
CONFIGURATION I
F/A Mixers
Control
Valves
1
Fan
CONFIGURATION II
Figure 105. Preliminary Installation Configurations
196
-------
The pressure drop results for the design point are shown in Figure 106
and the degree of fuel vaporization is shown in Figure 107. The fuel
(21)
vaporization predictions are based on the correlation previously de-
veloped. Since the liquid droplets in those experiments were relatively
large, the predictions for this mixer (in which the droplets will be
small) are believed to be conservative.
After completing the above analyses, the vaporization tests were
run on the gas turbine combustor fuel-air mixture supply reported above.
As indicated in Figure 97, most of the data indicated 93% or above of
the fuel was evaporated. The tests were run at atmospheric pressure
similar to Rankine burning conditions. The inlet air temperature was in
the range of 164 to 411°F. These temperatures are easily attainable
in a Rankine cycle air preheater required for an efficient engine. The
equivalence ratio of the data was 0.8, only slightly lower than contemplated
for the Rankine combustor. Thus, the test data can be seen as applicable
to the Rankine fuel-air mixture supply.
Reference to Figure 91 indicates that a tangential flow of primary
air is created in a central tube by variable louvers around the inlet
end. A "Sonicore" air-actuated fuel nozzle sprays gasoline in a 90°
conical spray down the center line of the central tube. The spray nozzle
creates a spray with a 20-50 micron mean drop diameter which quickly
evaporates due to the warm primary air introduced into the central tube.
197
-------
= 0.49 PPS
TA= 205 °F
PA = 16 PSiA
N = NO OF CUPS TOTAL
1
4 5 6 8 10 20
CUP PRESSURE DROP , AP/P $
30
40
60 80 100
Figure 106. Rankine Engine Fuel-Air Mixer Cup
Flow Characteristics
198
-------
VO
ion
90
80
70
60
O
s-
M 50
40
30
20 h
10
0
200
_L
T - 540 F
400
600
T
Primary Air Temperature
T - 350 F
T - 205 F
T - 100 F
Developed Flow Length
800
1000
Velocity, ft/sec
Figure 107. Estimated Vaporizer Performance
-------
The smaller drops follow the streamlines and are evaporated due to the
presence of the warm air. It takes only a configurational study to
adopt this tested concept to the Rankine combustor
BENCH TESTS
Test Apparatus
Burner and Auxiliary Equipment - To meet the design requirements of a
real engine, the individual modular burners will probably have to be other
than flat, but a satisfactory evaluation of the quantities of interest
here can equally well be made with a flat burner. An available unit made
of sintered copper shot and with a 5.5 in. x 5.5 in. burner surface,
about 0.4 in. thick and with about 35% porosity, was used in this work;
2
the burner area of 30.3 in is reasonable for laboratory bench tests
and is large enough for adequate precision in the desired measurements.
The scale factor relative to the specified engine parameters is approxi-
mately 1/30. Figure 108 shows an essentially identical sample unit,
made only half-filled with the sintered copper shot to illustrate its
construction.
A schematic diagram of the test set-up is shown in Figure 109. The
burner is secured to a flanged plenum, primarily to reinforce the rec-
tangular water manifolds of the burner which had been made entirely of
copper and thoroughly annealed during the sintering of the copper shot.
(The manifolds had to sustain a water pressure of as much as 120 psi to
suppress boiling, with the preheating required to avoid condensation of
pre-vaporized fuels.) The water preheater (Figure 109) is an identical
sintered copper burner, and both have manifolded parallel water tubes as
in Figure 108, rather than the monotubes in some of the proposed engine
designs.
200
-------
ro
O
». ^m
^l,...
.'
r
Figure 108. Half-Filled Copper Shot Burner, Illustrating Burner
Construction and Coolant Tube Manifolding
-------
Air (^ 150 psi)
Pressure
Gage
i—o—i
N>
O
N5
Flow
Meter
Propane or
Methane
Differential
Thermopile
NO Monitor
Critical
Flow
Orifice
Calrod
Heater
PDS
To Sample
Vac Flasks
Porous Plate
Test Section
Inlet
Temperature
T/C
Pressure
Gage
Figure 109. Schematic Diagram of Burner Test and Gas Sampling Arrangement
-------
_2
The water flow appropriate to this scale is about 1.1 x 10 Ibm/sec;
it was metered with a calibrated differential pressure orifice, and de-
livered from the pressurized reservoir shown. On the preheater, a methane
(or propane)-air mixture is burned at a composition and flow adjusted to
give the temperature rise required (up to 180°F) to match approximately
the preheated air temperature and to avoid condensation of the prevaporized
liquid fuel (for example, gasoline) in the burner pores. The air pre-
heater is a Variac controlled 1 KW Calrod immersion heater in about 2
/ *3
feet of 2-inch pipe. It can heat a steady flow of 1.76 x 10~ ft (STP)/
sec of air to 230°F.
The combustion air was fed from the 150 psi plant supply and metered
with a calibrated critical flow orifice meter; when propane was used as
the fuel, it was taken from a regulator on a standard container of chemically
pure (CP) propane. It was metered with a critical flow orifice, as were
also the fuel and air to the preheater when used.
When a liquid fuel was used, it was taken from the reservoir, pres-
surized to a few psi above system pressure with N_, through a rotameter
and delivered by the adjustable Zenity metering pump to a simple spray
nozzle (Monarch or Delavan; 0.6 gallons per hour). The rotameter was
calibrated with each of the three liquids used.
Many earlier attempts had been made to operate the system with pre-
vaporized fuel at temperatures up to 306°F, using the single spiral coolant
tube in another small porous-plug burner as the heater; it appeared to
be a satisfactory and readily adjustable source of the vaporization heat,
but pulsations seemed to be very difficult to avoid with any reasonable
feed-system modification, and steady flames were seldom obtained with
this arrangement. It was finally abandoned for the simpler arrangement
203
-------
discussed below. To ensure complete vaporization of the liquid fuels,
the air was usually preheated to about 212°F, as measured at the entry
to the plenum (Figure 109). All of the apparatus downstream of the fuel
injection was well insulated with glass wool, and the water temperature
at the burner inlet was maintained at about 212°F or higher to avoid
possible condensation of the fuel. The spray nozzle, (facing downstream),
mounted in the center of the pipe, discharged into the preheated air
stream ahead of the mixer. Satisfactory performance of the burner was
thus obtained; ignition was prompt and reliable to a steady flame of
uniform appearance, which extinguished promptly when the pump was stopped
with no evidence of residual unvaporized fuel.
