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

-------