United States Industrial Environmental Research EPA-600/7-80-017c
Environmental Protection Laboratory January 1980
Agency Research Triangle Park NC 27711
Advanced Combustion
Systems for Stationary
Gas Turbine Engines:
Volume III. Combustor
Verification Testing
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-017C
January 1980
Advanced Combustion Systems
for Stationary Gas Turbine Engines:
Volume III. Combustor Verification Testing
by
P.M. Pierce, C.E. Smith,
and B.S. Hinton
Pratt and Whitney Aircraft Group
United Technologies Corporation
P.O. Box 2691
West Palm Beach, Florida 33402
Contract No. 68-02-2136
Program Element No. INE829
EPA Project Officer: W.S. Lanier
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
FOREWORD
This report was prepared by the Government Products Division of the Pratt & Whitney
Aircraft Group (P&WA) of United Technologies Corporation under EPA Contract No.
68-02-2136, "Advanced Combustion Systems for Stationary Gas Turbine Engines." It is
Volume III of the final report which encompasses work associated with the accomplishment of
Phases HI and IV of the subject contract from 1 January 1978 through 12 April 1979. The
originator's report number is FR-11405.
Contract 68-02-2136 was sponsored by the Industrial Environmental Research Laboratory
of the Environmental Protection Agency (EPA), Research Triangle Park, North Carolina
under the technical supervision of Mr. W. S. Lanier.
The authors wish to acknowledge the valuable contributions made to this program by
Mr. W. S. Lanier, whose skillful management and insight have been a key factor in the success
of the Rich Burn/Quick Quench combustor design concept.
The Pratt & Whitney Aircraft Program Manager is Mr. Robert M. Pierce; the Deputy
Program Manager is Mr. Clifford E. Smith. Mr. Stanley A. Mosier is Technology Manager for
Fuels and Emissions Programs at the Government Products Division of Pratt & Whitney
Aircraft Group. Mr. Bruce S. Hinton has been a principal contributor to the technical effort in
Phases III and IV.
Special recognition is due Mr. E. R. Robertson of the Component Design and Integration
Group, who was responsible for all drafting, hardware fabrication, and data processing
activities. The skillful assistance of Mr. R. Taber of the Instrumentation Laboratory in setting
up and operating the gas analysis equipment is also acknowledged.
iii/i
IV
-------
CONTENTS
Section Page
SUMMARY xiii
1 INTRODUCTION 1
2 PHASE III COMBUSTOR DESIGN AND RIG PREPARATION 2
2.1 Review of Phase I and Phase II Results 2
2.2 Design Approach 3
2.3 Basic Design Concept 5
2.4 Initial Combustor Sizing and Selection of Basic Features 9
2.5 Primary Zone Liner Cooling 13
2.6 Residence Time Considerations 22
2.7 Primary Air Staging 24
2.8 Combustor Internal Aerodynamics 26
2.9 Premix Tube 31
2.10 Construction of the Full-Scale Combustor and Rig Hardware 38
3 PHASE IV VERIFICATION TESTING 47
3.1 Premix Tube Component Tests 47
3.2 Full-Scale Combustor Verification Tests 80
4 CONCLUSIONS FROM PHASES III AND IV 125
LIST OF SYMBOLS 126
REFERENCES 127
APPENDIX A DATA LISTINGS 129
APPENDIX B CONVERSION TO SI UNITS 137
-------
ILLUSTRATIONS (Continued)
Figure page
21 Identification of Stations Referred to in the Aerodynamic Model Calculations 34
22 Comparison of Predicted and Experimental Pressure Drop Characteristics of
the Bench-Scale Combustor 36
23 A Typical Centrally-Mounted Fuel Injector With a Mixing Device 40
24 Representative Multiple Fuel Injector Premix Tubes 41
25 Liquid Jet Penetration in Airstream 42
26 Radial "Spoke" Premix Tube Design 43
27 Final Layout of the Full Residence Time (FRT) Configuration of the Full-
Scale Combustor Prior to the Verification Test Program 45
28 Full-Scale Combustor During Assembly (FRT Version) 47
29 Full-Scale Combustor Fully Assembled (FRT Version) 48
30 Layout of the B-2 Rig Showing the FRT Combustor 49
31 Schematic Diagram of Rig Instrumentation System 51
32 B-2 Sample-Gas Analysis System 52
33 Measured Distribution (Normalized) of Liquid Obtained Using Simplex
Pressure Atomizing Fuel Injector 55
34 Liquid Distribution Pattern Produced by Centrally-Mounted Air-Blast Nozzle
(Nominal Design Point Air Velocity, Equivalent Ratio = 1.0) 55
35 Liquid Distribution Pattern Produced by Spray-Ring Injector (Nominal
Design Point Air Velocity, Equivalent Ratio = 1.0) 56
36 Combustor Test Configuration Centrally-Mounted Air-Blast Nozzle 57
37 Combustor Test Configuration Centrally-Mounted Air-Blast Nozzle in
Large Diameter Premix Tube 57
38 Combustor Test Configuration Simplex Pressure Atomizing Nozzle 58
39 Combustor Test Configuration Spray-Ring Injector 58
40 Variation in NO, Concentration With Equivalence Ratio for Prototype
Premising Tube With Centrally-Mounted Air-Blast Nozzle (Scheme
26-05A) 59
vn
-------
ILLUSTRATIONS (Continued)
Figure page
41 Variation in NOX Concentration With Equivalence Ratio for Prototype
Premixing Tube With Simplex Pressure Atomizing Nozzle (Scheme
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Premixing Tube Scheme 26-07A Showing Initial Design of the Spray-Ring
Fuel Injector
Intermediate Configuration of the Fuel Injector Spray Ring
Revised Design of the Splashplate/Injector Ring Arrangement
Comparison of Variation in NOX Concentration With Equivalence Ratio for
Revised Spraying Design and Centrally-Mounted Air-Blast Nozzle..
Basic Initial Premix Tube Configuration With Variable Damper Mechanism
Installed
Pressure Transverse Data for Basic Premixing Tube With Variable Damper
Modified Configuration of the Basic Initial Premix Tube Configuration
Burner Duct Modified to Eliminate Acoustic Resonance
Extended-Length Premix Tube
Premixed Flame Produced by Extended-Length Premixing Tube During
Ambient Operation (Nominal Design Point Air Velocity Equivalence
Ratio = 1.4 Nominal)
Premix Tube Airflow Calibration of Extended-Length Tube
Scheme 26-21A Original Premix Tube With Air-Blast Nozzle and Inlet
Swirl Vanes
Scheme 26-22A Original Premix Tube With Dual-Orifice Nozzle
Scheme 22-24A Short Premix Tube With Air-Blast Nozzle
Scheme 22-26A New Premix Tube With Air-Boost Nozzle and Inlet Swirl
Scheme 22-27A Original Premix Tube With Air-Boost Nozzle
Scheme 22-25A Short Premix Tube With Air-Boost Nozzle and Vortex
Spreaders
Scheme 22-28A Original Premix Tube With Air-Boost Nozzle and Vortex
Spreaders
Scheme 22-29A Original Premix Tube With Air-Boost Nozzle, Vortex
Spreaders and Variable Damper
61
61
62
63
64
65
68
69
69
70
71
75
76
76
77
77
78
78
79
Vlll
-------
ILLUSTRATIONS (Continued)
Figure Page
61 Scheme 22-19B Configuration for Flow Visualization Test of Six-Nozzle
Cluster Fuel Injector 79
62 Scheme 22-23A Original Premix Tube With Air-Boost No;:zle (No Inlet
Swirl) 80
63 Scheme 22-18A Original Premix Tube With Spray-Ring Injector and
: Segmented Splashplates 81
64 Scheme 22-20A Original Premix Tube With Low Delta P Spray Ring 81
65 Flame Observed Using Premix Tube (Scheme 26-29A) 84
66 Revised Premix Tube Design Incorporating Air-Boost Nozzle 86
67 Revised Premix Tube Design Incorporating "Spoke" Fuel Injector 86
68 Full-Scale Combustor Scheme FS-01A 88
69 Variation in Emission Concentrations With Overall Equivalence Ratio for
Tests Conducted With Scheme FS-01A 90
70 Original Arrangement of the Full-Scale Test Rig 92
71 Arrangement of the Full-Scale Test Rig Following Elevation of the Combustor 92
72 Full-Scale Combustor Scheme FS-02A 95
73 Variation in Emission Concentration With Overall Equivalence Ratio for
Tests Conducted With Scheme FS-02A 96
74 Comparison of Emission Data Obtained for Schemes PS-01A and FS-02A.... 97
75 Full-Scale Combustor Scheme FS-03A 98
76 Burner Scheme Definition (Scheme FS-03A) 100
77 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, First Test Series 101
78 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, Second Test Series 103
79 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, Third Test Series 105
80 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, Fourth Test Series ; 107
IX
-------
ILLUSTRATIONS (Continued)
Figure Page
81 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, Fifth Test Series 108
82 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme PS-03A, Sixth Test Series 109
83 Exit Temperature Profiles (Second Test Series, Probe at Mid-Span) Ill
84 Variation in Temperature Pattern Factor With Overall Equivalence Ratio... 112
85 Variation in Quick-Quench Section Pattern Factor (TPFQQ) With Overall
Engine Ratio (Second Test Series) 113
86 Variation in Liner Temperature Rise Factor (LTRF) With Overall
Equivalence Ratio and Fuel Type 114
87 Condition of Premix Tube Swirler Following Tests With Shale Derived DFM 115
88 Condition of Premixing Passage Following Tests With Shale Derived DFM.. 116
89 Condition of Premix Tube Swirler Following Tests With No. 2 Fuel 117
90 Full-Scale Combustor Scheme FS-04A 119
91 Burner Scheme Definition (Scheme FS-04A) 120
92 ECV Combustor During Assembly 121
93 ECV Combustor Fully Assembled Except for Variable Damper 122
94 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04A Firing No. 2 Fuel 125
95 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04A Firing No. 2 Fuel With 0.5% N 126
96 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04A Firing Shale DFM 127
97 Evidence of Fuel Leak Caused by Cracked Manifold 128
98 Full-Scale Combustor Scheme FS-04B 129
99 Premix Tube With Variable Damper Attached 130
100 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04B Firing No. 2 Fuel With 0.5% N 132
-------
ILLUSTRATIONS (Continued)
Figure Page
101 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04B Firing Shale DFM 133
102 Variation in Emission Concentrations With Overall Equivalence Ratio for
Scheme FS-04B Firing No. 2 Fuel 134
-103 Comparison of NOX Characteristics at the Idle and 50% Power Settings;
Showing Variation With Fuel Type 135
104 Composite Results Showing Use of the.Premix Tube Damper to Vary NOX
Characteristics of the Combustion 137
105 Variation in Liner Temperature Rise Factor (LTRF) With Overall
Equivalence Ratio for Tests Conducted With the ECV Combustor... 139
106 Variation in Minimum NOX Concentration With Primary Residence Time.... 141
XI
-------
LIST OF TABLES
Table Page
I Design Requirements for the Full-Scale Prototype Combustor - 10
II Comparison of Measured and Predicted Cooling Air Flowrates 22
III Aerodynamic Model Calculations for Full-Scale Prototype Combustor 33
IV Aerodynamic Model Calculations for Full-Scale Prototype Combustor 33
V Aerodynamic Model Calculations for Full-Scale Prototype Combustor 35
VI Aerodynamic Model Calculations for Full-Scale Prototype Combustor 35
VII The Effect of Important Parameters on Droplet Size 38
VIII Summary of Full Residence Time (FRT) Combustor Design Features. , 46
IX Bench Premix Tube Tests Performed in Support of the Full-Scale Com-
bustor Verification Test Program 73
X Premix Tube Component Tests of Alternative Fuel Injectors 83
XI Premix Tube Design Review Summary 85
XII Rig Test Conditions Simulating Various Engine Power Settings 129
xn
-------
SUMMARY
This report describes an exploratory development program to identify, evaluate, and
demonstrate dry techniques for significantly reducing production of NO, from thermal and
fuel-bound sources in burners of stationary gas turbine engines.
Duty cycle analyses were conducted to identify current and projected dominant operating
modes and requirements of stationary gas turbine engines. These analyses indicate that the
propensity for NO, to be generated in combustors of stationary gas turbine engines will
increase significantly in the future as compression ratios and turbine inlet temperatures are
increased to improve thermal efficiency and net plant heat rate. In ten years, uncontrolled
thermal NO, generation is predicted to double over today's levels; in 20 years, the factor is
predicted to triple.
An extensive survey was made of candidate combustor design concepts and an analytical
study was accomplished from which those concepts considered to have significant potential for
reducing production of NO, were identified. The initial compilation of 26 design concepts
included many variations of basic strategies such as fuel-rich combustion, ultra lean combus-
tion, heat removal, fuel prevaporization, and fuel-air premixing. An assessment of the NO,-
control effectiveness of each concept was made using a combustor streamtube computer code.
The code employs a modular approach in the prediction of combustor emissions (NO,, CO, and
unburned hydrocarbons), with submodels for the internal flow field, physical combustion
(including droplet vaporization and droplet burning), hydrocarbon thermochemistry, and NO,
kinetics.
The results of the computer studies were drawn upon to select a group of concepts for
experimental screening in a bench scale combustor test rig. An erector-set approach was
followed in the experimental program, making possible the rapid evaluation of many different
concepts and combinations of concepts. About half the NO, reduction techniques evaluated
were based on fuel-lean burning, and half were based on fuel-rich burning. Two successful
approaches were ultimately identified, and their performance relative to the program goals was
assessed. It was concluded that one of the two concepts, referred to by the descriptive name
"Rich Burn/Quick Quench," showed significant potential for application in stationary gas
turbine engines, and was capable of meeting or exceeding all program exhaust emission goals.
Based on this assessment, the Rich Burn/Quick Quench concept was selected for
implementation into the design of a full-scale (25 megawatt engine size) gas turbine com-
bustor. In carrying out the full-scale design, reference was made to parametric data generated
in the bench-scale experimental program which showed an inverse relationship between NO,
concentration levels and combustor primary zone residence time. Because direct scaling of
combustor features cannot be employed, it was necessary to execute a separate but parallel
design in larger scale, reproducing the essential processes of the basic Rich Burn/Quick
Quench concept.
Two configurations of the full-scale prototype combustor were designed and constructed.
The first provided a primary zone residence time about half as great as that utilized in the
bench-scale combustor, but greater than that available in a representative 25 megawatt engine
having on-board (in-line) burner cans. The second configuration was shorter in length, meeting
the basic envelope requirements of the representative engine. Tests of the two configurations
were conducted to verify proper implementation of the design concept, and to demonstrate the
exhaust emission characteristics attainable in the full-scale design.
xin
-------
The test results were very positive, showing that the Rich Burn/Quick Quench concept
can produce substantial reductions in NO, for both nitrogenous and non-nitrogenous petrole-
um distillate fuels. All program exhaust emission goals were met. Comparison of the general
emission characteristics to those documented earlier for the bench-scale combustor showed
good agreement and indicated the same general dependence of NO, concentrations on primary
zone residence time. Extrapolation of the results to greater values of residence time indicates
that further substantial reductions in NO, can be achieved given increased combustor length.
A second major implication of the test results is that the Rich Burn/Quick Quench concept
may be directly applicable to heavy fuels. Having demonstrated substantial reductions in the
quantities of NO, formed due to fuel-bound nitrogen (which may be present in coal-derived
and shale-derived feedstocks), the Rich Burn/Quick Quench concept may hold the answer to a
second major difficulty, lack of fuel volatility. Under fuel-rich burning conditions NO, formed
initially due to heterogeneous burning of nonvolatile droplets can be reduced to N2. Operation
of the existing prototype combustor on heavy fuels (coal-derived, shale-derived or petroleum
residual fuel) may show substantial reductions in NO, when these fuels are fired.
-------
SECTION 1
INTRODUCTION
Gas turbine engines currently in use by the electric utilities and by industry account for a
relatively small portion of the total quantity of oxides of nitrogen (NO,) emitted from
stationary sources in this country. On a local scale, however, the gas turbine can be a
significant contributor to air quality degradation, especially in the vicinity of engine installa-
tions where the NO, background level is already objectionably high. The impact of stationary
gas turbines may become even more significant in the future. Along with the present modes of
utilization, combined cycle and industrial cogeneration applications are being projected. In
these applications the advanced engine technology needed to provide higher cycle efficiencies,
and to accommodate the anticipated firing of coal-derived, shale-derived, and petroleum
residual fuels, will make it more difficult to meet proposed emission regulations.
Until recently, gas turbine combustors have been designed without regard for exhaust
emissions. Initial attempts to control NO, by modifying existing designs were generally
unsuccessful. Although water injection was identified as a potential solution, this approach is
expensive and ineffective when nitrogen-laden fuels must be burned. In light of these findings,
it was clear that new design concepts specifically addressing exhaust emissions should be
considered.
Under EPA Contract 68-02-2136, an exploratory development program was undertaken
to identify, evaluate, and demonstrate alternative combustor design concepts for significantly
reducing the production of NO, in stationary gas turbine engines. The investigations were
directed toward dry combustion control techniques suitable for use in a 25 megawatt (nominal)
engine. Operation on both petroleum distillate fuels (non-nitrogenous and nitrogen bound) and
low Btu* gaseous fuels was specified. Program goals were 50 ppmv NO, (at 15% 02) for
non-nitrogenous fuels (oil and gas), and 100 ppmv NO, (at 15% 02) for oil or gas containing
0.5% nitrogen by weight. The goal for CO was 100 ppmv (at 15% O2).
Accomplishment of the overall objective was effected via complementary analytical and
experimental programs. Intrinsic in the support activities were combustor analytical model and
engine duty cycle analyses, bench-scale screening tests of promising NO, reduction concepts
and, finally, full-scale evaluation tests of combustors incorporating the most promising NO,
reduction techniques.
The program was accomplished in four phases. The first phase consisted of an analytical
investigation of combustion concepts considered to have potential for reducing the production
of NO,. In the second phase of work, a number of promising low NO, production concepts were
bench-tested to select the best candidate for implementation into the design of a full-scale,
25-megawatt-size, utility gas turbine engine combustor. In Phase III, a full-scale low NO,
combustor was designed and fabricated. Verification testing of the prototype combustor was
conducted in Phase IV, and guidelines regarding the applicability of the demonstrated low NO,
design technology to stationary gas turbine engines were generated.
'Refer to Appendix B for SI unit conversion
-------
SECTION 2
PHASE III COMBUSTOR DESIGN AND RIG PREPARATION
In Phase III, the design of a full-scale combustor incorporating the successful
NO, reduction concept demonstrated in the bench-scale screening experiments of Phase II was
carried out. This section describes the analytical and experimental procedures used in
preparing the design, and describes the fabrication of the prototype combustor.
2.1 REVIEW OF PHASE I AND PHASE II RESULTS
In the first two phases of work, a review and analytical study had been conducted to
identify concepts that might have potential for reducing the production of NO, from thermal
and fuel-bound sources of nitrogen in stationary gas turbine engine combustors. The most
promising of these had been evaluated experimentally in bench-scale hardware. Of two
successful design concepts that emerged from this study, the Rich Burn/Quick Quench concept
was selected as the basis for the full-scale combustor design executed in Phase III.
The key elements of the Rich Burn/Quick Quench concept are identified in Figure 1. A
premixing chamber is provided in which the fuel is prevaporized and premixed with air to
form a homogeneous rich mixture. The prepared mixture is introduced into a primary zone
section of the combustor and burned without the further addition of airflow. The rich burning
process is terminated in a final step involving very rapid dilution, which provides the airflow
needed to achieve an overall lean exit plane equivalence ratio. The success of this concept,
which does not differ in its essential features from many previous proposals for a rich-burn,
quick-quench approach to NO, reduction, was largely a matter of execution, and of the
selection and refinement of techniques for achieving the idealized conditions called for in the
basic concept.
Liquid Fuel
With Bound
Nitrogen
(Fuel Rich)
(Fuel Lean)
W//A
Air
Figure 1. Rich Burning Concept Burner Components
-------
The arrangement of representative rich burner bench-scale hardware tested in Phase II is
shown in Figure 2. A single, high-velocity premixing passage was provided, terminating in a
swirler that served to stabilize the flame in the primary zone of the combustor. All the air
entering the primary zone cam'e through the premixing passage. At design point, the primary
zone operated fuel rich. It was followed by a dilution section designed for very rapid quenching
of the fuel-rich gases leaving the primary zone.
Primary
Zone
Dilution Zone
Quick-Quench Slots
Figure 2. Rich Burner Arrangements
Tests of the rich burner were conducted at elevated pressures and temperatures sim-
ulating actual engine operating conditions. In Figure 3, data are shown from tests conducted at
150 psia, at inlet air temperatures of 650°F and 750°F. By staging the amount of air that
entered the premixing tube, it was found that low NO, concentration levels could be achieved
over a range of overall (exit-plane) equivalence ratios. At the primary air settings shown (7 and
14%), NO, concentrations of 60 ppmv and lower were demonstrated using No. 2 fuel with
0.5% nitrogen (as pyridine). Even lower concentrations were demonstrated using
non-nitrogenous fuel. In Figure 4, representative bench-scale data points are presented for the
Rich Burn/Quick Quench concept, demonstrating low NO, concentration levels over a wide
range of operating conditions using No. 2 fuel.
2.2 DESIGN APPROACH
The objectives adopted for the design of the full-scale prototype combustor reflect the
requirements of conventional gas turbine combustion systems (temperature rise, pressure drop,
and others), as well as the stated emission goals of the current experimental development
program. It was intended that the NO, reduction technology generated in this program be
compatible with current state-of-the-art design practice for stationary gas turbine engines in
the 25-megawatt-size range. The design requirements for the full-scale combustor are pres-
ented in Table I.
-------
600
400
| 200
*o
°x,100
o
~ 60
| 40
0)
o
O
NJ
O
10
0
7% Primary Air
, at 750°F
14% Primary
Air at 650°F
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 3. Rich Burner Simulated Engine Cycle Characteristics (150 psia, 0.5%
Nitrogen)
<*uu
1 200
Q.
a
§ 100
o
o* 60
« 40
u
2
o 20
10
Nitrogen
r~ -i
r~~*
. i
l-J
ous No. 2 Fuel
i
Q n
Clean No. 2
a
=uel
0.1 0.2 0.3
Overall Equivalence Ratio
0.4
Figure 4. Rich Burn/Quick Quench Combustor Emission Trends
-------
TABLE I
DESIGN REQUIREMENTS FOR THE
FULL-SCALE PROTOTYPE COMBUSTOR*
Type Combustor: can (1 of 8, internally mounted)
Basic Dimensions: 10-in. dia, 20-in. length
Design Point Requirements:
(Baseload) (Idle)
Airflow - 31 lbm/sec 7.8 lbm/sec
Pressure - 188 psia 40 psia
Inlet Temperature - 722°F 285°F
Temperature Rise - 1160°F 625°F
Pressure Drop: 3% combustor, 2.5% diffuser
Lean Blowout: 0.006 fuel-air ratio (burner exit)
Exhaust Emissions (max. at any setting):
NO,
CO
(0% Fuel N)
50 ppmv
at 15% O2
100 ppmv
at 15% O2
(0.5% Fuel N)
100 ppmv
at 15% 02
100 ppmv
at 15% 02
Execution of the design of the full-scale combustor was based in large part upon the data
generated in the bench-scale combustor program. It is important to point out that while these
results may have provided a full characterization of the bench-scale combustor itself, they
could not be used directly to specify the complete design of the full-scale combustor. Scaling
criteria dictate that there can be no exact and complete correspondence between a prototype
combustor and its subscale model, with regard to physical dimensions, operating conditions,
and combustion performance. In lieu of direct scaling, a partial modeling approach was taken,
as described in this section. In the basic features of the full-scale combustor, and in the areas
of primary air staging (to control stoichiometry), combustor aerodynamics, liner cooling, and
residence times, an attempt was made to reproduce the essential processes of the rich-burning
concept, as identified and defined parametrically in the bench-scale test results. The design of
the full-scale combustor was executed separately, drawing upon analytical modeling techniques
and upon the bench testing of key components (particularly the full-scale premix tube) to
verify that the essential processes of the concept were successfully reproduced.
2.3 BASIC DESIGN CONCEPT
The basic features and demonstrated results (from Phase II bench-scale testing) of the
Rich Burn/Quick Quench concept, in summary form, are as follows:
Arrangement Two combustion zones are arranged in series: a fuel-rich
primary zone and a fuel-lean secondary zone, separated by a reduced diameter
"quick-quench" section. A diagram of the bench-scale configuration as tested,
is shown in Figure 2, with the major zones identified.
'Refer to Appendix B for SI unit conversion
-------
Critical Features Four key requirements for low exhaust emissions have
been identified, using distillate and low Btu gaseous fuels:
All the air entering the primary zone must be premixed
with fuel to prevent the formation of an interface between
the desired homogenous, fuel-rich mixture and air; in par-
ticular, liner cooling airflow cannot be discharged into the
primary combustion region. This interface is considered to
be a region where diffusion burning (combustion occurring
at near stoichiometric conditions) predominates and has
been shown to result in a substantial increase in NO, in the
combustor exhaust (refer to the results of bench scale
Scheme 29-22A, Volume II of this report).
For fuels with bound nitrogen, minimum NO, concentra-
tion levels are obtained at primary zone equivalence ratios
near 1.3. Non-nitrogenous fuels exhibit similar NO, charac-
teristics (at lower levels) up to the minimum point of the
NO, "bucket." Beyond this point, the NO, concentration,
corrected to 15% O2, remains essentially constant with
increasing equivalence ratio. However, it should be noted
that equivalence ratios approaching and exceeding 1.5
could produce unacceptable quantities of smoke. This has
the implication that primary zone airflow staging would
not be necessary for non-nitrogenous fuels over the range
of primary zone equivalence ratios from a value giving
acceptable NO, (perhaps 1.1 to 1.2) to a value where smoke
formation is still acceptable (perhaps 1.4 to 1.5). In a
typical 25 megawatt gas turbine this would represent an
operating range (for acceptable emissions) from about 60%
of baseload to peak load conditions. However, the optimum
value of primary zone equivalence ratio is near 1.3 for both
non-nitrogenous and nitrogen bearing fuels.
Quick-quench air is added at a single site; it must be
introduced in a manner that produces vigorous admixing,
approximating a step change in composition and tem-
perature.
To attain acceptable CO concentrations at the burner exit
plane, temperature (and therefore, fuel-air ratio) must be
maintained high enough within the secondary zone to
consume the large quantities of CO discharged from the
fuel-rich primary zone. The temperature also must not be
allowed to become excessive as appreciable NO, formation
in the fuel-lean secondary zone would result. The tem-
perature range generally accepted for the oxidation of CO
without appreciable NO, formation with the residence time
constraints of onboard gas turbine combustors is from
about 2100 to about 2800°F. This implies the need to
rapidly dilute the fuel-rich products from the primary
combustion volume to a lean stoichiometry with a tem-
perature in that range. Consequently, it is desirable to
-------
introduce only part of the remaining airflow (not entering
the primary zone) into the quick-quench region, leaving a
final quantity to be introduced later in the secondary zone
to achieve the desired combustor exit temperature. It is
also acknowledged that, for ideal operation of the Rich
Burn/Quick Quench concept at low power settings (low
overall equivalence ratios), the quick-quench airflow should
also be varied in proportion to the amount of flow restrict-
ed from the primary zone in an effort to maintain efficient
combustion in the secondary zone to consume CO.
Emission Characteristics The emission characteristics, or "signature," of
the basic concept are shown in Figure 5, as generated at a constant airflow
setting by varying the burner fuel flowrate. This "signature" has two notable
features:
A peak in the GO curve, to the right of which (at 0.2 exit
plane equivalence ratio and higher) measured concentra-
tion levels are low;
A minimum point or "bucket" in the NO, curve, which
corresponds approximately to a primary zone equivalence
ratio of 1.3. The NO, curve "bucket" represents the unique
low emission design point of the basic emission signature.
It should be noted that the data shown in Figure 5 are
representative of fuels containing significant bound nitro-
gen. Non-nitrogenous fuels exhibit the same CO character-
istics and the same NO, characteristics (but at a lower
level) up to the "bucket." Beyond that point, the
NO, emission remains nearly constant with increasing
equivalence ratio.
Variable Primary Zone Airflow Variable geometry can be employed to shift
the low emission design point over a broad range of exit plane equivalence
ratios, as shown in Figure 3. As described in Reference 1, the NO, "bucket" can
be shifted in this manner while maintaining an essentially fixed CO character-
istic. Again, variable geometry may not be necessary for clean fuels over a
limited operating range.
Residence Time Requirements The minimum NO, concentration levels
attained (at the bottom of the NO, curve "bucket") have been shown to decline
with increasing primary zone residence time. This characteristic results in
basic design tradeoffs among primary zone length (residence time), combustor
pressure drop, and resultant NO, concentration levels.
