EXPERIMENTS WITH A

HIGH PRESSURE WELL STIRRED REACTOR

Thesis

Submitted to

The School of Engineering of the

UNIVERSITY OF DAYTON

In Partial Fulfillment of the Requirements for

The Degree

Master of Science in Aerospace Engineering

By

Justin Tyler Gross

UNIVERSITY OF DAYTON

Dayton, Ohio

December, 2014
 EXPERIMENTS WITH A HIGH PRESSURE WELL STIRRED REACTOR

Name: Gross, Justin Tyler

APPROVED BY:

________________________________ ________________________________
Dr. Scott Stouffer, Ph.D. Dr. Jamie Ervin, Ph.D.
Advisory Committee Chairman Committee Member
Senior Research Engineer
Energy and Environmental Engineering

________________________________
Dr. Alejandro Briones, Ph.D.
Committee Member
Senior Research Engineer
University of Dayton Research Institute

________________________________ ________________________________
John G. Weber, Ph.D. Eddy M. Rojas, Ph.D., M.A., P.E.
Associate Dean Dean
School of Engineering School of Engineering

ii
 ABSTRACT

EXPERIMENTS WITH A HIGH PRESSURE WELL STIRRED REACTOR

Name: Gross, Justin Tyler

University of Dayton

Advisor: Dr. Scott Stouffer, Ph.D.

The existing WSR has been successfully used to evaluate pollutant emissions,

soot formation, combustion stability and alternative fuel performance at atmospheric

pressure. The HPWSR project extends the capabilities of this laboratory tool, allowing

the generation of benchmark quality data over a range of operating pressures. Lean blow

out data collected is correlated against loading parameter, forming fuel stability loops.

Combustion stability of JP-8 will be evaluated over a range of operating pressures from 4

psia to 5 atm. A new, internal vaporization technique is applied to allow operation at

higher pressures than with previous hardware. Lean emissions will be collected

simultaneously and used to further characterize the HPWSR operability and lean

combustion performance. This work will validate nitrogen dilution using true low

pressure testing compared to dilution in the same configuration, and further extend WSR

capabilities by operating at elevated pressures up to 20 atm.

iii
 ACKNOWLEDGEMENTS

Scott Stouffer, thank you for your support and guidance in making this research

possible. Our time spent guiding me through the engineering design process has helped

me to develop the skills and connections that will lead me to a truly rewarding career

To Craig Neuroth, Brad Day and David Baker, thank you for your support and

effort in integrating our hardware to the high pressure facility. Without your help and

expertise this experiment would not have been such a success.

To Alejandro Briones and Jamie Ervin, thank you for participating in my thesis

defense and for your time reviewing my thesis. Your comments helped strengthen the

presentation of my work, as well as gave useful feedback to help future writings.

To my wife, Kerianne, thank you for your support as I went through for school,

and for continuing to push me to finish this thesis even when I would have rather been

watching football.

Finally, a special thanks to funding from the Hons von Ohain Fellowship, which

made it possible to complete this research.

iv
 TABLE OF CONTENTS

ABSTRACT ....................................................................................................................... iii

ACKNOWLEDGEMENTS ............................................................................................... iv

LIST OF FIGURES ........................................................................................................... ix

LIST OF TABLES ............................................................................................................ xii

LIST OF SYMBOLS AND ABBREVIATIONS ............................................................ xiii

CHAPTER 1 INTRODUCTION ....................................................................................... 1

1.1. COMBUSTION STUDIES ........................................................................... 1

1.2. OBJECTIVES ............................................................................................... 2

1.3. CHAPTER OUTLINE .................................................................................. 3

CHAPTER 2 BACKGROUND ......................................................................................... 4

2.1. COMBUSTION STABILITY ...................................................................... 4

2.1.1. Loading Parameter .................................................................................... 4

2.1.2. Semi-empirical Model for Combustion at Sub-atmospheric Pressures .... 8

2.1.3. Evaluation of Alternative Fuels............................................................... 11

2.2. PREVIOUS WSR CONFIGURATIONS ................................................... 12

2.2.1. Spherical .................................................................................................. 12

2.2.2. Toroidal WSR ......................................................................................... 13

2.3. AFRL/UDRI ATMOSPHERIC WSR ........................................................ 13

2.4. PREVIOUS WSR RESEARCH ................................................................. 15

v
 2.4.1. Emissions and Soot Formation................................................................ 16

2.4.2. Lean Blowout .......................................................................................... 18

2.5. TOTAL TEMPERATURE PROBE ........................................................... 20

2.5.1. Importance of This Study ........................................................................ 21

2.5.2. Theoretical Development ........................................................................ 22

2.5.3. Temperature Probe Correction ................................................................ 25

2.5.4. Comparison to Laser Diagnostics ........................................................... 28

2.5.5. Related Work........................................................................................... 29

CHAPTER 3 EXPERIMENTAL SETUP AND PROCEDURE ..................................... 31

3.1. ATMOSPHERIC EXPERIMENTS WITH A METAL REACTOR .......... 31

3.2. HIGH PRESSURE WELL STIRRED REACTOR DESIGN ..................... 33

3.2.1. Autoignition Issues .................................................................................. 37

3.2.2. Fuel Vaporization .................................................................................... 39

3.2.3. Vaporizer Safety...................................................................................... 41

3.3. INTERNAL VAPORIZATION .................................................................. 45

3.3.1. Experiments with a WSR at Elevated Pressure ....................................... 46

3.4. FINAL ASSEMBLY .................................................................................. 48

3.5. TEST PLAN AND PROCEDURE ............................................................. 50

3.5.1. Experiments at Atmospheric Pressure..................................................... 50

3.5.2. Alternative Fuel Testing .......................................................................... 52

3.5.3. Nitrogen Dilution .................................................................................... 55

3.6. EXPERIMENTS WITH AN EXTERNAL VAPORIZER ......................... 57

3.6.1. Sub-atmospheric ...................................................................................... 60

vi
 3.6.2. High Pressure Operation ......................................................................... 61

3.7. EXPERIMENTS USING THE INTERNAL VAPORIZATION

TECHNIQUE .............................................................................................. 62

3.7.1. High Pressure Operation ......................................................................... 62

3.7.2. Sub-atmospheric ...................................................................................... 63

3.8. TOTAL TEMPERATURE PROBE ........................................................... 64

3.8.1. Thermocouple Comparisons ................................................................... 66

3.8.2. Laser Diagnostics .................................................................................... 67

CHAPTER 4 EXPERIMENTAL RESULTS AND DISCUSSION ................................ 70

4.1. ALTERNATIVE FUEL EVALUATION ................................................... 70

4.2. COMBUSTION STABILITY .................................................................... 74

4.2.1. Lean Blowout at Atmospheric Pressure .................................................. 74

4.2.2. Experiments at Low and Elevated Pressures........................................... 76

4.2.3. Low Pressure Simulation with Nitrogen at Atmospheric Pressure ......... 83

CHAPTER 5 CONCLUSION.......................................................................................... 87

5.1. PROBLEM SUMMARY ............................................................................ 87

5.2. CONCLUSIONS......................................................................................... 88

5.3. FUTURE WORK ........................................................................................ 90

BIBLIOGRAPHY ............................................................................................................. 92