The heat flux to the burner was obtained from the water flow, its
specific heat and the temperature rise; calibration of the ten-element
thermopile with which the temperature rise was measured agreed well with
standard copper-constantan tables. One of the 1/8 inch diameter stain-
less steel sheathed junctions of this device contains a separately wired
single-junction thermocouple. This end was mounted at the burner inlet
to monitor the water temperature as delivered by the preheater. The
temperature rise across the main burner was frequently as much as 144°F,
so that the exit water was usually flashing to steam beyond the back pres-
sure valve (Figure 109). When a steady flame could be established quickly
(for example, with propane-air mixtures), the system would come to equili-
brium in a few minutes.
Burned-gas temperature measurements were made with 0.005 in. butt-
Virtis Corp., Gardiner, N.Y.
204
-------
welded platinum platinum-(10 percent) rhodium, silica cooled thermocouples;
the methods used in their preparation and in computing the radiation
correction were those described by Kaskan . In the case of the EPA
gasoline, such gas temperature measurements were made for the entire
range of conditions of interest. For a few flames, the thermocouple was
placed at many locations above the burner surface to learn whether there
were significant lateral variations in temperature attributable to velocity
nonuniformity; none larger than about 18 to 36°F were found even though
the appearance of near-lifted flames that occur with weak mixtures, for
example, did show an observable effect of the proximity of the cooling
tubes to the surface of the sintered copper. In all the reported measure-
ments, the thermocouple was located at or as close as possible to the
point of maximum temperature in the direction of the hot gas flow, usually
about 0.2 to 0.4 in. above the burner surface.
Some of these conditions, particularly wherein the burned gas tem-
perature exceeds the melting point of platinum, are rather severe for
the survival of these thermocouples, and it is fairly tedious to make
them. The gas temperature was frequently obtained from the heat flux
and a careful enthalpy balance; when it was measured in both ways, there
was good agreement between them and with previous data .
Gas Sampling and Analysis - The burned gas was withdrawn through a quartz
probe 0.04" I.D. x 0.12" O.D. x 1.18" long; it was joined to a 1.18 in.
length of 0.4 in. O.D. quartz tubing, which was connected and sealed to
a 10 in. length of 1/4 in. O.D. thin-walled metal tubing. The height of
the probe tip above the burner could be adjusted and measured; its lateral
position could also be varied, keeping the quartz-to-metal seal outside
the flame gas column. From the dimensions and approximate temperature
of the probe and the sample flow rate, it was estimated that the sample
205
-------
cooling rate was approximately 1800°F/msec. The metal tube was jacketed
with plastic tubing and fittings for water-cooling, and was connected
by polyethylene tubing to a glass trap (to remove liquid condensed from
the sample), and to the inlet of the sampling system. By placement of
the probe at various distances from the burner, the hot-gas residence
time of the sampled gas could be varied and computed from the burned gas
velocity (which was calculated from the mass velocity and the burned
gas density) and the distance from the probe to the burner.
The sample could be pumped to the various analytical devices as
shown in Figure 109, using a small diaphragm compressor in the Stack
Sampling Unit that accompanied the NO monitor referred to later.
A
3
Typically, the probe sampling rate was about 0.2 in /sec. This unit
also contained an aftercooler for the compressor, a gas distribution
manifold and a small rotameter which monitored the flow actually de-
livered for analysis. This flow was generally about 0.1 in /sec, the
balance of the sampled flow being discarded beyond the compressor.
The 0_ content of the gas was measured with a Beckman E-2 analyzer.
It was done primarily to confirm the actual equivalence ratios, for ex-
ample, when the input fuel flow was uncertain. The °?/N2 rati° was also
determined by gas chromatography together with carbon monoxide using a
13-X column, helium carrier and a thermal conductivity detector. It had
been calibrated with CO/O^/N mixtures prepared from metered flows of
the gases.
206
-------
A second gas chroraatograph was used to examine the gases for hydro-
carbons, employing a Poropak-N column and a flame ionization detector.
It had been calibrated with a number of light hydrocarbons; the sensitivity
was equivalent to 1 ppm or less of, for example, CH , C H- or C H,.
For nitrogen oxide measurement by the phenol-disulfonic (PDS) acid
3 3
procedure, the pumped samples were taken into 180 in or 300 in evacuated
flasks containing the required hydrogen peroxide - sulfuric acid solution.
The gas sample pressure was brought to 0.5 atmosphere and the pressure
was then raised to 1 atmosphere with oxygen. The analysis was made by
(22)
the PDS procedure
Measurements of total NO were also made with an electrochemical
x
NO monitor that was available. The calibration or span gas used was
X
stated to be 196 ppm NO in N?. The instrument ±8 convenient since it
gives a result in a few minutes and is far less tedious to use than the
well-accepted PDS procedure. The sample pumping and conditioning assembly
(provided for passing the sample continuously to the monitor) could also
be used for all of the other analytical sampling as well.
The instrument is supposedly immune to interferences from other gas
components including possible hydrocarbons. In the course of earlier
work, it was noted, however, that the electrochemical monitor does respond
strongly to both acetylene and ethylene, and if present they will seriously
interfere. In a few experiments with unintentionally rich flames in
the present work, this effect could be quite clearly seen. Therefore,
Dynasciences Corp., Chatsworth, Calif.; model NX-130
207
-------
it was used only as a guide in confirming trends. Though its indications
generally did not grossly disagree with the PDS determinations, they did
tiend to he higher by about 20 percent. This could not have been due to
hydrocarbons, since they were eventually shown to be entirely negligible;
it now seems likely that the calibration or span gas used was in error
(low in NO) by this amount, since a subsequent PDS analysis of this gas
suggested that its NO content in N was only about 150 ppm. It is,
therefore, possible that the instrument itself was performing satisfactorily.
Emissions measurements of practical interest would actually be made
on the engine exhaust, or in this case at least downstream of the boiler,
and therefore on gas which has been quenched to some extent at a rate
not easily determined. Such quenching should be very effective on the
NO formation process owing to its large temperature dependence, and the
result will be close to what is actually present in.the hot gas before
the boiler, and also to what is measured by the sampling method used
(23)
here. Freezing the CO concentration is much less likely , and a
measurement on the boiler exhaust is certain to be lower than on samples
obtained with much faster cooling, such as that in the sampling pro-
cedure used here. But in any case, the measured concentration may well
be low.
It was not feasible nor considered essential to provide a cooler
equivalent to the boiler for the present work, so the [CO] determinations
no doubt represent levels intermediate between the true values at the
high temperature and the values of ultimate practical interest. Thus,
they represent upper limits to the emissions levels that would be ob-
served in practice, but lower limits to what is actually present at the
burned gas conditions.