-------
1000
Q.
a
O
x
>0
£
o
0.1 0.2 0.
Overall Equivalence Ratio
Figure 5. Rich Burner Characteristics (50 psia, 600°F, 0.5% Nitrogen
-------
2.4 INITIAL COMBUSTOR SIZING AND SELECTION OF BASIC FEATURES
To initiate the design of the full-scale combustor, studies were conducted to identify the
critical features of the bench-scale combustor and to determine what methods might be
employed to reproduce the essential processes of the Rich Burn/Quick Quench concept. As
stated, the bench-scale combustor hardware cannot be "scaled-up" directly to produce a
full-scale design. However, the parametric data from the bench-scale combustor can be used to
characterize the essential processes of the basic concept. To achieve emission characteristics in
full scale comparable to those demonstrated in bench scale, it is necessary to execute a second
design (in larger scale), reproducing the critical features of the smaller combustor and setting
up the same basic physical processes.
Preliminary design activity was directed toward defining the size and basic features of
the full-scale Rich Burn/Quick Quench combustor. Initial sizing calculations indicated that a
primary zone length roughly twice that of a representative engine combustor might be
necessary to achieve NO, concentration levels in the 50 ppmv range, when No. 2 fuel with
0.5% nitrogen is burned. This conclusion was based on bench-scale data in which the tradeoff
between primary residence time and the minimum achievable NO, concentration level had
been documented by varying the diameter and the length of the primary zone of the
combustor. The bench-scale combustor configurations tested are shown in Figure 6. Primary
zone diameters of three, five, and six inches were included. Two lengths were tested, 9 and 18
in. (measured from the premix tube swirler to the centerline of the quick-quench slots), and in
one configuration an enlarged premixing tube (designed to pass 70% more airflow) was
evaluated. The results obtained are presented in Figure 7 in terms of the tradeoff between the
minimum attainable NO, concentration* levels and the primary zone residence time.**
I
(^ -.
M-l_fl
Figure 6. Bench-Scale Test Configurations Used to Determine Primer Zone Residence Requirements
* The "minimum attainable NO, concentration" is measured at the bottom of the "bucket" in the characteristic
NO, curve of the Rich Bum/Quick Quench concept, illustrated in Figure 5.
** cold flow residence time
primary zone length
cold flow reference velocity
9
-------
Minimum NOX Concentration - ppmv at 15% C>2
|xj Ji. 05 00 . C
D O O 0 0 C
^^
\,
^-~
>
0°
^^^
600°F,50psia
4% Nominal Liner Pressure Loss (Except as Indicated)
Fuel:
No. 2 With 0.5% N
O Neat No. 2
^^
^O
-»._
o
0
0.04 0.08 0.12 0.16 0.20
Primary Zone Residence Time, sec (Cold)
0.24
0.28
0.32
Figure 7. Variation in Minimum NO* Concentration With Primary Zone Residence Time for Bench-Scale Tests of the Rich Burn/Quick
Quench Concept
-------
In the design of the full-scale combustor, the general relationship between residence time
and NO, concentration levels depicted in Figure 7 was adopted. It was assumed that the
absolute levels demonstrated in the bench-scale combustor (50 to 60 ppmv over a broad range
of operating conditions, as illustrated in Figure 4) were ultimately achievable in the full-scale
combustor. To select a design-point value of the primary zone residence time, several factors
were considered:
1. If an absolute value of residence time equal to that which had been utilized
in the bench-scale combustor were adopted, a primary zone length about
2.5 times greater than the nominal length available in a representative
25-megawatt engine combustor would be required.
2. Primarily because of an inherently lower surface-to-volume ratio, it was
reasoned that the full-scale combustor might not require the full value of
residence time established for the bench-scale combustor.
3. For an initial configuration, a value equal to half the residence time
utilized in the bench-scale combustor was selected.
4. Because more than one value of primary zone residence time would be
required to establish the exact residence time dependence (and to establish
whether the data being obtained fall on the negative slope portion or the
flat portion of the curve in Figure 7) it was decided that a second
configuration of the full-scale combustor, differing in primary zone length,
should also be tested.
Consideration was also given to the front-end configuration of the full-scale combustor.
By varying the number of premix tubes, it was possible to trade system complexity against
overall combustor length. A single premixing tube was less complex because of more straight-
forward variable-geometry actuating requirements (the valving of airflow would be required in
conjunction with only one premixing passage). On the other hand, multiple premixing tubes
(six, for example) would require a more complex mechanical system, but could offer reduced
length and might be expected to produce a more uniform fuel-air distribution within the
primary zone.
The preliminary design activity led to a "first cut" configuration of the combustor, shown
in Figure 8, which had the following basic features:
1. A single centrally mounted premix tube having a velocity versus length
schedule similar to that of the smaller tubes employed in the bench-scale
test program. Variable inlet vanes (not shown) were provided at the
premixing tube entrance to regulate the primary zone airflow. The premix
tube was offset slightly with respect to the centerline of the combustor in
order to be in-line with an engine diffuser passage.
2. An extended length primary zone (65% of the overall combustor length)
was provided for increased residence time.
3. A primary liner cooling scheme was provided that did not call for the
discharge of spent cooling air into the combustion region of the primary
zone. Airflow from the primary-liner convective cooling passage was dis-
charged into the aft dilution section through openings in the wall at the
dump plane of the combustor.
11
-------
Figure 8. "First-Cut" Configuration of Full-Scale Combustor
-------
4. The quick-quench section was designed to provide strong mixing, so that
an abrupt termination of the primary zone rich-burning process could be
achieved. An area ratio of 2.8 to 1 was adopted in the "necked-down"
section of the combustor, matching the optimum value determined in the
bench-scale tests.
5. The aft dilution section of the combustor was extended into the engine
transition duct to maximize the available length for oxidation of CO, while
still maintaining an extended-length primary zone in the interest of
achieving low NO,.
After consideration had been given to the sizing and basic features of the combustor,
subsequent design work was concentrated in two areas: the final refinement and verification of
the primary liner cooling scheme; and the evaluation of possible design tradeoffs that might be
made in the interest of reducing combustor overall length to conform to the available space
within a representative engine envelope. In the following paragraphs, a brief review of the work
performed in the primary liner cooling study is presented, and bench-scale rig data indicating
the tradeoff between secondary zone length and CO concentration levels are discussed.
2.5 PRIMARY ZONE LINER COOLING
An analytical effort was undertaken to design a convectively cooled combustor liner
compatible with the Rich Burn/Quick Quench concept. A scheme was required that did not
call for the discharge of spent cooling air into the combustion region of the primary zone. To
meet this requirement, the feasibility of utilizing impingement cooling was investigated
initially. Preliminary calculations were performed for the bench-scale burner rather than the
full-scale burner so that model predictions could be verified by bench-scale experimental data.
The heat load to the primary liner, under its most severe operating condition (unity
equivalence ratio), was predicted using a liner design computer program, which took into
account both convective and radiative heat transfer. At operating conditions of 50 psia and
600°F, the predicted heat load was 5 X 10* Btu/ft2 hr. In subsequent analyses, a second
computer code was used to predict the convective, heat transfer coefficient resulting from a
given impingement hole size, spacing and gap, assuming an allowable 1500°F metal tem-
perature on the outer surface of the liner. Intial results indicated that a hole diameter of 0.060
in., and a transverse hole and row spacing of 0.5 in. would be sufficient to cool a 10-in. length
of the primary liner without dumping any cooling air into the combustion region. This hole
pattern required approximately 25% of the burner airflow. The spent cooling flow would be
subsequently discharged into the burner as dilution air at the throat of the quick-quench
section.
A feasibility test of the impingement cooling concept was carried out in the bench-scale
rig. The burner configuration, shown in Figure 9, consisted of a double-wall primary liner
made up of concentric cylindrical/conical pieces. The outer piece contained a plurality of small
holes through which the liner cooling airflow entered, impinging on the surface of the inner
piece. The annular passage between the pieces led to the necked-down section of the burner,
where spent cooling air was discharged through the quick-quench slots. In the initial test of
this configuration, failure of the inner liner wall occurred. Examination of the hardware
indicated-;that the longitudinal ribs separating the inner and outer pieces had constrained the
inner wall, preventing thermal expansion. As a result, buckling of the inner wall occurred, and
the effectiveness of the impingement cooling technique was compromised leading to failure. A
second configuration was built up without longitudinal ribs. Upon retesting, the new liner also
exhibited signs of buckling, this time in the axial direction, which compromised the effective-
ness of the film-cooling process, and once again led to failure of the inner liner.
13
-------
Figure 9. Impingement Cooling Scheme Implemented in the Rich Burn/Quick Quench Bench-Scale Combustor
-------
Despite these outcomes, analytical predictions continued to indicate that the impinge-
ment cooling technique could meet the cooling requirements of the Rich Burn//Quick Quench
combustor. However, a review of the bench-scale test results indicated that other questions
remained to be answered, and that additional analyses and experimental verification tests
should be undertaken to verify that an adequate flowrate of cooling air had been provided. In
particular, it appeared that an expanded analysis of the aerodynamic characteristics of the
convective cooling channel was needed. A revised analytical procedure was formulated after
the analysis of Reference 2. The aerodynamic effects treated by the model included the
pressure losses arising from friction, heat and mass addition, and sudden expansion. Both
convective and radiative heat transfer processes were included. Predictions made using the
expanded model indicated that the proper flowrate within the convective cooling passage
(roughly 30% of the total burner airflow) could be readily achieved under the impingement
cooling scheme only if the cooling airflow were discharged at low velocity (to avoid an excessive
loss due to sudden expansion).
It was also indicated that the required cooling might be accomplished without the use of
impingement jets if other means of achieving adequate turbulence within the cooling passage
could be provided. One alternative suggested by the analysis was the use of swirling flow
within the passage. According to this concept, swirl vanes would be provided at the entrance to
the convective cooling passage. A "first-cut" configuration of this arrangement, shown in
Figure 10, illustrates the swirl cooling concept. To verify the results of the analytical studies,
and to assess the effectiveness of the swirl cooling technique, a short series of bench-scale
experiments was carried out. The data generated in these experiments were used as a standard
of comparison for the analytical model predictions with regard to the influence of burner
airflow rate, inlet pressure, and inlet air temperature on the primary liner wall temperature
level. Two liner convective cooling schemes were evaluated: a swirling scheme and a nonswirl-
ing scheme, shown in Figures 10 and 11. Both schemes were the same except at the entrance to
the cooling passage. A photograph of the experimental hardware is shown in Figure 12. The
tests conducted indicated that there was no appreciable difference in the cooling effectiveness
of the two schemes. Based on these results, the swirl cooling scheme was dropped from further
consideration.
15
-------
CONNECTIVE PASSAGE
SWIRL VANES
Figure 10. Liner Convective Cooling Scheme With Inlet Swirl for Increased
Turbulence (Scheme 29-73A)
CONVECTIVE PASSAGE
Figure 11. Liner Convective Cooling Scheme Nonswirling Case (Scheme 29-76A)
16
-------
Figure 12. Bench-Scale Combustor Configuration Used in Heat-Transfer Model Verification
-------
During testing with the nonswirl cooling scheme, five different cases were investigated to
assess the effect of pressure, mass flow and inlet temperature, and to compare experimental
data with data from the analytical model. Table II presents the five cases investigated and
compares the experimental cooling flow with the calculated cooling flow predicted by the
model. The cooling flow was determined experimentally with total and static pressure probes
mounted in the cooling passage. Agreement between experimental and analytical values was
within 10%. The inlet temperature effect on wall temperature is presented in Figure 13. The
experimental wall temperatures were determined by averaging two wall thermocouples located
5 and 7 in. downstream from the dome. The two temperatures measured are believed to be
indicative of the overall average liner temperature. Agreement between the model and the
experimental data is very close. An increase in inlet temperature from 400° to 600°F roughly
increased the maximum wall temperature (at an overall FA of 0.070) from 1300 to 1600°F.
TABLE II
COMPARISON OF MEASURED AND PREDICTED
COOLING AIR FLOWRATES*
Case
1
2
3
4
5
PB
(psia)
50
100
50
50
500
Temperature
CF)
600
600
600
600
400
Liner
Pressure Drop
(pet)
2.2
1.4
4.8
6.0
4.6
Cooling Air
Measured
(pps)
0.80
1.36
1.19
1.29
1.31
Flowrate
Calculated
(pps)
0.78
1.26
1.17
1.31
1.27
The mass flow effect on wall temperature is shown in Figure 14. Although the agreement
here is less satisfactory, the trends are believed accurate. As the mass flow was increased, the
convection heat transfer was increased roughly in proportion. However, the radiation from the
hot gas to the hot wall remained nearly the same. Since radiation accounts for a large
percentage of the heat transferred into the wall, and convective cooling accounts for most of
the heat removed from the hot wall, a trend of decreasing wall temperature with increasing
mass flow was considered logical. The pressure effect on wall temperature is presented in
Figure 15. The increased wall temperature due to increased burner pressure predicted by the
model was not verified by experimental data.
A common technique for enhancement of cooling effectiveness is to increase the surface
area on the cooling side of the hot wall. Cooling fins were analyzed to determine their
effectiveness. Heat transfer calculations performed for cast fins on the cooled side of the inner
liner indicated that a primary zone wall temperature of 1536°F could be achieved if 43.2% of
the total burner airflow could be made available to cool the primary liner. In order to ensure
that this relatively high percentage could be provided, it was necessary to make a revision to
the "first-cut" configuration of the Full-Scale Combustor (Figure 8) to allow discharge of the
primary cooling air through the quick-quench slots rather than through the sudden expansion
dump farther downstream. This arrangement, shown in Figure 16, was adopted in combination
with the cast-fin inner liner as the best available alternative for primary liner cooling.
'Refer to Appendix B for SI unit conversion
18
-------
1800
1700
50 psia Rig Pressure
OTlnlet = 600°F
AP/PT = 2.2%
ATlnlet = 400°F
AP/PT = 4.6%
Predicted
Curve
at 600°F
Predicted
Curve
at 400°F
Scheme 29-76A
I
1100
1000
0.04
0.05
0.06 0.07 0.08
Primary Zone Fuel-Air Ratio
0.09
0.10
Figure 13. Variation in Measured and Predicted Liner Temperatures With Inlet
Air Temperature and Fuel-Air Ratio (Bench-Scale Rig Data)
19
-------
1800
Tlnlet = 60° F; 50 psia Rig Pressure
Predictions:
AP/PT = 2.2%
Ap/PT = 4.8%
AP/PT = 6.0%
AP/PT = 2.2%
AP/PT = 4.8%
[7] AP/PT = 6.0%
Effected Through
Variations in
Burner Air Flowrate
Scheme 29-76A
1000
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Primary Zone Fuel - Air Ratio
Figure 14. Variation in Measured and Predicted Liner Temperatures With
Burner Air Flowrate and Fuel-Air Ratio (Bench-Scale Rig Data)
20
-------
1900
1800
1700
1600
2
1500
1400
1300
1200
1100
Tlnlet=600F
Predicted
Curve at
100 psia
Predicted Curve
at 50 psia
Q50 psia, AP/PT= 2.2%
A100 psia, AP/PT = 1.4%
Scheme 29-76A
I
0.04
0.05
0.06 0.07 0.08
Primary Zone Fuel-Air Ratio
0.09
0.10
Figure 15. Variation in Measured and Predicted Liner Temperatures With Rig
Pressure and Fuel-Air Ratio (Bench-Scale Rig Data)
21
-------
Figure 16. Final Configurations of the Full-Scale Prototype Combustor
2.6 RESIDENCE TIME CONSIDERATIONS
The "first-cut" configuration of the full-scale combustor represented a compromise
solution to the problem of achieving low concentration levels of NO, and CO within the limited
length of a representative engine combustor compartment. The configuration allocated most of
:the available combustor length to the primary zone in the interest of achieving lower NO,. It
was understood that this arrangement might result in higher CO concentration levels because
the rear section of the combustor had been radically truncated and the secondary zone had
been combined with the engine transition piece.
A brief series of bench-scale rig tests was conducted to generate data showing the tradeoff
between secondary zone length and CO concentration levels. The combustor configuration
tested (Figure 17) provided full power-range primary airflow (20% of total), and had no
dilution section except for the dump piece at the end of the slotted quick-quench section. Gas
sample measurements were taken at the dump plane (2 '/2 in. downstream of the slot
centerline), and at a location in the exit duct (9'/2 in. downstream of the slot centerline). Data
were already on hand for gas sample measurements taken at a far-downstream position
(approximately 8 ft downstream of the combustor, where a "fully-mixed-out" sample was
routinely measured).
Figure 17. Bench-Scale Burner Configuration
Length Studies (Scheme 29-77A)
Used in Secondary Zone
22
-------
The measurements taken at the dump plane showed very high CO concentration levels
(above the 3000 ppmv maximum analyzer range), indicating that the oxidation of CO had only
begun at this plane, as might be expected. The data recorded at the other locations are shown
in Figure 18. There is clear evidence that the CO oxidation process is a gradual one, which has
not been fully completed at the "upstream" sample location. By increasing the residence time
(through a lowering of the reference velocity), a lower concentration level was achieved.
However, this level still did not match the "fully-reacted" concentration level (approximately 8
ppmv) measured at the far-downstream probe.
400
300
Effect of Sample Probe Position and Reference Velocity
a.
a.
B 200
T3
0)
o 100
Upstream
Sample,
Ref = 60
Upstream
Sample,
VRef = 40 fps
. i Far Downstream
J>>f Sample
S/S.. ,-^ Jxv«,
0.3 0.4
(Schemes 29-57 A
and 29-77A)
Figure 18. Carbon Monoxide Characteristics of Rich Burn/Quick Quench
Combustor
These results indicated that the truncated dilution section of the "first-cut" en-
gine-retrofit combustor configuration was too short for the completion of the CO oxidation
process. This conclusion was in agreement with streamtube analytical model predictions, which
indicated that a 10- to 15-in. length would be required to produce CO concentrations in the 10
ppmv range. The exact tradeoff between secondary zone length and CO concentration levels
would, of course, have to be determined in rig tests of the full-scale combustor.
23
-------
With regard to the design compromises involved in reducing the combustor overall length
to conform to the available space within a representative engine combustion section envelope,
it was decided to adopt the tradeoff incorporated in the "final" configuration shown in Figure
16. The very short length of this engine-compatible version (ECV) of the full-scale combustor
was, however, viewed as an item of concern. In order to meet the reduced-length requirement,
residence times in the primary zone and, in particular, in the dilution section had been
decreased below the design-point values derived from bench-scale rig data. While these lower
residence times might eventually prove to be adequate for achieving the program exhaust
emission goals, a better demonstration of the basic Rich Burn/Quick Quench design concept
could almost certainly be gained by conducting tests of a "stretched" configuration of the
full-scale combustor. It was decided that a second configuration of the full-scale combustor
hardware, of different overall length, should be assembled and tested. In addition to the
engine-compatible version described (Figure 16), the full residence-time (FRT) configuration
was designed as shown in Figure 19.
The FRT and ECV combustor configurations differed in two main areas: (1) primary
zone length in the ECV was 12.5 in. compared to 18 in. in the FRT combustor; (2) a
louver-cooled dilution piece was added just downstream of the quick-quench section in the
FRT combustor, yielding an increase of 8 in. in the length of the secondary zone. It should be
noted that, even though the FRT combustor provided an 18-in. long primary zone and an
extended length secondary zone in comparison to the ECV combustor, neither configuration
was an optimum design in terms of attainable NO, emission levels. An increase in primary zone
length beyond that provided in the FRT configuration may exhibit further reductions in NO,
emissions.
2.7 PRIMARY AIR STAGING
The bench-scale test results from Phase II had consistently shown that minimum NO,
concentration levels were achieved when the primary zone equivalence ratio was maintained
near a value of 1.3. In order to achieve this value over a broad range of combustor exit plane
equivalence ratios (engine power settings), a method of varying the amount of airflow admitted
to the primary zone was required. At the baseload setting, slightly more than 20% of the total
combustor airflow is called for in the primary zone; at idle, approximately 10% is required.
The method of primary air staging selected for the full-scale combustor is depicted in
Figure 20. A variable damper, consisting of two sets of vanes (one movable, one fixed) was
mounted at the inlet plane of the premix tube. The variable damper can be adjusted to achieve
a 2:1 variation in premix tube airflow. At the full-open setting, only a nominal pressure drop
(less than 0.1%) was incurred by airflow passing through the vanes. A large number of narrow
vanes was employed to minimize wake formation in the incoming airflow. In going from the
full-open to the full-restricted setting, the total damper travel required was only about 10 deg
(or 0.25 in. at the maximum diameter).
24
-------
to
en
Figure 19. Final Residence Time Configuration of the Full-Scale Prototype Combustor
-------
Figure 20. Premix Tube Variable Damper Mechanism
2.8 COMBUSTOR INTERNAL AERODYNAMICS
The combustor internal airflow distribution is determined by several factors, which
include the relative areas of openings in the combustor liner, the pressure/velocity distribution
of the approach airflow, and the combustor internal geometry (cross-sectional area as a
function of length). The full-scale prototype combustor must meet a prescribed schedule of
internal equivalence ratios, and, therefore, must be designed for a specific internal airflow
distribution.
The Rich Burn/Quick Quench concept calls for a "necked-down" shape that produces
locally high velocities in a quick-quench section for the purpose of vigorous mixing. An
analysis of the effect of these high velocities on the combustor pressure drop and airflow
distribution showed that significant "mixing losses" are incurred in the quick-quench section.
These losses must be considered in tailoring the liner hole pattern to achieve the required
airflow splits (these mixing losses are believed to be desirable and, in general, to be indicative
of the high rate of mixing achieved in that section of the combustor).
To ensure an accurate determination of the liner hole areas required in the full-scale
prototype combustor, a computer model was formulated to simulate the aerodynamic processes
described above. The model accepts as input, a prescribed fractional airflow distribution, the
inlet air temperature and pressure, the fuel flowrate, and the required liner pressure drop. The
cross-sectional area profile of the combustor is also input, and an external pressure distribu-
tion may be specified. The calculation is performed in a downstream-marching fashion,
beginning with an initial guess for the premix tube airflow in Ib/sec. At each of several stations
along the length of the burner, the pressure drops associated with various components and
processes are computed. These pressure drops include the following: (1) premix tube entrance
and blockage losses (both at the variable damper and at the fuel injector); (2) swirler pressure
loss; (3) momentum pressure loss; (4) mixing loss in the quick-quench section; (5) mixing loss
in the dilution zone. At the exit plane, a check is made on the overall pressure drop. If it agrees
with the specified input value, the solution is complete. Otherwise, a new value for the premix
tube airflow rate is assumed, and the computation is repeated. The final solution includes the
total airflow that can be passed through the combustor for a given overall pressure drop and
specified distribution, and the schedule of hole areas required to achieve that distribution.
26
-------
Several cases were run with the aerodynamic model for the purpose of sizing the holes in
the quick-quench section of the combustor and in the dilution zone. The results verified that a
major source of combustor pressure drop is the "mixing loss" in the quick-quench section. The
model computes as "mixing loss" the total pressure drop due to mass addition (from the
one-dimensional momentum equation). In the quick-quench section, the mass added through
the penetration holes is assumed to have zero axial velocity. This flow must be accelerated,
along with the approach flow from the primary zone, to a uniform axial velocity consistent
with the cross-sectional area of the "necked-down" (quick-quench) section of the burner. The
smaller the diameter of the "necked-down" section, the greater the required acceleration, and
the greater the resultant total pressure drop.
The full-scale prototype combustor design called for a 6-in. dia quick-quench section (in
conjunction with a 10-in. dia primary zone section). The pressure drop incurred in this section
was substantial, according to the aerodynamic model. In order to pass the quantity of airflow
required in a representative engine, the model predicted that an overall combustor pressure
drop of 5.5% would be required. At the combustor design-point pressure drop of 3%,
calculations indicated that only 66% of the design-point airflow would pass through the
combustor. The controlling factor in these results was the "mixing loss" incurred in the
quick-quench section of the combustor. This section has a throttling effect on the combustor
flowrate. The higher the axial velocity in the "necked-down" passage (i.e., the smaller the
diameter), the lower the quantity of airflow (from both primary and quick-quench sources)
that can pass through that section without an increase in burner pressure drop.
To illustrate the results described, four of the cases run with the aerodynamic model have
been summarized and are presented in Table III through Table VI. In Table III, predictions
for the prototype combustor operating at 3% pressure drop and at a baseload power setting
are shown. The data include computed flow properties at selected stations along the length of
the combustor. The stations are identified in Figure 21. It may be seen from the tables that
there is a progressive decline in total pressure caused by the losses incurred at the various
stations. Table III and Table IV show cases for 3 % pressure drop, (at idle it was assumed that
the premix tube damper is adjusted to provide higher blockage). Table V and Table VI show
cases for 5.5% pressure drop. Note that the total airflow passed by the combustor at 3%
pressure drop, as shown in Table III (22 pps), is only about two-thirds the amount required
(31 pps) in a representative engine test. On the other hand, the amount passed at 5.5%
pressure drop (29.7 pps) closely approaches the requirement.
The predicted results were verified experimentally in tests of the bench-scale combustor,
as shown in Figure 22. Good agreement with the experimental data was demonstrated. The
predictions indicated that the selected diameter of the quick-quench section (6-in.) was too
small to pass the airflow required in an engine-compatible design. If ultimately substantiated
by test results, these results would dictate an increase in the diameter of the quick-quench
section. Calculations performed using the model also indicated that an increase in diameter to
8 in. would be required to provide full design-point airflow at 3% pressure drop. There was
another alternative as well. Full design-point airflow could be achieved using the selected
geometry if a pressure drop of 5.5% was available. This value coincides with the total
combustion system pressure drop (diffuser plus combustor) of the representative engine. By
placing ram scoops at the entrance to the primary liner cooling passage (which carries airflow
to the quick-quench slots), and by positioning the premix tube to capture high-velocity air at
the diffuser dump plane, it was reasoned that recovery of most of the compressor-exit total
pressure might be achieved, thereby making available to the primary and secondary zones of
the combustor a pressure drop nearly equal to the 5.5% value required.
27
-------
TABLE III.
AERODYNAMIC MODEL CALCULATIONS FOR FULL-SCALE
PROTOTYPE COMBUSTOR*
Configuration for 39f Pressure Drop
Baseload Power Setting (Damper Open)
Station
1
8
Wa(cum) pps 4.50 4.50 4.50 4.50 4.50 13.60 13.60 22.064
Equivalent ratio (local) 0.0 1.300 1.300 1.300 1.300 0.430 0.430 0.265
T, °F 722 722 722 3686 3686 2494 2494 1878
Ps psia 187.61 186.63 186.85 186.77 186.32 180.99 182.39 179.14
PT psia 188.00 187.41 186.95 186.85 186.85 184.14 182.74 182.36
Velocity fps 91.7 121.7 46.7 78.5 202.2 420.1 139.2 377.5
Mach No. 0.055 0.073 0.028 0.026 0.067 0.163 0.054 0.164
TABLE IV.
AERODYNAMIC MODEL CALCULATIONS FOR FULL-SCALE
PROTOTYPE COMBUSTOR*
Configuration for 3% Pressure Drop
Idle Power Setting (Damper at Minimum Setting)
Station
1
8
Wa(cum) pps 0.5558 0.5558 0.5558 0.5558 0.5558 3.286 3.286 5.536
Equivalent ratio (local) 0.0 1.300 1.300 1.300 1.300 0.220 0.220 0.131
Tt °F 285 285 285 3406 3406 1318 1318 921
P8 psia 39.98 39.58 39.59 39.58 39.55 38.58 38.81 38.24
PT psia 40.00 39.61 39.59 39.58 39.58 39.10 38.87 38.80
Velocity fps 33.5 44.2 17.4 38.2 102.8 285.6 97.1 262.1
Mach No. 0.025 0.033 0.013 0.013 0.034 0.141 0.048 0.146
* Refer to Appendix B for SI unit conversion
-------
w\A-m \H \-i V»T v*l * -i.x 4.*.*rr.i_«.e-r f z.r
23 4567
Figure 21. Identification of Stations Referred to in the Aerodynamic Model Calculations
-------
TABLE VI.
AERODYNAMIC MODEL CALCULATIONS FOR FULL-SCALE
PROTOTYPE COMBUSTOR*
Configuration for 5.5% Pressure Drop
Idle Power Setting (Damper at Minimum Setting)
Station
1
8
Wa(cum) pps
0.7394 0.7394 0.7394 0.7394 0.7394 4.3794 4.3794 7.4494
Equivalent ratio (local) 0.0 1.300 1.300 1.300 1.300 0.219 0.219 0.129
T, °K 285 285 285 3406 3406 1317 1317 914
Ps psia 39.97 39.23 39.24 39.23 39.18 37.40 37.R1 36.76
PT psia 40.00 39.29 39.25 39.24 39.24 38.36 37.93 37.80
Velocity fps 44.2 58.9 22.7 52.9 141.0 392.2 129.4 363.6
Mach No. 0.033 0.044 0.017 0.018 0.048 0.194 0.064 0.203
TABLE V.