APPENDIX A TOTAL TEMPERATURE PROBE RESULTS ...................................... 97

A.1. THERMOCOUPLE COMPARISONS ................................................... 97

A.2. LASER DIAGNOSTICS....................................................................... 100

vii
 A.3. PROBE MEASUREMENTS WITH RADIATION-CONDUCTION

CORRECTION ................................................................................................... 102

APPENDIX B VAPORIZER STRESS ANALYSIS ...................................................... 107

viii
 LIST OF FIGURES

Figure 2.1: Stability curves at varying pressures ............................................................... 5

Figure 2.2: Stability loop correlated with loading parameter ............................................. 6

Figure 2.3: Lower half of the toroid with the jet ring in the WSR housing ...................... 15

Figure 2.4: Set-up and instrumentation points of the ceramic total temperature probe .... 22

Figure 3.1: HPWSR with metal reactor assembly ............................................................ 32

Figure 3.2: Reactor stack up with sealing rings ................................................................ 34

Figure 3.3: HPWSR hardware with pressure vessel ......................................................... 36

Figure 3.4: Jet ring inlet design......................................................................................... 37

Figure 3.5: Air assist atomization nozzle.......................................................................... 39

Figure 3.6: Impact of droplet size on vaporization ........................................................... 40

Figure 3.7: Dewpoint pressure dependence for JP-8 ........................................................ 42

Figure 3.8: Mixture temperature vs. dewpoint at maximum vaporizer pressure .............. 43

Figure 3.9: Rankine-Hugoniot plot with annotated regions.............................................. 44

Figure 3.10: Fuel injection points for internal vaporization ............................................. 47

Figure 3.11: Final HPWSR hardware installation ............................................................ 49

Figure 3.12: Flow path schematic for experiments with an external vaporizer ................ 59

Figure 3.13: Laboratory set-up of the ceramic total temperature probe ........................... 65

Figure 3.14: Probe orientation for TDLS investigation .................................................... 69

ix
Figure 4.1: Lean blowout of alternative fuels in a metal reactor ...................................... 71

Figure 4.2: Sensitivity of lean blowout equivalence ratio to DCN .................................. 72

Figure 4.3: Effect of ceramic and metal reactors on LBO ................................................ 73

Figure 4.4: Lean blowout in an atmospheric WSR ........................................................... 75

Figure 4.5: Comparison of previous LBO data in a ceramic WSR .................................. 76

Figure 4.6: Combustion stability curve for JP-8 ............................................................... 77

Figure 4.7: Blowout points taken with internal and external vaporization ....................... 78

Figure 4.8: Pressure dependence of JP-8 lean blowout in a WSR .................................... 79

Figure 4.9: CO emissions for JP-8 at 5 atm ...................................................................... 80

Figure 4.10: Minimum CO emissions for JP-8 at 5 atm ................................................... 81

Figure 4.11: CO emissions of JP-8 at 1 and 5 atm ........................................................... 82

Figure 4.12: Emissions of CO2 and H2O at 5 atm ............................................................ 83

Figure 4.13: JP-8 combustion stability curve using simulated low pressure .................... 84

Figure 4.14: Correlation of true low pressure to nitrogen dilution data ........................... 86

Figure A.1: WSR exhaust temperature comparison ......................................................... 97

Figure A.2:Temperature comparison of McKenna burner flame ..................................... 98

Figure A.3:Comparison of temperatures in a hydrogen Hencken burner flame ............... 99

Figure A.4: Comparison of temperatures in an ethylene Hencken burner flame ........... 100

Figure A.5: Intrusive impact of a probe on a Hencken burner flame ............................. 101

Figure A.6: Comparison of corrected probe hydrogen flame temperatures ................... 103

Figure A.7: Comparison of corrected probe ethylene flame temperatures ..................... 104

x
Figure A.8:WSR exhaust temperature with probe radiation-conduction correction ...... 105

Figure A.9: McKenna burner ethylene flame temperature with radiation-conduction

correction ............................................................................................................ 106

xi
 LIST OF TABLES

Table 3.1 Atmospheric Blowout Test Matrix ................................................................... 52

Table 3.2: Selected Alternative Fuels ............................................................................... 53

Table 3.3: Test Conditions for Alternative Fuels.............................................................. 54

Table 3.4: Conditions for Nitrogen Dilution Testing........................................................ 57

Table 3.5: Test Conditions for Sub-Atmospheric Blowout with External Vaporization .. 61

Table 3.6: Test Conditions for High Pressure Blowout with External Vaporization ....... 62

Table 3.7: Test Conditions for High Pressure Blowout with Internal Vaporization ........ 63

Table 3.8: Test Conditions for Sub-Atmospheric Blowout with Internal Vaporization ... 64

Table 3.9: Test Conditions for Ceramic Probe to Thermocouple Comparison ................ 67

Table 3.10: Test Conditions for Hencken Burner Test ..................................................... 68

Table 3.11: Test Conditions for Intrusive Characterization.............................................. 69

xii
 LIST OF SYMBOLS AND ABBREVIATIONS

English Symbols
A = Arrhenius reaction rate expression
C = constant of integration
D = diameter
E = activation energy
h = heat transfer coefficient (W/m2-K)
k = thermal conductivity (W/m2-K)
= mass flow rate (lb/min)

m’’’ = mass flux

mo = mass flow rate of oxidizer
mf = mass flow rate of fuel
MW = molecular weight (g/mol)
q = rate of heat transfer (W)
q” = heat flux (W/m2)
T = temperature (K)
U = velocity (m/s)
x = distance (m)

Greek Symbols
γ = ratio of specific heats
ρ = density (kg/m2)
τ = residence time time (s)
Φ = equivalence ratio

xiii
Subscripts
1 = location in flame
2 = downstream location
4 = turbine inlet
∞ = freestream
cond = conduction
condi = condition
conv = convection
rad = radiation
stoich = stoichiometric
t = total
x = location

Abbreviations
AFRL = Air Force Research Laboratory
HPCRF = High Pressure Combustion Research Facility
SLPM = Standard Liters Per Minute
WSR = Well Stirred Reactor

xiv
 CHAPTER 1

INTRODUCTION

1.1. COMBUSTION STUDIES

Modern gas turbine combustors have become increasingly regulated to reduce

both fuel consumption and pollutant emissions. Growing concerns over the supply of oil

currently used to make transportation fuel has spurred the development of alternative

fuels that could be suitable for gas turbine engines. These alternative fuels seek to use

commonly available feed stock to produce viable transportation fuels to help reduce the

dependence on oil. More efficient combustors can be designed to operate fuel-leaner to

reduce fuel consumption. As the fuel to the combustor is reduced, the flame is more

highly stressed. In order for next generation combustors to optimize efficiency and

operate in the most fuel-lean conditions, a complete understanding of the stability limits

is required. Moreover, as new alternative fuels are developed, characterization of their

physical and chemical properties is required to determine if the new fuel is feasible in a

gas turbine combustor, and if it can meet the emissions and stability requirements of

modern engines. All current gas turbine engines are designed to operate on oil-based jet

fuel. Testing of alternative fuels is needed to demonstrate that their combustion is stable,

generates adequate power and burns cleanly to make alternative fuels viable replacements

for existing jet fuels.