208
-------
Fuels - CP propane was taken from a standard 100 Ib liquified petroleum
gas container and the toluene was also CP reagent. Prior to receipt of
the specified EPA gasoline, some work was done with AMOCO Super Premium
(no lead, no phosphate). Its specific gravity was 0.76 at 76°F suggest-
ing a considerable content of aromatic material; it was not analyzed for
hydrogen-carbon (H/C) ratio, but based on a subsequent inquiry of the
manufacturer's representative, it was taken to be about 50% toluene,
50% paraffins, with a gross composition equivalent to C7H for approxi-
mate stoichiometry estimates. Similarly, the specific gravity of the
EPA gasoline was 0.755, and though its aromatic content is specified to
be lower, its allowable olefin content is much higher (about 30%), so
its combustion stoichioraetry was expected to be very close to that of
AMOCO, which also accords with subsequent 0 analyses of the combustion
products.
These three liquid fuels were analyzed for bound nitrogen with these
results:
AMOCO Super Premium 11+5 ppm N
Toluene 24+5 ppm N
EPA Gasoline 33+5 ppm N
This N content could account for at most a few ppm of NO in the
burned gases , which would be within the likely error of the analysis
and was ignored in considering the emission results.
Near the conclusion of this work, some NO formation measurements,
X
also reported here, were made with one other liquid fuel (CP n-heptane)
2
using a smaller (15.5 in ) but similar water-cooled burner, in which the
burned-gas temperature was directly measured. The intent was to examine
209
-------
for NO formation some flames of a fuel with H/C ratio near that of
K
propane, but of molecular weight in the range of the other fuels.
Data Reduction and Correction - Where preheated air was used, a correction
to the measured heat flux was applied to obtain the values presented
which are referred to an initial condition of 76°F. This correction can
be shown to be:
6F = 2.13 V (T -76) Btu/hr-ft2 (12)
with sufficient accuracy. The unit of the terms in Eq. 12 are V?_ in
en/sec and T in °F. For example, with T = 230°F and V = 25 cm/sec,
Ai r AT r 2.J
2
6F = 8200 Btu/hr-ft . Although it is at most only a few percent of the
total enthalpy flux, it is a significant fraction of the measured heat
flux. (The burner heat flux, F, is only 10 to 30% of the total combus-
tion flux, but it must contain all of the effect of the preheat, so long
as the burned gas temperature depends only on the mass velocity (V_,.);
6F/F can therefore be as high as 25%.)
The heat-balance calculation of the burned gas temperature from the
heat flux thus corrected was done as follows. For the calculated adiabatic
flame temperature (T ) of the mixture initially at 76°F, a value of the
3
enthalpy of the mixture was obtained, from which was subtracted this
correction, Ah:
Ah = (Fb/V25) x ^ = hT - hT (Btu/lb gas) (13)
m a b
in which F, is the corrected burner heat flux, m is the average molecular
b
weight of the burned gas and the constant is a conversion factor to ob-
, tain the result in the engineering units usually tabulated. The enthalpy
at the calculated adiabatic flame temperature is h , while the enthalpy
a
210
-------
(24)
at the burned gas temperature is h . These tables have been com-
Tb
piled for various fuel-air mixtures which generally do not correspond
•»
exactly to the fuels and compositions encountered here. But we are
only interested in enthalpy differences which do not vary significantly
with the H/C ratio of the fuel for a given equivalence ratio, R, and
temperature interval, and it is readily shown that the temperature decre-
ment is obtained in this way with ample accuracy. Thus, with the quanti-
ties h and Ah, h is obtained and the corresponding burned gas tem-
a b
perature, T , is read or interpolated from the table.
In view of the fairly low H/C ratios of toluene and the two gaso-
lines, some attention was given to estimates of the adiabatic flame tem-
peratures, or at least how much they should differ from that of propane-
air for a given equivalence ratio. The difference (mainly due to the
stoichiometry) would not be expected to be large; but owing to the great
sensitivity of NO formation rate to temperature and the way in which
X
T was used when the actual gas temperature was estimated from the heat
3
flux, it seemed desirable to compute the difference with some care. The
unit heating value of the stoichiometric mixture of air with the vapors
2
of each of the three liquid fuels was computed to be 94.5 Btu/ft + 0.11
(76°F, 1 atm) for all of them, or 4.5% higher than the corresponding
3
value for propane-air (90.6 Btu/ft ).
A correlation can be made between the trends of these unit heating
values with accurately calculated adiabatic flame temperatures, for
various classes of hydrocarbons of varying H/C ratio and molecular weight.
It shows that the 4.5% difference should result in a 45 to 54°R higher
value of T ; for stoichiometric propane-air (R = 1), T is 4091°R, so
for the three liquids at R = 1 in air, a value of 4140°R is estimated.
211
-------
There are also available a few calculated values for toluene, benzene,
(25)
cyclohexane, etc., mostly for various rich mixtures (R = 1.05 to 1.17)
that can be compared with the corresponding values for propane-air; this
comparison also indicates a difference of about 45°R. The value of 4140°R
for R = 1 was therefore assumed, and with similar estimates of the dif-
ference for lean mixtures, a plot of T vs. R was made from the computed
cl
values for propane-air mixtures. In no case is the estimate believed
to be in error by more than +^ 9°F.
Measurements of Heat Flux and Burned Gas Temperatures
Figures 110 and 111 summarize the heat flux data, corrected to 76°F
in the manner previously described, for propane-air mixtures and for EPA
gasoline-vapor air mixtures.
The burned-gas temperatures, largely computed from the heat flux
data, are shown in Figure 112 for propane and EPA gasoline, and other
data derivable in part from these plots are also summarized in Table 14.
Similar data giiTnma-Hp.fi for the toluene and AMOCO gasoline-air mixtures
are also shown in Table 14. In the case of the EPA gasoline, direct
measurements of the gas temperature by thermocouple, as well as the
indirect measurements from the heat flux data, were generally made; on
the whole, these were in very good agreement (within 54°R) and the lines
shown in Figure 112 average these temperatures.
The data for propane-air mixtures in Figure 112 can be compared with
those of Kaskan and the agreement is excellent. There are no data
in the literature with which the results for the other fuels could be
compared. The heat flux data were in this case extended further toward
the adiabatic flame condition, well beyond the range of interest for this
application. Although it has no practical consequences in the present
212
-------
40,000
N>
I-"
U>
\->
jz
P
OJ
M
PQ
QJ
O
4-1
30,000
20,000
10,000
id
0)
R = 1.0
= 0.9
R = 0.8
10
20
30
Superficial Velocity, V , era/sec
Figure 110.