AERODYNAMIC MODEL CALCULATIONS FOR FULL-SCALE
PROTOTYPE COMBUSTOR*
Configuration for 5.5% Pressure Drop
Baseload Power Setting (Damper Open)
Station
1
8
Wa(cum) pps 6.035 6.035 6.035 6.035 6.035 18.105 18.105 29.658
Equivalent ratio (local) 0.0 1.300 1.300 1.300 1.300 0.433 0.433 0.265
Tt °F 722 722 722 3686 3686 2505 2505 1875
Pa psia 187.30 185.72 185.94 185.78 184.97 175.15 177.74 171.71
PT psia 188.00 186.94 186.12 185.93 185.93 180.98 178.38 177.66
Velocity fps 123.4 163.4 61.7 108.7 271.6 578.4 191.1 524.4
Mach No. 0.074 0.098 0.037 0.036 0.090 0.224 0.074 0.228
'Refer to Appendix B for SI unit conversion
30
-------
Pressure
Drop - psi
^ Predicted
Experimental
O AP = 1.7psi
AP = 4 psi
2
1
12345 6 7
Station
Figure 22. Comparison of Predicted and Experimental Pressure Drop Character-
istics of the Bench-Scale Combustor
2.9 PREMIX TUBE
Good fuel preparation (effective prevaporization and premixing) is important in the
design of the full-scale combustor. If the airflow entering the primary zone has not been
sufficiently admixed with fuel to form a reasonably homogeneous mixture, it is possible that
diffusion burning could take place between the incoming air and the droplets or localized
pockets of fuel-rich gases already present. Because diffusion burning proceeds at near-peak
flame temperatures, it is the authors' opinion that significant concentration levels of NO, could
be formed in the primary zone under these circumstances. Such levels may not be reduced to
molecular nitrogen later in the combustion process.
In order to provide uniform premixing (and prevaporization) of the fuel and air that are
introduced into the primary zone, a number of candidate designs for the full-scale premix tube
were proposed and evaluated (both analytically and experimentally) during the Phase III
design effort. In the course of these evaluations, a considerable body of design data was
gathered. These data were assembled to form a premix tube design system. In this section, a
brief description of the design system is presented. Several of the premix tube designs
described in the discussions of the design system were evaluated in tests of the full-scale
combustor, and will be described further in Section 3.
31
-------
2.9.1 Atomization
Atomization of the liquid fuel and optimization of droplet sizes is important for two
reasons. First, fuel vaporization is dependent on fuel drop size; the smaller the fuel droplet, the
faster it vaporizes. Because vaporization is usually one of the attainable goals of a premix
system, atomization determines the premixing length required for vaporization. Second, even if
complete vaporization is not attained, it can be expected that very small droplets (< 20/um)
will behave like vapor in the combustion process as vaporization of a small droplet occurs
nearly instantaneously as it approaches the flame-front. Thus, small premixed fuel droplets in
air can approach the performance of a perfectly premixed, prevaporized system.
?.
'-''Of the various atomization techniques available, air atomization has perhaps the greatest
potential for producing fine droplets in premix tubes. In order to optimize air atomization,
thr.ee general directions of fuel injection or combinations thereof can be used:
1. Downstream axial injection (low fuel velocities)
2. Upstream axial injection
3. Cross-stream (radial or tangential) fuel injection.
All these types of injection provide a high relative velocity between the fuel and air, thus
promoting good atomization.
The empirical correlation for droplet size that follows was derived from References 3-12,
which include theoretical analyses and experimental data for liquid jets, sheets and droplets.
The correlation has the form:
SMD = K(df)>f)V,)cU)d(p.)°(V.)f
where K, a, b, c, d, e, and f are constants.
The correlation is a function of the following variables:
«f viscosity of fuel
a, surface tension of fuel
p, density of fuel
p, density of air
V. velocity of air (relative to fuel)
df characteristic initial dimension of fuel (diameter, thickness, etc.); in
this case, dt was taken as the diameter of the fuel orifice corrected
for the discharge coefficient (about 0.6).
W
.' atomizing airflow to fuel flow ratio.
W,
In.the references cited, other parameters have been shown to have a negligible influence on the
Sauter Mean iDiameter (SMD).
32
-------
The last parameter, W./W,, is a droplet interference and interaction term that can, under
certain circumstances, be eliminated from the list. If all of the airflow passing through the
premix tube is used in the atomization process, the air-to-fuel ratio can be expected to range
from about 10 in fuel-rich premix tubes (0 = 1.3) to about 20 in fuel-lean premix tubes. It has
been shown (Reference 3) that for values greater than five, the air-to-fuel ratio does not play a
significant role in the atomization process. In the design of full-scale combustor hardware, the
term Wa/W, was eliminated from the equation.
Table VII gives a list of the exponents a through f from the various references. In
reviewing the references, it was apparent that some of the constants were remarkably
consistent (particularly b, c, and f) while others varied. By the use of dimensional analysis,
three exponents can be calculated from three selected exponents. The following equation was
derived:
SMD
T/-/J \0.
K(df)
5/ \0.376/ \-0.125/ \~0.5/\T \-
(a,) (p,) (p.) (V.)
(1)
TABLE VII.
THE EFFECT OF IMPORTANT PARAMETERS
ON DROPLET SIZE *
Drop
2?
2f
2r
MMD
2f
2f
SMD
SMD
SMD
»
2f
SMD
a
0.5
0.5
0.16
0.375
0.375
' 0.166
0.375
b
0.33
0.66
0.25
0.34
0.25
0.25
0.333
0.25
c
0.16
0.33
0.25
0.41
0.50
0.375
0.33
0.375
0.33
0.50
0.375
d
-0.16
-0.33
-0.25
-0.84
-0.5
-0.375
-0.37
0.25
-0.16
-0.125
-0.125
e
-0.33
-0.66
-0.25
-0.25
-0.3
-0.875
-0.16
-IK66
-0.5
/
-0.66
-1.33
-0.75
-1.33
-1.0
-0.75
-1.0
-1.0
-0.75
-0.66
-1.33
-1.0
Reference
3
3
4
5
6
7
8
2
9
10
11
SMD = K(d,)-
The proportionality constant K was determined to be 48 in Reference 8. Equation (1) allows
the designer to predict actual SMD values, provided that the value of d, is known. Also,
equation (1) allows the designer to evaluate the effects of changing pertinent parameters. It
should be noted that air velocity is the single most important parameter in the atomization of
a liquid fuel. As a typical example, an air velocity of 400 fps at ambient pressure is predicted
to shatter a thin kerosene jet (0.062 in.) into droplets with a SMD of 16 nm.
Refer to Appendix B for SI unit conversion
33
-------
2.9.2 Distribution
In addition to atomization, the proper distribution of fuel in a premix tube must be
achieved. Poor fuel distribution results in incomplete atomization due to droplet interaction
effects, slower vaporization, and mixture nonuniformity. If a premix system is properly
optimized, the fuel must be uniformly distributed throughout the airstream by the time the
mixture enters the main combustor.
In tests of smaller bench-scale premix tubes (1-in. dia), experience has shown that
centrally mounted pressure-atomizing fuel nozzles are capable of properly distributing the fuel.
In tests of larger, full-scale premix tubes (3-in. dia), two techniques appear to offer potential
for a uniform fuel distribution. First, a centrally mounted injector can be used in combination
with an inlet-plane mixing device such as a swirler. An example of this type of fuel distribution
system is shown in Figure 23. The swirling airstream produced by preswirl vanes centrifuges
larger droplets outboard and transports smaller droplets by turbulence. Care must be exercised
in the design of this type distribution system both, in the avoidance of reverse flow zones and
the avoidance of excessive wall wetting by the fuel. Second, multiple injection sources can be
used with or without mixing devices. Figure 24 shows two premix tubes using multiple
injectors, one with and one without a mixing device.
Preswirl Vanes
Air Boost or Other
Fuel Nozzle
Figure 23. A Typical Centrally-Mounted Fuel Injector With a Mixing Device
Simple radial fuel injectors mounted on the wall of a cylindrical premix tube can also be
employed. This approach offers the advantage of providing a uniform fuel distribution without
the complexity of a mixing device. Radial injection also eliminates all internal blockage and
provides a "clean" premix tube design. However, the provisions for fuel penetration must be
carefully determined to properly distribute the fuel without excessive wall wetting. Designs of
this type can be undertaken using the three penetration design curves for radial fuel injection
from References 13, 14 and 15. These are shown to be in fairly good agreement in Figure 25.
Data from Reference 5 are also plotted in Figure 25.
34
-------
Sprayring Injector
Swirl Vanes
Centerbody
(a) Without Mixing Device
Swirl
Vanes
Radial Fuel
I njector
(b) With Mixing Device (Preswirl Vanes)
Figure 24. Representative Multiple Fuel Injector Premix. Tubes
35
-------
Schetz and Padhye
Kolpin, Horn and
Reichenbach
Ingebo and
Foster
Chelko
= Penetration of Distance
= Diameter of Jet
= Liquid Density
= Air Density
= Liquid Velocity
= Air Velocity
I
40
60
1/2
v°
Figure 25. Liquid Jet Penetration in Airstream
A promising candidate design for optimum fuel distribution was the radial "spoke" design
shown in Figure 26. Each spoke has multiple orifice injectors which tangentially feed the fuel
into the airstream. The injection system shown has 12 spokes and 36 individual orifices spaced
on an equal area basis. Reference 16 employed a similar fuel injection system and obtained
excellent premixing results. This design was evaluated extensively in tests of the full-scale
combustor described in Section 4.
36
-------
Spoke Fuel
Injector
Swirl
Vanes
3.2 in. dia
Figure 26. Radial "Spoke" Premix Tube Design
2.9.3 Pressure Loss
In order to design a premix tube that passes the desired airflow and meets the require-
ment for overall combustor pressure drop, an assessment was made of the pressure losses of
the various parts of the premix tube. Three major sources of pressure loss were identified in
premix tubes of the design shown in Figure 26: internal blockage loss, diffuser boundary layer
loss, and swirler dump loss.
In sizing the premix tubes used in this program, internal blockage loss was calculated
from the one-dimensional momentum equation. Diffuser boundary layer loss was calculated
from diffuser pressure recovery maps available in the literature. Swirler dump loss was
calculated from the one-dimensional equation of motion assuming a one-dynamic head loss
based on the discharge area of the swirler. By summing the losses of the various components
and iterating to a specific overall loss, the required "size" of the premix tube was determined.
2.9.4 Candidate Premix Tube Designs
The premix tube design system described was compared to the results of previous
development activity, both in-house and that reported in the literature, involving many
alternative premix tube configurations. Designs incorporating various fuel injectors (central-
ly-mounted, wall-mounted, spray bars, air-boost, air-blast, and pressure-atomizing), various
flameholding devices (inlet and exit swirlers, bluff bodies), and various provisions for
fuel-spreading (inlet swirl, vortex generators, and multiple-point sources) were represented in
the background data. It was found that while the design system outlines the principles and
techniques that should be employed in the execution of various designs, it does not provide a
means of selecting a specific combination of design features from the many alternatives
available.
37
-------
To ensure that a combination of features best suited for the particular application
intended are selected, it is essential that candidate premix tube designs be evaluated
experimentally. Ideally, a number of alternative configurations are selected for evaluation in
component tests prior to the actual mating of the premix tube to the combustor. In these tests,
the predicted performance of the designs can be verified, and a basis of demonstrated
performance can be established for selecting design features.
A number of alternative configurations for the premix tube of the full-scale combustor
were selected and evaluated experimentally as part of the design effort. In Section 4, the
chronological development of this component verification activity is documented. Among the
designs subsequently built and tested, are those depicted in Figures 23, 24 and 26. The premix
tubes, shown in Figures 16 and 19 in conjunction with the ECV and FRT combustor designs,
were the initial configurations proposed. The approach taken (experimental verification of a
number of proposed designs) ultimately produced a superior premix tube design, and provided
a means for identifying and correcting deficiencies in the configurations initially proposed.
2.10 CONSTRUCTION OF THE FULL-SCALE COMBUSTOR AND RIG HARDWARE
The purpose of the Verification Testing planned in Phase IV was to ensure proper
implementation of the Rich Burn/Quick Quench concept in full-scale burner hardware, and to
demonstrate (at intermediate pressure and under ideal air-feed conditions) a level of per-
formance consistent with program goals for exhaust emissions and conventional performance
requirements. The full-scale combustor hardware was constructed to facilitate the mod-
ifications that were anticipated during the test program. Bench-scale parametric data had
indicated that combustor residence times in excess of those available in current in-line engine
combustors might be required if very low concentration levels of CO and NO, were to be
achieved. Therefore, the combustor hardware was constructed in such a manner that two
configurations of different overall length could be assembled and tested. The primary liner was
made up of 4-in. long cast liner sections that could be welded together to form any desired
total length. Tests of both an engine-length combustor (the ECV configuration) and an
extended-length combustor (the FRT configuration) were planned. Modifications to the
configuration of the premix tube in the course of the test program were also anticipated,
because of the difficult design requirements for this component. A modular approach was
adopted allowing the premix tube and fuel injectors to be bolted to the combustor as a unit.
The final layout drawing of the FRT combustor prior to the start of the verification test
program is shown in Figure 27. The ram scoop shown at the entrance to the primary liner
cooling passage (which carries airflow to the quick-quench slots) was added as a means of
recovering a greater fraction of compressor-exit total pressure. This technique, in .conjunction
with good premixing tube pressure-recovery characteristics, would provide an effective com-
bustor pressure drop of nearly 5.5 %. As discussed in Section 2.8, analytical model predictions
had indicated that a pressure drop on that order would be required to overcome mixing losses
in the quick-quench section of the prototype combustor.
The design features of the full-scale (FRT) combustor are summarized in Table VIII. A
photograph of the FRT combustor during construction is shown in Figure 28. The premixing
tube and primary liner shroud were not attached in this figure. The fully-assembled configura-
tion is shown in Figure 29, except for the premixing tube damper mechanism. A view of the
damper is.shown in Figure 20. The ECV configuration of the full-scale combustor, which was
constructed by modifying the FRT combustor hardware, is described in Section 3.10.
38
-------
CO
CO
Figure 27. Final Layout of the Fail Residence Time (FRT) Configuration of the Full-Scale Combustor Prior to the Verification Test
Program
-------
TABLE VIII
SUMMARY OF FULL RESIDENCE TIME (FRT)
COMBUSTOR DESIGN FEATURES*
Type Combustor
Length (Primary)
Length (Dilution)
Length (Overall)
Outer Diameter
Inner Diameter
Combustor Reference Area (Primary)
Type Nozzle
Swirler
Combustor Material
Outer Liner
Inner Liner
Combustor Wall Thickness
Outer Liner
Inner Liner
Design Point Conditions
Fuel-Air Ratio
Volumetric Heat Release Rate Based
on:
Inlet Pressure
Combustor Airflow
Combustor Reference Velocity
(PrimaJry)
Combustor Total Pressure Loss
Combustor Can, Convective Primary
Zone Cooling, Film Dilution Zone
Cooling
19.0 in.
8.0 in.
45.0 in.
11.25 in.
9.8 in.
75.4 in. sq
Dual sprayring with 16 holes (0.030
dia)
4.874 in. OD, 0.56 in. ID, 20 vanes
with centerbody
Type 347 SST
Stellite 31 (X40)
0.0625 in.
0.125 in. on dia with 0.125 high fins
0.0189
2.05 X 106 Btu/(ft3-hr-atm)
188 psia
31.5 Ib/s
29.0 f/s
5.5%
*Refer to Appendix B for SI unit conversion
40
-------
Figure 28. Full-Scale Combustor During Assembly (FRT Version)
41
-------
Figure 29. Full-Scale Combustor Fully Assembled (FRT Version)
42
-------
2.10.1 Experimental Rig Hardware and Test Stand Preparation
A layout diagram of the combustor test rig is presented in Figure 30. The combustor was
mounted in a large-diameter cylindrical duct or plenum case and fitted to a sector-annular exit
transition liner. Exhaust flow from the combustor was discharged through the transition liner
into a traverse case, which contains a moveable probe with gas sample, temperature and
pressure instrumentation. Downstream of the traverse case, an exit transition duct was
provided, with a viewing port for monitoring the burner during intermediate-pressure testing.
A remotely operated backpressure valve was located in the exhaust duct to permit various
operating pressure levels to be set.
A continuous mixed-out gas sample was abstracted from the rig exhaust stream at a
location approximately six feet from the exit plane of the combustor. The abstraction of gas
samples was also provided for through the exit traverse probe.
Rig instrumentation was provided to measure pertinent airflow rates, the pressure and
temperature of the inlet air, the combustor pressure drop, the exit temperature pattern, wall
temperatures, wall static pressures, the fuel flowrate, as well as combustor exhaust emissions.
A schematic diagram of rig instrumentation is shown in Figure 31. A venturi meter was
provided for the total rig inlet airflow, covering a range from 5 to 25 pps with measurement
uncertainties of ±0.5%. Combustor inlet total temperatures and pressures are measured in the
inlet plenum at near-stagnation conditions. Three shielded chromel/alumel thermocouples and
three static pressure ports are provided. The thermocouple readings have associated uncertain-
ties of ±0.7% including test stand circuitry. Fuel flowrates were measured by turbine-type
flow transducers and were displayed on digital voltmeter readouts. Total uncertainty for these
instruments is ±0.5%. Rotameters were provided in the lines for approximate readings.used in
setting test-point conditions.
Combustor exit total temperature and total pressure measurements, along with con-
tinuous gas sampling for emission analysis, were taken with the traverse probe shown in Figure
30. Nine platinum/platinum-rhodium thermocouples were equally spaced between ten sample
ports on the traverse probe. For exit total pressure measurement, the gas sample line was
closed and a pressure transducer was used to measure the pressure. The gas sample traverse
probe was air-cooled to maintain a proper sample temperature.
The analysis of gaseous emissions from the combustor was accomplished using the system
shown in Figure 32. The gas sample was cooled in the probe to approximately 300°F, thereby
quenching high-temperature oxidation reactions, but maintaining an amount of heat adequate
to prevent the loss of unburned hydrocarbons by condensation. The gas sample was conducted
through an electrically heated transfer line to the gas-sample analysis system. The sample
transfer time was less than 2 sec. Instruments were provided for analyzing the different
constituent gases. Concentrations of unburned hydrocarbons were measured using a Beckman
402 flame-ionization detector. Concentrations of carbon monoxide and carbon dioxide were
measured by nondispersive infrared analyzers. Determinations of nitric oxide and total
NO, concentrations were made using a Thermoelectron chemiluminescent-type analyzer.
Concentrations of oxygen were determined using a Beckman Model Series 742 polargraphic
analyzer. All temperatures and pressures necessary for monitoring the operation of the
gas-samplng system were measured using instrumentation maintained in the gas analysis cart.
Filters and gas driers were located within this system to ensure the proper conditioning of the
exhaust gas sample. The calibration gases were traceable to National Bureau of Standards
reference material. Check calibrations of the testing standards against the primary standards
were made periodically to ensure their continued accuracy.
-------
Figure 30. Layout of the B-2 Rig Showing the FRT Combustor
-------
Gas Sample Traverse Probe-
Burner Skin Temperature (BST1-20)-
Plenum Total Temperature (TT3-1, 2, 3)-
Plenum Static Pressure (PS3-1, 2, 3, 4).
Venturi Throat Metal
Temperature (V5VMT).
Total Pressure Upstream of
5 in. Venturi (V5PTI1) (V5PTI2)
Gas
Combustor Exit Total Temperature (TT41-49)
Exit Transition Duct
Sample Rake
Spray Water
Stand Duct (Ambient)
Total Temperature Upstream of
5 in. Venturi (V5VTT1. 2)
Traverse Case Heated Transfer Line j /Sample Gas
Temperature
(TSG1-3)
Static Pressure at
Venturi Throat (V5VPS-1. 2)
Primary Fuel
Manifo]d_Pressure PF1
Primary Fuel
Temperature (TF1, TF2)-
in. Venturi Flow Conditioner
Emission Sampling Mobile Cart-
Scanivalve Control-
Counter
At Control Room
Combustor Primary Fuel Inlet Pressure (PF1)
Combustor Primary Fuel Flowrate (WF1, WF2)
Combustor Primary Fuel Temperature (TF1, TF2)
Preheater Fuel Flowrate (WFHB)
18 - Pin Cables (14)
Lewis Switch
Sample Gas
Temperature
(TSG4)
DVM
Control Box (In Control Room)
Figure 31. Schematic Diagram of Rig Instrumentation System
-------
3-Way Valves
(Air-Operated)
Q Temperature Sensor
(p) Pressure Sensor
GN Out
Samplt **w"
Parteuiat* Trap
CaMxatad Orificm
Ov*n Tamparatur*
Hydrocarbon AnatyM*
to>
HO. An«yi«>
Moor*
Carbon UonoiMl* Anatyz
Carbon DM»K>* Anafyr**
P*rm*ai>on Tub* |0>>«»
Ourrv Manifold
GMton Frtiar
Oiygw Ar
Figure 32. B-2 Sample-Gas Analysis System
Combustor instrumentation was provided to measure primary zone airflow, primary zone
cooling airflow, primary zone skin temperatures, and dilution zone skin temperatures. Primary
zone airflow was calculated using the static pressure measured at the throat of the premix
tube, the upstream total pressure measurement, and the calibrated cross-sectional area of the
premix tube. Cooling shroud airflow was determined from total and static pressures measured
in the cooling passage. Approximately twelve chromel/alumel thermocouples were mounted on
the primary zone liner wall to monitor liner wall temperatures. The dilution zone was
instrumented with approximately eight chromel/alumel thermocouples.
46
-------
SECTION 3
PHASE IV VERIFICATION TESTING
In Phase IV, the experimental evaluation of the full-scale combustor was accomplished.
As in the bench scale test program, tests were conducted at a nearly constant airflow setting
(constant pressure drop) while fuel flow was varied to map an emission "signature," in an
attempt to study the basic characteristics of the combustor. In an engine, pressure, tem-
perature, and airflow vary with power setting. Both the Full-Residence-Time Version (FRT)
and the Engine Compatible Version (ECV) of the combustor were tested to obtain an emission
"signature" at several points over a range of conditions spanning the operating requirements of
a commercially available 25 megawatt stationary gas turbine engine. Three fuels were used in
the test program: No. 2 distillate; No. 2 with 0.5% N (as pyridine), and a distillate cut shale
oil.
The experimental program consisted of two parallel parts: component tests of various
configurations of the full-scale premix tube (involving cold flow calibration and preliminary
combustion tests), and verification testing of the complete full-scale combustor.
Component tests of the various candidate premix tube designs were conducted initially in
support of the full-scale combustor design effort. As described in Section 2.9.4; preliminary
component tests of this type serve to verify the predicted performance of a candidate premix
tube design. In the course of the full-scale combustor verification test program, additional
component testing of modified or alternative premix tube configurations was also carried out
to verify proper functioning of revised designs.
In Section 3.1 the experimental procedures used in the evaluation of the premix tube
designs are described in their entirety. The verification tests of the complete full-scale
combustor and the related data analysis are then described in their entirety in Section 3.2. The
discussion of each of the two parallel verification efforts is arranged in chronological order
within its own subsection.
3.1 PREMIX TUBE COMPONENT TESTS
Tests were conducted to verify the performance of candidate premix tube designs prior to
their use in the full-scale combustor. This extensive experimental effort was conducted
initially in support of the combustor design effort, and later in support of the full-scale
combustor verification test program. In this subsection the premix tube tests are described in
chronological order and grouped according to the immediate objectives of the experiments.
3.1.1 Initial Design Verification
Several configurations of the preliminary premix tube designs proposed for use in the
ECV and FRT combustors (depicted in Figures 16 and 19) were evaluated experimentally in a
series of cold flow and combustion tests following initial fabrication. These preliminary tests
were conducted to verify the general performance of the initially proposed configurations.
First, an investigation was conducted to determine the patterns of dispersion of liquid
produced in the premixing passage by three candidate fuel injectors. Two centrally mounted
fuel nozzles were evaluated: (a) an air-blast design; and (b) a simplex pressure atomizing
nozzle (85 GPH, 80 deg cone angle). An initial concern in the design of the premixing tube had
47
-------
been the possibility that a centrally-mounted fuel injector might not disperse fuel over the
entire cross section of the premixing tube. To help ensure that a design providing complete
dispersion would be available, an alternative configuration, consisting of a spray ring injector,
was also fabricated and tested.
The three injectors were evaluated in cold flow tests of a wooden premix tube model at
ambient air temperature and pressure, using water as the test fluid. Air velocities were set to
match the engine design requirements. The distribution of liquid in the flow field (measured at
the premix tube exit plane, with the swirler removed) was determined by collecting water
samples at various radial locations using a point source probe. The results of these measure-
ments were as follows:
1. The centrally mounted pressure atomizing nozzle did not fully disperse the
liquid over the entire cross section of the flow field. As shown in Figure 33,
a sharply center-peaked profile was obtained.
2. The centrally mounted air-blast nozzle produced a similar pattern, with a
somewhat higher concentration of liquid toward the center. Photographic
documentation of this result is shown in Figure 34.
3. The sprayring injector produced a nearly uniform distribution over the
entire cross-section, as shown in Figure 35.
Diametrical Traverse
Exit Plane of
Premixing Tube
(No Swirler)
Nominal Design
Point Air Velocity
I
0
-2
50 -2.00 -1.50 -1.00 -0.50 Q 0.50 1.00 1.50 2.00 2.50
Radial Position - in.
Figure 33. Measured Distribution (Normalized) of Liquid Obtained Using
Simplex Pressure Atomizing Fuel Injector
48-
-------
Figure 34. Liquid Distribution Pattern Produced by Centrally-Mounted Air-
Blast Nozzle (Nominal Design Point Air Velocity, Equivalent
Ratio = 1.0)
Figure 35. Liquid Distribution Pattern Produced by Spray-Ring Injector (Nomi-
nal Design Point Air Velocity, Equivalent Ratio = 1.0)
49
-------
Dispersion pattern test results clearly indicated that the sprayring design produced a
more uniform distribution than either of the centrally mounted injectors. Based on these
findings, effort was focused on the refinement of the sprayring design, with a view toward its
ultimate use in the full-scale combustor.
Combustion tests of the premixing tube and swirler were also conducted with each of the
three fuel injectors. In these tests, the premixing tube was secured to an 8-in. dia sheet-metal
liner as shown in Figure 36. The other configurations evaluated were as follows.
1. The air-blast nozzle in conjunction with a larger diameter premixing tube
and swirler (Figure 37). An increased diameter became necessary in light of
the analytical predictions described in Section 2.8, which indicated that
the entire burner pressure drop of three percent would be unavailable to
the premixing tube (because of the "mixing loss" incurred in the
quick-quench section), and that a larger effective swirler flow area would
be required.
2. The simplex pressure atomizing nozzle (Figure 38) in the large diameter
premixing tube.
3. The sprayring injector (Figure 39) in the large diameter premixing tube.
Figure 36. Combustor Test Configuration
Nozzle
Centrally-Mounted Air-Blast
50
-------
Figure 37. Combustor Test Configuration Centrally-Mounted Air-Blast
Nozzle in Large Diameter Premix Tube
Figure 38. Combustor Test Configuration
Nozzle
Simplex Pressure Atomizing
51
-------
Figure 39. Combustor Test Configuration Spray-Ring Injector
The premixing tube assembly was mounted in the bench-scale rig, and tests were
conducted at 50 psia and 600°F inlet air pressure and temperature. In these tests the entire
complement of rig airflow was passed through the premixing tube; there was no dilution of the
primary zone exhaust products. Exhaust emission data were recorded over a range of
equivalence ratios from about 0.4 to 1.4.
Representative NO, data are shown in Figure 40 for the air-blast nozzle, and in Figure 41
for the simplex nozzle. The curves each exhibit peak concentrations near an equivalence ratio
of 1.0, as expected. At fuel-rich equivalence ratios moreover, the two curves are nearly
identical. However, the air-blast injector exhibited a much sharper rate of increase in NO, at
fuel-lean equivalence ratios, and a higher peak concentration. These results were interpreted as
an indication that the air-blast injector had provided slightly better premixing (therefore
exhibiting a steeper NO, curve indicative of premixed burning as opposed to diffusion
burning).