1
 Research of fundamental combustion characteristics can be achieved in a Well

Stirred Reactor (WSR). The well stirred reactor is a laboratory idealization of a gas

turbine combustor that utilizes pre-mixed, pre-vaporized fuel and air flow to isolate

chemical kinetic effects from the fluid dynamic effects. The WSR is primarily used to

evaluate particulate and pollutant emissions and the combustion stability of hydrocarbon

fuels.

1.2. OBJECTIVES

In a real gas turbine engine, the combustor is operating at elevated pressures and

temperatures. To fully understand how the combustion behaves in a gas turbine

combustor, characterization at similar pressures and temperatures is required. Lefevre [21]

states that increasing the temperature and pressure of the combustion gases improvers the

atomization of fuel droplets, increases the turbulent flame speed, and increases the

reaction rate of combustion. The results is that increased temperature and pressure yield

more efficient combustion that produces less pollutants and is more stable than

combustion occurring at lower temperatures and pressures. Gas turbine combustors

could also exhibit sub-atmospheric pressures if the combustor were to blow out while at

cruising altitude [21, 23]. It is critical to the safety of the plane that the stability of

combustion at low pressures be characterized.

The objective of the current study is to use the existing geometry of the WSR and

adapt it for experiments at elevated and sub-atmospheric pressures that are more

representative of the conditions present in a gas turbine engine. The expanded capability

of the WSR will allow characterization of the effects of increased pressure on the

2
combustion stability and efficiency of the flame at higher pressures. Additionally, the

capability of the WSR to achieve sub-atmospheric pressure will allow the study of the

limits of combustion at low pressures.

1.3. CHAPTER OUTLINE

Chapter 2 elaborates on the background of the WSR, and the history of the design and

past experiments using the geometry. The section also contains explanation of important

parameters to combustion, as well as a method to simulate low pressure reactions using

nitrogen dilution.

Chapter 3 details the design and experimental setup and procedure of the high pressure

WSR. Discussion includes many of the design challenges when dealing with fuel

vaporization at high temperatures and pressures, as well as the experimental points

covered during the campaign.

Chapter 4 presents and discusses the experimental results of studies using the WSR at

elevated and subatmospheric pressure as well as the lean blowout stability of alternative

fuels, and characterization of a ceramic temperature probe.

Chapter 5 presents the conclusions, and suggests possibilities for future study.

3
 CHAPTER 2

BACKGROUND

2.1. COMBUSTION STABILITY

Fuel consumption in a modern gas turbine engine can be greatly decreased by

operating the combustor at fuel-lean conditions. As the fuel flow in decreased, however,

the fuel approaches its lean flammability limit where combustion can no longer be

sustained. The equivalence ratio at which combustion ceases is referred as lean blowout.

Operating an engine near lean blowout, however, requires a solid understanding of the

flame stability over the range of air flow rates, pressures and temperatures that will be

experienced during flight. Therefore, it is important that the stability characteristics of the

combustor be defined so that all operating conditions can be shown to fall within the

operating envelope. [39, 21,23]

2.1.1. Loading Parameter

Often in engine testing, the blowout equivalence ratio is plotted against the chamber

reference velocity, due to the relationship between bulk flow velocity and flame speed.
[21]
If the velocity of the gas stream through the combustor is higher than the turbulent

flame speed, the flame will move downstream and the combustion will not be stable. [21]

However, plotting stability loops in this way requires that a new curve be created for each

4
different pressure. Figure 2.1 shows typical stability curves as a function of combustor

velocity. As combustor velocity increases, more fuel is required to sustain combustion.

The stability curve will continue to increase in fuel/air ratio as combustor velocity

increases until the flame reaches the stoichiometric fuel/air ratio. The stoichiometric

fuel/air ratio is the most stable condition for combustion because it is the condition of

highest heat release, and it represents the peak on the stability curve. Any more fuel

added to the system past the stoichiometric fuel/air ratio will decrease the combustion

temperature, and therefore the stability. In addition to the fuel/air mixture, the combustor

operating pressure impacts the combustion stability. As the pressure is increased, the

flame speed and reaction rate will increase, improving the stability of the flame by

allowing the flame front to exist in a higher velocity air stream. For each different

operating pressure, a new stability loop is defined.[21]

Figure 2.1: Stability curves at varying pressures [21]

Early work performed by Longwell and Weiss [41] demonstrated that flame

stability could be correlated with the air loading parameter. The air loading parameter is

derived from reaction rate theory, and is defined in Equation 1.

5
 (1)

In Equation 1, n is the effective order of the reaction, is the combustor mass flow rate,

V is the reactor volume and P is the reaction pressure. Longwell and Weiss [41] determined

n to be 1.8 for isooctane. Further experimental work by Kretschmer and Odgers[4]

recommended that the reaction order can be related to the combustor equivalence ratio,

. The correlations by Kretschmer and Odgers[4] showed that for lean mixtures

and for rich mixtures.

Figure 2.2 shows similar stability curves correlated with the air loading

parameter. The benefit of using the loading parameter is that the x-axis is a function of

pressure and mass flow, allowing data collected at different pressures, combustor

volumes, or flow rates to be plotted on a single curve.

Figure 2.2: Stability loop correlated with loading parameter [21]

6
 In Figure 2.2, many different operating points are correlated to a single curve,

representing the region in which combustion can be sustained in the combustor. To

increase the loading parameter, the mass flow into the combustor can be increased.

However, only so much mass flow may be passed for a given volume. As the mass flow

through the reactor is increased, the residence time of the fluid in the reactor volume is

decreased [21]. Because the volume of the burner is fixed, decreased residence time

increases the fluid velocity. As shown in Figure 2.1, the stability of a flame decreases

with increased velocity. To continue increasing the mass flow rate, the combustor volume

can be increased to maintain the flame in a stable burning range. Additionally, from the

definition of loading parameter, the operating pressure can be lowered to increase the

loading parameter. In general, lowering the operating pressure decreases the range of

operating conditions over which the combustor can be operated. Lower operating

pressures decrease the turbulent flame speed [39], therefore high mass flow rates lead to

flame blowout. All of these combinations are represented in a stability loop plotted

against loading parameter, making the loading parameter a valuable tool for combustor

characterization.