Burner Heat Flux Back to the Burner from
the Flame; Propane-Air Mixtures
40
-------
40,000
30,000
I
R = 0.9
M
OJ
C
M
0)
.C
id
PQ
cd
0)
a:
20,000
10,000
R = 0.8
R = 0.7
10
20
Superficial Velocity, V , cm/sec
30
40
Figure 111« Burner Heat Flux Back to the Burner from the
Flane; Gasoline (EPA)-Air Mixtures
-------
100
Absolute Temperature, T, °R
u
•H
U
o
n)
•rJ
0
O
§• 10
o
o
o
I
o
o
CM
CO
o
o
o
P-I
Propane-Air
o
r»
r^
CM
Gasoline-Air
Data From:
Thermocouple Heat Flux
Gasoline (EPA)-Air Flames R = 0.7
R » 0.8
R =• 0.9
O
D
Propane-Air Flames
R " 0.8
R - 0.9
I
I
4.5
5.0
5.5
6.0
6.5
Reciprocal Absolute Temperature, [1/T(°K) x 10 ]
Figure 112. Comparison of Gasoline (EPA) and Propane-Air Burned Gas
Temperatures
215
-------
Table 14„ RANKINE MODEL BURNER TEST RESULTS
Fuel
C3H8
Toluene
AMOCO
EPA
Gasoline
R
0.83
0.95
1.0
0.68
0.78
0.86
0.88
1.0
0.70
0.80
0.90
Btu/ft3
76.1
86.0
90.6
64.4
73.7
81.3
83.1
94.5
66.2
75.6 .
85.1
F
raax _
Btu/hr-ft
23200
31500
36500
15900
23900
34500
27900
41800
19200
25200
37200
V25,ra
cm/sec
15
18
20
12
16
18
17
19
15
17
19
p
0.18
0.17
0.17
0.18
0.18
0.20
0.17
0.20
0.17
0.17
0.20
max
°R
3330
3550
3640
3060
3310
3490
3510
3600
3130
3310
3492
T
a
°R
3760
4010
4090
3330
3710
3890
3940
4140
3440
3730
3960
S
u
cm/sec
37
42
43
24
35
38
41
43
30
39
43
E
Btu/
( Ibm-mole)
99000
97200
95400
108000
110000
97200
108000
89100
99000
97200
93100
Stoichiometric (R = 1.0) F/A ratios by weight:
Propane 0.0641 (0.0420 by volume)
Toluene 0.0740
Gasolines 0.0703
-------
context, an interesting incidental observation is that the heat flux
apparently does not tend toward zero as V_,. approaches the adiabatic
burning velocities for these mixtures. There would evidently be a resi-
2
dual heat flux of about 8200 Btu/hr-ft which is no doubt due mainly to
radiation from the flame to the cold burner surface. From the emissivity
of such flames (estimated as about 0.03), the radiative flux in such a
geometry should be approximately of this magnitude. If it is computed
for each condition and subtracted from the measured values, the remainder
then does tend to zero as the unburned gas velocity, V-,, approaches the
adiabatic burning velocity, S .
In Table 14, h is essentially the lower heat of combustion of the
(gaseous) fuel per unit volume of the mixture (76°F, 1 atm). F
D ^ITlcOt
and Voc are respectively the maximum heat flux and the unburned mix-
/j,m
ture velocity at which it occurs. The fraction 6 ( F, /h )
m b.max v,...
25,m
is the fraction of the enthalpy flux rejected to the burner at the maxi-
mum heat flux. The maxima, as seen in Figures 110 and 111, are of course
rather flat, so that the designation of Voc and 3 are somewhat arbi-
25,m m
trary; they are included as possibly useful for design purposes. T
max
is the flame temperature at which the heat flux back to the burner is
a maximum.
T is the calculated adiabatic flame temperature and S is the normal
H U
burning velocity of the mixture (that is, Su is the value of V25 which occurs
at the temperature, T ) obtained by extrapolation (Figure 112) and
a
noted as a matter of general interest. E is the apparent "activation
energy" indicated by the slopes of these lines: E = -4606 R (d(log _V )/
d(l/T)). R is the universal gas constant, R = 1.9859 Btu/°R Ibm-mole.
The last three columns would suffice to determine the temperature de-
217
-------
pendence of V . For a given V and R, there is little difference in
the actual burned-gas temperature for a given fuel. At least, for the
hotter flames (near R = 1), the higher initial enthalpy content (h) of
the higher molecular weight fuels seems mostly to enhance the burner
heat flux rather than the burned gas enthalpy.
It is worth noting that, where both were measured, comparison of
the directly measured temperatures and those computed from the heat flux
( 6 )
shows little, if any, of the kind of disagreement noted by Kaskan in
comparing his results with other data computed from heat flux measure-
ments . To the extent that combustion is not entirely complete in
the flame itself, it might be expected that those temperatures calculated
from the heat flux would be somewhat higher; but if there is such a con-
sistent difference here, it appears to be no more than about 54°R which
is about the expected uncertainty in the temperature anyway. Similarly,
any calorimetric losses associated with the heat flux measurements would
also result in a higher computed gas temperature. It appears that neither
source of error could have been significant.
Emissions Measurements
Carbon Monoxide - The analytical results are given in Table 15 as the
measured concentrations in samples taken about 20 msec from the flame.
Many of them are at or below the equilibrium concentrations corresponding
to the burned gas conditions. Some analyses were also made on samples
taken closer to the flame where [CO] was found to be considerably higher
than equilibrium, and the expected decay behavior could be qualitatively
observed as shown for a few cases in Figure 113,
No quantitative interpretation or kinetic analysis of the measure-
ments was attempted for reasons discussed below; but the measurements do
218
-------
Table 15c RANKINE MODEL BURNER EMISSION RESULTS
W
d[NO]
dt
ppm/msec
C H
AMOCO
Toluene
Gasoline
(EPA)
R
Dim.
0.93
0.93
0.93
0.83
0.83
0.83
0.96
0.88
0.95
0.86
0.78
0.68
0.89
0.88
0.86
0.77
0.78
0.76
0.78
0.70
0.68
v
V25
cm/sec
25
20
15
25
20
15
25
25
—
20
20
20
25
20
15
25
20
15
15
25
20
T
Xb
°R
3690
3570
3430
3540
3440
3330
3780
3690
—
3470
3370
3260
3650
3560
3380
3490
3350
3200
3130
3420
3220
[Measured]
Ppra
n D Prompt [NO]
NO
Meas.
0.95
0.25
0.17
0.56
0.27
0.05
8.0
7.5
—
1.8
1.0
0.3
3.4
—
—
1.0
—
0.06
—
—
"NO
Calc.
0.64
0.26
0.05
0.22
0.10
0.03
1.2
0.86
—
0.23
0.07
0.02
0.47
—
—
0.19
—
0.02
—
—
Meas.
(Approx.)
25
15
10
7
6
5
—
—
—
30
15
10
25
—
—
35
—
15
— -
—
Pred.
1 Est.
15
12
10
5
5
5
25
20
—
15
*\j j
% 0
20
—
—
•HO
—
% 5
—
—
ppra
[NO]
20M
sec
ppm
[CO]
PDSa) (DYN)b)20M sec
44
20
14
19
12
6
160
150
—
65
35
15
90
—
—
55
—
18
—
—
(38)
—
—
—
—
—
—
—
—
(69)
(31)
(31)
(77)
(77)
(51)
(59)
(38)
(25)
(38)
(21)
500
150
1500
1500
1200
800
400
600
400
< 100
200
^ 100
[CO]ea
eq.