NO, emission data were not obtained for the sprayring injector because of flashback
conditions encountered during the test of that piece. Flame was held upstream of the swirler,
leading to its complete destruction. The cause of the failure was determined to be an incorrect
angle of divergence of the aft section of the premixing tube. Although called out as 6 deg, the
angle was measured and found to be 11 deg. This excessively large angle of divergence was
believed to have caused flow separation and flashback. To correct this problem, the aft section
of the premixing tube was rebuilt to the correct specifications.
52
-------
OOU
320
£ 280
6s" 240
a*
- 200
| 160
o
£
fc 120
0
0* 80
z
40
0
(
./
Y
£
c~v
/
f
T
\
G
\2
^X
>^
U
) 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.<
Equivalence Ratio
Figure 40. Variation in NO, Concentration With Equivalence Ratio for Pro-
totype Premixing Tube With Centrally-Mounted Air-Blast Nozzle
(Scheme 26-05A)
320
280
240
200
160
1
Q.
Q.
CM
O
in
o
o
o
X
9 40
80
GT
0.2 0.4 0.6 0.8 1.0
Equivalence Ratio
1.2
1.4
1.6
Figure 41. Variation in NO, Concentration With Equivalence Ratio for Pro-
totype Premixing Tube With Simplex Pressure Atomizing Nozzle
(Scheme 26-06A)
53
-------
Further tests were conducted to verify proper functioning of the rebuilt hardware, and to
evaluate other features of the premixing tube design, particularly those associated with the
sprayring injector. Initially, six series of cold flow tests were performed, in which the profiles of
total and static pressure were measured at various locations in the flow field of the premixing
tube. It was verified that the rebuilt aft section of the premixing tube did not generate
separated flow. However, several features of the sprayring injector produced wake regions of
appreciable size in the flowstream. These included the fuel injector ring, which was made of
0.32 in. OD tubing, and to a degree, the inner and outer splash plates (see Figure 42).
Modifications were made to both these features as follows (see Figure 43):
1. The fuel injector ring (0.32-in. diameter) was replaced by two stacked fuel
injector rings of 3/16-in. diameter. A wedge-shaped trailing-edge piece was
also added to help prevent separation.
2. The outer splash plate was replaced by an outer conical partition extend-
ing upstream nearly to the inlet plane, and downstream to the throat. It
was reasoned that fully developed flow would be established on both sides
of the partition, preventing the formation of a wake at the trailing edge. At
the same time the partition would still function as a splash plate, and
would prevent fuel wetting of the premix tube wall.
3. The inner splash plate was removed altogether. It was reasoned that
jet-on-jet impingement occurring in the high-velocity airstream at the
center of the premixing passage might provide adequate atomization.
Fuel Injector Ring With
ID and OD Splashplates
Figure 42. Premixing Tube Scheme 26-07A Showing Initial Design of the Spray-
Ring Fuel Injector
-------
Dual Tubes With
Trailing Edge Piece
(Scheme 26-07C)
Conical Partition
Figure 43. Intermediate Configuration of the Fuel Injector Spray Ring
The modifications described were among those evaluated in subsequent cold flow tests.
Pressure traverse data indicated that the wakes associated with the fuel injector ring had
effectively been eliminated. However, there were still local regions of low velocity associated
with the trailing edge of the conical partition. It appeared that a mismatch of the velocities on
either side of partition tended to develop as a result of small nonuniformities in the approach
airflow.
Subsequent modifications to the conical partition produced no apparent improvement,
and the piece was finally eliminated in favor of an outer splash plate similar to the original
design. Pressure traverse data indicated that no significant wake regions were generated by the
new splash plate/injector arrangement.
The revised configuration, shown in Figure 44, was evaluated in combustion tests. The
premixing tube was mounted in the bench-scale rig, and tests were conducted at 50 psia and
600° F inlet air pressure and temperature. The entire complement of rig airflow was passed
through the premixing tube; again there was no dilution of the primary zone exhaust products.
Exhaust emission data were recorded over a range of equivalence ratios from about 0.4 to 1.4.
55
-------
Figure 44. Revised Design of the Splashplate/Injector Ring Arrangement
Representative NO, data are shown in Figure 45. Included for reference are the data
previously shown for the original centrally mounted airblast fuel nozzle. The two sets of data
are nearly identical at fuel-rich equivalence ratios. However, the sprayring design produced a
significantly lower NO, concentration at the single fuel-lean condition tested. These results
were taken to indicate that the sprayring injector provided better premixing than the airblast
nozzle during fuel-lean operation. More significant, however, was the fact that refinements
made to the premixing tube itself (correcting the angle of divergence), and to the sprayring
successfully eliminated the flashback problem previously encountered.
In the initial experiments described, proper functioning of the premix tube with regard to
flashback-free operation, an even fuel distribution, and good exhaust emission characteristics
had been verified. The tests had been performed using the basic premixing tube without the
variable damper piece. Furthermore, the combustion tests had been made at elevated pressure
in a rig chamber having no provision for visual observation of the general flame appearance.
Further tests were conducted in which cold flow pressure measurements were made with
the variable damper installed, and combustion tests were conducted at atmospheric pressure in
an ambient discharge rig allowing visual observation of the premixed flame. In addition to the
basic premix tube previously tested, an extended-length version, providing higher throat
velocities for improved atomization and longer length for increased residence time to allow
more fuel vaporization to occur, was also designed and evaluated.
56
-------
360
320
280
E
Q.
?- 240
CM
O
In 200
160
u
CD
_
cS 120
80
40
0
O
Revised Sprayring
O Air Blast Nozzle Data '
(Previously Reported)
0 0.2 0.4 0.6 0.8 1.0 1.2
Equivalence Ratio
1.4 1.6
Figure 45. Comparison of Variation in NO* Concentration With Equivalence
Ratio for Revised Spraying Design and Centrally-Mounted Air-Blast
Nozzle
The experiments carried out and pertinent results were as follows:
1. The variable damper device was evaluated in conjunction with the basic
premix tube in the configuration shown in Figure 46. First, cold flow
measurements were made of the profiles of total and static pressure in the
premixing passage (at the throat of the venturi). In these tests the fuel
injector was removed so that flow field characteristics due to the variable
damper could be determined. With the damper in the full-open position it
was found that there was no serious disruption of the flow field. However,
with the damper fully restricted, there was evidence of reverse flow, as may
be seen from the data presented in Figure 47. The test piece was examined
and a slight misalignment was found between the fixed and movable
damper plates. This misalignment was believed to have produced a non-
uniform circumferential distribution in the flow field, and may have been
responsible for the reverse flow observed. Another possible contributing
factor was a step discontinuity that was present in the wall of the
premixing tube, at the inlet plane (see Figure 46). Both these conditions
were corrected prior to subsequent combustor tests with the damper in the
fully restricted position.
57
-------
, Step Discontinuity in Wall of
Premixing Tube
(Scheme 26-09A)
Figure 46. Basic Initial Premix Tube Configuration With Variable Damper
Mechanism Installed
2. Combustion tests were performed initially using the configuration tested in
the cold flow experiments, with the damper in the fully open position.
Visual observations of the flame indicated that there was a slight concen-
tration of fuel toward the center of the premixing passage. There were also
indications of isolated regions of liquid burning (regions of luminous
flame). Otherwise, the appearance of the flame was generally acceptable.
3. The premixing tube was modified as shown in Figure 48. An inner
splashplate was added to the fuel injector sprayring in an attempt to
eliminate the central region of high fuel concentration observed in the
previous scheme. The wall of the premixing passage was also reworked to
eliminate the step discontinuity at the inlet plane. Combustion tests were
performed with the damper in the open position. It was found that the
central region of high fuel concentration had been dispersed slightly,
taking on an annular rather than a cylindrical form (as evidenced by the
appearance of the region of luminous flame). The fraction of the flame
observed to be luminous was somewhat diminished with respect to the
previous scheme.
58
-------
5yo
D
low
-^
0
u
n
u u
-40
-60
-80
-100
-1.75 -1.50 -1.25 -1.00 -0.75 -0.50 -0.25 CL 0.25 0.50 0.75 1.00 1.25 1.50 1.75
Radial Position - in.
Figure 47. Pressure Transverse Data for Basic Premizing Tube With Variable Damper
-------
False Wall Added to
Eliminate Step Discontinuity
Inner Splashplate Added
8 in. Dia Duct
(Scheme 26-10A)
Figure 48. Modified Configuration of the Basic Initial Premix Tube Configuration
-------
4. Combustion tests were performed again using the same premixing tube
configuration, after modifications had been made to the test rig to
eliminate a condition of acoustic resonance. (In atmospheric combustion
tests resonant conditions often occur and can affect the combustion
process; changes in the dimensions of the burning duct usually eliminate
the problem.) The modified configuration is shown in Figure 49. Tests
conducted with this arrangement were free of acoustic resonance. There
was no apparent change in the combustion process with respect to the
previous test series.
10 in. Dia Duct
(vs 8 in. Dia Previously)-
(Scheme 26-11 A)
Figure 49. Burner Duct Modified to Eliminate Acoustic Resonance
5. An extended-length premixing tube was also evaluated in combustion tests
at atmospheric pressure. The configuration, shown in Figure 50, differed
from the short-length design previously evaluated in that a longer diver-
gent section was provided. Given the same swirler diameter and the same
angle of divergence (these two specifications are nominally the same for
the short and long premixing tubes), the longer divergent section made it
possible to specify a smaller diameter for the premixing tube venturi
throat. The smaller diameter was expected to produce a higher mixture
velocity and to result in better fuel atomization and a wider margin against
flashback. The increased length also provided a longer residence time for
more complete fuel vaporization. The extended-length configuration was
constructed as an alternative (and more conservative) approach to provid-
ing high-quality fuel-air mixture preparation. Visual observations of the
61
-------
flame made during initial combustion tests indicated that very high quality
premising had been achieved. There was a total absence of luminous flame
at fuel lean equivalence ratios. At fuel-rich settings (slightly beyond the
design point equivalence ratio of 1.3) the flame became orange while
retaining the same texture and semitransparent quality associated with a
blue flame under fuel-lean conditions. These observations indicated
near-perfect premising and a high degree of fuel vaporization.
(Scheme 26-12A)
Figure 50. Extended-Length Premix Tube
A photograph of the flame under fuel-rich conditions is presented in
Figure 51. (Although reproduced in this report in black and white, the
uniform consistency and absence of luminous flame are apparent.)
The extended-length premixing tube was also tested with the variable
damper fully restricted (simulating idle conditions). Visual observations
indicated the same excellent premixed flame reported in the previous test
series.
Based on the very good overall performance of the extended-length premix tube, the
configuration shown in Figure 50 was selected for use in the full-scale combustor verification
test program.
Further component tests were conducted to calibrate the airflow capacity of the premix
tube as a function of the pressure differential between inlet stagnation pressure and the static
pressure at the throat of the premixing passage. The measurements taken provided an
indication of the total pressure loss taken across the variable damper, and were necessary for
use in the computation of the premix tube airflow at various damper settings during the
full-scale combustor verifications tests. A calibration graph was generated (Figure 52) and
incorporated into the data reduction program.
62
-------
Figure 51. Premixed Flame Produced by Extended-Length Premising Tube
During Ambient Operation (Nominal Design Point Air Velocity
Equivalence Ratio =1.4 Nominal)
63
-------
0.1?
0.10
0.08
0.06
CO
Q_
0.04
0.02
Damper Closed
PT = Constant
Damper Open
Nomenclature
P-p - Combustor Inlet Total Pressure
Pg - Static Pressure at Throat
of Premix ing Passage
Figure 52. Premix Tube Airflow Calibration of Extended-Length Tube
-------
3.1.2 Bench Tests in Support of the Full-Scale. Combustor Test Program
During initial verification tests of the full-scale combustor (described in Section 3.2.1)
inadequate performance of the extended-length premix tube was observed in terms of
poor-quality fuel preparation and the occurrence of flashback. To determine whether the
difficulties encountered were due to deficiencies in the design of the premix tube (and not due
to rig-related causes) further component tests of the extended-length configuration were
conducted.
Both flow visualization and combustion tests were performed, as summarized in Table
IX. Initially, combustion tests were conducted using Scheme 26-13A, the actual configuration
tested in the full-scale combustor. These tests were intended to resolve the apparent dis-
crepancy between the poor premixing quality observed in the full-scale combustor test and the
excellent quality observed in the original bench tests of the same premix tube (Figure 51).
Visual observations made during the repeat premix tube test indicated that the flame quality
was poor, having a luminous appearance indicative of diffusion burning. This result matched
that of the full-scale combustor test, and was much worse than the original bench test results.
Because contaminated fuel had at one point inadvertently been introduced into the full-scale
combustor rig, resulting in partial plugging of the fuel injector sprayring, it was postulated that
some foreign material might still be present (even though the sprayring had been thoroughly
back-flushed following the incident with contaminated fuel and visual spray tests performed
which indicated that all jets were flowing), causing a maldistribution of fuel in the premixing
passage. Rather than cut apart the sprayring to perform a thorough examination, it was
decided to fabricate a new fuel injector assembly identical to the original.
TABLE IX
BENCH PREMIX TUBE TESTS PERFORMED IN
SUPPORT OF THE FULL-SCALE COMBUSTOR
VERIFICATION TEST PROGRAM*
Scheme
Type
Test
Purpose
Results
26-13A Combustion
Retest of full-scale comb-
ustor rig premix tube
26-14A Combustion New sprayring. No
possibility of contami-
nation.
26-13B Flow Check fuel atomization
Visualization and distribution
26-14B Flow Check fuel atomization
Visualization and distribution
26-15A Combustion Centrally mounted
simplex nozzle 85 GPH,
SOdeg.
26-16A Combustion Centrally mounted
simplex nozzle 35 GPH,
90deg
26-17A Combustion Centrally mounted
simplex nozzle 12 GPH,
90 deg.
Poor flame quality (lumi-
nous, opaque) matching
B-2 rig results
Poor flame quality (lumi-
nous, opaque) matching
B-2 rig results
No discernable deteriora-
tion in spray character-
istics
No discernable deteriora-
tion in spray character-
istics
Poor flame quality (lumi-
nous, opaque, and
concentrated in center
Flame quality better than
26-15A, but still unaccep-
table
Acceptable flame quality
(traces of luminous
flame)
Refer to Appendix B for SI unit conversion
65
-------
The second fuel injector assembly, designated Scheme 26-14A, was also subjected to
combustion tests. Once again, the flame quality was poor, matching the results obtained in the
full-scale combustor rig test. This outcome did not shed any light on the reasons for poor
functioning of the premix tube, and served to perpetuate rather than resolve the original
discrepancy between the good initial component test results and the later very poor full-scale
combustor results. In fact, it had been found impossible to duplicate the initial bench-rig test
results which had shown excellent flame quality.
In an attempt to find a cause for the apparently consistent and repeatable deterioration
in premixing quality associated with the sprayring design of the full-scale premixing tube, a
series of flow visualization tests were subsequently conducted. Both the original sprayring
injector and the second duplicate sprayring were evaluated. In these tests there was found to
be no difference (within the limits of visual observation) between the two sprayrings (Schemes
26-13B and 26-14B). The general results (atomization and distribution of the liquid in the
premix tube airstream) were judged comparable to those obtained during flow visualization
tests performed initially for the original sprayring injector. These observations shed no new
light on the question of deteriorated premixing performance.
Having found no explanation for the observed poor flame quality in the above-described
combustion and flow visualization tests (no anomaly in the functioning of the sprayring for
example, and no evidence that the original tests may have been erroneous or nonrepresen-
tative), it was decided that the most productive approach leading to the restoration of
excellent quality premixing and flame appearance consisted in the evaluation of alternative
fuel injector designs. It was reasoned that the performance of the sprayring injector might have
been marginal all along, providing high quality premixing on some occasions, and poor quality
on others, in response to changes in secondary factors.
3.1.3 Verification Tests of Alternative Fuel Injectors
The testing of alternative fuel injectors was begun with three candidate configurations
having centrally mounted fuel injectors of the same basic type (pressure atomizing with 12, 35,
and 85 GPH nozzle designations). Combustion tests were conducted to determine the effect of
rated flow capacity (and the resultant variation in fuel droplet diameter distribution that
occurs when the three nozzles are compared at the same flowrate) on flame appearance. It was
found that flame appearance improved (a lower incidence of opaque luminous flame) as lower
capacity nozzles (better atomization) were inserted and tested. The smallest nozzle (12 GPH)
exhibited generally acceptable flame appearance, while the two larger nozzles were judged
unacceptable because of excessive luminous flame. The largest nozzle (85 GPH) was also
judged unacceptable because of a poor fuel distribution pattern (fuel concentrated in the
center of the passage).
The choice of a centrally mounted fuel injector for testing as an alternative to the original
sprayring design was predicated upon the need for improved fuel atomization. The substantial
presence of opaque luminous flame observed in the component and full-scale combustion tests
had indicated that the burning of sheets of fuel or very large droplets had taken place. By the
substitution of a centrally mounted simplex (pressure atomizing) or airblast fuel injector of
known performance, good initial atomization of the fuel (upon injection into the premix
passage) could be assured.
Not only the atomization of the fuel, but also the distribution (in an even pattern across
the premix passage) must be provided in an acceptable fuel injector design. The original
sprayring design had provided good fuel distribution by virtue of its ring arrangement (which
allows the fuel to be introduced through 16 jets into equal-area sectors). Centrally mounted
fuel injectors on the other hand introduce the fuel at a single point source, and rely upon the
66
-------
penetration of the spraycone (radially outward across the premix passage) to provide an
even-pattern distribution. Because small droplets do not penetrate well in a high velocity
airstream, there is a trade-off between penetration (distribution) and the degree of atomization
associated with a centrally mounted fuel nozzle. It was believed possible that an aerodynamic
method of spreading the fuel across the premix passage might be required in conjunction with
centrally mounted injectors.
Preliminary designs were prepared of several candidate fuel injectors, including:
(a) centrally mounted air-blast and pressure-atomizing nozzles in conjunction with inlet swirl
vanes (which provide aerodynamic spreading of the fuel); (b) centrally mounted air-boost
nozzles (requiring an external compressor) in conjunction with swirl-vane and vortex aero-
dynamic spreading devices; and (c) modified sprayring injectors.
The specific configurations proposed and evaluated were as follows:
(a) Air-Blast and Pressure-Atomizing Nozzles
As stated, the choice of centrally-mounted fuel injectors predicated upon the apparent
need for improved atomization. By the substitution of a centrally mounted pres-
sure-atomizing or air-blast fuel injector of known performance, good initial atomization of
the fuel (upon introduction into the premix passage) could be assured. To enhance the
prospects for achieving an even distribution, two aerodynamic methods of spreading
small fuel droplets across the premix passage were proposed: (1) moderate inlet swirl (5
to 10 deg vanes); and (2) vortex spreaders (multiple small swirlers 4 or 8 in number
mounted in a ring around the central fuel injector). The two methods are illustrated in
the configurations tested: inlet swirl Schemes 26-21A, 26-22A, 26-24A, 26-26A,
26-27A (Figures 53, 54, 55, 56, and 57); vortex spreaders Schemes 26-25A,
26-28A, and 26-29A (Figures 58, 59, and 60). In an alternative approach to the use of
aerodynamic devices for fuel spreading, a design employing multiple pressure atomizing
nozzles (six nozzles in a hexagonal arrangement) was also proposed and evaluated
(Scheme 26-19B, Figure 61).
Inlet Swirler (5 deg Vanes)
Figure 53. Scheme 26-21A Original Premix Tube With Air-Blast Nozzle and Inlet Swirl Vanes
67
-------
Dual-Orifice Nozzle
Pressure Atomizing
2.5deg Inlet Swirl Vanes
Figure 54. Scheme 26-22A Original Premix Tube With Dual-Orifice Nozzle
Inlet Swirler (5 deg Vanes)
Figure 55. Scheme 22-24A Short Premix Tube With Air-Blast Nozzle
68
-------
7.5 deg Inlet Swirl Vanes
Sonicore Model 250T
Figure 56. Scheme 22-26A New Premix Tube With Air-Boost Nozzle and
Inlet Swirl
7.5 deg Inlet Swirl Vanes
Sonicore Model 250T
Figure 57. Scheme 22-27A Original Premix Tube With Air-Boost Nozzle
-------
Sonicore Model 250T
Vortex Swirlers
(4 Places)
Figure 58. Scheme 22-25A Short Premix Tube With Air-Boost Nozzle and
Vortex Spreaders
Sonicore Model 250T
Vortex Swirlers
(8 Places)
Figure 59. Scheme 22-28A Original Premix Tube With Air-Boost Nozzle
and Vortex Spreaders
70
-------
Figure 60. Scheme 22-29A Original Premix Tube With Air-Boost Nozzle,
Vortex Spreaders and Variable Damper
-12 GPH Pressure-Atomizing
Nozzle, 6 Places
1
J L .....
V
Plexiglas Premix Tube
Figure 61. Scheme 22-19B Configuration for Flow Visualization Test of
Six-Nozzle Cluster Fuel Injector
71
-------
(b) Air-Boost Nozzles
To provide greater atomization capability, energy from an external source can be utilized.
"Air-boost" atomization calls for the use of compressed gas as a convenient means of
providing a localized source of high energy at the point of fuel injection. A Sonicore
nozzle, model 250T, was selected for evaluation in bench testing. Configurations utilizing
both inlet swirl vanes and vortex spreaders to promote an even fuel distribution were
constructed, as shown in Figures 56 through 60 (Schemes 26-25A through 26-29A). A
larger capacity Sonicore nozzle, Model 281T, was also selected for evaluation, and
subsequently tested with no provision for aerodynamic spreading of the fuel (Scheme
26-23A, Figure 62).
Sonicore Model 281T
Figure 62. Scheme 22-23A Original Premix Tube With Air-Boost Nozzle
(No Inlet Swirl)
(c) Modified Sprayring Injectors
The sprayring-type injector employed in the original full-scale premix tube design had
inherently good fuel distribution characteristics (because of the basic ring arrangement
which allows fuel to be introduced through 16 jets into equal-area sectors of the premix
passage) but did not provide acceptable atomization of the fuel in initial tests. Two
modified designs were proposed: (1) a sprayring having "segmented" or multiple
individual splashplates designed to eliminate the pooling of liquid from adjacent fuel jets
which had occurred on the surface of the original full-ring splashplate (pooling had been
observed in some flow visualization tests); (2) a sprayring with no splashplates, designed
to operate a low pressure drop (thereby producing low-velocity fuel jets that do not
penetrate to the premix tube wall) the number of fuel jets was increased from 16 to 64.
The two sprayring injector configurations, Schemes 26-18A and 26-20A are shown in
Figures 63 and 64.
Component tests were conducted to evaluate the various fuel-injector designs described
in the previous section.
72
-------
A-A
Figure 63. Scheme 22-18A Original Premix Tube With Spray-Ring Injec-
tor and Segmented Splashplates
Low Delta P Spraying
Figure 64. Scheme 22-20A Original Premix Tube With Low Delta P Spray
Ring
73
-------
The tests performed and the highlights of the results obtained are summarized in Table
X. All tests (both flow visualization and combustion tests) were conducted at ambient
pressure, and at nominal design-point premixing passage velocities. The rig inlet air was
preheated to 600° F in the combustion tests.
As indicated in Table X, the results obtained with segmented splashplates (Scheme
26-ISA) indicated no significant improvement in flame quality with respect to the previous
full-ring splashplate used on the original sprayring injector. Similarly, the six nozzle cluster
design (Scheme 26-19B, utilizing 12 GPH pressure atomizing nozzles) was found to produce a
poor distribution of fuel in the flow visualization tests conducted.
The low-delta-P sprayring (Scheme 26-20A) was a generally successful design, producing
a good quality flame with only traces of luminous burning.
Among the three types of centrally-mounted fuel nozzles that were evaluated (pressure
atomizing, air-blast, and air-boost), flame quality ranged from acceptable (Scheme 26-23A,
employing an oversize Sonicore nozzle with no flow-spreading device) to excellent
(Schemes 26-28A and 26-29A, Sonicore nozzle with 8 vortex spreaders). In all the cen-
tral-nozzle schemes except 26-28A and 26-29A, there were local concentrations of fuel (and
traces of luminous flame) in the center of the primary combustion flow field.
The most promising configuration (Scheme 26-29A), which was ultimately selected for
evaluation in the full-scale combustor, produced an excellent quality flame (no luminous
burning) and a uniform fuel distribution. A photograph of the flame obtained with this scheme
is presented in Figure 65. The use of an air-boost (Sonicore) nozzle was viewed as a potential
drawback, because of the implied requirement for an external boost compressor. However, the
very fine atomization produced by the Sonicore nozzle was believed to be a factor in the
outstanding premixing performance obtained with this scheme. The selection of Scheme
26-29A for use in the full-scale combustor was made with a view toward establishing a limiting
case in which the emission characteristics achievable with very good premixing could be
demonstrated.
3.1.4 Revised Premix Tube Designs
In the second verification test series performed using the complete full-scale combustor
the air-boost premix tube (Scheme 26-29A, shown in Figure 60) was employed. As described in
section 3.2.2, low NO, concentration levels were demonstrated, a result attributed primarily to
the superior atomization characteristics of the air-boost nozzle. At the same time however,
preignition of the fuel took place inside the premixing passage and damage to the premix tube
swirl vanes was incurred. The damage was similar that encountered in the initial tests of the
full-scale combustor.
Subsequent technical activity was directed toward the elimination of preignition in the
full-scale premix tube. An in-house design review was held, and it was concluded that
modifications to the basic premix-tube configuration specifically aimed at reducing the
likelihood of preignition should be made. Two alternative premix tube designs were proposed.
Both incorporated modifications specifically aimed at reducing the likelihood of preignition.
74
-------
TABLE X
PREMIX TUBE COMPONENT TESTS OF ALTERNATIVE
FUEL INJECTORS
Scheme
Type
Test
Purpose
Results
26-ISA
Combustion
Evaluate sprayring with
segmented splashplates
Generally poor quality flame (lumi-
nous, opaque) but slightly better
than baseline Scheme 26-13A.
26-18B
Flow
Visualization
Check fuel atomization and
distribution
Segmented splashplates eliminate
some local concentrations of liq-
uid.
26-19B
Flow
Visualization
Evaluate 6-Nozzle Cluster Excessive wetting of wall.
26-20A Combustion Evaluate low delta P spray-
ring (no splashplates)
Generally good flame quality (trace
of luminous flame).
26-20B
Flow
Visualization
Check fuel atomization and
distribution
Good atomization; acceptable dis-
tribution (fuel spreads almost to
wall and to center of passage).
26-21A Combustion Evaluate air-blast nozzle
with 5 deg inlet swirl vanes
Good flame quality (trace of lumi-
nous flame) slight concentration
in center.
26-21B
Flow
Visualization
Check fuel atomization and
distribution
Coarse spray produced by nozzle
but atomized by droplet shatter-
ing; liquid does not quite spread
to wall.
26-22A Combustion Evaluate dual orifice nozzle
with 2.5 deg inlet swirl
vanes
Good flame quality using primary
orifice only (secondary too large
for ambient testing).
26-22B
Flow
Visualization
Check fuel atomization and
distribution
Secondary orifice flowed and found
to have wider spraycone, may
cause wetting of wall at high
pressure.
26-23A
Combustion Evaluate Sonicore nozzle
Acceptable flame quality but con-
centration of luminous flame in
center.
26-24A Combustion Evaluate air-blast nozzle
with 5 deg inlet swirl vanes
in short premix tube
Good flame quality (trace of lumi-
nous flame) slight concentration
in center.
26-25A
Combustion
Evaluate Sonicore nozzle
with 4 vortex spreaders
Good quality flame except slight
concentration (trace of luminous
flame) in center.
26-26A Combustion Evaluate Sonicore nozzle
with 7.5 deg inlet swirl
vanes in new premix tube
Excellent quality flame except
slight concentration in center.
Swirl strength improved.
26-27A
Combustion
Evaluate Sonicore nozzle
with 7.5 deg inlet swirl
vanes
Good quality flame except slight
concentration in center.
26-28A
Combustion
Evaluate Sonicore nozzle
with 8 vortex spreaders
Excellent quality flame with good
distribution.
26-29A
Combustion
Same as 26-28A but with
inlet damper installed
75
Same as 26-28A.
-------
Figure 65. Flame Observed Using Premix Tube (Scheme 26-29A)
The damage to the premix tube swirler incurred during the second series of full-scale
combustor tests provided a strong indication that the basic design of the diffusing section of
the full-scale premix tube had been a contributing factor in the occurrence of preignition. As
part of the design review, a summary comparison was made of the premix tube diffusing
passage design and established criteria in the areas of autoignition, passage velocity, and flow
separation. In Table XI the key elements of the comparison are identified. It may be seen from
Table X that the original premix tube designs met both the autoignition and flow-separation
criteria, but failed to satisfy the third criterion of maintaining a mass-average minimum
velocity greater than 200 fps. This value was based on experience gained from tests of a variety
of different premixing devices, and was considered generally consistent with an alternative
criterion that minimum local velocities in the premixing passage be maintained at values
greater than 130 fps (a conservative calculation of expected turbulent flamespeed under the
design-point conditions specified for the premixing passage of the full-scale combustor was
used to determine the value of 130 fps). Because the minimum mass-average velocity in the
original basic premix tube design (which occurs at the leading edge of the swirler) is only 130
fps, it was reasoned that lower local velocities were probably present in the flowstream (along
the wall or due to profiles in the free stream) during tests of the full-scale combustor, and that
conditions favorable for flame stabilization were set up in these regions.