Many studies have been conducted using various laboratory combustors that

explore the stability characteristics of fuel at atmospheric pressures. Ballal et al. [5]

conducted a combustion stability experiment using a research reactor that simulated the

main features of an annular gas turbine combustor, with propane as a fuel. Lean blowout

was measured as a function of air loading parameter. Experimental data were compared

to predictions from the Perfectly Stirred Reactor (PSR) theory. Ballal et al.[5]

demonstrated, through optical access to the combustor that blowout is preceded by a

7
specific series of events which indicate that the LBO mechanism is very complex

phenomenon. Comparison of experimental data showed that the research combustor

behaves like a PSR over an air loading parameter range of 0.1 – 100

Furthermore, the results illustrated that blowout data may be correlated with the air

loading parameter, as derived from PSR theory due to the fact that the loading parameter

accounts for the effects of residence time and flow velocity as well as pressure.

2.1.2. Semi-empirical Model for Combustion at Sub-atmospheric Pressures

Extensive testing is performed to ensure that the flame is statically stable over a

large range of operating conditions, ensuring a safe flight. When engaging in combustion

stability testing, it is important to evaluate the system at the extremes of the operating

range. One limitation observed in the previous studies is that combustion at atmospheric

pressure does not yield a significant variation in blowout equivalence ratio (when

correlated with loading parameter) to construct a full stability loop. When combustion

stability is plotted against the air loading parameter, as shown in Figure 2.2, the variation

in the range of conditions is dependent upon the mass flow rate and operating pressure.

Atmospheric testing in a WSR operates at a fixed pressure and velocity, thus limiting

mass flow rate and therefore the loading parameter. In a combustion system, the worst

fuel stability is fully exposed at the lowest combustion pressures [1]. In aviation, low

operating pressure occurs at high altitude due to the decrease in atmospheric pressure as

altitude is increased. [23] To understand low pressure combustion, the most highly

stressed condition, it is necessary to recreate these conditions in a laboratory

environment. To mitigate the high cost and complexity of full scale, true low pressure

8
combustion test, and several methods have been proposed to simulate low reaction

pressure.

A study by Lefebvre and Halls [1] used a water injection technique to simulate low

combustion pressure and increase the loading parameter of the combustor. In a practical

gas turbine, as altitude increases, the pressure decreases, and, correspondingly, the heat

release rate of the flame. The addition of water vapor into the system lowers the

combustion temperature and thereby simulates the lower heat release of low pressure

flames. In their study, fuel and air flow rates were held constant, and water was gradually

mixed with the fuel until extinction occurred. The process was repeated over a range of

flow rates, allowing a full stability loop to be constructed. Lefevbre and Halls concluded

that the water injection technique allows the point of peak stability to be clearly

defined.[1]

Additional work in the simulation of low pressure, high loading parameter flames

was conducted by Sturgess et al. [10] The study proposed that a simulation of low pressure

via nitrogen dilution was achieved when the reaction rates were matched. The steps for

the simulation proposed by Sturgess et al. [10] is shown below. The simulation began with

a bimolecular global reaction for a propane-air flame with excess nitrogen, given by,

(2)

Where B is the combustion efficiency, varying from 0 to 1 and is the excess

nitrogen defined by,

(3)

9
 K is the ratio of the mass flow of the excess nitrogen to the fuel flow rate at a

given condition. Atmospheric nitrogen will always be present in the combustion gas

when using air, however, when K is greater than 0, additional nitrogen has been added to

the flow as a diluent. When K = 0, the reactants are solely propane and air. An Arrhenius

reaction rate expression for a bimolecular single step reaction, proposed by Longwell et

al. [41] is assumed for lean mixtures of the form,

(4)

Equation 4 can be rearranged in terms of air loading parameter and takes the

form,

(5)

Equation 5 can be used to represent the reaction rate for a propane-air flame with

excess nitrogen, if K is greater than 0, or a pure propane-air flame, if K = 0. A reaction

rate simulation is achieved when the simulated loading parameter is equal to the desired

loading parameter. That is,

(6)

In Equation 6, R is the reaction rate expression from the right hand side of

Equation 4 for the desired case. In the simulation, the pressure is 1 atm and excess

nitrogen is present (i.e., K > 0) and in the desired case, the pressure is subatmospheric

and there is no excess nitrogen (i.e., K = 0).

Sturgess et al. [10] used a propane fueled diffusion flame combustor operating at

constant air and fuel flow rate to test the effect of excess nitrogen on lean blowout

10
characteristics. Nitrogen was added into the air incrementally until lean blowout was

reached. Sturgess et al. [10] showed that the nitrogen dilution technique allowed the air

loading parameter to be increased over three orders of magnitude. Additionally, the data

were shown to correlate well with PSR numerical calculations.

2.1.3. Evaluation of Alternative Fuels

Currently, the rising cost of conventional oil derived fuels as well as concerns

over the depleting oil reserve have led to rapidly expanding research into alternative

energy sources. Of main concern to the aerospace industry is the development of gas

turbine fuels from renewable sources. Commercial aviation is the largest consumer of

aviation fuel, with the US accounting for 44% of the global demand.[7] Recent efforts

have targeted development of a jet fuel derived from a renewable source. Potential JP-8

substitutes can be derived from feed stock of coal, natural gas or various biomass sources

such as algae, wood pulp or commercial fat, oil and grease waste.[7] A paper by Blakey et

al. [35] in 2011 detailed the process used to create the currently available biofuels.

Alternative transportation fuels can be created through the Fisher-Tropsch process or

hydro-treating of a bio mass feed stock. In the Fischer-Tropsch process, coal is gasified,

and turned into syngas, a mixture of hydrogen and carbon monoxide. The syngas passes

through a series of catalysts and hydro-cracking cycles, producing a heavy liquid

hydrocarbon fuel with properties similar to JP-8. Hydro-treated renewable jet fuel (HRJ)

is derived from oil biomass through hydro-treating, a process of adding hydrogen to

unsaturated hydrocarbon chains. Blakey et al. [35] showed that in a life cycle analysis,

biomass oils such as camelina, palm and soy produce less CO2 emissions per MJ of

energy than current petroleum based jet fuels.

11
 2.2. PREVIOUS WSR CONFIGURATIONS

The WSR has been used for many years to study combustion reactions. A highly

turbulent premixed air fuel mixture is injected into the reactor volume, continuously

mixing the combustion products with incoming reactants. The high degree of mixing

results in near uniform temperature and species profiles across the reactor volume,

yielding the closest approximation of a PSR achievable in a laboratory. Moreover, the

reactor walls are constructed from ceramic to allow the reactor to operate in a near

adiabatic state. The combination of the high rate of mixing and adiabatic nature of

ceramic test sections yields a reactor that is ideal for testing the purely kinetic aspect of

fuels, de-coupling fluid dynamic phenomena. The WSR removes the fluid mechanic

effects of fuel mixing and vaporization associated with combustion in an engine,

allowing evaluation solely of the fuel oxidation chemistry in question. Sections 2.2.1-

2.2.3 discusses some of the WSR designs that were used in the past, as well as an

overview of the combustion conditions used.