(T }
(v
1100
700
3000
1200
1200
720
300
600
350
145
200
65
a)
Phenol disulphonic acid method
b)
Dynasciences electrochemical NO monitor
X
-------
2200
2000
1800
e
o.
Q.
O
1600
1400
120C-
A
A
Gasoline/Air
Equivalence Ratio
R * 0.9
25 cm/sec, T * 3600°R
c
0)
o
c
o
100(
•o
0)
3
W)
80(
60C-
D
D
Propane
R ~ 0.9
D
D
20(
A
A
Propane/Air
I
5 10 15 20
Residence Time, Milliseconds
Figure 113, CO Analyses from Two Flames
220
25
-------
adequately show that, with this hot gas residence time, and a practically
attainable quenching or cooling rate in the boiler, the CO emission would
be less than the specified level, which is equivalent for the average
operating condition to about 800 ppm. A few of the measured values in
Table 15 apparently exceed this level, but they too would certainly fall
below 800 ppm in the boiler; the values given represent an upper limit
to what can be expected of the engine.
(jfi\
In this connection, it has been shown in other work with a
geometrically similar burner system, that a large reduction in [CO] does
occur across such a downstream heat exchanger; it was found to decrease
by more than an order of magnitude to values typically corresponding to
equilibrium at temperatures lower than 3240°R. Whatever the analysis
or interpretation of this process may be, as a practical matter, a
similar reduction must occur in any similar device.
The CO data given in Table 15 was converted to the Bnission Index
and is plotted for all the fuels tested in Figure 114. The bulk of the
data falls below the level corresponding to the original 1976 Federal
Standard, especially at the lower values of superficial velocity (V~c)
or lower power levels. At some of the higher values of superficial
velocity (V25) and hence higher power levels, the measurements are above
the level of the Standard for gasoline burned at an equivalence ratio
of 0.9. However, it is expected that the CO emissions averaged over a
duty cycle would be below the Standard since a large fraction of the
duty cycle occurs at the lower power levels.
Hydrocarbons - Most of the sampling for this purpose was done at residence
times of 10 to 20 msec; in no case with any of the fuels was any hydro-
carbon detected; it is concluded that with any reasonable excess air
there will be essentially zero (<1 ppm) emission of hydrocarbons from
221
-------
100
10
1.0
FUEL
• C3H8
V AMOCO
D GASOLINE (EPA)
• GASOLINE (EPA)
[| GASOLINE (EPA)
1 ATMOSPHERE
I I
EQUIVALENCE
RATIO
.83
.93
.96
.68-.70
.76-.78
.86-.89
CO FEDERAL STANDARD
(3.4 GM/MI)
GASOLINE
I
I
I
10 15 20
SUPERFICIAL VELOCITY, V25/ CM/SEC
25
Figure 114. CO Experimental Measurements on Water-Cooled Burners
with 20 Milliseconds Residence Time
222
-------
burners of this kind. In fact, it was generally observed that the normal
background CH, concentration (1.0 to 1.2 ppm) was, if anything, decreased
by the flame.
Oxide of Nitrogen - Most of the results summarized in Table 15, were ob-
tained at conditions similar to but not necessarily identical with those
in the thermal tests just described, since they were not usually done
simultaneously. The main interest here is in the NO formation data,
compared in Table 15 and Figure 115 with the computed values to be ex-
pected by the thermal or hot-air mechanism; with the assumption of [0] =
[0] , the equation of the solid line was obtained; for combustion products
at 1 atm with 0_ mole fraction (N ):
18 -136.500
RJJQ E d[NO]/dt = 3'3 * 10 e RT (NQ )1/2 (ppm/millisec) (14)
from the same basic kinetic data used by Fenimore . The present estimate
of the NO formation rate in Eq.14 is about 30% higher than the original
estimate given in Figure 3; this correction was made after more careful
consideration of the relevant kinetic data. The line (practically straight
in this temperature interval) was calculated for an oxygen mole fraction,
N = 0.02, and the data were adjusted or normalized to this oxygen content
U2
to facilitate comparisons at a fairly representative excess air condition.
This oxygen content corresponds to an equivalence ratio of R = 0.91.
The individual rates were obtained from the slopes of linear plots
of the analytical results versus the time from the burner, and the "prompt
NO" was obtained from the intercept at zero time in essentially the manner
described by Fenimore . The "predicted estimate" of the prompt NO was
made from Fenimore's data with the assumption that it would be the same for
all hydrocarbon fuels at the same equivalence ratio.
223
-------
Temperature, °K
00
o o
in »»
IN CM
0
o
ro
O
O
CM
CM
O
O
O
O
o
O
o
O
o
00
Q.
Q.
0)
4J
10
C
O
10
e
o
u,
to
10
1.0
0.1
0.01
Equation 14
Original Line y
- From Fig. 3 A
Data Adjusted to
2% 02 Mole Fraction
O Propane
Q Heptane
^ Toluene
X EPA Gasoline
AMOCO
4.5
5.5
Reciprocal Burned Gas Temperature, [1/T(°K) x 10 ]
Figure 115. Variation of NO Formation Rate from "Hot-Air" Mechanism as
a Function of Reciprocal Absolute Temperature. NO Formation
Rate Adjusted to 2% Molar 0^
224
-------
The errors in both the thermal NO derived from the slopes, and in
x * '
the intercepts are no doubt large. There were usually too few observa-
tions to warrant any assumption other than linearity or constant R^ at
a given condition; in no case were there more than three sampling posi-
tions (for example, at 1.0, 2.0 and 3.0 cm from the burner) and frequently
only two were taken. The uncertainty in R^ may be as much as 50%
and the intercepts may be in error by 5 or 10 ppm, (though there was
no case in which there appeared to be no intercept at all), but the
tabulated values at 20 msec in Table 15 would be in reasonable accord
with a measurement at that residence time (taken as appropriate for the
present application and representing what would be found in the engine
exhaust), whatever the distribution between prompt and thermal NO may
A
be. The uncertainty in these values of [NO ] at 20 msec is estimated
X
as no more than 30%.
With these reservations concerning their accuracy, the data may be
compared with the calculated line, bearing in mind this also has a likely
uncertainty of about + 30%. The data for propane-air flames may be said
to agree reasonably well; they lie within a factor of two of the calculated
values, and this may be little outside the combined errors.
For the other four fuels, a single line could represent well enough
all the data, but everywhere it would be just about an order of magnitude
higher; alternatively, the disagreement with the predicted line is equiva-
lent to about 270°R for a given rate, or by 180°R relative to the propane
data. In any case, it is too large to be accounted for by experimental
errors in either the analyses or in the temperature measurements. (The
possible contribution of fuel nitrogen had been anticipated, but the
analyses mentioned earlier showed that it could account for at most only
225
-------
a few ppm in the burned gases.)