76
-------
TABLE XI
PREMIX TUBE DESIGN REVIEW SUMMARY*
Parameter
Autoignition
Passage
Velocity
Flow
Separation
Criteria
T res <1 29 ms at
800°F
Vmin » ST max
~ 130 fps local min
~200 fps avg min
Conical half-angle of
6 deg or less
Original
Premix Tubes (1 & 2)
T res = 3.6 ms
130 fps average
minimum
5.6 deg
Redesign
T res = 0.8 ms
T res = 1.0 ms
350 fps average
minimum
5.6 deg
To provide an additional margin against preignition, it was proposed that passage
velocities be increased substantially. To accomplish this objective the length of the diffusing
passage and the diameter of the passage at the discharge plane were both reduced. In Figure
66, the revised premix tube diffusing passage design is shown in conjunction with a central-
ly-mounted air-boost fuel nozzle. A second proposed version of the modified premix tube is
shown in Figure 67. This design featured a radial "spoke" spraybar and a smaller throat
diameter (to achieve a higher air velocity for improved fuel atomization). In order to meet the
design criterion for flow separation (half-angle less than 6 deg) it was necessary to increase the
length of the diffusing section in this design. In order to maintain the required premix tube air
pressure drop (avoiding an increase due to the adoption of a swirler having a smaller diameter)
it was necessary to reduce the swirler vane angle from 45 to 26 deg in both designs, thereby
providing the same effective flow area.
The proposed air-boost premix tube configuration was assembled utilizing a Sonicore
nozzle and 5 deg inlet swirl for aerodynamic fuel spreading. This configuration, shown in
Figure 68, was combustion tested in the component rig facility to determine its general
performance. During these tests a bistable mode of flameholding was observed. Near lean
blowout the premix tube flame was entirely blue and was anchored at the centerbody of the
swirler. At slightly higher equivalence ratios (still fuel lean) the flame had the appearance of
being locally fuel-rich in the center. As higher operating equivalence ratios were approached,
the mixture apparently exceeded the local (rich) flammability limit in the center of the
combustion duct, and the flame lifted from the swirler and became stabilized just downstream
at the rig discharge plane where the entrainment of ambient air could take place. Because of
the observed fuel-rich region of flame in the center of the combustion duct, and the related
lifted-flame phenomenon, it was concluded that the method of fuel spreading employed (inlet
swirl) was ineffective at the low fuel flowrates required under bench-test (atmospheric
pressure) operating conditions. At low fuel flows the Sonicore nozzle is a very effective
atomizer. The small fuel droplets produced tend to follow local air patterns and remain in the
stream tubes in which they were deposited by the fuel injector. The influence of the centrifugal
force field set up by the inlet swirler is less pronounced for small droplets than for larger
droplets, with the result that less spreading of the fuel occurs. At higher fuel flows, the
Sonicore nozzle performance declines (larger values of SMD are encountered) with the result
that droplet spreading may improve. This result is indicative of a generally undesirable
trade-off between atomization and distribution associated with the use of air-boost nozzles.
The second configuration (Scheme 26-33A, Figure 67) was also combustion tested.
Results showed that both the atomization and distribution of fuel provided by the radial
spraybars were excellent. Because of these results, and because of the greatly increased margin
against flashback provided by this design, Scheme 26-33A was selected for testing in the
full-scale combustor.
'Refer to Appendix B for SI unit conversion
77
-------
Figure 66. Revised Premix Tube Design Incorporating Air-Boost Nozzle
: " Figure 67. Revised Premix Tube Design Incorporating "Spoke" Fuel Injector
78
-------
r~
I
Ram Capture Piece
Extended-Length
Premixing Tube
H*
L
fct
Figure 68. Full-Scale Combustor Scheme FS-01A
79
-------
3.2 FULL-SCALE COMBUSTOR VERIFICATION TESTS
Verification Testing of the complete full-scale combustor was accomplished in four parts,
each consisting of the evaluation of a separate configuration. The first three configurations
were variations of the Full-Residence-Time (FRT) version of the combustor utilizing different
premix tubes. A major portion of the Phase IV development effort was devoted to evaluation
and development of the premix tube. Testing of each of the three FRT combustor configura-
tions was preceded by extensive premix tube component testing (see the discussion of that
effort in Section 3.1). The fourth and final configuration consisted of the short-length,
engine-compatible version (ECV) of the full-scale combustor tested in conjunction with the
premix tube from the third configuration of the FRT combustor.
A general description of the full-scale combustor rig hardware and related equipment was
given in Section 2.10.
In this subsection, the full-scale combustor verification tests are described in chronologi-
cal order and grouped according to the four configurations evaluated.
3.2.1 Initial FRT Configuration (Scheme FS-01A)
Verification testing of the full-scale combustor was initiated as part of a general checkout
of rig systems. A brief test of the basic FRT combustor including the extended-length premix
tube (Scheme 26-12A, Figure 50), was conducted. The configuration, designated Scheme
FS-01A, is shown in Figure 68.
There were several objectives in the initial tests, including the checkout of rig systems,
calibration of the combustor with regard to internal airflow distribution (for comparison to the
aerodynamic model predictions described in Section 2.8), determination of the general operat-
ing characteristics of the combustor, and determination of a basic emission signature.
The results of the tests showed generally satisfactory functioning of rig systems. The
combustor internal airflow distribution, as determined by total and static pressure measure-
ments in the primary liner cooling shroud, and by static pressure measurements at the throat
of the premix tube, closely matched the analytical model predictions. Problems were identified
in the operation of the premix tube, and in the basic combustor emission signature. The test
results indicated that the degree of fuel preparation provided by the full-scale premix tube had
been substandard, and that local ignition of the fuel-air mixture upstream of the swirl vanes
had occurred. Visually the flame (as seen on the video monitor) appeared very luminous with
characteristics indicative of diffusion burning. Post-run observations showed locally heavy
carbon deposits on the swirler and primary liner as if fuel had run out of the premixing tube
along portions of its wall. Also, the center sections of seven swirler vanes were burned through,
as a result of the fuel preignition.
The exhaust emission data, shown in Figure 69, indicated that the emission signature
associated with the Rich Burn/Quick Quench concept in bench-scale tests had been duplicated
in some respects in the full-scale combustor (the NO, curve, in particular, exhibited a peak and
a minimum region or "bucket" similar to the curve shown in Figure 5 for the bench-scale
combustor). However, there were significant differences, including the absence of a peak in the
CO curve (CO concentrations increased asymptotically as overall equivalence ratio setting was
reduced, apparently toward lean blowout of the combustor), and the relatively high NO,
concentration (56 ppmv at 15% O2, compared to levels of 36 ppmv and lower obtained for the
bench-scale combustor) measured in the minimum region of the NO, curve. These differences
in the basic emission signature were consistent with the observed deficiencies in premix tube
performance. For example, the apparent asymptotic increase in CO toward lean blowout at low
80
-------
overall equivalence ratios was consistent with the observed condition of very poor fuel
preparation in the primary zone (efficient burning would have been possible only after a
sufficient quantity of fuel had been introduced to support combustion in the secondary zone;
hence, lean blowout would have been overcome only at the higher overall equivalence ratio
settings). The increased NO, concentration measured at bottom of the NO, curve bucket (in
comparison to bench-scale results) was also consistent with the occurrence of fuel preignition
in the premix tube, and, in general, with burning under nonpremixed conditions in the primary
zone.
400
50 psia
525° F
No. 2 Fuel
0
0.1 0.2
Overall Equivalence Ratio
Figure 69. Variation in Emission Concentrations With Overall Equivalence
Ratio for Tests Conducted With Scheme FS-01A
The data obtained for Scheme FS-01A during the initial test series are presented in
Appendix A, as part of the complete test results for the full-scale combustor. The tables
contain the major parameters necessary to specify combustor operating conditions, and contain
liner temperature and exhaust emission data.
81
-------
Several potential causes of the occurrence of flameholding within the premixing passage
were proposed and investigated: the possible separation of flow along the wall of the diffusing
passage; severe swirler vane separation; fuel collection and flameholding in recirculation zones
caused by wakes from the damper; nonuniform approaching airflow; and a possible TEB leak
("TEB" or triethyl borane, a pyroforic liquid, was injected through the premix tube to effect
ignition of the combustor). Two of the proposed causes (nonuniform approach airflow and
TEB leak) would have accounted for both the observed internal flameholding and the
apparent deterioration in the quality of premixing.
After examination of the data and the test rig, no immediate conclusion was reached.
Because of the relatively low velocity of airflow entering the plenum section of the rig (in
which the combustor was mounted), serious distortions in the airflow approaching the
combustor appeared unlikely. Flow separation in the diffusion passage of the premix tube or at
any appreciable distance downstream of the damper were^ considered unlikely because these
conditions had not been observed in the initial premix 'tube component verification tests.
Swirler vane separation was viewed as an insufficient cause of the observed fuel preignition
without other contributing factors, such as the presence of regions of stagnation of reverse flow
inside the premixing passage (separation of the flow passing over the swirler vanes is an
accepted occurrence in other proven premix tube designs). The possibility of a steady TEB
leak, which would have served as a source of continuous ignition for fuel in the premixing
passage, could not be dismissed on the basis of available evidence.
In preparing for further tests of the full-scale combustor, several steps were taken to
ensure that the various potential problem areas identified (even though subsequently dis-
counted) would no longer be a factor in the operation of the combustor. Included were the
following rig modifications:
a. The front of the burner was elevated so that the burner axis made about
a five-degree angle with the plenum case axis. This change was meant to
promote a more direct flowpath between the rig entrance duct and the
premix tube, and reduce airflow distortions at the premix tube inlet. A
comparison of rig configurations before and after this change may be seen
in Figures 70 and 71.
b. The rig direct-fired air preheater was relocated to a position two feet
upstream of the original location in the duct leading to the rig plenum.
This change was made to allow more time for any preheater-induced flow
distortions to "wash out," and also give a more uniform temperature
profile when the preheater is used.
c. The method of TEB injection was changed to provide for introduction
directly into the primary zone of the combustor rather than through the
premixing passage. In the initial tests of the full-scale combustor, the
possible leakage of TEB into the premixing passage during steady-state
operation was postulated as a likely cause of damage to the premix tube
swirler. Relocation of the TEB line eliminated this type leakage as a
factor in future testing.
d. Additional instrumentation was provided to ascertain airflow profiles into
. the premix tube. Twelve total pressure probes were added at four
circumferential and three radial locations.
82
-------
Jt
fvfetr**-'.:. -?> » .-««»« t i-t-»-f)4-|
Figure 70. Original Arrangement of the Full-Scale Test Rig
Figure 71. Arrangement of the Full-Scale Test Rig Following Elevation of the
Combustor
83
-------
Two further steps were taken: (1) diagnostic tests of the premix tube were conducted (as
described in Section 3.1.2) in the component rig to determine further potential causes of the
observed preignition and apparent deterioration in the quality of premixing; (2) at the same
time, alternative fuel injector designs were formulated for subsequent evaluation in the
component rig (described in Section 3.1.3).
3.2.2 Second FRT Configuration (Scheme FS-02A)
Testing of the full-scale combustor was resumed using an air-boost premixing tube.
Diagnostic tests of the initial premix tube design (Figure 50) had indicated that an alternative
fuel injector might be required to achieve adequate premixing. A number of modified premix
tube configurations incorporating various fuel injectors and fuel-spreading devices were
evaluated in subsequent component tests, as described in Section 3.1.3. The air-boost design
(Scheme 26-29A, shown in Figure 60) was selected as having outstanding premixing per-
formance (based on observations of flame quality). Because of the inherent penalty associated
with the use of air-boost nozzles in a gas turbine combustor (the auxiliary equipment required
is bulky and expensive), Scheme 26-29A was chosen with a view toward establishing a limiting
case in which the emission characteristics achievable with very good premixing (made possible
by the very fine atomization of the air-boost nozzle) could be demonstrated.
The complete combustor configuration, designated Scheme FS-02A, is shown in Figure
72. Scheme FS-02A was identical to Scheme FS-01A, except for substitution of the air-boost
premix tube. There were also minor changes in the mounting arrangements for the combustor
and the configuration of the rig, as described in Section 3.2.1. A single series of tests was
conducted in the full-scale combustor plenum-rig facility at an operating pressure of 50 psia
(nominal) and a combustor inlet air temperature of 525°F, using neat No. 2 fuel. The 'exhaust
emission data measured are presented in Figure 73. In Figure 74, a comparison of these results
to those obtained in the initial test series is shown. A very low NO, concentration of 29 ppmv
was measured at the bottom of the "bucket" in the NO, curve. This level was 60% of the
program goal of 50 ppmv. No staging of the primary-zone airflow had been attempted in the
tests performed. However, it was anticipated that staging could be employed to shift the
"bucket" in the NO, curve to the left and to the right in the manner demonstrated for the
bench-scale Rich Burn/Quick Quench combustor, thereby establishing a low NO, "corridor"
over the operating range from idle to full power. Bench-scale rig results had also indicated that
the effect of increased operating pressure on NO, concentration levels would not be significant.
A moderate increase in NO, would be expected, however, due to combustor inlet air tem-
perature levels.
The CO curve in Figure 73 exhibited the same characteristic shape observed in tests of
the bench-scale Rich Burn/Quick Quench combustor. The peak in the curve, however, was not
as high as the levels measured for the bench-scale combustor (about 300 ppmv, compared to
levels as high as 900 ppmv for the bench-scale combustor). The location of the peak also
represented a variation in the full-scale combustor results with respect to those obtained for
the bench-scale combustor (in representative bench-scale tests, the peak occurred in the
vicinity of 0.1 overall equivalence ratio compared to 0.2 in Figure 73). It was expected that the
location of the peak in the CO curve could be varied by adjusting the stoichiometry of the
secondary zone of the combustor (independent control of the CO characteristics of the
combustor, without any appreciable influence on NO, characteristics, had been demonstrated
in this manner in the bench-scale program). By reducing the total quantity of air admitted to
the secondary zone (i.e., by reducing the sum of quick-quench airflow plus premix-tube
airflow), it was anticipated that a leftward shift of the CO curve would be possible. If this
adjustment were made, lower CO concentration levels would be expected in the range of
equivalence ratios above 0.2. A reduction in CO concentration levels would also be expected as
a result of increasing the inlet-air temperature level.
84
-------
Figure 72. Full-Scale Combustor Scheme FS-02A
85
-------
400
300
CM
O
LO
O
4-*
Oi
i
O
O
c
O
E
LU
200
100
0
No. 2 Fuel
50 psia
525°F
5.5% P/PT
O CO
400
0.10 0.20
Overall Equivalence Ratio
0.30
0.10 0.20
Overall Equivalence Ratio
0.30
Figure 73. Variation in Emission Concentration With Overall
Equivalence Ratio for Tests Conducted With
Scheme FS-02A
Figure 74. Comparison of Emission Data
Schemes PS-01A and FS-02A
Obtained for
-------
The evaluation of Scheme FS-02A of the full-scale combustor was limited to a single
series of tests because of a repeat occurrence of damage to the premix tube swirler. Inspection
of the combustor following the tests, for which data are presented in Figure 73, revealed that
half the swirl vanes were missing or severely damaged, apparently due to preignition of the
fuel inside the premixing passage. The damage was similar to that incurred in the initial tests
of the full-scale combustor.
The very low NO, concentration levels obtained in the tests of Scheme FS-02A were
attributed to the superior fuel atomization characteristics of the air-boost nozzle, and to the
effectiveness of the fuel-spreading devices employed (vortex swirlers surrounding the fuel
nozzle). At the same time, however, the very promising emission results achieved were
seriously compromised by the recurrence of the preignition phenomenon observed in Scheme
FS-01A. Subsequent to the second test series, it was decided that modifications to the
aerodynamic design of the basic premix tube (which had been the same in the two initial
schemes) should be made. Following an in-house review, two revised designs were formulated,
as described in Section 3.1.4. In both designs, internal premixing passage velocities were
increased substantially; one configuration employed an air-boost nozzle, the other a number of
radial "spoke" spraybars. The second, nonair-boost design was ultimately selected for eval-
uation in Scheme FS-03A of the full-scale combustor.
3.2.3 Third FRT Configuration (Scheme FS-03A)
Testing of the FRT combustor was resumed following a review of the aerodynamic design
of the basic premix tube, and the subsequent formulation and preliminary testing of a revised
premix tube design. The revised design, which featured higher premixing passage velocities
(350 FPS minimum at the full-power setting), and incorporated a radial "spoke" fuel injector,
is shown in Figure 67. The combustor configuration (Scheme FS-03A, shown in Figure 75) was
identical to that tested previously except for substitution of the redesigned premix tube. In the
tests conducted, a constant premix tube airflow setting was maintained. The premix tube
variable damper was not used.
Figure 75. Full-Scale Combustor Scheme FS-03A
87
-------
3.2.3.1 First Test Series
The experimental evaluation of Scheme FS-03A was accomplished in three parts. In the
initial test series, the combustor was tested at 50 psia rig pressure and 450° F inlet air
temperature using No. 2 fuel and No. 2 fuel with 0.5% nitrogen (as pyridine). Examination of
the combustor following the tests revealed damage to the premix tube swirler, and the
presence of a metal instrumentation tag in the premixing passage. The position of the tag, and
the pattern of the metal discoloration in the premix tube wall (due to the uneven heating
associated with internal flameholding), indicated that the tag had been ingested into the
premix tube (having broken free at an upstream site inside the rig) and had lodged against the
fuel injector spraybars. The resulting wake inside the premixing passage caused flameholding
and damage to seven of the fifteen swirler vanes.
3.2.3.2 Second Test Series
Repairs were made to the premix tube, and the initial test series was partially repeated:
data was obtained at 50 psia rig pressure and 400°F inlet air temperature using No. 2 fuel with
0.5'Y. nitrogen (as pyridine). The emission data may be found in Table II of Appendix A for
comparison purposes. Because there was no significant change in the emission characteristics
of.the combustor in these repeat tests, it was decided that the tests using non-nitrogenous No.
2 fuel need not be repeated. A single data point at 100 psia rig pressure was obtained during
the second test series prior to a test stand malfunction (U-tube failure resulting in mercury
contamination of the control room) which forced the shutdown of the rig.
3.2.3.3 Third Test Series
Data was obtained at 100 psia rig pressure and 575°F inlet air temperature using No. 2
fuel and No. 2 fuel with 0.5% nitrogen (as pyridine) in a third test series. Examination of the
combustor before and after the tests showed no further distress to the premix tube swirl vanes.
The combustor liner was found to be in good condition except for minor deterioration of the
flamespray coating at the entrance to the quick-quench zone. It was noted that a metal band
or collar on the combustor had come loose during the third test series. When in place, this
band prevents the direct entry of air from the rig plenum into the quick-quench zone of the
combustor, forcing it to follow an alternative path through the primary liner cooling passage.
In the displaced position, some airflow was allowed to enter the quick-quench section without
passing through the cooling shroud. In a separate incident, damage to the rig exit traverse
probe was sustained midway through the third test series because of interrupted cooling water
flow (due to the failure of a bellows section inside the rig). Five of the nine gas-stream
thermocouples were destroyed, and pattern factor data were unavailable for the last five test
points.
3.2.3.4 Fourth Test Series
The fourth test series consisted of an evaluation of the operation of the combustor on
shale-derived DFM and verification of the previous test results obtained using No. 2 fuel (to
determine whether the loosened collar on the quick-quench section of the combustor may have
affected performance or emission characteristics).
The data obtained in the four test series conducted are presented in Appendix A of this
report. The tables in Appendix A contain the major parameters necessary to specify combustor
operating conditions, and contain liner temperature and exhaust emission data for Scheme
FS-03A.
88
-------
A scheme definition sheet for the configuration evaluated (Scheme FS-03A) is presented
in Figure 76, showing the design point airflow distribution of the combustor, and the location
of liner skin thermocouples. Calculated values of the primary and secondary airflow rates are
also presented in Table IV of Appendix A for each test point.
3.2.3.5 Exhaust Emission Data
The exhaust emission data generated during the first of the three test series are
presented in Figure 77. The curves obtained for both NO, and CO exhibit the characteristic
shapes documented in numerous tests of bench-scale combustor during Phase II. A minimum
NO, concentration of 26 ppmv (corrected to 15% 02) was achieved using neat No. 2 fuel. This
level compares favorably to the minimum levels achieved in tests of the bench-scale combustor
(20 to 36 ppmv at 15% O2 depending on primary zone residence time) using neat No. 2 fuel. It
should be noted, however, that a direct comparison of these results is not possible because of
differences in the inlet air temperature (450°F in the current results vs 600°F in the
bench-scale data).
The NO, curve obtained for No. 2 fuel with 0.5% nitrogen is less complete. However, a
minimum concentration of 75 ppmv (corrected to 15% O2) was documented. This level is
higher than the minimum levels achieved in tests of the bench-scale combustor (33 to 55 ppmv
at 15% O2 depending on primary zone residence time) using No. 2 fuel with 0.5% nitrogen.
Differences in primary zone residence time (60 ms, on a cold-flow basis, for Scheme FS-03A vs
values of 80 to 170 ms for various schemes of the bench-scale combustor) may account for the
observed increase in the NO, concentration level. Similarly, differences in the quality of
premixing and in the effectiveness of the quick-quench section may exist in the full-scale
combustor relative to the bench-scale combustor, and may be a factor in the higher
NO, concentration level.
The CO concentration levels obtained in the first test series are lower than those
documented in the bench-scale combustor test program (300 ppmv in Figure 77, at the peak of
the CO curve, compared to values as high as 900 ppmv for the bench-scale combustor).
Generally, the magnitude of the peak in the CO curve is believed to be related to the
effectiveness of mixing in the quick-quench section. A lower peak concentration implies
reduced mixing effectiveness. In Figure 77, the relatively low peak concentrations in the CO
curves, and the slightly higher minimum NO, concentrations already noted, were both
consistent with the view that the rate of mixing achieved in the quick-quench section of the
full-scale combustor (Scheme FS-03A) may have been somewhat less than that achieved in the
bench-scale combustor.
The data generated during the first test series (Figure 77) also reflected the influence (if
any) of damage to the premix tube swirl vanes. It was believed that ingestion of the metal
instrumentation tag (which caused flameholding in the premixing passage and resultant
damage to the vanes) occurred near the end of the first test series during the runs conducted
using No. 2 fuel with pyridine additive (after completion of tests with neat No. 2 fuel, and
after a brief interruption to replenish the stand fuel supply). The effect of the damage (and of
burning in the premixing passage) on the exhaust emission data in Figure 77 was unknown.
However, a comparison of the results in question with those generated in subsequent testing
(in the second test series described below) had shown only minor differences in the basic
emission characteristics and concentration levels.
Exhaust emission data generated during the second test series are presented in Figure 78.
The second test series was conducted to determine whether the data from the first series may
have been biased by damage to the premix tube. A comparison of the curves for NO, and CO
in Figure 78 to those in Figure 77, showed general agreement with regard to the characteristic
shapes and the emission concentration levels, with the following exceptions.
-------
C D E'FG HIJ K
Al A2 A B
LB
46.14
AREF
88.20
VOLREF
2590.0
ACOSUM
24.12
STATION
Al
A2
A
B
C
0
E
F
G
H
I
J
K
L
M
AX
13.847
4.335
8.038
75.391
28.260
28.260
72.346
72.346
72.346
72.346
72.346
72.346
72.346
72.346
39.337
ACD
0.0
0.0
4.978
0.0
0.0
10.854
0.0
0.420
0.523
0.447
5.049
0.224
0.829
0.792
0.0
WACUM
0.0
0.0
20.641
20.641
20.641
65.649
65.649
67.391
69.559
71.413
92.349
93.278
96.716
100.000
100.000
PHI
0.0
0.0
1.285
1.285
1.285
0.404
0.404
0.394
0.381
0.371
0.287
0.284
0.274
0.265)
0.265
HEADER AXIAL LOC RAO IOC CIRCUN LOG
TLIN 1
TLIN 2
TLIN 3
TLIN 4
TLIN 5
TLTN 6
7.72
12.74
17.59
22.44
31.97
36.67
4.80
5.00
5.00
5.00
4.80
4.80
0.0
0.0
0.0
0.0
0.0
0.0
Figure 76. Burner Scheme Definition (Scheme FS-03A)
90
-------
500
With 0.5% N
CO No. 2 Fuel
With 0.5% N
500
0.1 0.2 0.3
Overall Equivalence Ratio
0^400
in
c
o
300
c
o
o
o
c
O
'en
200
100
0
(Runs FS-03A-11-*-FS-03A-20)
I I
NOX No. 2 Fuel With 0.5% N
CO No. 2 Fuel With 0.5% N
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 77. Variation in Emission Concentrations With Over-
all Equivalence Ratio for Scheme PS-03A, First
Test Series
Figure 78. Variation in Emission Concentrations With Over-
all Equivalence Ratio for Scheme PS-03A, Second
Test Series
-------
First, the peaks in the CO and NO, curves in Figure 78 were broader (covering a
somewhat wider range of equivalence ratios on the abscissa) than those in Figure 77. Similarly,
the bucket in the NO, curve in Figure 78 was broader than the one in Figure 77. It was
believed that these differences were the result of slightly lower inlet air temperatures in the
second test series (400°F compared to 450°F in the first test series, as shown in Table I of
Appendix A). Inlet air temperature was not independently controllable at the full-scale
combustor test facility without the use of a direct-fired heater burner. To avoid the introduc-
tion of heater burner emissions as an unknown element in the initial test results, the
temperature of the rig inlet air was allowed to vary in accordance with the levels available from
the stand (slave engine) supply. A lower inlet temperature can be expected to promote the
formation of CO (due to uniformly lower gas temperatures and increased quenching in the
mixing regions of the combustor), and to reduce the rate of formation of thermal NO,,
resulting in generally higher CO concentration levels, and generally lower NO. levels over the
entire range of equivalence ratios tested. These changes would give the appearance of an
increase in breadth to both the CO and NO, curves.
%
A lower inlet air temperature might also be expected to cause a reduction in the degree of
fuel prevaporization achieved in the premix tube of the combustor, and, as a result, serve to
further reduce the slope of the NO, vs equivalence ratio curve (the rate of formation of NO, is
more sharply responsive to changes in the burner equivalence ratio when premixed than under
nonpremixed conditions). This effect may also have contributed to the increased breadth of
the NO, curve peak in Figure 78.
The second difference that can be noted in comparing the data of Figure 78 to those of
Figure 77 is the shift in location of the NO, curve bucket (from 0.22 equivalence ratio in
Figure 77 to 0.27 equivalence ratio in Figure 78). The shift implies an increase in premix tube
airflow in the second test series of about 20%. This result is consistent with the view that
premix tube airflow in the first test series was too low as a result of the blockage created by
ingestion of the metal instrumentation tag, and the resultant increase in pressure drop created
by burning in the premixing passage.
Exhaust emission data generated during the third test series are shown in Figure 79.
Tests were conducted at 100 psia rig pressure and 575°F inlet air temperature. The combustor
configuration (Scheme FS-03A) was the same as that tested in the first two test series;
however, it is likely that some change in operating characteristics may have resulted from the
loosening of the metal band on the combustor liner. This incident occurred at some point
during the third test series. Comparison of the NO, and CO data for No. 2 fuel with 0.5% N in
Figure 79 to those generated at 50 psia rig pressure and 400°F inlet air temperature during the
second test series (Figure 78), shows the following similarities and differences.
1. The peaks and buckets in the NO, curves occur at the same values of
overall equivalence ratio. The peak NO, concentration in Figure 79 is
substantially higher (442 ppmv) than that in Figure 78 (226 ppmv) in
keeping with the higher rig pressure and higher inlet air temperature. The
minimum NO, concentration in Figure 79 (79 ppmv) is only slightly higher
than that in Figure 78 (70 ppmv) indicating the absence of any appreciable
effect of increased pressure and increased inlet air temperature at the
bottom of the NO, curve bucket. This result is consistent with the
bench-scale data, which also indicated only a slight increase in the min-
imum attainable NO, concentration with increased rig pressure and inlet
air temperature (see Table VI of Appendix A).