2.2.1. Spherical

In 1955, Longwell and Weiss [41] designed the first WSR to study high

temperature kinetics of hydrocarbon fuels. The reactor used by Longwell and Weiss was

made of two spherical halves of insulating firebrick. The upper temperature limit of the

firebrick was 1900 K. The reactor was enclosed in a nickel test section that allowed

instrumentation that included optical access, emissions probes and thermocouples. The

combustion products exhausted through holes drilled in the reactor. The work from

12
Longwell and Weiss using the spherical WSR suggested that lean blowout data may be

correlated to the loading parameter, derived from the kinetic theory. Longwell and Weiss

assumed an overall reaction order (n) to be equal to 1.8 for iso-octane. The data

correlation suggested that the reaction order was a good estimate.

One of the main disadvantages of the spherical reactor design was the material

properties of the firebrick. The high thermal conductivity resulted in high heat loss

through the walls, and the material quickly degraded when exposed to the high

temperatures and pressures of the test. [41]

2.2.2. Toroidal WSR

In 1984, Nenniger et al. designed the toroidal WSR. [27] The reactor had the same

internal volume (250 mL) as the spherical reactor used by Longwell and Weiss. [41] The

toroidal reactor was designed to achieve a high rate of mixing. 42 inlet jets were angled

20° off the radius to promote jet penetration and encourage bulk swirl tangential to the

toroid. Nenniger et al. used a tracer dye and laser-induced fluorescence (LIF) to visualize

the flow characteristics in the reactor. [27]

2.3. AFRL/UDRI ATMOSPHERIC WSR

A toroidal WSR, originally designed by Nenniger et al [27], as described above,

was used and improved by Zelina [42] then Stouffer et al. [33,36] The reactor is composed of

two half-toroid sections forming a reactor volume of 250 mL. Several ceramic

compounds have been used in previous tests, including alumina, fused silica, and zirconia

ceramic. The reactor halves sandwich an Inconel jet ring. The jet ring consists of 48 jets

drilled at 20° off the radius. The offset promotes swirl in the reactor and increases the

13
turbulence of the fluid. High turbulence increases molecular level mixing of the fuel and

air, establishing the well stirred condition. The jet ring is fed to tubes through which the

fuel/air mixture is introduced to the reactor. Additionally, the jets are sized such that the

passages are aerodynamically choked at the flow rates encountered during testing.

Choking the jets accomplishes two things. First, the high velocity of the jets enhances the

mixing in the reactor, and second chocking isolates the air supply from downstream

pressure fluctuation, allowing steady flow. Assuming one-dimensional compressible flow

( , the reactor pressure will be less than 52.8% of the upstream pressure if the jet

ring is choked.

The fuel-air mixture is reacted in the toroid and exits through a single exhaust port

in the center of the reactor volume. Due to the swirl induced by the jets, a flow

straightener is placed in the exhaust port to avoid entrainment of outside air into the

reactor caused by the recirculation established by the swirling flow.

The lower half of the reactor is instrumented through four holes. The typical

instrumentation used is a type-B thermocouple for measuring the reaction temperature, a

pressure tap to measure reactor pressure, an emission probe and an igniter. In several

studies, a stepper motor was used to sweep the thermocouple through the reactor,

generating a temperature profile along the minor radius of the toroid. The igniter uses

spark ignition and is fed from an aircraft exciter box. The igniter is recessed from the

main flow to preserve the material. Other designs use a pneumatic actuator to retract the

igniter form the flow after ignition is established. The pressure tap is made from ceramic

and sits flush with the top of the reactor. The location of the typical instrumentation is

14
called out. The WSR setup illustrated in Figure 2.3 was used in the studies by Zelina, [42]

Stouffer et al. [33] and Bohan. [2]

Figure 2.3: Lower half of the toroid with the jet ring in the WSR housing [2,28]

Additional thermocouples are installed on the jet ring to monitor metal

temperature. High jet ring temperatures can be caused by inadequate cooling by the

incoming mixture, high reactor temperature or cracked ceramic, allowing combustion

gases to escape the test section. In several studies, supplemental jet ring cooling was

achieved through a cooling ring that allowed nitrogen flow to cool the exterior jet ring.

The nitrogen provided cooling and the inert environment surrounding the reactor,

providing additional safety.[2,28]

2.4. PREVIOUS WSR RESEARCH

The WSR has been used extensively over the years to evaluate fundamental

combustion performance. Specifically, the WSR has been used to evaluate the gaseous

15
emissions, soot formation and combustion stability. The following section details some of

the previous studies conducted with a WSR.

2.4.1. Emissions and Soot Formation

As regulation and environmental concerns over gaseous emissions continue to

grow, the importance of understanding and optimizing the combustion process grows as

well. The well stirred reactor is used extensively in the evaluation of emissions

characteristics and soot formation of transportation fuels.

A 1994 study by Zelina et al. [13] used a WSR to evaluate the emissions of CO,

CO2, NOx and unburned hydrocarbons (UHC) for ethane, methane and a hydrocarbon

blend. Zelina et al.[13] showed that the CO emissions varied little among the three fuels

tested, and reached minima around . The CO concentration increased

dramatically as the equivalence ratio was increased. Similarly, the UHC emissions

showed little variation between the three fuels, reaching minima at . The UHC

emissions increased in both leaner and richer mixtures due to decreasing reaction

temperatures. NOx is strongly temperature dependent, and Zelina et al. [13] showed a peak

NOx concentration at an equivalence ratio near unity, where the combustion temperatures

are highest. NOx emissions were highest (in descending order) for the HC blend, ethane

and methane. The NOx concentrations correspond with the flame temperature as well,

with the HC blend being the hottest, and methane the coldest. The higher NOx was

formed when buring the HC blend due to the higher reaction temperatures increasing the

thermal NOx production. Zelina and Ballal [14] also calculated the combustion efficiency

for these fuels, using the WSR. Emissions of CO, NOx and UHC display the inefficiency

of the combustion process, and are, therefore, useful for calculating the combustion

16
efficiency. Zelina and Ballal [14] showed that a ceramic WSR had a combustion efficiency

of greater than 99% for all fuel-lean equivalence ratios up to stoichiometric. For rich

mixtures, the combustion efficiency falls dramatically 70% at . In rich mixtures,

the production of CO and UHC are increased due to lower equilibrium flame

temperatures, lack of oxygen and surplus of fuel. Zelina and Ballal showed little variation

in the efficiency for the three fuels tested. [14]

In 1997, Blust and Ballal [11] investigated the emissions characteristics of liquid

and gaseous fuels using a WSR. While gaseous fuels can provide a basis for modeling

and fundamental combustion research, liquid fuels are used almost exclusively in

transportation, and therefore their emissions characteristics are important. Methane,

Ethane, Cyclohexane, n-Heptane, Toluene, Ethylbenzene, n-Dodecane and an aromatic

blend were used in the experiment, and the emission index of CO and UHC was

measured. The emission index (EI) is the mass of the species of interest divided by the

mass of fuel taken to produce the sample. Blust and Ballal [11] showed that alkanes (with

the exception of methane) produced more CO as carbon number (CN) increased.