Thus, the measured formation rate of NO, at least for the fuels of
high average molecular weight exceeds by about an order of magnitude the
rate expected or calculated with the assumption that [0] = [0] . The
conclusion seems inescapable that the actual [0] averaged over the resi-
dence time (^ 20 msec) must in fact be higher by a factor of five or
more. Since the oxygen concentration must be decreasing toward equili-
brium and would be expected to attain [0] in this time, the oxygen
concentration nearer the flame would have to be still larger. It should
be noted that the plotted data do not include the contributions, shown
in Table 15, of the intercepts or prompt NO. Had they been arbitrarily
included in averaging the rate, the results would be somewhat higher
still. It seemed reasonable to separate its contribution however, since
it does appear to be real and generally agrees satisfactorily with data
though the possibility that it too results from excess [0] cannot be en-
tirely ruled out by these results.
There have been many investigations of the emission of nitrogen
oxides from combustors and engines with a variety of fuels, including
some similar to those used in this work; unfortunately, there appears
to have been none with preraixed flames of hydrocarbons higher than
propane, with which the data given here can be directly compared. There
f 26 ^
has been reported one rather detailed experimental study and analysis
of NO formation in a preraixed, quasi-one-dimensional propane-air flame
system,, With near-adiabatic flames of one equivalence ratio (R = 0.8),
estimates were made of the variation of the rate with time for up to 15
msec from the main zone of flame reactions. The maximum instantaneous
rate a few msec from the flame appeared to be an order of magnitude or
226
-------
more higher than would be predicted from Eq.14, generally decreasing
thereafter.
For the same burned gas temperature, and for roughly comparable
burned gas conditions, an average d(NO)/dt computed over a residence
time comparable with that considered here (about 20 msec) would be about
a factor of two higher than that measured in the present work; it would
more nearly agree with the data for higher hydrocarbons (Figure 115 and
Table 15). In view of the differences in equipment and approach, the
agreement may be reasonable; but from either set of results, it has to
be concluded that the NO formation rate is considerably higher than
Eq.14 would suggest, whatever the detailed interpretation may be.
The failure of the approximation or assumption that [0] = [0] in
the burned gases of hydrocarbon flames has been discussed by others in
the context of NO formation; it formed the basis for interpretation of
(2ft)
the data just mentioned, and in the following generally similar dis-
(27)
cussion, its connection with CO and its afterburning is considered.
That higher-than-equilibrium concentrations of free radicals do
occur in post-flame gas has been shown in many experimental studies.
The nature of the chemical processes involved, the partial equilibrium
state, and the subsequent relatively slow radical recombinations are
/2g\
well established and need not be elaborated here; but it is worth
notinp, that the slow decay or "afterburning" of CO in lean hydrocarbon-
air flame gases is intimately connected with such excess radical con-
centrations. It is known that the afterburning is really a recombination
(27 29)
process ' in which CO is in effect one of the radicals in excess
of equilibrium. It is connected with the others through the equilibrium
227
-------
constant for the balanced reaction:
CO + OH =• CO + H (15)
And through similar relations for the other balanced reactions, in
effect:
CO + 02 = C02 + 0 (16)
is also balanced. In the products of lean flames [CO] must then be just
proportional to [0] or [0]/[0] = CO/[CO] and will always be greater than
unity during the afterburning of carbon monoxide; the rate of formation of NO
will then be larger by this ratio also.
For reasons discussed previously, the [CO] measurements summarized
in Table 15 may have no really quantitative meaning; the true [CO] in
the burned gas must be higher than that measured owing to uncertainties
in the sampling procedure. But even from these data it is qualitatively
clear that [CO] does, as expected, decrease from levels considerably
higher than equilibrium.
(29)
There appears to have been only one investigation in which ex-
cess radical concentrations were determined and directly related to [CO]
and its decay in the burned gas of a lean hydrocarbon flameo It was an
ethylene-air flame at 3528°R at an equivalence ratio of 0.82, and [CO]/[CO]e
about 1 msec after the main reaction zone of the flame was approximately
10, decreasing to approximately 5 in the 6 msec period observed. This
ratio might have approached unity in roughly the time of interest here
(20 msec), during which its average value and therefore [0]/[0]
evidently would have been about 5. • ...
If it is assumed that this ratio (deduced from only one set of data)
applies to the lean burned gases of other hydrocarbons, and that the
228
-------
ratio has no temperature dependence, a line parallel to the predicted
line (Figure 115) could be drawn; the data would generally lie within a
factor of two of this adjusted prediction. Within the likely errors of
the data, this could then be considered fair agreement.
The fairly consistent difference between the results for propane
and the other fuels (Figure 115) is still unexplained. Assuming the
validity of and extending the argument above, it would be inferred that
[CO]/[CO] following the main reaction zone is appreciably higher in
flames of these higher hydrocarbons. There is some indication of such
a tendency in data obtained here; but it would be worthwhile to investi-
gate this point (using, for example, optical techniques ' ) that
would permit an unequivocal determination of [CO] or its equivalent in
the actual burned gas conditions of such flames.
The NO data shown in Table 15 was converted to Emission Index and
X
is plotted for all the fuels in Figure 116, as a function of superficial
velocity, V , and equivalence ratio. Some of the individual NO values
for gasoline burned at an equivalence ratio of 0.9 are above the level
of Federal Standard at high values of V ; however,the NO values at the
lower values of V - (or lower power) are well below that of the Standard,
and it is expected that the N0_ emissions averaged over a duty cycle would
be below the Standard since a large fraction of the duty cycle occurs
at the lower power levels.
Summary and Conclusions From Bench Tests
The thermal and emissions characteristics of some non-adiabatic
hydrocarbon-air flames at 1 atmosphere have been determined for use in
the design of a transpiration (cooled porous metal) burner for the com-
bustor boiler of a low-NO Rankine cycle engine. Flames of mixtures
X
229
-------
10
o
o
1.0
2;
O
UJ
CSI
O
0.1
[J
FUEL
C3H8
SH8
AMOCO
AMOCO
TOLUENE
TOLUENE
TOLUENE
GASOLINE (EPA)
GASOLINE (EPA)
GASOLINE (EPA)
I
EQUIVALENCE
RATIO
N0? FEDERAL STANDARD
CO. 4 GM/MILE)
GASOLINE
SYMBOLS WITH FLAGS C U ) ARE NO
DATA MEASURED WITH DYNASCIENCE
ELECTROCHEMICAL MONITOR
I
10
15
20
25
SUPERFICIAL VELOCITY, V CM/SEC
Figure 116. NC>2 Experimental Measurements on Water-Cooled Burners With
20 Milliseconds Residence Time
230
-------
of air with propane and with prevaporized toluene, heptane and two gaso-
2
lines were stabilized on a 30.3 in water-cooled porous copper burner.