92
-------
500
FS-03A-32)
NOX No. 2 Fuel
CO No. 2 Fuel
NOX No. 2 Fuel
With 0.5% N
No. 2 Fuel
With 0.5% N
I
0
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 79. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme PS-03A, Third Test Series
2. The peak CO concentration in Figure 79 is 116 ppmv, substantially lower
than the peak value of 322 ppmv in Figure 78. This difference was
expected because of the increased rig pressure and increased inlet air
temperature in the third test series. As discussed earlier in this section, the
peak CO concentrations measured at 50 psia rig pressure (in the first and
second test series, Figures 77 and 78) were substantially lower than those
documented at the same rig pressure in the bench-scale test program. In
the third test series, these characteristically lower values were further
reduced (due to increased rig pressure, and inlet air temperature) to the
extent that six of eight measured CO concentrations were less than the
program goal of 100 ppmv.
93
-------
In Figure 80, the variation in emission concentrations with overall equivalence ratio for
tests of the fourth test series conducted with Scheme FS-03A firing shale DFM is shown; these
data, obtained at 50 psia rig pressure and 475°F inlet air temperature, can be compared to the
results for Scheme FS-03A firing No. 2 fuel and No. 2 fuel with 0.5% nitrogen shown in
Figures 77 and 78. The following observations concerning the two sets of data can be made: (1)
NO, concentrations measured at the bottoms of the NO, curve "buckets" may be seen to vary
with fuel nitrogen content in the expected manner (26 ppmv for No. 2 fuel with 0% nitrogen,
64 ppmv for shale DFM with 0.24% nitrogen, and 75 ppmv for No. 2 fuel with 0.5% nitrogen);
(2) the peaks in the NO, and CO curves for the two sets of data coincide (occur at the same
values of overall equivalence ratio on the abscissa); and (3) the peak concentration in the CO
curve in Figure 86 (data for shale DFM) is somewhat higher than the peak concentration
measured for No. 2 fuel (360 ppmv compared to 300 ppmv, both corrected to 15% 02). Taken
as a whole, these results indicate that the emission characteristics of the combustor obtained
during the firing of shale DFM conformed generally to expectations. The slight increase in CO
concentration levels in the shale DFM tests, which is not a significant difference, may be
attributable to minor variations in the configuration of the combustor hardware or other
rig-related factors or may be due to fuel-related effects.
500
(Runs FS-03A-33-*-FS-03A-39)
50 psia
475°F
Shale DFM
0
0
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 80. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme PS-03A, Fourth Test Series
94
-------
In Figure 81, the data generated at 100 psia rig pressure and 570°F inlet air temperature
for Scheme FS-03A firing shale DFM are presented; these results can be compared to the data
for Scheme FS-03A firing No. 2 fuel and No. 2 fuel with 0.5% nitrogen presented in Figure 79.
It may be seen that comments made previously concerning results obtained at 50 psia rig
pressure apply to the 100 psia data as well: (1) the NO, curve "buckets" vary with fuel nitrogen
content in the expected manner (44 ppmv for No. 2 fuel, 64 ppmv for shale DFM, and 79
ppmv for No. 2 fuel with 0.5% pyridine); (2) peaks in the NO, and CO curves coincide; and (3)
CO concentration levels in the shale DFM tests are somewhat higher than the levels obtained
for No. 2 fuel.
500
(Runs FS-03A-40-FS-03A-46)
I I
100 psia
570°F
Shale DFM
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 81. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme PS-03A, Fifth Test Series
The last two comments also apply to the data generated at 100 psia rig pressure and
570°F inlet air temperature for Scheme FS-03A firing neat No. 2 fuel (shown in Figure 82).
These data can be compared to the results shown in Figure 79, obtained for the same
(nominal) burner configuration and the same fuel. NO, concentration levels at the bottoms of
the NO, curve "buckets" were essentially the same (44 ppmv vs 45 ppmv) in the two test
series. The only notable difference in the data is the generally higher CO concentration level
95
-------
obtained during the repeat tests. Because the combustor configurations tested were identical
except for the loosened quick-quench collar, which had been a factor in the initial test series, it
was concluded that the mixing effectiveness of the quick-quench section may have been
compromised in the initial tests resulting in a less distinct peak in the CO curve (a uniformly
lower CO concentration level). This result, which did not constitute a major difference, was the
only apparent effect of the loosened quick-quench collar.
500
100psia
570°F
No. 2 Fuel
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 82. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme PS-03A, Sixth Test Series
3.2.3.6 Exit Temperature Profiles and Combustor Liner Temperatures
Combustor exit gas stream temperatures were measured using a rig radial traverse probe.
Thermocouples are provided at nine locations equally spaced over the circumference of the
annular exit transition piece. In the tests concluded, readings were taken at a single radial
position near mid-span. The radial traverse capability of the probe was not used in order to
maximize the run time available for generating a basic emission signature of the combustor.
Representative exit thermocouple data are presented in Figure 83. A strong central peak
is evident in the circumferential profile, with peak-to-peak minimum differentials as great as
1400° F. Values of temperature pattern factor (peak-to-average temperature differentials
normalized to overall temperature rise) are presented in Table V of Appendix A and Figure 84.
The range of values obtained (0.3 to 0.7) is substantially higher than the generally accepted
96
-------
target range of 0.2 to 0.3. Examination of the curves in Figure 83 indicates that the peak
temperatures occur in a region occupying about one-third of the circumference (three of nine
thermocouples) at the exit of the annular transition duct. These peak temperatures appear to
represent the same "top-hat" profile that exists at the discharge plane of the quick-quench
section (where a jet occupying about one-third the local cross-sectional area enters the aft
dilution section). It can be expected that this top-hat temperature profile will persist in the
flow as it passes through the aft dilution section and through the transition duct (a distance of
about 2.5 jet diameters), and will appear at the exit plane of the combustor. In Figure 83, lines
indicating the magnitude of the top-hat profile at the end of the quick-quench section
(assuming complete mixing within the jet) are superimposed on the exit-plane circumferential
profiles. It can be seen that there is close agreement between the measured peak exit
temperatures and the ideal quick-quench section temperatures. This result has two important
implications. First, the high values of temperature pattern factor in Figure 84 and Table V of
Appendix A appear to be the result of ineffective mixing in the aft dilution section of the
combustor. Because of high-velocity flow in the center of the passage, it is not unexpected that
penetration and mixing in this section may be ineffective. Second, it is noteworthy that no
temperature reading at the exit plane of the combustor exceeds the ideal (mixed out)
temperature of the quick-quench section by a significant amount. To illustrate this feature of
the data, computations of temperature pattern factor were performed for gases at the
quick-quench section on the assumption that the peak temperatures measured at the com-
bustor exit plane are equal to those that exist at the end of the quick-quench section. Values of
this parameter, TPFQQ, are plotted in Figure 85 as a function of overall equivalence ratio. The
maximum value obtained, 0.096, is well below the generally accepted target range of 0.2 to 0.3.
Although this parameter has been computed indirectly, and may, therefore, be subject to error
(for example, some of the values of TPFQQ are slightly negative, indicating that the peak exit
temperature was actually lower than the quick-quench section temperature due to mixing in
the aft dilution section), the very low values obtained indicate that excellent mixing was
achieved in the quick-quench section of the combustor.
The effect of fuel type on combustor liner temperature levels in the fourth test series is
illustrated in Figure 86. Data from skin thermocouples attached to the outer surface of the
primary combustor liner (parameters TL,N1 through TLIN 6 in Table IV; T/C locations shown in
Figure 7) have been used to compute values of the liner temperature rise factor (LTRF). This
parameter, defined in Figure 86, provides a basis of comparison for No. 2 fuel and
shale-derived DFM in terms of the overall average liner temperature rise (normalized to
burner ideal temperature rise). Results indicated that the shale DFM produced a slightly
higher liner temperature rise (a difference as great as 4% of the burner ideal temperature rise
at some equivalence ratio settings).
A further effect of the use of shale DFM is illustrated in Figures 87 and 88. Photographs
of the premix tube swirler and premixing passage show a minor buildup of carbon on the
surfaces of these parts. For comparison, the condition of the swirler following tests with No. 2
fuel is shown in Figure 89. It is believed that the deposits shown were the result of DFM fuel
contamination (the distillate fuel used in this test program contains heavy earth waxes that
were acquired at the refinery when processed fuel was placed in tanks originally used for crude
shale) and are not a characteristic of shale-derived fuels in general.
97
-------
3000
2800
2600
2400
2200
2000
1800
1600
1400
1200
1000
800
600
400
I I I I I
O Temperatures for Runs
FS-03A-11 Through FS-03A-19
(Ascending Temperature Level)
- -Mixed-OutTemperature Levels
at Quick-Quench Section
0.2703
0.2427
0.2224
0.2064
4567
Circumferential Position
8
Figure 83. Exit Temperature Profiles (Second Test Series, Probe at
Mid-Span)
98
-------
1.0
0.8
o
as
CO
a.
a
a>
0.6
0.2
O RunsFS-03A-1
[D RunsFS-03A-11
RunsFS-03A-21
FS-03A-10
FS-03A-19
FS-03A-27
0.8
0.1 0.2 0.3
Overall Equivalence Ratio
Figure 84. Variation in Temperature Pattern Factor With
Overall Equivalence Ratio
0.6
0.4
a
a
LL.
Q.
0.2
0
D RL
TPFQ
'TQC
ATQC
nsFS-03A-r
TMax
Q " Al
t r-> Ideal Ten
in Quid
Section
1 ~ TTQQ "
-"-FS-03A-
-TTQQ
QQ
nperature
c-Quench
rT3
JZk- ir
g
]
0.1 0.2 0.3
Overall Equivalence Ratio
0.4
Figure 85. Variation in Quick-Quench Section Pattern Factor
(TPFQQ) With Overall Engine Ratio (Second Test
Series)
-------
1.0
0.8
0.6
0.4
0.2
0
(Runs FS-03A-40-*-FS-03A-54)
Shale DFM
QNo.2
LTRF
TLAVG'TTIN
ATIDEAL
r* Average of all Liner
T/C Readings
100psia
570°F
0.1 0.2 0.3
Overall Equivalence Ratio
0.4
Figure 86. Variation in Liner Temperature Rise Factor (LTRF) With Overall
Equivalence Ratio and Fuel Type
100
-------
Figure 87. Condition of Premix Tube Swirler Following Tests With Shale Derived DFM
-------
g
Figure 88. Condition of Premixing Passage Following Tests With Shale Derived DFM
-------
8
Figure 89. Condition of Premix Tube Swirler Following Tests With No. 2 Fuel
-------
3.2.4 Evaluation of the ECV Configuration (Scheme FS-04A)
Following tests of the full-residence-time (FRT) configuration, the combustor hardware
was' reworked to the short-length engine-compatible version (ECV). Testing of this configura-
tion was accomplished in two test series, one directed toward the establishment of a basic
emission signature, and the other providing a demonstration of the use of variable geometry to
achieve low NO, concentration levels over a wide range of combustor operating conditions.
Three fuels were fired: No. 2 fuel; No. 2 with 0.5% nitrogen (as pyridine); and shale-derived
DFM. The results obtained indicate that the emission signature of the ECV combustor is
similar to that obtained previously for the FRT version. However, the NO, concentration levels
measured were the same to slightly lower than those measured for the FRT combustor. This
result was unexpected, and apparently occurred due to better placement of the penetration air
jets in the aft dilution section of the combustor. Operation of the premix tube variable damper
was successfully accomplished, and NO, levels less than the program goals were demonstrated
over the entire power range. Toward the end of the final test series, pieces of the premix tube
damper broke loose (due to fatigue failure of tack welds). There was some damage to the
premix tube swirler as a result. Because the problem experienced was mechanical in origin, this
occurrence did not indicate any deficiency in the aerodynamic design of the premix tube or the
variable damper. Aside from the damage to the premix tube, the combustor was found to be in
good condition. Complete details of the tests are presented in this section.
In the initial tests, the premix tube variable damper was not used. The configuration,
Scheme FS-04A, is shown in Figure 90. A scheme definition sheet, showing the design point
airflow distribution of the combustor and the location of liner skin thermocouples, is shown in
Figure 91. Comparison of these figures can be made to Figures 75 and 76, in which details of
the FRT combustor (Scheme FS-03A) previously tested are given. The FRT and ECV
combustor configurations differed in three main areas: (1) primary zone length in the ECV was
12.5 in. compared to 18 in. in the FRT combustor; (2) the louver-cooled dilution piece just
downstream of the quick-quench section in the FRT combustor was removed, yielding a
reduction of 8 in. in the length of the secondary zone; (3) the final dilution airflow, which was
introduced at Station I in Scheme FS-03A (see Figure 76), was introduced through axial-
ly-directed holes in the wall of the dump section in Scheme FS-04A (Station E in Figure 91).
Photographs of the ECV combustor hardware are shown in Figures 92 and 93.
The data obtained in the two test series are presented in Tables I through V of Appendix
A. The tables contain the major parameters necessary to specify combustor operating condi-
tions, and contain liner temperature and exhaust emission data.
104
-------
Figure 90. Full Scale Combustor Scheme FS-04A
-------
Al A2
LB
33.50
AREF
88.20
L/D
3.16
VOLREF
1505.0
ACDSUM
27.26
STATION
Al
A2
A
B
C
D
E
F
G
AX
13.847
A.335
8.038
75.391
28.260
28.260
72.346
72.346
39.337
ACD
0.0
0.0
5.468
0.0
0.0
10.953
10.049
0.792
0.0
HA CUM
0.0
0.0
20U>58
20.058
20.058
60.234
97.095
100.000
100.000
PHI
0.0
0.0
1.285
1.285
1.285
0.428
0.266
0.258
0.258
HEADER
AXIAL LOG RAD LOG CIRCUM LOG
TL1N1
TL1N2
TL1N3
TL1N4
TL1N5
TL1N6
TL1N7
TL1N8
7.50
10.50
14.80
14.80
14.80
21.50
21.50
21.50
4.8
5.0
.0
.0
.0
3.4
3.4
3.4
5.
5.
5.
Avg
180
90
180
270
0
90
270
Figure 91. Burner Scheme Definition (Scheme *FS-04A)
106
-------
Figure 92. ECV Combustor During Assembly
107
-------
Figure 93. ECV Combustor Fully Assembled Except for Variable Damper
108
-------
3.2.4.1 First Test Series
Emission-signature data were generated at 100 psia rig pressure and 560°F inlet air
temperature for Scheme FS-04A in the initial test series. The results for No. 2 fuel, No. 2 fuel
with 0.5% nitrogen (as pyridine), and shale-derived DFM are shown in Figures 94 through 96.
Comparison of the data in Figure 94 (for No. 2 fuel) can be made to those shown in Figure 82
for the FRT combustor. The two sets of data are similar in that the peaks in the CO curves
occur at approximately the same value (about 0.20) of overall equivalence ratio. Minimum
NO, concentrations for the two combustors also occur within the same basic range of overall
equivalence ratios (0.20 to 0.27). The principal differences exist in the CO concentration levels
measured (594 ppmv at the peak of the curve for the ECV combustor vs 224 ppmv for the
FRT combustor, both corrected to 15% 02), and in the minimum NOX concentrations recorded
(38 ppmv for the ECV combustor vs 45 ppmv for the FRT combustor). Taken as a whole, the
results obtained indicate that the two combustors have comparable emission characteristics (as
expected), and that the anticipated increase in NO, in the ECV configuration (because of a
reduced primary zone residence time) did not take place. As shown, there was instead a
general increase in the CO concentration level, along with the unexpected decline in the
minimum achievable NO, concentration level. The initial interpretation of these results was
that a tradeoff had been effected between NO, and CO in the aft dilution section of the
combustor. It was reasoned that in the previous FRT configuration (Scheme FS-03A), partially
mixed gases in the region of jet-induced recirculation at the dump plane of the quick-quench
section may have supported combustion reactions that contributed to the formation of
NO,. The direct introduction of penetration air into this region in the ECV combustor
(Scheme FS-04A) would have terminated these reactions, resulting in a net decline in NO, and
an increase in CO. This hypothesis implies that the mixing process initiated within the
quick-quench section is incomplete at the dump plane of the combustor (not an unreasonable
assumption because the dump plane is only one-half inch downstream of the trailing edge of
the penetration jets). Subsequent findings, however, have made it necesary to modify the
hypothesis. Examination of the combustor following the first test series revealed a crack in the
premix tube fuel manifold. During the tests in question, fuel had leaked onto the outer surface
of the combustor dome (see Figure 97) and had been ingested into the primary liner cooling
passage. Fuel entering the passage would ultimately be discharged through the quick-quench
slots. The introduction of raw fuel into the combustor at this location under highly turbulent,
overall fuel-lean conditions would account for the increased CO concentration levels observed
in the first test series. An increase in unburned hydrocarbon concentration levels would also be
expected, and as may be seen in Figures 94 through 96, did also occur (concentration levels as
high as 33 ppmv were measured compared to the usual levels of 5 ppmv or less). In the second
test series (after the manifold had been repaired), CO concentration levels were found to be
lower and generally comparable to those obtained for the FRT combustor. The conclusion
drawn from these results was that the net decline in NO, concentration levels in the shorter
(ECV) combustor has been the result of a more effective quenching process brought about by
the introduction of a substantial portion of the combustor airflow (36%) through axial-
ly-directed holes in the wall of the dump section of the combustor. It appears that this airflow
may have purged the dump region of the partially mixed reacting gases which have con-
tributed to the production of NO, in the FRT combustor. The fact that a net reduction in
overall NO, concentration levels could result from this change indicates that the local
reduction achieved was substantial enough to offset any increase in NO, production due to a
shorter primary zone residence time.
109
-------
700
Emission Concentrations - ppmv at 15% 02
-» K> CO -P» Ul O> >
O O 0 0 O 0 C
oooooooc
(R
uns FS-04
<
q
Q
A-11-*-F
T
I
N/
r\
:S-04A-13
Q N
)
0.
<3>co
O UHC
_,_ 100 nsia
560°F
No. 2 Fuel
700
(Runs FS-04A-1-FS-04A-8)
0 0.1 0.2 0.3 0.4 0.5
Overall Equivalence Ratio
Figure 94. Variation in Emission Concentrations With Over-
all Equivalence Ratio for Scheme FS-04A Firing
No. 2 Fuel
I
NOX
CO
UHC
100psia, 560°F;
No. 2 Fuel With
0.5% N
0.1 0.2 0.3 0.4
Overall Equivalence Ratio
0.5
Figure 95. Variation in Emission Concentrations With Over-
all Equivalence Ratio for Scheme FS-04A Firing
No. 2 Fuel With 0.5% N
-------
700
(RunsFS-04A-9, 10
andFS-04A-14-»H7)
O UHC
100psia
560°F
Shale DFM
0.1 0.2 0.3 0.4
Overall Equivalence Ratio
Figure 96. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme FS-04A Firing Shale DFM
111
-------
Figure 97. Evidence of Fuel Leak Caused by Cracked Manifold
112
-------
3.2.4.2 Second Test Series
In the second test series, the premix tube damper was installed and tests were performed
at inlet conditions representing three engine power settings (see Table XII). The configuration
evaluated, Scheme FS-04B, is shown in Figure 98. In Figure 99, a photograph of the premix
tube assembly with the variable damper mechanism attached is shown.
TABLE XII
RIG TEST CONDITIONS SIMULATING
VARIOUS ENGINE POWER SETTINGS
Idle
50% Power
100% Power
Notes:
Tn
CF)
320*
550
7802
PT3 Primary Airflow
(psia) (%)
40
96
1003
11
16
21
Direct-fired rig heater burner required, resulting in vitia-
tion of inlet air.
' Highest rig pressure. Engine value is 188 psia.
Figure 98. Full-Scale Combustor Scheme FS-04B
Test results obtained firing No. 2 fuel with 0.5% nitrogen (as pyridine) are shown in
Figure 100. At the idle setting, a minimum NO, concentration of 95 ppmv (corrected to 15%
O2) was measured at 0.18 overall equivalence ratio. This concentration is somewhat higher than
the levels measured at the 50 and 100% power points (84 and 81 ppmv, respectively),
indicating that the low inlet air temperature associated with the idle setting (320°F nominal)
has a detrimental effect on fuel vaporization (causing an increase in the occurrence of droplet
burning, and resultant higher NO,). The low inlet air temperature and the low rig pressure
(40 psia) also contributed to an increase in CO concentration levels (753 ppmv near the peak of
the curve, corrected to 15% O2) at the idle setting.
113
-------
Figure 99. Premix Tube With Variable Damper Attached
114
-------
800
I I I
(Runs FS-04B-1 to 5, 15 to 18, 19 to 22)
NOX CO
Damper Setting
Idle
50% Power
100% Power
No. 2 Fuel With 0.5% N
A V
0.1
0.2 0.3 0.4
Overall Equivalence Ratio
0.5
0.6
0.7
Figure 100. Variation in Emission Concentrations With Overall Equivalence
Ratio for Scheme FS-04B Firing No. 2 Fuel With 0.5% N
115
-------
The intermediate and baseload power settings (50 and 100% power) are meant to differ
in primary zone airflow (Table XII) as well as rig inlet conditions. Although the damper
setting was varied in going from the 50 to the 100% power points, it appears that no
appreciable increase in primary airflow was effected (the minimum point in the NO, curve for
50% power occurs at 0.19 overall equivalence ratio compared to 0.20 for 100% power).
Apparently the residual blockage of the damper device in its full-open position caused a
reduction in premix tube airflow with respect to the quantity that can be passed when the
device is completely removed (in Figure 95, data obtained with the damper removed show a
minimum point in the NO, curve at 0.24 overall equivalence ratio, indicating an increase in
primary airflow from about 15% to about 18% of the total combustor airflow). Aside from the
intended difference in primary airflow setting, the 50 and 100% power points differ primarily
in inlet air temperature (550°F vs 780°F, per Table XIII). At the higher temperature, the rig
direct-fired heater burner is operated. The rig supply limit of 100 psia precludes testing at 188
psia, the full baseload pressure; therefore, the difference between the 50 and 100% point rig
pressure conditions is only 4 psi.
In Figure 100, it may be seen that the NO, curves for the 50 and 100% power points are
nearly identical, reflecting the similarity in primary zone airflow rates and rig pressure, and
indicating that the increase in inlet air temperature had no appreciable effect (the
NO, concentrations at the 100% power point were corrected by subtracting the 12 ppmv
contribution of the heater burner, separately measured, from the raw data). Figure 100 also
shows that the maximum CO concentration measured at the 100% power point was 86 ppmv,
compared to 182 ppmv at the 50% power point, a decrease due almost entirely to the higher
inlet air temperature (CO concentrations at the 100% power point were also corrected by
subtracting the 37 ppmv contribution of the heater burner from the raw data).
Test results obtained firing shale-derived DFM are shown in Figure 101. Data points
were recorded at the bottom of the NO, curve bucket for the idle setting and for the 50%
power setting (minimum NO, concentrations were ascertained by monitoring the gas analyzer
reading while adjusting rig fuel flow). A minimum concentration of 80 ppmv (corrected to 15%
O2) was documented at idle; at the 50% power setting, 75 ppmv (corrected to 15% O2) was
achieved. Data were not recorded at the 100% power setting because of the close similarity of
that point to the 50% power point (there had been no appreciable difference in
NO, concentrations measured at the two points in the previous tests conducted with
pyridine-spiked No. 2 fuel). The CO data shown in Figure 101 are comparable to those shown
in Figure 100 for pyridine-spiked No. 2 fuel.
In Figure 102, test results obtained firing non-nitrogenous No. 2 fuel are presented. As in
the case of shale-derived DFM, data were recorded only at the idle and 50% power points.
Minimum NO, concentrations of 49 and 43 ppmv (corrected to 15% O2) were demonstrated at
the idle and 50% power settings. The CO characteristics were comparable to those obtained
for the other two fuels.
For purposes of comparison, the NO, characteristics obtained for the three fuels at the
idle and 50% power points are summarized in Figure 103. Minimum concentrations of 49, 80,
and 95 ppmv (corrected to 15% 02) were measured at idle for No. 2 fuel (0% nitrogen), shale
DFM (0.24% nitrogen), and pyridine-spiked No. 2 fuel (0.5% nitrogen), respectively. At 50%
power, the minimum concentrations were 43, 75, and 84 ppmv (corrected to 15% 02)
respectively, for the same three fuels.
116
-------
(Runs FS-04B-8, 13, & 14)
600
500
CM
O
<5
I 400
0.
01
.1
a
c
o 300
o
O
c
o
en
10
*" 200
100
0
NOX CO Damper Setting
O O ld|e
A V 5QO/° Power
Shale DFM
»>--
^
O A-
\
3 0.1 0.2 0.:
Overall Equivalence Ratio
Figure 101. Variation in Emiaxion Concentrations With Ov-
erall Equivalence Ratio for Scheme FS-04B Fir-
ing Shale DFM
700
Q.
Q.
on
O
600
CN
O
Lf>
Z 500
03
400
300
c
0>
o
o
O
o 200
100
0
(R
jns FS-04B-6, 7, and 9 to 12)
i i i
NOX CO Damper Setting
O
A
9
i
i
i
i
i
i
i
i
if
&O^
0
V
No. 2
7
Idle
50% Pox
Fuel
/ver
0 0.1 0.2 0.3 0.4
Overall Equivalence Ratio
0.5
Figure 102. Variation in Emission Concentrations With Ov-
erall Equivalence Ratio for Scheme FS-04B Fir-
ing No. 2 Fuel
-------
400
OD
CM
O
# 300
in
Q.
Q.
I 200
O)
u
o
O
c
O
.£ 100
E
0
Idle
O No. 2
£ No. 2 with 0.5% N
n Shale F
O
0-0
400
300
200
100
0.1
A No. 2
A No. 2 with 0.5% N
O Shale DFM
0.2 0.3 0
Overall Equivalence Ratio
Figure 103. Comparison of NO, Characteristics at the Idle and 50% Power Settings; Showing Variation With Fuel Type
-------
Composite results showing the use of the premix tube damper to vary the NO, character-
istics of the combustor are presented in Figure 104. The data shown are from tests conducted
using No. 2 fuel with 0.5% nitrogen. Because of the close proximity of the 50 and 100% power
settings (due to the lack of full modulation capability of the premix tube damper at these two
points, and the nearly identical rig pressure levels), composite results are shown for data
generated using both Scheme FS-04A and FS-04B (in Scheme FS-04A, the absence of the
damper resulted in greater premix tube airflow, providing data representative of a higher
power setting), as well as Scheme FS-04B (which had the damper attached) alone. Using data
from both schemes, the movement of the NO, curve bucket from 0.17 to 0.235 overall
equivalence ratio can be demonstrated.
Examination of the combustor following the second test series indicated that pieces of the
premix tube damper had broken loose during the test (due to fatigue failure of tack welds).
One piece was ingested into the premix tube where it lodged against several spraybars. There
was some damage to the premix tube swirler, as a result of flameholding inside the premix
passage, in the wake of the ingested part. This occurrence was due to mechanical failure and
does not, we believe, reflect any deficiency in the aerodynamic design of the premix tube.
Otherwise, the burner was found to be in good condition, with the exception of some
deterioration in the flamespray coating and the failure of several tack welds on the guide
chutes in the quick-quench section.
3.2.5 Liner Temperatures
The effect of fuel type on liner temperature levels in the FRT combustor was reported in
Section 3.4.3. Data from skin thermocouples attached to the outer surface of the combustor
liner were used to compute values of the liner temperature rise factor (LTRF). This parameter
provides a basis of comparison for two fuels (in this case No. 2 fuel and shale-derived DFM) in
terms of the overall average liner temperature rise (normalized to burner ideal temperature
rise).
Although only five liner thermocouples were available, and although the "liner tem-
perature rise" computed can be expected to vary in absolute value with the number and
placement of thermocouples, with the movement of the flamefront inside the combustor, and
with other factors, LTRF "is a useful indicator of the relative change in liner temperatures
when identical tests (same combustor configuration and operating conditions) are conducted
using two separate fuels. Results for the FRT combustor indicated that the shale DFM
produced an increase in liner temperature rise, as great as 4rf of the burner ideal temperature
rise greater at some settings, when compared to No. 2 fuel.
During the rework of the combustor hardware (from the FRT to the ECV configuration),
the five original skin thermocouples were destroyed. They were replaced by ten thermocouples
on the ECV combustor, at the locations shown in Figure 91. It was planned that these
thermocouples would provide more complete liner temperature data, including absolute
readings at additional locations and a greater base for the LTRF.
Data for the ECV combustor are presented in Table IV, Appendix A. The maximum
individual temperature recorded was 1908°F; however, this reading was taken just prior to
failure of the cable leading to the thermocouple in question and may not be accurate (the
output from several thermocouples was lost due to the battering of cables on the outside of the
rig in hot gas flowing from a leaking gasket). Several other readings of 1800 to 1860°F were
also recorded.