Additionally, it was shown that increasing molecular weight between alkanes (such as n-

heptane and n-dodecane) did not have an effect on CO production. The aromatics

produced less CO as the CN was increased. Finally, it was shown that methane produced

the most CO of all fuels tested, mainly due to its low molecular weight in the EI

calculation. In addition, methane produced the most UHC, followed by toluene. The

UHC production of alkanes rapidly decreased with increasing CN. Aromatics produced

less UHC than alkanes, with the exception of ethylbenzene, which behaved similar to the

alkanes for the same CN. [11]

17
 Additional studies have been performed using a WSR to characterize soot

formation. While CO and UHC directly correlate to combustion efficiency and are

undesired pollutants, soot emission is becoming increasing regulated as well. Soot, in

general, is the emission of particulate carbon as a result of locally rich combustion. A

study in 2005 by Stouffer et al. [33] used a WSR to investigate methods for soot reduction.

Six different additives were used, viz., nitromethane, nitroethane, nitropropane,

cyclohexane, pyridine and quinoline. Collected carbon mass was compared to fuel

without additives. Stouffer et al. explored the effect of chemistry and temperature on fuel

additives. The inlet temperature of the baseline fuel was adjusted to account for the

additional heat release of the nitroalkanes. With concentrations of 2 and 4 %, the

nitroalkanes showed a reduction in carbon mass across all equivalence ratios tested, with

4% showing the largest reduction. Nitroalkanes were the most effective at soot reduction

of all the additives tested; however, they were accompanied by an increase in NOx

emission. Cyclohexane decreased soot production over the entire range of conditions.

Stouffer et al. concluded that the soot reduction on the lower end of the equivalence ratio

range was chemical, while reduction at higher equivalence ratios was due to increased

flame temperature. The least effective additive used by Stouffer et al. in the study was

quinolone, which increased the carbon mass over all of the test conditions. [33]

2.4.2. Lean Blowout

Many studies have been conducted using various laboratory combustors that

explore the stability characteristics of fuel at atmospheric pressures. Ballal et al. [5]

conducted a combustion stability experiment using a research reactor that simulated the

main features of an annular gas turbine combustor, and used propane as a fuel. Lean

18
blowout was measured as a function of air loading parameter. Experimental data were

compared to prediction from Perfectly Stirred Reactor (PSR) theory. Ballal et al. [5]

showed, through optical access to the combustor, that blowout is preceded by a

systematic and detailed series of events which indicate that the LBO mechanism is very

complex. Comparison of experimental data showed that the research combustor behaved

like a PSR over an air loading parameter range of 0.1 – 100 . [5]

A study by Stouffer et al. [36] used a well stirred reactor (WSR) to evaluate the

lean blowout characteristics of standard JP8 and a synthetic fuel derived from the Fisher-

Tropsch process. Data collected for both JP8 and Fisher-Tropsch fuels were correlated

against air loading parameter and showed a 2.8% difference in the blowout equivalence

ratio between the two fuels. A full combustion stability loop could not be created, and the

blowout equivalence ratio varied on 1% as the loading parameter was doubled. The

testing conditions yielded a loading parameter range of 0.6-1.4 [36]

In 2012, Karalus et al. [25] conducted a lean blowout study with premixed

hydrogen-air flame. The study aimed to characterize the mechanism of lean blowout in a

premixed flame, and compared the experimental results to CFD predictions. Karalus et al.

showed that lean blowout did not occur when the flame was spatially homogeneous. The

flame was stabilized by a recirculation zone of combustion products near the base of the

inlet jet. Additionally, Karalus et al. [25] achieved lean blowout by maintaining air flow

rate constant while lowering fuel flow. The study showed that this method to achieve lean

blowout maintained near constant turbulence intensity parameter as lean blowout was

approached. Each reduction in the fuel flow was followed by a 15 minute wait time to

allow the reactor to achieve thermal equilibrium. Karalus et al.[25] stated that if lean

19
blowout is approached too rapidly, the hotter reactor walls will temporarily support

combustion at a lower equivalence ratio due to residual heat. Therefore, thermal

equilibrium is crucial for recording an accurate measurement of the lean blowout

equivalence ratio of a fuel. [25]

2.5. TOTAL TEMPERATURE PROBE

Accurate measurement of the flame temperature is critical to evaluating chemical

kinetics. [39] The majority of the flame temperature measurement is done with

thermocouples. Thermocouples use two lengths of dissimilar metal that generate a

voltage when presented with a temperature gradient. Thermocouples offer ease of

calibration and installation, while providing reasonable accuracy and response time. The

main limitation, however, is the maximum temperature that a thermocouple can withhold.

For near stoichiometric combustion, the flame temperature can exceed 4000 °F. For lean

blowout studies where the equivalence ratio is maintained at or below approximately 0.6,

the flame temperature can reach 2600-2800 °F. [39] A type K thermocouple uses nickel-

chromium and nickel-aluminum leads and has a maximum temperature of 2285 °F. [2]

The low maximum service temperature makes a type K thermocouple unsuitable for most

combustion environments. Type K thermocouples are used extensively in monitoring

metal temperatures and hardware health. Type B thermocouples are routinely used for

measurement of flame temperatures. Type B thermocouples use platinum (30%)-rhodium

and platinum (6%)-rhodium leads. The maximum operating temperature of a type B

thermocouple is 3092 °F. [42] The high temperature capability of the type B thermocouple

allows it be used in studies evaluating lean blowout or soot formation in fuel-rich

20
conditions. However, a type B thermocouple cannot be used for all conditions,

specifically near stoichiometric conditions, ranging from approximately 0.8 to 1.3. An

additional limitation to thermocouple use is fragility. To decrease response time and

reduce unnecessary flow perturbation, thermocouple wires are selected as thin as

possible. In high velocity conditions, the thermocouple can easily break or otherwise no

longer provide accurate readings. The issue of fragility and longevity is especially

evident when measurements are to be taken in full scale engines.

To address the limitations of thermocouples in a combustion environment, a new

probe was designed. The goals of the development were:

 To increase the maximum operating temperature of the probe.

 To increase the strength of the probe for withstanding high flow rates in

full scale testing.

 To ease implementation and data reduction.

The following sections discuss the theoretical development of the measurement

technique as well as some non-intrusive techniques used in the combustion field. These

same non-intrusive techniques are used to evaluate the performance of the new

temperature probe.

2.5.1. Importance of This Study

The current study aims to develop and characterize a diagnostic tool that can be used

to measure temperature and emissions species in combustion environments that were

previously accessible only through laser diagnostics. Additionally, very little literature

exists detailing the quantitative effects of introducing a probe into a flow stream. In

21
combustion, thermocouples and emissions probes are used extensively. The current study

acquires laser measurements simultaneously with the probe to show the perturbations that

intrusive diagnostics impart to a flame, shedding light into the discrepancy between

conventional and laser diagnostic temperature measurements.