Burned gas temperature as a function of mixture mass velocity and equiva-
lence ratio was determined for each fuel, and other related flame para-
meters were computed and summarized. The temperature was obtained either
from direct thermocouple measurements or from the measured burner heat
flux and a heat balance; when both were made on the same flame, they
agreed well.
Determinations of NO , CO and hydrocarbons were made on probe
samples with hot gas residence times up to about 20 millisec. Hydro-
carbons were found to be <1 ppm and thus negligible from any flame ex-
amined (R £ 0.95). The absolute concentration levels of carbon monoxide
as measured are of doubtful significance, but the decay of [CO] from
values much higher than equilibrium could be observed. The levels es-
timated for practical engine exhaust conditions are expected to be
satisfactorily Iower0
From the measured NO concentrations, the rate of the thermal or
x
hot-air formation of NO in the burned gas was estimated and correlated
with the measured burned-gas temperatures. It was compared with the
prediction from accepted kinetic data, with the assumption that [0] =
[0] ; the measured rates are all higher than predicted, and for the
equ
fuels of higher molecular weight, by about an order of magnitude at all
observed temperatures (3240°R-3780°R).
It is concluded that the average [0] must correspondingly exceed
[0] and the probable connection between this excess and [CO] is dis-
eq
cussed.
231
-------
REFERENCES
(1) Amann, C.A., W.R. Wade, and M.K. Yu. Some Factors Affecting Gas
Turbine Passenger Car Emissions. SAE Paper 720237. 1972.
(2) Wade, W.R., P.I. Shen, C.W. Owens, and A.F. McLean. Low Emissions
Combustion for the Regenerative Gas Turbine: Part 1 - Theoretical and
Design Considerations. ASME Paper 73-GT-ll. 1973.
(3) Azelborn, N.A., W.R. Wade, J.R. Secord, and A.F. McLean. Low
Emissions Combustion for the Regenerative Gas Turbine: Part 2 - Experi-
mental Techniques, Results, and Assessment. ASME Paper 73-GT-12. 1973.
(4) Wohl, K. Quenching, Flashback, Blow-off - Theory and Experiment.
Fourth Symposium on Combustion. 1953. p. 68-89.
(5) Botha, J.P. and D.B. Spalding. The Laminar Flame Speed of Propane
Air Mixtures with Heat Extraction from the Flame. Proc. Roy. Soc.
(London). A225; 71-96, 1954.
(6) Kaskan, W.E. The Dependence of Flame Temperature on Mass Burning
Velocity. Sixth Symposium on Combustion. 1957. p. 134-143.
(7) Fenimore, C.P. Formation of Nitric Oxide in Premixed Hydrocarbon
Flame. Thirteenth Symposium on Combustion. 1970. p. 373-380.
(8) Barnett, H.C. and R.R. Hibbard, (ed.). Basic Considerations in the
Combustion of Hydrocarbon Fuels with Air. NACA Report 1300. 1957. p. 135.
(9) Schneider, P.J. Conduction Heat Transfer. Addison-Wesley, 1955.
p. 218-221.
(10) Barnett, H.C. and R.R. Hibbard, (ed.). Basic Considerations in the
Combustion of Hydrocarbon Fuels with Air. NACA Report 1300. 1957. p. 1-31.
232
-------
(11) Wukusick, C.S. and J.F. Collins. An Iron-Chromium-Aluminum Alloy
Containing Yttrium. Materials Research and Standards. December 1964.
(12) Cundiff, J.W. Superalloy Improves Anti-Smog Devices for Autos.
Metal Progress. September 1972.
(13) Madsen, P. and R.M. Rusnak. Oxidation Resistant Porous Material
for Transpiration Cooled Vanes. NASA CR-1999. 1972.
(14) Adarasky, R.F. Oxidation of Silicon Carbide in the Temperature
Range 1200°F to 1500°F. J. Physical Chemistry. £3: 305-307. 1959.
(15) Jorgensen, P.J., et al. Oxidation of Silicon Carbide. J. Am.
Cer. Soc. 42; 613-616. 1959; Effects of Oxygen Partial Pressure on
the Oxidation of Silicon Carbide. J. Am. Cer. Soc. 43; 209-212. 1960.
(16) Automobile Gas Turbine-Optimum Cycle Selection Study. General
Electric Report GESP 730 FS. June 1972.
(17) Fenimore, C.P. Formation of Nitrogen Oxides from Fuel Nitrogen
in Ethylene Flames. Combustion and Flame, 19; 289. 1972.
(18) Siegler, M. and G.E. Moore. Flame Recombination of Oxygen and
Hydrogen. Chem. Engrg. Prog. Symp., ,66.:.l. 1970.
(19) U.S. Patent 3,589,184 (1971) (Continuous Flow Gas Calorimeter).
G.E. Moore.
(20) U.S. Patent 3,672,839 (1972) (Exothermic Gas Generator). G.E. Moore.
(21) Bahr, D.W. Evaporation and Spreading of Iso-octane Sprays in High
Velocity Air Streams. NACA-RME 531.14. November 1953.
(22) Oxides of Nitrogen in Gaseous Combustion Products (Phenoldisulfonic
Acid Procedure). ASTM Test Method Specifications Designation D-1608-GO,
reapproved 1967.
(23) Halpern, C. and F.W. Ruegg. A Study of Sampling of Flame Gases.
J. Res. NBS, 6_0: 29-37. 1958.
233
-------
(24) Fremont, H.A. et al. Properties of Combustion Gases - System
C H -Air. General Electric Co., Cincinnati, Ohio. 1955 and Kutzko, G.G.
Properties of Combustion Gas-System CH.-Air. General Electric Company.
R68AEG161. 1968.
(25) Fristrom, R.M. and A.A. Westenberg. Flame Structure. McGraw-Hill,
1965. p. 22-23.
(26) Williams, G.C., A.F. Sarofim, and N. Lambert. Nitric Oxide Forma-
tion and Carbon Monoxide Burnout in a Compact Steam Generator. Symp.
on Emissions from Continuous Combustion Systems. Plenum Press, 1972.
p. 141.
(27) Schott, G.L. Comments on Article by Robert F. Sawyer and Trilochan
Singh. Co-reactions in the Afterflame Region of Ethylene/Oxygen Ethane/
Oxygen Flames. Thirteenth Symposium on Combustion. 1970. p. 403-415.
(28) Fenimore, C.P. Chemistry in Premixed Flames. Macmillan, 1964.
(29) Kaskan, W.E. Excess Radical Concentrations and the Disappearance
of Carbon Monoxide in Flame Gases from Some Lean Flames. Combustion and
Flame (London). J3:49-60. 1959.