119
-------
400
(Data From Figure 13]
400
300
200
100
0
I I
Scheme FS - 04A and FS -04B
(Data From Figures 8 and 13)
0.1
0.2
0.3 o
Overall Equivalence Ratio
Figure 104. Composite Results Showing Use of the Premix Tube Damper to Vary NO, Characteristics of the Combustion
0.3
-------
Computations of the LTRF were performed as planned; however, the values obtained are
significantly different from those presented in Figure 86 (for the FRT combustor). Two of the
five readings used in the prior FRT computation were taken from thermocouples located in the
aft dilution section of the combustor where the liner temperatures are relatively low. The other
three were located in sections of the primary zone that appear relatively insensitive to
flamefront movement (in some other sections of the combustor, readings can actually decline
as the firing rate is increased, due to the shifting of zones of high heat release). As a result of
this particular placement of thermocouples in the FRT combustor, values computed for the
average liner temperature rise tend to be low, and then increase in direct proportion to the
burner ideal temperature rise. LTRF for the FRT combustor was essentially constant at a
value of about 0.4 (see Figure 86).
By contrast, values of LTRF computed for the ECV combustor, which are shown in
Figure 105, are considerably higher (0.4 to 1.4) and vary inversely with the combustor overall
equivalence ratio. Examination of the temperature data in Table IV shows that all
thermocouples exhibit high readings at some or all of the overall equivalence ratio settings (low
temperatures measured on the liner of the aft dilution section, which was removed in the ECV
combustor, are no longer present). Parameters TLIN1 through TLIN5 also show trends opposite to
the burner ideal temperature rise, presumably the result of flamefront movement. As a
consequence, the values of average liner temperature rise computed from these data are higher
than those obtained for the FRT combustor, and more nearly invariant with burner ideal
temperature rise. When normalized to the ideal burner temperature rise in the computation of
LTRF, the average liner temperature rise declines sharply with increasing overall equivalence
ratio. The data in Figure 105 thus indicate that relatively high temperatures exist in some
portions of the combustor liner even at low overall equivalence ratios, and that shifting of the
temperature pattern occurs as the setting is increased. The spread between the maximum liner
temperatures measured at the low-power and full-power settings does not appear to be great.
Interpretation of the LTRF data in Figure 105 to determine the effect of fuel type (No. 2
fuel vs shale DFM) on liner temperature rise was not possible because of scatter in the ECV
combustor data. The emergence of scatter in comparison to the previous data obtained for the
FRT combustor may have been a result of the strong dependence of LTRF on overall
equivalence ratio in the case of the ECV combustor.
3.2.6 Residence Time Effects
Bench-scale data indicating the dependence of the minimum attainable NO, concentra-
tion on primary zone residence time were presented in Figure 7. In Figure 106, full-scale
combustor data for the FRT and ECV configurations are compared to the previous bench-scale
results. For the purposes of these comparisons, effective primary zone volumes of 0.818 ft3 and
0.568 ft3 were assumed for the FRT and ECV configurations, respectively. Values of the
primary airflow rate, inlet air temperature, and pressure were taken from the data tables in
Appendix A. For the FRT combustor, test points FS-03A-26 and FS-03A-54 were selected as
representative; FS-04A-6 and FS-04A-13 were selected for the ECV combustor. The residence
time values shown are based on the cold flow characteristics of the combustor, and were
computed as follows for test FS-04A-6:
air density at 564°F and 100.4 psia = 0.265 lb/ft3
primary zone volume = 0.568 ft3
primary zone airflow = 3.618 Ib/sec
rm = (0.265) (0.568)73.618 = 0.042 sec
121
-------
1.5
(RunsFS-04A-1 FS-04B-22)
1.3
1.1
0.9
-------
100
CM
O
in
4->
(O
a.
o.
x
O
.a
tO
Full Scale Combustor Data
O No. 2 Fuel
I I
No. 2 Fuel With 0.5% Nitrogen
A = 10
PPMV
A = 10
PPMV
Bench Scale Data
No. 2 Fuel With 0.5% Nitrogen
Bench Scale Data, No. 2 Fuel
ECV FRT
II I
20
0.04 0.08 0.12 0.16 0.20
Primary Zone Residence Time (Cold Flow) - sec
Figure 106. Variation in Minimum NO, Concentration With Primary Residence Time
0.24
0.28
-------
The results in Figure 106 indicate that the full-scale combustor data points lie above the
curves for the bench-scale combustor. It is noteworthy that there is a decline in the minimum
attainable NO, concentration for the ECV combustor (42 msec) compared to the FRT
combustor (61 msec) when firing No. 2 fuel. This result, which was discussed earlier in this
section, has been attributed to the purging effect of the axially-directed penetration airflow
that was introduced through the wall of the dump section of the ECV combustor (at Station E
in Figure 91). By eliminating a region of recirculating gases that may have contributed to
thermal NO, formulation in the FRT combustor (fuel-nitrogen NO, formation is less likely to
depend upon a region of increased residence time), this change appears to have produced a
decline of about 10 ppmv (in 15% 02 units) in the minimum attainable NO, concentration. As
indicated in Figure 106, the same 10 ppmv increment matches the separation in curves that
can be projected through the two data points for No. 2 fuel with 0.5% nitrogen. It is
reasonable to expect that a decline in thermal NO, due to the altered airflow distribution
would appear in these results as well, and that the fuel-nitrogen NO, characteristics would be
largely unaffected.
124
-------
SECTION 4
CONCLUSIONS FROM PHASES III AND IV
With the completion of Phases III and IV of the program, several conclusions were
drawn:
1. The Rich Burn/Quick Quench combustor concept was successfully trans-
ferred 'from subscale to a size representative of a 25 megawatt (Mw) gas
turbine engine (GTE) combustor. Indicative of this transformation was the
demonstration of the same emission trends in the larger size combustor as
seen in the subscale combustors of Phase II.
2. Substantial emission reductions, representing improvements better than
the emission goals of the program, were demonstrated while operating
on both non-nitrogenous and nitrogen bearing fuels at pressures up to
nearly seven atm. Because Phase II results showed that the NO, emissions
of this combustion concept are independent of pressure level, it is reason-
able to expect that similar emission levels wold be achieved at pressure
levels typical of full-power conditions of a 25-Mw GTE.
3. Two lengths of the Rich Burn/Quick Quench combustor were tested in
Phase IV: one with about twice the length of a typical 25-Mw GTE
combustor; the other, sized to fit a typical in-line engine case envelope.
Both lengths of the combustor met the emission goals of the program.
4. Variable geometry was successfully employed to vary the airflow admitted
into the primary combustion volume. This demonstrated the ability to
meet the program emission goals over the range of operating conditions
experienced in a typical 25-Mw GTE.
5. The method of final dilution air addition was shown to be important in
NO, formation within the secondary zone.
6. The Rich Burn/Quick Quench combustor also met the program emission
goal while operating on a shale-derived diesel fuel marine. This indicates
the potential for handling other alternative fuels (both shale oil and coal
derived) by this combustion concept.
7. From the data gathered in Phase IV, the following areas of further
development were indicated:
Improvements in the exit temperature pattern factor. .
Primary zone liner cooling techniques and advanced mate-
rials for the primary zone liner.
Alternative fuel preparation devices to handle heavy fuels
and allow easier control of airflow.
Operation of the combustor on other alternative fuels and
at full engina conditions.
125
-------
LIST OF SYMBOLS
The following symbols are used in the test data summaries contained in Tables I through V.
Units
o de-
termined from metered fuel and air flowrates
Symbol
EQR %
Definition
Combustor overall fuel-air equivalence ratio de-
FflN Combustor inlet total pressure psia
TTIN Combustor inlet total temperature °F
WA Total combustor airflow rate pps
LPL Combustor total pressure loss 'V
FUEL Fuel type. "2" designates No. 2 fuel oil. "2P"
designates No. 2 fuel with pyridine (0.5%N)
PHIP Primary zone equivalence ratio
NOX15 NOX concentration corrected to 15'V 02 ppmv
NO15 NO concentration corrected to 15fr 02 ppmv
C015 CO concentration corrected to 15','r 02 ppmv
LJHC15 Unburned hydrocarbon concentration corrected to ppmv
15f>r 02
CO2 CO2 concentration, uncorrected, as measured pctv
O2 O2 concentration, uncorrected, as measured pctv
CFHAC Carbon balance parameter; total carbon out divided
by total carbon in
EFFGA Combustion efficiency from gas analysis mea- ''i
surements
TLIN1 Combustor liner temperatures, measured at locations °F
Through defined in Figure 5
TLIN6
WAPRI Primary zone airflow rate pps
WASEC Secondary zone (quick-quench) airflow rate pps
FA Overall fuel-air ratio determined from metered fuel
and air flowrates
TPF Temperature pattern factor
126
-------
REFERENCES
1. Mosier, S. A., "Advanced Combustion Systems for Stationary Gas Turbines,"
EPA-600/7-77-073e, July 1977, Presented at Second Stationary Source Combustion
Symposium, August 1977.
2. Lefebvre, A. H. and Herbert, M. V.; "Heat Transfer Processes in Gas Turbine Combus-
tion Chambers," Proceedings of the Institute of Mechanical Engineers (London), Vol.
174, No. 12, 1960, pp. 463-478.
3. Rizkalla, A. A., and A. H. Lefebrve, "The Influence of Air and Liquid Properties on
Airblast Atomization," Joint Fluids Engineering and ASME Conference, Montreal,
Quebec, 13-15 May 1974.
4. Adelberg, M., "Mean Drop Size Resulting from the Injection of a Liquid Jet Into a
High-Speed Gas Stream (Including Corrections to August 1967 Paper)," AIAA Journal,
Vol. 6, No. 6, June 1968.
5. Ingebo, Robert D., and Hampton H. Foster, "Drop-Size Distribution for Crosscurrent
Breakup of Liquid Jets III Airstreams," NACA Technical Note 4087, October 1957.
6. Weiss, Maldem A., and Charles H. Worsham, "Atomization in High Velocity Airstreams,"
ARS Journal, Vol. 29, No. 4, April 1959.
7. Nukiyama, S. and Y. Tanasawa, "Experiments on the Atomization of Liquids in an Air
Stream," Droplet-Size Distribution in an Atomized Jet, transl. by E. Hope, Rept. 3,
18 March 1960, Defense Research Board, Department of National Defense, Ottawa,
Canada; transl. from Transactions of the Society of Mechanical Engineers (Japan), Vol.
5, No. 18, February 1939.
8. Kurzius, S. C., and F. H. Raab, "Measurement of Droplet Sizes in Liquid Jets Atomized
in Low-Density Supersonic Streams," Rept. TP 152, March 1967, Aerochem Research
Labs., Princeton, N. J.
9. Lorenzetto, G. E. and A. H. Lefebrve, "Measurements of Drop Size on a Plain-Jet
Airblast Atomizer," AIAA 1976.
10. Ingebo, Robert D., "Effect of Airstream Velocity on Mean Drop Diameters of Water
Sprays Produced by Pressure and Air Atomizing Nozzles," Gas Turbine Combustion and
Fuels Technology, ASME, 27 November through 2 December 1977. Edited by E. Karl
Bastress.
11. Dombrowski, N., and W. R. Johns, "The Aerodynamic 'Instability and Disintegration of
Viscous Liquid Sheets," Chem. Eng. Sci., Vol. 18, 1963.
12. Wolfe, H. E., and W. H. Andersen, "Kinetics, Mechanism, and Resultant Droplet Sizes of
the Aerodynamic Breakup of Liquid Drops," Aerojet - General Corporation, Downey,
California, Report No. 0395-04 (18) SP/April 1964/Copy 23.
127
-------
13. Donaldson, Coleman, Snedeker, and Richard, "Experimental Investigation of the Struc-
ture of Vortices in Simple Cylindrical Vortec Chamber," ARAP Report No. 47, December
1962.
14. Chelko, Louis, "Penetration of Liquid Jets into a High Velocity Airstream," NACA
E50F21, 14 August 1950.
15. Koplin, M. A., K. P. Horn, and R. E. Reichenbach, "Study of a Liquid Injectant Into a
Supersonic Flow," AIAA Journal, Vol. 6, No. 5, May 1968, pp. 853-858.
16. Tacina, Robert, "Experimental Evaluation of Premixing/Prevaporizing Fuel Injection
Concepts for a Gas Turbine Catalytic Combustor," Gas Turbine Combustion and Fuels
Technology, ASME, 27 November through 2 December 1977, Edited by E. Karl Bastress.
128
-------
APPENDIX A
DATA LISTINGS
129
-------
TABLE I
COMBUSTOR OPERATING PARAMETER DATA
Tent No.
FS-01A-1
FS-01A-2
FS-01A-3
FS-01A-4
FS-01A-5
FS-02A-1
FS-02A-2*
FS-02A-3*
FS-02A-4*
FS-02A-5*
FS-02A-6*
FS-02A-7*
FS-02A-8*
FS-02A-9*
FS-02A-10*
FS-02A-11*
FS-02A-12*
FS-02A-13*
FS-02A-14*
FS-02A-15*
FS-03A-1
FS-03A-2
FS-03A-3
FS-03A-4
F8-03A-5
FS-03A-6
FS-03A-7
FS-03A-8
FS-03A-9
FS-OliA-10
FS-03A-11
FS-03A-12
FS-03A-13
FS-03A-14
FS-03A-15
FS-03A-16
FS-03A-17
FS-03A-18
FS-03A-19
FS-03A-20
FS-03A-21
FS-03A-22
F8-03A-23
FS-03A-24
FS-03A-25
FS-03A-26
FS-03A-27
FS-03A-28
FS-03A-29
FS-03A-30
FS-03A-31
FS-03A-32
FS-03A-33
FS-03A-34
FS-03A-35
FS-03A-36
EQR
0.1381
0.2108
0.1817
0.2456
0.1148
0.1715
0.2006
0.2253
0.2529
0.2776
0.3023
0.3285
0.1279
0.1439
0.1570
0.1701
0.1831
0.1962
0.2296
0.2616
0.1323
0.1613
0.1831
0.2035
0.2296
0.2485
0.2544
0.2369
0.2151
0.1977
0.0959
0.1264
0.1526
0.1773
0.2064
0.2224
0.2427
0.2703
0.2980
0.1294
0.1264
0.1570
0.1846
0.2006
0.2311
0.2645
0.2878
0.1831
0.2209
0.2282
0.2587
0.2907
0.1098
0.1344
0.1546
0.1864
PTIN
55.1
54.9
54.9
54.6
54.4
16.1
16.1
16.1
16.1
16.1
16.1
16.1
50.5
50.5
50.5
50.5
50.5
50.5
50.5
50.5
49.9
50.8
50.9
50.2
50.2
50.8
50.6
50.0
51.2
50.1
50.4
49.9
49.9
50.2
49.9
50.0
50.4
50.1
50.3
98.4
99.9
100.0
100.1
100.1
100.4
98.9
98.1
100.5
100.1
99.7
100.5
99.6
50.2
50.2
50.4
50.0
TTIN
501.0
509.7
512.3
516.3
514.7
488.0
488.0
488.0
488.0
488.0
488.0
488.0
539.7
539.7
539.7
539.7
539.7
539.7
539.7
539.7
453.0
456.0
460.0
461.0
453.0
468.3
470.0
470.0
470.0
470.0
395.0
398.0
399.0
402.0
403.0
405.0
402.0
402.0
404.0
549.0
568.0
574.0
576.0
577.0
580.0
581.0
580.0
578.0
579.0
580.0
579.0
582.0
460.0
473.0
479.0
483.0
WA
7.298
6.913
6.886
6.764
7.294
2.501
2.501
2.501
2.501
2.501
2.501
2.501
9.354
9.354
9.354
9.354
9.354
9.354
9.354
9.354
8.315
8.158
8.184
8.477
8.296
8.108
7.984
8.070
7.804
7.599
8.867
8.565
8.395
8.227
8.015
8.255
8.108
8.230
8.081
17.310
16.272
15.705
15.683
16.348
15.991
15.538
15.742
15.688
14.905
16.194
15.851
15.447
9.221
9.158
9.184
8.824
LPL
3.57
3.22
3.31
3.06
3.43
4.78
4.78
4.78
4.78
4.78
4.78
4.78
5.90
5.90
5.90
5.90
5.90
5.90
5.90
5.90
5.51
5.41
5.50
5.87
5.77
5.61
5.53
5.70
5.37
5.78
5.75
5.61
5.61
5.48
5.22
5.40
5.26
5.39
5.18
5.79
5.61
5.40
5.50
5.55
5.53
5.36
5.71
4.99
4.81
5.76
5.62
5.52
5.38
5.28
5.36
5.30
Fuel
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2P
2P
2P
2P
2P
2P
2P
2P
2P
2P
2P
2P
2P "
2P
2P
2P
2P
2P
2P
2P
2P
2
2
2
2
2
S
S
S
S
/
Test No.
FS-03A-37
FS-03A-38
FS-03A-39
FS-03A-40
FS-03A-41
FS-03A-42
FS-03A-43
FS-03A-44
FS-03A-45
FS-03A-46
FS-03A-47
FS-03A-48
FS-03A-49
FS-03A-50
FS-03A-51
FS-03A-52
FS-03A-53
FS-03A-54
FS-04A-1
FS-04A-2
FS-04A-3
FS-04A-4
FS-04A-5
FS-04A-6
FS-04A-7
FS-04A-8
FS-04A-9
FS-04A-10
FS-04A-11
FS-04A-12
FS-04A-13
FS-04A-14
FS-04A-15
FS-04A-16
FS-04A-17
FS-04B-1
FS-04B-2
FS-04B-3
FS-04B-4
FS-04B-5
FS-04B-6
FS-04B-7
FS-04B-8
FS-04B-9
FS-04B-10
FS-04B-11
FS-04B-12
FS-04B-13
FS-04B-14
FS-04B-15
FS-04B-16
FS-04B-17
FS-04B-18
FS-04B-19
FS-04B-20
FS-04B-21
FS-04B-22
EQR
0.2066
0.2326
0.2616
0.1113
0.1402
0.1633
0.1879
0.2211
0.2370
0.2514
0.1192
0.1337
0.1613
0.1875
0.2020
0.2282
0.2413
0.2602
0.1090
0.1352
0.1642
0.1817
0.2078
0.2355
0.2660
0.2514
0.1850
0.2124
0.1977
0.2035
0.2689
0.2326
0.2760
0.2110
0.1647
0.1773
0.1032
0.2253
0.1453
0.1032
0.1061
0.1497
0.1532
0.1439
0.1701
0.1962
0.2238
0.2168
0.1922
0.2267
0.1904
0.1701
0.1453
0.2006
0.2296
0.2631
0.1846
PTIN
50.0
50.2
50.1
100.3
99.6
99.7
100.8
100.3
100.3
100.5
100.4
100.0
100.5
101.1
97.2
98.1
98.3
99.1
99.4
100.9
99.4
98.5
99.3
100.4
100.3
99.9
99.3
99.9
99.8
96.2
97.5
100.9
100.9
99.2
99.2
40.0
39.8
41.4
39.1
41.1
39.4
39.2
41.2
92.6
93.8
94.0
95.2
95.6
95.4
95.6
93.4
92.2
93.0
97.6
99.2
99.4
99.4
TTIN
485.0
487.0
488.0
561.0
566.0
568.0
569.0
570.0
570.0
570.0
568.0
567.0
567.0
568.0
569.0
568.0
569.0
570.0
563.0
564.0
563.0
564.0
563.0
564.0
562.0
561.0
565.0
565.0
564.0
563.0
564.0
573.0
575.0
574.0
574.0
309.0
317.0
318.0
318.0
318.0
314.0
315.0
327.0
549.0
556.0
559.0
562.0
561.0
562.0
562.0
563.0
563.0
562.0
782.0
778.0
777.0
783.0
WA
8.907
8.718
8.563
18.188
17.326
17.415
17.301
16.558
17.086
16.939
17.085
18.373
17.797
17.309
18.187
17.856
17.740
17.558
18.674
18.356
17.540
17.963
17.657
17.332
16.844
17.114
17.724
17.310
16.575
17.700
16.823
16.939
16.277
17.126
17.136
7.479
8.091
7.451
7.496
6.759
7.705
7.519
8.040
17.555
17.273
16.859
16.658
16.802
17.070
16.833
17.149
17.421
17.464
16.702
16.255
16.255
15.952
LPL
5.50
5.48
5.39
5.63
5.47
5.57
5.46
5.24
5.63
5.52
5.09
5.99
5.62
5.25
6.32
6.26
6.15
6.05
5.93
5.55
5.39
5.73
5.49
5.43
5.14
5.31
5.79
5.56
5.07
6.38
5.79
5.60
5.36
6.09
6.14
4.91
6.05
4.51
6.53
3.35
6.11
5.64
5.25
5.73
5.39
5.23
5.16
5.04
5.15
4.98
5.42
5.59
5.33
5.39
6.14
5.73
5.19
Fuel
S
S
S
S
S
S
S
S
S
S
2
2
2
2
2
2
2
2
2P
2P
2P
2P
2P
2P
2P
2P
S
S
2
2
2
S
S
S
S
2P
2P
2P
2P
2P
2
2
S
2
2
2
2
S
S
2P
2P
2P
2P
2P
2P
2P
2P
Values listed are approximate only emissions and fuel
flow were read, airflow was maintained nearly constant.
130
-------
TABLE II
EMISSION CONCENTRATION DATA
Test No.
FS-01A-1
KS-OIA-2
FS-01A-3
KS-01A-4 '
FS-OIA-5
FS-02A-1
KS-02A-2*
FS-02A-3*
KS-02A-4*
FS-02A-5*
FS-02A-6*
FS-02A-7*
FS-02A-8*
FS-02A-9*
FS-02A-10*
FS-02A-11*
FS-02A-12*
FS-02A-13*
FS-02A-14*
FS-02A-15*
FS-03A-1
FS-03A-2
FS-03A-3
FS-03A-4
FS-03A-5
FS-03A-6
KS-0:tA-7
FS-();iA-H
FS-o;tA-9
FS-03A-10
KS-():iA-ll
FS-03A-12
FS-03A-13
FS-03A-14
FS-03A-I5
FS-03A-16
FS-03A-17
FS-03A-18
FS-03A-19
FS-03A-20
FS-03A-21
FS-03A-22
FS-03A-23
FS-03A-24
FS-03A-25
FS-03A-26
FS-03A-27
FS-03A-28
FS-03A-29
FS-03A-30
FS-03A-31
FS-03A-32
FS-03A-33
FS-03A-34
FS-03A-35
FS-03A-36
FS-03A-37
EQR
0.1381
0.2108
0.1817
0.2456
0.1148
0.1715
0.2006
0.2253
0.2529
0.2776
0.3023
0.3285
0.1279
0.1439
0.1570
0.1701
0.1831
0.1962
0.2296
0.2616
0.1323
0.1613
0.1831
0.2035
0.2296
0.2485
0.2544
0.2369
0.2151
0.1977
0.0959
0.1264
0.1526
0.1773
0.2064
0.2224
0.2427
0.2703
0.2980
0.1294
0.1264
0.1570
0.1846
0.2006
0.2311
0.2645
0.2878
0.1831
0.2209
0.2282
0.2587
0.2907
0.1098
0.1344
0.1546
0.1864
0.2066
PHIP
0.7485
1.2629
1.0080
1.6139
0.5954
0.7114
0.8333
0.9350
1.0501
1.1518
1.2534
1.3618
0.6141
0.6893
0.7520
0.8147
0.8773
0.9400
1.0967
1.2533
0.6158
0.7279
0.8398
0.9301
1.1185
1.1305
1.0775
1.0138
0.9528
0.8095
0.4681
0.6155
0.7421
0.8918
0.9989
1.0806
1.1646
1.3063
1.4452
0.6482
0.5650
0.7001
0.8121
0.9125
1.0351
1.1917
1.2880
0.8462
1.0007
1.0385
1.1595
1.2977
0.5434
0.6642
0.7623
0.9154
1.0317
NO,,:,
150.3
60.7
84.9
55.6
104.7
37.4
40.2
36.7
27.7
22.6
19.9
19.4
37.1
74.5
75.9
71.3
62.2
51.9
32.9
31.1
91.8
177.5
128.4
73.0
25.9
29.8
107.8
89.9
75.2
136.0
171.8
193.8
226.0
185.2
129.4
96.3
73.1
69.8
80.8
258.3
265.8
442.0
294.2
180.0
104.9
79.2
95.9
203.6
81.9
67.3
48.2
43.7
'114.5
196.9
230.9
138.4
93.3
NO,,
137.1
56.1
77.8
53.2
86.6
37.0
40.2
36.4
27.4
22.6
19.9
19.4
28.8
65.3
61.1
54.8
55.3
46.5
30.9
30.3
83.0
150.3
92.5
49.1
15.4
20.9
85.7
68.7
53.1
105.6
170.4
186.1
211.5
183.7
102.6
67.5
45.2
44.0
62.8
232.5
246.7
401.4
254.2
150.5
82.2
67.6
87.8
162.6
59.8
48.2
37.3
39.9
111.6
182.7
210.9
109.2
65.2
CO,,
251.0
192.5
220.5
154.8
348.8
451.2
608.9
671.8
516.3
328.4
189.4
143.4
92.5
144.5
248.2
309.5
343.7
332.4
244.0
121.5
72.9
175.9
255.9
296.2
291.4
202.3
160.9
218.3
284.7
249.5
18.1
33.8
79.3
189.0
296.5
322.4
298.4
256.6
181.4
47.5
53.1
79.8
105.6
115.6
97.6
60.7
42.0
103.4
88.0
96.4
69.8
37.5
50.1
100.6
191.0
325.4
362.9
UHCIS
7.8
1.4
1.5
1.0
20.7
7.3
7.1
4.1
3.0
3.5
2.8
2.8
2.8
2.9
3.0
6.4
4.9
3.7
3.1
2.7
1.9
1.4
1.2
1.1
4.4
21.8
14.1
9.3
11.4
6.1
4.0
3.5
3.8
3.0
2.7
2.3
1.8
15.3
8.9
6.3
3.5
2.6
Test No.
FS-03A-38
FS-03A-39
FS-03A-40
FS-03A-41
FS-03A-42
FS-03A-43
FS-03A-44
FS-03A-45
FS-03A-46
FS-03A-47
FS-03A-48
FS-03A-49
FS-03A-50
FS-03A-51
FS-03A-52
FS-03A-53
FS-03A-54
FS-04A-1
FS-04A-2
FS-04A-3
FS-04A-4
FS-04A-5
FS-04A-6
FS-04A-7
FS-04A-8
FS-04A-9
FS-04A-10
FS-04A-11
FS-04A-12
FS-04A-13
FS-04A-I4
FS-04A-15
FS-04A-16
FS-04A-17
FS-04B-1
FS-04B-2
FS-04B-3
FS-04B-4
FS-04B-5
FS-04B-6
FS-04B-7
FS-04B-8
FS-04B-9
FS-04B-10
FS-04B-I1
FS-04B-12
FS-04B-13
FS-04B-14
FS-04B-15
FS-04B-16
FS-04B-17
FS-04B-18
FS-04B-19
FS-04B-20
FS-04B-21
FS-04B-22
EQR
0.2326
0.2616
0.1113
0.1402
0.1633
0.1879
0.2211
0.2370
0.2514
0.1192
0.1337
0.1613
0.1875
0.2020
0.2282
0.2413
0.2602
0.1090
0.1352
0.1642
0.1817
0.2078
0.2355
0.2660
0.2514
0.1850
0.2124
0.1977
0.2035
0.2689
0.2326
0.2760
0.2110
0.1647
0.1773
0.1032
0.2253
0.1453
0.1032
0.1061
0.1497
0.1532
0.1439
0.1701
0.1962
0.2238
0.2168
0.1922
0.2267
0.1904
0.1701
0.1453
0.2006
0.2296
0.2631
0.1846
PHIP
1.1501
1.3075
0.5758
0.7352
0.8431
0.9761
1.1320
1.2275
1.2944
0.6145
0.6865
0.8170
0.9468
1.0369
1.1557
1.2243
1.3115
0.5242
0.6620
0.7949
0.8879
1.0062
1.1301
1.2649
1.1870
0.9369
0.9405
1.0417
1 .3399
1.1988
1.3543
1.0679
0.9126
__
NO,,,
67.4
64.0
161.4
392.5
292.7
141.8
72.4
63.5
65.7
169.2
330.6
290.1
128.9
81.0
51.0
46.2
44.9
194.2
364.4
289.2
201.0
143.4
83.5 .