Once the ceramic temperature probe is fully characterized, it can be used to

measure pressure, temperature and emissions in combustors that do not have optical

access. Moreover, the hardware required to support the probe is common and the probe

assembly is relatively inexpensive compared to existing laser diagnostics. These aspects

will allow deployment in the field without the cost and complexity of laser diagnostics.

2.5.2. Theoretical Development

The ceramic temperature probe uses flow continuity and the isentropic gas

relations to calculate a flame temperature from gathered temperature and pressure data

across a series of orifices. Figure 2.4 illustrates the location and numbering of the stations

used in the theoretical development.

Sample to
Vacuum
Pump
Figure 0.5: Set-up and instrumentation points of the Total Temperature Probe

Figure 2.4: Set-up and instrumentation points of the ceramic total temperature
probe

22
Mass continuity states that the flow rate entering the system must be the same as that

exiting, as represented in Equation 7. [23]

(7)

Since the gas flowing through the system is combustion products well above the

condensation temperature, it is assumed that it behaves as an ideal gas. The density in

Equation 7 can be rewritten using the ideal gas law, [23] shown below in Equation 8.

(8)

In Equation 7, it is necessary to know the area at the fluid entry and exit planes, the fluid

velocity, pressure and temperature. The pressure can be measured at a pressure tap in the

probe, and atmospheric pressure is assumed for the flame.

The velocity of the fluid may be difficult to determine for flow through a tube, however,

if the system consists of two orifices, a sufficient pressure drop will provide aerodynamic

choking. Therefore, the velocity is known to be the local speed of sound, [23] and

Equation 8 can be rewritten as;

(9)

which simplifies to;

(10)

In Equation 9, station 1 and 2 represent the orifice in the flame and the downstream

orifice, respectively. A thermocouple inserted into the flow at the downstream orifice can

be used to measure the temperature at station 2, while the temperature at station 1 is the

flame temperature, and is the value of interest. The gas constant, R, is assumed to be the

same between the two orifices in the system, and therefore cancels out from the equation.

23
The composition of the products is assumed to be frozen once it enters the probe.

However, Cp, and, therefore, will vary with temperature. Since the area of the orifices

used is known, the final parameter to define is the fluid temperature. With the above

assumptions, Equation 9 can now be solved for the temperature at the orifice in the flame.

(11)

The flame total temperature is determined using the isentropic relations. Both the

pressures and temperatures in Equation 11 apply to a fluid at sonic velocity. The

isentropic relations[23] for pressure and temperature are shown below.

(12)

(13)

Finally, it is assumed that the static and total temperatures are equivalent at each

station in the probe, due to the low flow velocities. A large probe with small orifice will

result in sonic velocities through the orifice, but the fluid will quickly expand in the body

of the probe and slow such that the static and total temperatures can be assumed equal.

The total temperature is within 99% of the static temperature for Mach numbers less than

0.15. The Mach number in the probe can be readily calculated from isentropic relations,

knowing the ratio between the probe internal area and the orifice area, . [23]

When applied to a real world system, the geometric area is not representative of

the flow area. Often, the area is qualified with a discharge coefficient. During the

development of the temperature probe, the discharge coefficient of the orifice in the

probe could not be directly measured. Instead, the probe orifice and downstream orifice

24
were flowed together to generate an effective area. With the geometric areas known, the

ratio of the discharge coefficients could be backed out. This ratio of the discharge

coefficients in the system was included in the calculation of the flow temperature.

2.5.3. Temperature Probe Correction

The majority of the conventional temperature measurement techniques contain an

error in the reading from the impact of the probe on the flow. The metal thermocouple

bead at combustion temperature experiences increased radiation losses. Heat transfer

theory would also suggest that since the base of the thermocouple is cooler than the

flame, heat is removed from the measurement point through conduction. If an accurate

measurement is desired, it is necessary to correct for the heat transfer losses. Fortunately,

the main losses are through radiation and conduction. A study by West and Westwater[40]

treats a thermocouple as a fin and applies a correction for heat transfer losses. They

compared the results of a combined heat transfer correlation with temperature probe

corrections derived by adding the independently estimated changes due solely to

radiation and conduction. Their results showed a reasonable correlation. Thus, the current

study uses the combined radiation-conduction correction proposed by West and

Westwater[40] to adjust the probe measured temperatures.

In practice, the ceramic temperature probe is inserted perpendicular to the flow,

and connected to the sampling lines with a series of fitting and adapters. The cantilever

nature of the probe coupled with the constant cross-sectional area allows the probe to be

modeled as a fin. Additionally, the length of the probe, the low fluid flow rate through the

system and the number of stainless steel fittings used allow the assumption that the base

25
of the probe is at the wall temperature and will be relatively independent of flame

temperature. The development of a combined conduction and radiation correction, as

proposed by West and Westwater is shown below. [40]

The heat equation can be applied to a differential element along the length of the

probe.

(14)

The heat transfer terms shown in Equation 14 can be expanded with detailed heat

flows, shown in Equation 15, where FA and Fe are the radiation angle factor and radiation

emissivity factor respectively. [40]

(15)

Equation 15 cannot be directly integrated. The radiation term must instead be

written in an alternative form, as suggested by Fourier. [40]

(16)

Equation 16 is equivalent to expressing the radiation heat transfer coefficient as

the following:

(17)

Substituting Equation 17 into Equation 15 and rearranging yields:

(18)

26
Equation 18 can be directly integrated if the probe has a constant perimeter and cross-

sectional area, yielding a general solution. [40]

(19)

Where

Boundary conditions must be applied to solve for the constants in Equation 19. As

previously mentioned, it is assumed that the base of the probe is equal to the wall

temperature, and that there is no heat flux at the tip, .

The probe is used to measure the temperature at the choked orifice. Since the orifice is

significantly smaller than the diameter of the probe, the heat transfer across the orifice

negligible. Therefore, the assumption of no heat flux at the orifice is valid. Applying the

boundary conditions yields: [40]

(20)

(21)

where . The value of the constants can be substituted into Equation 19,

resulting in a general equation for the temperature at any point, x along the probe.

However, the point of interest is the orifice in the flame, which occurs at . Applying

yields the correction equation for the probe. [37]

(22)

27
In Equation 22, is the actual gas temperature, is the wall temperature and is the

measured temperature. [40]

2.5.4. Comparison to Laser Diagnostics

Laser diagnostic measurements allow a less intrusive method for the evaluation of

many parameters in a combustion environment. When any intrusive technique is used

(i.e., emission probe, thermocouple, etc.,) the flame is disturbed. The perturbation caused

in the flame by the measurement technique has the potential to skew the accuracy of the

measurement, especially if a large probe is used to sample from a small flame. Fluid

temperature and species concentrations are the main parameters that are measured with

current laser diagnostic techniques

TdLAS, or tunable-diode laser absorption spectroscopy, uses a tunable diode to

tune the laser over the desired characteristic absorption lines in the flame. When the laser

impacts a species of the same emission wavelength, a reduction in the signal intensity is

recorded. The intensity of the beam through the flame can be correlated to the species

concentration through the Beer-Lambert law [17]. The flame temperature is measured by

sweeping the absorption spectra over two different absorption lines for the same species.