(30) Kaskan, W.E. The Source of the Continuum in Carbon Monoxide-
Hydrogen-Air Flames. Combustion and Flame (London). _3:39-48. 1959.
234
-------
APPENDIX A
COMBUSTOR TEST FACILITY
The combustor test facility is located in the Gas Dynamic Facility
No. 1 of Bldg. D and is shown in simplified schematic form in Figure A-l
The facility was designed for l/5th or full scale automotive gas
turbine combustor testing. The design criteria are as follows:
1. Primary Air Flow 0.031 - 0.4 pps to 400°F
2. Secondary Air Flow .75 - 2.25 pps to 1400°F
3. Fuel Flow (Gasoline) - 6 Ib/hr to 75 Ib/hr ambient temperature
The plant air supply (100 psig) is fed through a filter from which
it is split to provide air to both the primary and secondary flow metering
stations. Pressure regulating valves ahead of the metering tubes (sharp
edge orifice type) control the orifice upstream pressures. The primary
metering system consists of three orifice metering tubes in parallel to
cover the 13:1 flow range required. Remote operated pneumatic actuated
valves are used to direct air flow through the proper orifice. A down
stream flow control valve can be used in conjunction with the combustor
back pressure valve to control combustor pressure and primary air flow.
Shown in Figure A-2are the test cell flow metering sections and the pri-
mary and secondary flow heaters.
The primary heater is electrically powered from an adjacent substation.
The power control system consists of ignitions with firing rate varied
to control voltage and current to the resistance type air heaters.
The secondary air system is basically the same as the primary, how-
235
-------
Figure A-l
GDFI
-------
N3
Figure A-2. Test Cell Heaters and Flow Metering Sections
-------
ever only one orifice metering station is required to cover the air flow
range.
Cooling water is injected into the corabustor discharge gas stream
prior to entering the butterfly type back pressure valve. Water flow is
automatically controlled to maintain a maximum temperature of 500°F
entering the back pressure valve. An exhaust silencer discharges the
gases and water vapor vertically to the atmosphere.
The fuel system consists of a 250 gallon storage tank equipped with
a float for level indication and located outside of the test area. A
saall BOtor driven gear type pump with pressure regulating valve and by-
pass return to the tank Is located adjacent to the storage tank. Fuel is
fed into the test cell to a control station where a remote operated valve
controls the flow to the combustor. A turbine type flowmeter senses fuel
flow. Shown in FigureA»U3 is the fuel storage and pumping system.
The control room houses all controls and instrumentation. Operation
is controlled completely from the control consoles in this area. Shown
in Figures A-A and A-5are the heater and facility control consoles re-
spectively. All pertinent parameters are indicated and if necessary re-
corded. An automatic fuel-air ratio control system is capable of holding
a constant fuel-air ratio over a range of primary air flows. The ratio
type system also allows changes in the fuel-air ratio to be made. The
system will then hold the new fuel-air ratio over a range of air flows.
To operate a combustor of the l/5th scale size, the inner valves of
pressure regulating and control valves are changed and smaller orifice
plates are installed in the flow metering tubes. The turbine type fuel
flow sensor is also replaced with a smaller unit. The portable Scott
Research Laboratory Inc. Model 108H Dilute Exhaust Gas Analysis System
238
-------
U>
Figure A-3. Fuel Storage and Pumping
-------
N3
•e-
o
Figure A-4. Heater Control Console
-------
Figure A-5. Facility Control Console
-------
purchased by General Electric shown in Figure A-6 is available for use
in the Gas Dynamic Facility No. 1 (GDF-I). It has provisions for measur-
ing NO, N02, CO, C0_, 0- and total hydrocarbon emissions.
The facility has been given a mechanical checkout. Both heater units
have been tested and operate satisfactorily. A complete checkout will be
made after the installation of a combustor.
242
-------
NJ
Figure A-6. Scott Exhaust Gas Analyzer
-------
ACKNOWLEDGEMENT
Contributions to the work reported herein were made by a number of
people in the Energy Systems Programs and the Corporate Research and
Development components of the General Electric Company. The main con-
tributions are listed below:
Combustor Loading and Emission Analysis
and Report Editing
Combustor Concept Feasibility Development
Testing Operations
Fuel-Air Mixture Supply Development
Porous Plate Combustor Fabrication Development
Combustor Configuration and Engine
Integration Design
Preliminary Burner-Vapor Generator Design
Bench Tests
C.W. Deane
J.A. Bond
G.C. Wesling
R.A. Fuller
H. Bradley
J.A. Bond
B.L. Moor
R.G. Frank
J.F. White
J.F. Collins
A.W. Schnacke
J.R. Peterson
G.E. Moore
B.E. Cans
These contributions are gratefully acknowledged.
The contributions to the program by T.S. Mroz and W.C. Cain, the
EPA Project Officers are also gratefully acknowledged.
244
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
EPA-A60/3-73-001
3. RECIPIENT'S ACCESSION-NO.
4. TITLE ANDiuBTITLE
Development of Low Emission Porous-Plate Combustor
for Automotive Gas Turbine and Rankine Cycle Engines
5. REPORT DATE
September 1973
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Richard '^J. Rossbach
9. PERFORMINqOR~ANIZATION NAME AND ADDRESS
General Ellectric Company
Energy Syjstems Programs
P. 0. Box! 15132
Cincinnati , Ohio ^52 15
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING',AGENCY NAME AND ADDRESS
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Alternative Automotive Power Systems Division
Ann Arbor,' Michigan ^8105
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The purpose of this contract was to evaluate analytically and experimentally the use
of the porous-plate combustor for use in the gas-turbine or Rank!ne-cycle advanced
automobile engines to control exhaust emissions. As regards the gas turbine applica-
tion, this report contains analytical results on the burner area requirements for the
various operating conditions of the Baseline Engine as well as exhaust emission pre-
dictions. The design concept of an air-cooled, variable-area combustor for this engine
is presented. Operational and emissions data on several experimental combustors are
presented along with the fabrication development leading to these combustors. Finally
the demonstration results for a full-scale fuel-air mixture system are presented. With
regard to the automotive Rankine engine application, heat load and emissions data are
presented for propane and four liquid fuels. Although the fuel-air mixture system
developed for the gas turbine combustor is directly applicable, alternate systems were
investigated. Finally recommendations for the development of both gas turbine and Ran-
kine combustors are presented.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Held/Group
Air polIut ion
Development
Combustion chamber
Automotive engines
Rankine cycle
Fabri cat i on
Exhaust emissions
Hydrocarbons
Carbon monoxide
Propane
Gasoli ne
N i ch rome
Oxi des of ni trogen
Porous plate combustor
Burner-vapor generator
Fuel atomizer
13B
2lB
21E
3. DISTRIBUTION STATEMENT
UnIi mi ted
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
245
------- |