110.6
101.6
87.0
65.3
59.0
46.4
37.9
70.3
93.7
57.8
100.7
95.1
185.4
113.2
115.6
173.fi
51.0
49.2
80.0
181.1
73.4
48.2
43.0
81.2
75.4
94.1
83.6
120.0
239.3
81.1*
97.8*
143.3*
95.5*
NO,,
45.0
47.5
128.6
342.3
244.4
101.2
47.3
42.1
47.0
141.2
289.7
240.9
94.1
50.0
31.3
29.7
32.9
192.4
359.5
184.7
93.0
56.8
31.2
61.7
45.0
28.1
22.0
18.2
15.0
18.4
33.7
74.8
17.3
28.8
79.2
174.5
100.1
103.8
166.4
44.2
39.1
41.8
155.1
48.2
28.5
28.9
57.0
52.5
80.7
58.2
86.6
196.3
67.7*
93.0*
140.6*
82.7*
co,s
320.2
213.4
81.3
124.4
174.2
195.5
165.6
149.9
118.3
89.4
122.1
166.8
183.1
224.8
185.2
151.9
105.3
55.3
86.1
363.4
481.7
481.1
481.0
306.9
402.0
641.5
585.3
561.6
594.1
294.5
291 .6
81.4
622.6
698.5
748.8
218.0
517.5
752.8
183.4
65.9
494.0
601 .6
171.2
200.0
174.5
134.3
143.8
174.6
105.8
161.0
178.R
181.6
85.9
56.0
34.8
93.0
l/WC,,
2.1
1.5
9.7
5.0
3.1
2.2
1.6
1.4
1.3
5.0
4.0
2.2
1.3
1.3
1.1
1.0
0.9
10.0
5.1
8.9
11.4
10.7
10.8
3.1
5.5
33.8
21.6
23.7
24.4
5.4
7.8
2.2
20.3
32.5
7.1
10.fi
5.1
8.3
.
_..
_.
._
_-_
Corrected for oxides of nitrogen from vitiation of inlet air.
131
-------
TABLE HI
GAS ANALYSIS PARAMETER DATA
Test No.
FS-01A-1
FS-01A-2
FS-01A-3
FS-01A-4
FS-01A-5
FS-02A-1
FS-02A-2*
FS-02A-3*
FS-02A-4*
FS-02A-5*
FS-02A-6*
FS-02A-7*
FS-02A-8*
FS-02A-9*
FS-02A-10*
FS-02A-11*
FS-02A-12*
FS-02A-13*
FS-02A-14*
FS-02A-15»
FS-03A-1
FS-03A-2
FS-03A-3
FS-03A-4
FS-03A-5
FS-03A-6
FS-03A-7
FS-03A-8 .
FS-03A-9
FS-03A-10
FS-03A-11
FS-03A-12
FS-03A-13
FS-03A-14
FS-03A-15
FS-03A-16
FS-03A-17
FS-03A-18
FS-03A-19
FS-03A-20
FS-03A-21
FS-03A-22
FS-03A-23
FS-03A-24
FS-03A-25
FS-03A-26
FS-03A-27
FS-03A-28
FS-03A-29
FS-03A-30
FS-03A-31
FS-03A-32
FS-03A-33
FS-03A-34
FS-03A-35
FS-03A-36
FS-03A-37
EQR
0.1381
0.2108
0.1817
0.2456
0.1148
0.1715
0.2006
0.2253
0.2529
0.2776
0.3023
0.3285
0.1279
0.1439
0.1570
0.1701
0.1831
0.1962
0.2296
0.2616
0.1323
0.1613
0.1831
0.2035
0.2296
0.2485
0.2544
0.2369
0.2151
0.1977
0.0959
0.1264
0.1526
0.1773
0.2064
0.2224
0.2427
0.2703
0.2980
0.1294
0.1264
0.1570
0.1846
0.2006
0.2311
0.2645
0.2878
0.1831
0.2209
0.2282
0.2587
0.2907
0.1098
0.1344
0.1546
0.1864
0.2066
CO,
2.04
3.13
2.65
3.65
1.67
2.36
2.70
3.09
3.54
3.90
4.27
4.67
2.16
2.44
2.61
2.81
3.07
3.27
3.34
4.44
2.01
2.44
2.78
3.04
3.50
3.79
3.75
3.44
3.08
2.85
1.32
1.69
2.00
2.33
2.74
2.97
3.28
3.58
3.95
1.79
1.88
2.34
2.72
2.98
3.40
4.02
4.36
2.79
3.34
3.47
3.89
4.39
1.55
1.88
2.17
2.59
2.90
0,
18.22
16.60
17.25
15.83
18.45
16.75
16.45
15.86
15.26
14.81
14.27
13.76
17.55
17.32
16.84
16.52
16.15
15.86
15.05
14.27
18.63
17.89
18.39
17.09
17.09
16.08
15.84
16.13
16.51
16.73
19.37
18.87
18.30
17.80
17.27
16.86
16.29
15.87
15.31
18.40
18.79
17.86
17.32
16.98
16.28
15.58
14.98
16.88
16.15
15.92
15.37
14.57
19.18
18.75
18.44
17.88
17.44
CFRAC
1.0583
1.0691
1.0499
1.0699
1.0407
0.9918
0.9745
0.9948
1.0122
1.0142
1.0194
1.0271
1.1931
1.2059
1.1879
1.1821.
1.2014
1.1941
1.2028
1.2156
1.0752
1.0773
1.0802
1.0640
1.0927
1.0931
1.0531
1.0372
1.0196
1.0247
0.9677
0.9480
0.9339
0.9466
0.9555
0.9645
0.9782
0.9577
0.9576
0.9857
1.0599
1.0608
1.0579
1.0614
1.0575
1.0892
1.0888
1.0903
1.0843
1.0887
1.0821
1.0854
0.9854
0.9889
0.9946
0.9923
0.9973
EFFGA
99.67
99.76
99.73
99.81
99.51
99.46
99.27
99.19
99.38
99.60
99.77
99.83
99.89
99.83
99.70
99.63
99.59
99.60
99.71
99.85
99.89
99.77
99.68
99.64
99.64
99.75
99.80
99.73
99.65
99.69
99.96
99.94
99.89
99.76
99.63
99.61
99.64
99.69
99.78
99.93
99.86
99.86
99.84
99.82
99.86
99.91
99.94
99.86
99.88
99.87
99.91
99.95
99.89
99.85
99.75
99.60
99.56
Test No.
FS-03A-38
FS-03A-39
FS-03A-40
FS-03A-41
FS-03A-42
FS-03A-43
FS-03A-44
FS-03A-45
FS-03A-46
FS-03A-47
FS-03A-48
FS-03A-49
FS-03A-50
FS-03A-51
FS-03A-52
FS-03A-53
FS-03A-54
FS-04A-1
FS-04A-2
FS-04A-3
FS-04A-4
FS-04A-5
FS-04A-6
FS-04A-7
FS-04A-8
FS-04A-9
FS-04A-10
FS-04A-11
FS-04A-12
FS-04A-13
FS-04A-14
FS-04A-15
FS-04A-16
FS-04A-17
FS-04B-1
FS-04B-2
FS-04B-3
FS-04B-4
FS-04B-5
FS-04B-6
FS-04B-7
FS-04B-8
FS-04B-9
FS-04B-10
FS-04B-11
FS-04B-12
FS-04B-13
FS-04B-14
FS-04B-15
FS-04B-16
FS-04B-17
FS-04B-18
FS-04B-19
FS-04B-20
FS-04B-21
FS-04B-22
EQR
0.2326
0.2616
0.1113
0.1402
0.1633
0.1879
0.2211
0.2370
0.2514
0.1192
0.1337
0.1613
0.1875
0.2020
1 0.2282
0.2413
0.2602
0.1090
0.1352
0.1642
0.1817
0.2078
0.2355
0.2660
0.2514
0.1850
0.2124
0.1977
0.2035
0.2689
0.2326
0.2760
0.2110
0.1647
0.1773
0.1032
0.2253
0.1453
0.1032
0.1061
0.1497
0.1532
0.1439
0.1701
0.1962
0.2238
0.2168
0.1922
0.2267
0.1904
0.1701
0.1453
0.2006
0.2296
0.2631
0.1846
CO,
3.24
3.71
1.55
2.00
2.30
2.65
3.12
3.37
3.59
1.71
1.95
2.34
2.72
2.97
3.35
3.54
3.84
1.63
2.02
2.39
2.65
3.04
3.43
3.86
3.64
2.64
3.00
2.86
3.01
3.96
3.44
4.00
3.09
2.33
2.40
1.38
3.13
1.92
1.38
1.47
2.20
2.22
2.06
2.49
2.86
3.29
3.17
2.87
3.29
2.89
2.45
2.11
3.64*
4.03*
4.61*
3.34*
0,
17.23
16.36
18.96
18.32
17.94
17.47
16.73
16.31
15.98
18.64
18.37
17.83
17.30
16.95
16.48
16.14
15.67
19.86
19.80
19.74
17.39
17.95
17.17
16.08
16.33
17.83
17.53
17.67
17.53
16.75
16.91
16.13
17.37
18.34
17.81
18.97
16.60
18.11
18.86
18.78
17.60
17.77
17.77
17.32
16.74
16.09
16.32
16.79
16.15
16.85
17.23
17.77
15.71*
15.08*
14.16*
15.86*
CFRAC
0.9943
1.0096
0.9805
1.0029
1.0014
1.0049
1.0035
1.0085
1.0135
1.0154
1.0317
1.0316
1.0298
1.0475
1.0473
1.0515
1.0572
1.0572
1.0617
1.0454
1.0532
1.0585
1.0502
1.0491
1.0462
1.0253
1.0180
1.0493
1.0686
1.0620
1.0606
1.0334
1.0515
1.0220
0.9828
0.9586
1.0046
0.9584
0.9585
0.9818
1.0610
1.0425
1.0202
1.0482
1.0451
1.0555
1.0407
1.0660
1.0418
1.0813
1.0254
1.0399
1.0303
1.0393
1.0689
1.0301
EFFGA
99.61
99.74
99.87
99.83
99.78
99.76
99.80
99.82
99.85
99.88
99.84
99.79
99.78
99.73
99.77
99.81
99.87
99.90
99.88
99.53
99.38
99.39
99.38
99.62
99.50
99.11
99.22
99.25
99.20
99.63
99.62
99.89
99.18
99.05
99.08
99.70
99.36
99.07
99.49
99.92
99.41
99.28
99.80
99.76
99.79
99.84
99.83
99.79
99.87
99.81
99.79
99.78
99.92
99.94
99.96
99.91
Includes effect of vitiated inlet air.
132
-------
TABLE IV
COMBUSTOR LINER TEMPERATURE DATA
Test No.
FS-01A-1
FS-01A-2
FS-01A-3
FS-01A-4
FS-01A-5
FS-02A-1
FS-02A-2*
FS-02A-3*
FS-02A-4*
FS-02A-5*
FS-02A-6*
FS-02A-7*
FS-02A-8*
FS-02A-9*
FS-02A-10*
FS-02A-11*
FS-02A-12*
FS-02A-13*
FS-02A-14*
FS-02A-15*
FS-03A-1
FS-03A-2
FS-03A-3
FS-03A-4
FS-03A-5
FS-03A-6
FS-03A-7
FS-03A-8
FS-03A-9
FS-03A-10
FS-03A-11
FS-03A-12
FS-03A-13
FS-03A-14
FS-03A-15
FS-03A-16
FS-03A-17
FS-03A-18
FS-03A-19
FS-03A-20
FS-03A-21
FS-03A-22
FS-03A-23
FS-03A-24
FS-03A-25
FS-03A-26
FS-03A-27
FS-03A-28
FS-03A-29
FS-03A-30
FS-03A-31
F8-03A-32
FS-03A-33
F8-03A-34
FS-03A-35
FS-03A-36
FS-03A-37
FS-03A-38
FS-03A-39
FS-03A-40
FS-03A-41
FS-03A-42
FS-03A-43
F8-03A-44
EQR
0.1381
0.2108
0.1817
0.2456
0.1148
0.1715
0.2006
0.2253
0.2529
0.2776
0.3023
0.3285
0.1279
0.1439
0.1570
0.1701
0.1831
0.1962
0.2296
0.2611
0.1323
0.1613
0.1831
0.2035
0.2296
0.2485
0.2544
0.2369
0.2151
0.1977
0.0959
0.1264
0.1526
0.1773
0.2064
0.2224
0.2427
0.2703
0.2980
0.1294
0.1264
0.1570
0.1846
0.2006
0.2311
0.2645
0.2878
0.1831
0.2209
0.2282
0.2587
0.2907
0.1098
0.1344
0.1646
0.1864
0.2066
0.2326
0.2616
0.1113
0.1402
0.1633
0.1879
0.2211
(Dome
Avg)
TUN,
1066
1410
1366
1279
1130
1025
1115
1033
1075
1123
1136
1120
1065
1097
1078
1210
1103
871
895
964
1023
1080
1090
1102
1134
1122
1025
1045
1107
1173
1214
1255
1291
1287
1193
1295
1247
1275
1296
943
948
960
1004
1047
1104
1177
989
1076
1113
1168
1228
BST3
TLIN.
960
1226
1175
1301
916
1193
1136
1208
1280
1358
1399
1430
972
942
902
989
1228
810
BST6
TLIN,
1030
1403
1305
1496
915
1244
1166
1304
1431
1501
1537
1552
1171
1160
1126
1244
1452
1054
1142
1240
1338
1429
1461
1498
1532
1605
1385
1398
1501
1570
1666
1724
1761
1717
1633
1745
1711
1751
1666
1237
1351
1426
1623
1566
1637
1666
1366
1642
1601
1684
1777
BST7
TLIN,
1021
1392
1289
1513
900
1192
1106
1314
1427
1502
1533
1547
1298
1308
1280
1445
1486
1108
1221
1331
1403
1434
1490
1540
1304
1263
1259
1391
1465
1503
1575
1645
1661
1480
1605
1566
1622
1628
1137
1269
1364
1463
1502
1696
1699
1233
1390
1459
1532
1657
BST8
TLIN,
556
637
600
688
556
570
602
505
522
533
537
538
578
573
565
572
599
428
440
451
463
472
473
480
490
519
590
609
626
640
635
647
671
695
637
650
648
663
694
499
520
533
644
663
664
686
597
611
618
628
645
BST9 BST10 BST12
TLIN, TLIN, TLIN.
659
787
728
828
640
691
701
592
644
677
695
709
783
770
755
757
727
481
520
555
582
616
624
651
661
689
660
679
711
742
757
804
R54
868
781
790
801
847
882
543
570
690
612
627
648
679
632
655
669
688
717
133
-------
TABLE IV
COMBUSTOR LINER TEMPERATURE DATA (Continued)
Test No.
FS-03A-45
FS-03A-46
FS-03A-47
FS-03A-48
FS-03A-49
FS-03A-50
FS-03A-51
FS-03A-52
FS-03A-53
FS-03A-54
FS-04A-1
FS-04A-2
FS-04A-3
FS-04A-4
FS-04A-5
FS-04A-6
FS-04A-7
FS-04A-8
FS-04A-9
FS-04A-10
FS-04A-11
FS-04A-12
FS-04A-13
FS-04A-14
FS-04A-15
FS-04A-16
FS-04A-17
FS-04B-1
FS-04B-2
FS-04B-3
FS-04B-4
FS-04B-5
FS-04B-6
FS-04B-7
FS-04B-8
FS-04B-9
FS-04B-10
FS-04B-11
FS-04B-12
FS-04B-13
FS-04B-14
FS-04B-15
FS-04B-16
FS-04B-17
FS-04B-18
FS-04B-19
FS-04B-20
FS-04B-21
FS-04B-22
EQR
0.2370
0.2514
0.1192
0.1337
0.1613
0.1875
0.2020
0.2282
0.2413
0.2602
0.1090
0.1352
0.1642
0.1817
0.2078
0.2355
0.2660
0.2514
0.1850
0.2124
0.1977
0.2035
0.2689
0.2326
0.2760
0.2110
0.1647
0.1773
0.1032
0.2253
0.1453
0.1032
0.1061
0.1497
0.1532
0.1439
0.1701
0.1962
0.2238
0.2168
0.1922
0.2267
0.1904
0.1701
0.1453
0.2006
0.2296
0.2631
0.1846
(Dome
Aug)
TUN,
1224
1234
1027
1041
1085
1120
1124
1191
1128
1171
1211
1265
1277
1279
1133
1143
1411
1394
1387
1299
1251
1182
1052
1166
1171
1093
892
1155
998
921
858
978
1040
1288
1371
1424
1418
1488
1456
1401
1478
1460
1383
1504
1488
1490
1492
BST3
TUN,
1049
1157
1211
1251
1300
980
725
729
1239
1129
1213
1090
950
969
834
1077
1337
919
807
862
865
793
791
901
903
1095
1192
1258
1177
1246
1281
1206
1334
1347
1295
1309
1323
1248
1366
BST6
TUN,
1788
1806
1421
1474
1550
1587
1599
1651
1683
1727
1195
1318
1371
1419
1463
1416
1179
1228
1475
1490
1494
1422
1167
1423
1037
1264
1459
1193
1043
1265
1138
1141
1057
1180
1278
1435
1494
1526
1391
1455
1580
1541
1523
1478
1422
1490
1588
1551
1650
BST7
TLINt
1690
1713
1290
1347
1428
1473
1480
1555
1602
1656
1223
1387
1464
1492
1534
1218
920
920
1533
1478
1498
1414
1096
1138
885
1497
1522
1276
1022
1224
1163
1105
1168
1231
1290
1477
1549
1535
1382
1468
1619
1422
1526
1546
1531
1510
1315
1234
BST8
TUN.
655
663
607
610
618
628
629
638
645
660
1254
1358
1405
1388
1440
944
828
848
1335
1213
1144
858
921
1123
924
1512
1439
1185
880
1293
1080
893
939
1095
1087
1309
1430
1479
1488
1588
1524
BST9
TLIN,
728
744
649
656
'673
691
696
715
728
747
1438
1581
1706
1786
1793
1783
1834
1827
1783
1791
1763
1726
1642
1860
1858
1785
1740
1172
1258
1179
1513
1246
1236
1512
1543
1692
1750
1775
1783
1837
1823
1367
1500
1605
1670
1604
1750
1720
1797
BST10
TLIN,
1403
1581
1578
1643
1715
1715
1695
1746
1679
1708
1671
1640
1583
1622
1339
1550
1603
_
BST12
TUN,
1423
1576
1684
1707
1706
1693
1804
1771
1665
1703
1641
1600
1639
1908
--
.
134
-------
TABLE V
PERFORMANCE PARAMETER DATA
Test No.
FS-01A-1
FS-01A-2
FS-01A-3
FS-01A-4
FS-01A-5
FS-02A-1
FS-02A-2*
FS-02A-3*
FS-02A-4*
FS-02A-5*
FS-02A-6*
FS-02A-7*
FS-02A-8*
FS-02A-9*
FS-02A-10*
FS-02A-11*
FS-02A-12*
FS-02A-13*
FS-02A-14*
FS-02A-15*
FS-03A-1
FS-03A-2
FS-03A-3
FS-03A-4
FS-03A-5
FS-03A-6
FS-03A-7
FS-03A-8
FS-03A-9
FS-03A-10
FS-03A-11
FS-03A-12
FS-03A-13
FS-03A-14
FS-03A-15
FS-03A-16
FS-03A-17
FS-03A-18
FS-03A-19
FS-03A-20
FS-03A-21
FS-03A-22
FS-03A-23
FS-03A-24
FS-03A-25
FS-03A-26
FS-03A-27
FS-03A-28
FS-03A-29
FS-03A-30
FS-03A-31
FS-03A-32
FS-03A-33
FS-03A-34
FS-03A-35
FS-03A-36
FS-03A-37
FS-03A-38
FS-03A-39
F8-03A-40
F8-03A-41
FS-03A-42
FS-03A-43
FS-03A-44
FS-03A-45
EQR
0.1381
0.2108
0.1817
0.2456
0.1148
0.1715
0.2006
0.2253
0.2529
0.2776
0.3023
0.3285
0.1279
0.1439
0.1570
0.1701
0.1831
0.1962
0.2296
0.2616
0.1323
0.1613
0.1831
0.2035
0.2296
0.2485
0.2544
0.2369
0.2151
0.1977
0.0959
0.1264
0.1526
0.1773
0.2064
0.2224
0.2427
0.2703
0.2980
0.1294
0.1264
0.1570
0.1846
0.2006
0.2311
0.2645
0.2878
0.1831
0.2209
0.2282
0.2587
0.2907
0.1098
0.1344
0.1546
0.1864
0.2066
0.2326
0.2616
0.1113
0.1402
0.1633
0.1879
0.2211
0.2370
WAPRI
1.350
1.155
1.243
1.031
1.411
0.603
0.603
0.603
0.603
0.603
0.603
0.603
1.956
1.956
1.956
1.956
1.956
1.956
1.956
1.956
1.790
1.805
1.787
1.858
1.702
1.781
1.885
1.888
1.767
1.861
1.828
1.766
1.732
1.630
1.660
1.696
1.688
1.700
1.666
3.446
3.642
3.527
3.552
3.603
3.564
3.455
3.521
3.397
3,290
3.564
3.528
3.466
1.874
1.846
1.858
1.791
1.789
1.763
1.715
3.515
3.317
3.358
3.318
3.233
3.308
WASEC
2.861
2.613
2.667
2.537
2.818
0.994
0.994
0.994
0.994
0.994
0.994
0.994
3.857
3.857
3.857
3.857
3.857
3.857
3.857
3.857
3.217
3.262
3.210
3.189
3.344
3.198
3.245
3.170
3.164
3.030
3.545
3.899
3.839
3.844
3.704
3.701
3.783
3.707
3.650
7.768
7.179
7.008
6.940
6.990
6.880
6.484
6.521
6.604
6.338
6.633
6.598
6.502
4.005
3.975
4.022
3.899
3.836
3.783
3.778
7.766
7.708
7.680
7.277
7.258
7.424
VREF
16.7
14.5
15.6
13.1
18.0
25.2
25.2
25.2
25.2
25.2
25.2
25.2
27.5
27.5
27.5
27.5
27.5
27.5
27.5
27.5
23.2
23.1
22.9
24.2
21.9
23.1
24.6
24.9
22.8
24.5
22.0
21.6
21.2
19.9
20.4
20.8
20.5
20.8
20.3
25.1
26.6
25.9
26.1
26.5
26.2
25.8
26.5
24.9
24.3
26.4
25.9
25.7
24.3
24.3
24.5
23.9
24.0
23.6
23.0
25.4
24.2
24.6
24.0
23.5
24.1
EFFMB
121.1
125.9
124.4
127.6
121.0
84.4
-jj
98J6
-
116.0
110.6
113.6
115.3
116.4
104.4
103.4
108.9
104.1
109.1
111.2
107.6
103.5
105.6
104.8
109.4
109.5
109.1
108.9
113.4
108.1
108.8
110.5
111.3
109.7
110.3
108.7
101.6
103.3
104.2
104.3
103.9
140.6
144.3
148.3
151.9
157.3
158.2
166.5
139.1
151.3
152.0
157.3
161.3
166.4
TPF
0.53
0.52
0.51
0.56
0.53
0.91
0.70
0.48
0.34
0.36
0.43
0.43
0.40
0.37
0.29
0.36
0.34
0.45
0.43
0.38
0.38
0.45
0.46
0.48
0.48
0.50
0.53
0.54
0.54
0.49
0.35
0.49
0.69
0.48
136
-------
TABLE V
PERFORMANCE PARAMETER DATA (Continued)
Test No.
FS-03A-46
FS-03A-47
FS-03A-48
FS-03A-49
FS-03A-50
FS-03A-51
FS-03A-52
FS-03A-53
FS-03A-54
FS-04A-1
FS-04A-2
FS-04A-3
FS-04A-4
FS-04A-5
FS-04A-6
FS-04A-7
FS-04A-8
FS-04A-9
FS-04A-10
FS-04A-11
FS-04A-12
FS-04A-13
FS-04A-14
FS-04A-15
FS-04A-16
FS-04A-17
FS-04B-1
FS-04B-2
FS-04B-3
FS-04B-4
FS-04B-5
FS-04B-6
FS-04B-7
FS-04B-8
FS-04B-9
FS-04B-10
FS-04B-11
FS-04B-12
FS-04B-13
FS-04B-14
FS-04B-15
FS-04B-16
FS-04B-17
FS-04B-18
FS-04B-19
FS-04B-20
FS-04B-21
FS-04B-22
EQR
0.2514
0.1192
0.1337
0.1613
0.1875
0.2020
0.2282
0.2413
0.2602
0.1090
0.1352
0.1642
0.1817
0.2078
0.2355
0.2660
0.2514
0.1850
0.2124
0.1977
0.2035
0.2689
0.2326
0.2760
0.2110
0.1647
0.1773
0.1032
0.2253
0.1453
0.1032
0.1061
0.1497
0.1532
0.1439
0.1701
0.1962
0.2238
0.2168
0.1922
0.2267
0.1904
0.1701
0.1453
0.2006
0.2296
0.2631
0.1846
WAPRI
3.296
3.312
3.579
3.508
3.438
3.549
3.528
3.490
3.477
3.895
3.745
3.615
3.671
3.639
3.618
3.540
3.628
3.502
3.901
3.461
3.377
3.283
3.325
3.392
3.088
'
WASEC
7.477
7.571
7.974
7.813
7.674
7.745
7.643
7.599
7.574
6.476
6.856
5.962
6.755
6.711
6.622
6.509
6.549
6.928
6.860
7.088
6.739
6.764
7.885
8.340
7.706
7.736
2.924
3.434
3.092
3.078
2.719
3.112
2.882
3.008
6.684
6.618
6.402
6.380
6.486
6.598
6.924
6.571
7.256
7.451
6.581
6.442
6.319
6.417
VREF
24.0
24.0
26.1
25.4
24.8
26.6
26.2
25.9
25.6
28.4
27.0
26.4
27.1
26.6
26.2
25.6
26.3
25.7
28.4
26.1
25.2
23.8
24.2
25.1
22.8
EFFMB
168.9
144.4
152.6
158.0
165.0
166.6
174.4
172.1
172.9
119.1
107.6
103.8
110.4
113.2
130.1
140.3
133.8
125.1
132.8
129.9
135.3
146.0
119.1
109.1
114.9
101.4
147.1
132.9
152.7
135.3
TPF
136
-------
APPENDIX B
SI UNIT CONVERSION TABLE
S7 Multiply by
°C °C = (5/g)(°F-32)
cm 2.54
cm2 0.1550
liters 0.0164
m 0.3048
m2 0.0929
m3 0.0283
m/sec 0.3048
N/m2 3.3863
kg/sec 0.4535
kg/hr 0.4535
m3 0.003785
w/m2 315.24808
N/m2 6894.7572
137
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-80-017C
2.
3. RECIPIENT'S ACCESSION NO.
4 TITLEANDSUBT1TLE Advanced Combustion Systems for
Stationary Gas Turbine Engines: Volume 3.
Combustor Verification Testing
5. REPORT DATE
January 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R.M. Pierce, C.E. Smith, and B.S. Hinton
8. PERFORMING ORGANIZATION REPORT NO.
FR-11405
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Pratt and Whitney Aircraft Group
United Technologies Corporation
P.O. Box 2691
West Palm Beach, Florida 33402
10. PROGRAM ELEMENT NO.
INE829
11. CONTRACT/GRANT NO.
68-02-2136
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AISID PERIOD COVERED
Final; 1/78 - 4/79
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES IERL-RTP project officer is W.S. Lanier, Mail Drop 65, 919/541-
2432.
. ABSTRACT
rep0rts des cribe an exploratory development program to identify, eval-
uate, and demonstrate dry techniques for significantly reducing NOx from stationary
gas turbine engines. (Volume 1 describes Phase I research activities to compile a
series of combustor design concepts which could potentially meet the program goals ,
and Volume 2 describes the Phase n bench-scale evaluation of those techniques: the
rich-burn/quick-quench (RB/QQ) concept was found to be effective in limiting pollu-
tant emissions when burning either clean fuels or fuels containing significant amounts
of chemically bound nitrogen. ) Volume 3 describes the scaleup of the RB/QQ model
to a full-scale (25 MW) gas turbine combustor, and documents test results from the
full-scale evaluations. Test results were very positive, showing that the RB/QQ
concept can reduce NOx to approximately 45 ppm (at zero % O2) for clean distillate
oil and to approximately 75 ppm for a distillate oil doped to 0. 5% nitrogen, as pyri-
dine. CO emissions below the 100 ppm program goal were also demonstrated. These
tests also indicate that the new combustor concept may be capable of low emission
performance on petroleum residual oil and synthetic liquid fuels such as SRC II or
shale oil. Results from testing on those fuels is included in Volume 4, an addendum.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Atomizing
Gas Turbine Engines Shale Oil
Stationary Engines
Nitrogen Oxides
Carbon Monoxide
Combustion
Combustion Chambers
Pollution Control
Stationary Sources
Combustor Design
Staged Combustion
Dry Controls
Fuel Preparation
Fuel-bound Nitrogen
13B
2 IE
2 IK
07B
2 IB
13H
2 ID
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
152
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
138
------- |