The integrated absorbance of the selected species is a function solely of the flame

temperature. [2]

Another laser diagnostic tool is Coherent anti-Stokes Raman Spectroscopy

(CARS), which uses three laser beams, viz., a pump, a Stokes and a probe beam. The

optical signal generated by the interaction between the three beams and the sample is at

the anti-Stokes frequency, or the sum of the frequencies of the original three beams. If the

28
difference between the pump beam and the Stokes beam is equivalent to the Raman

resonance frequency, the optical signal is resonantly-enhanced.[27, 28, 29]

The main application of CARS in combustion is the measurement of temperature.

The strength of the optical signal generated in with CARS diagnostics is a function of the

difference between the number of ground-state species and the vibrationally-excited

species. The number of excited species in a given state is a function of the temperature-

dependent Boltzmann distribution. Therefore, the strength of the signal may be correlated

to a temperature, allowing remote measurement of combustion gases. [30, 31,32]

Temperature measurement with CARS is well covered in the literature. A study

by Hancock et al. [32] measured the temperature of various flames over a Hencken burner

with CARS. The Hencken burner is a small, laboratory flat flame burner that produces

consistent profiles suitable for research using a sensitive method such as CARS.

Hydrogen-air and ethylene-air flames were studied. Hancock et al. showed that the

temperature of the Hencken burner flame matched closely with the equilibrium flame

temperature, and that the flame temperature was uniform across the width of the flame

front. [32]

2.5.5. Related Work

There is minimal work in the literature that addresses the limitations of

thermocouple measurements as presented in the current study. The concept used,

however, has been utilized in a similar form for enthalpy probes for measurements in

plasma. Due to the extremely high temperatures in plasma, conventional temperature

measurement techniques are not applicable. A study by Destefani et al.[3] utilized an

29
enthalpy probe to measure the temperature and enthalpy of microwave induced argon

plasma. A water cooled stainless steel probe was used to draw a plasma sample through

the inner diameter. The water temperature was measured at the inlet and exit planes. The

temperature rise in the water is due to the plasma flow. In order to calculate the plasma

temperature, tare data was collected with the probe in the plasma stream and no flow

through the probe. The change in temperature between the tare and experiment conditions

was used to calculate the enthalpy of the plasma from a simple enthalpy balance. The

study achieved positive results; however, the accuracy of the enthalpy measurement was

highly dependent on the flow rate through the probe. [3]

Another study by Magre et al. [30] compared intrusive measurements to CARS

measurements of the same flow field. A stainless steel probe was used to measure the

temperature of a Mach 6 stream of hydrogen-air combustion products. The probe was

introduced into the flow and quickly retracted to prevent melting. The study showed the

intrusive measurement does have an impact on the flow field. However, CARS

measurements were not taken simultaneously with the probe, so the cause of the

measurement discrepancy could not be quantified.[30]

With the help of research partners, a set of experiments was performed to

characterize the presence of the probe on the flame.[17] Tunable Diode Laser Absorption

Spectroscopy (TdLAS) was used to measure the path averaged temperature over a

Hencken burner flame.[17] Measurements were taken with the probe at two different

heights above the flame front, both with and without probe sample being collected.

30
 CHAPTER 3

EXPERIMENTAL SETUP AND PROCEDURE

3.1. ATMOSPHERIC EXPERIMENTS WITH A METAL REACTOR

The first iteration of high pressure WSR operation was designed using a metal

reactor. The reactor retained the toroidal shape of the atmospheric set-up, [36] but it was

made from Inconel 625 with an alumina thermal barrier coating. While the thermal

properties of ceramics make them suitable for a combustion environment, the low

thermal conductivity and tensile strength make them prone to cracking. Atmospheric

operation demonstrated that the life of a ceramic reactor was very short, with most only

withstanding a single cycle of combustion followed by cooling to ambient before

cracking. For experiments at atmospheric pressures, cracks in the reactor did not present

detrimental effects. There was minimal force on the reactor to open the cracks to the

extent that the combustion gases would communicate with the interior of the test rig.

When experimenting at high pressure, however, the pressure difference could cause the

cracks to propagate through the entire thickness of the reactor and allow combustion

products to escape the toroid. The use of a metal reactor would prevent cracking and

provide a positive seal of the combustion cavity, while a ceramic thermal barrier coating

would allow protecting the metal from temperature in excess of its melting temperature.

31
 Due to the high thermal conductivity and relatively low melting temperature of

metals the reactor, however, required active cooling. The upstream half of the toroid is

water cooled. A clover leaf tube is set into grooves on the upstream instrumentation plate.

Water is fed to the tube, maintaining the cooling plate at a near constant temperature. The

upstream cooling plate is called out in Figure 3.1.

Water Cooled
Upstream Side

Air
Air Cooled
Cooled
Exhaust
Exhaust Side
Side

Figure 3.1: HPWSR with metal reactor assembly

The exhaust side of the reactor sits on a flange built into the exhaust housing. Air

passages are machined into the plate allowing air flow through the majority of the area.

Conduction from the hot reactor half to the air cooled flange provides cooling. The

exhaust section mates with an instrumentation flange to allow the reactor assembly to be

installed in a pressure vessel.

The upstream water cooled plate has four holes taped for instrumentation. The

holes accommodate an igniter, emissions probe, thermocouple and pressure tap.

Compression fittings seal each of the instrumentation holes. In addition to the reactor

32
instrumentation, the exterior metal temperature was measured by type-K thermocouples.

The steady state temperature of the reactor indicates if the cooling set-up is sufficient for

the combustion heat release.

Since the use of a metal reactor was a departure from the set-up established during

previous WSR research, initial experiments were conducted to serve as a proof of

concept. All of operating points during the initial experimental phase with the metal

reactor were at atmospheric pressure. The cooling was seen to be sufficient to maintain a

safe metal temperature at the lean conditions seen near blowout. As the reactor

equivalence ratio was increased beyond 0.5 the cooling struggled to keep up with the heat

release. Initial lean blowout conditions taken in the reactor showed an elevated blowout

equivalence ratio for the same flow rates, as compared to previous data [31] with a ceramic

reactor. The increase in blowout equivalence ratio is due to the heat loss from the reactor.

The cooling that is necessary for reactor health makes the reactor less adiabatic than

reactors constructed from ceramic.

3.2. HIGH PRESSURE WELL STIRRED REACTOR DESIGN

The HPWSR designed was modeled after the successful atmospheric WSR

design. The same 250 mL toroidal reactor section was used.[36] The toroidal section

promotes consistent mixing throughout the reactor volume. Previous experience showed

that while a metal reactor provides a consistent, positive seal, the combination of high

thermal conductivity and the need for cooling results in skewed blowout results. The heat

removed from the reactor to maintain structural integrity elevates the blowout

equivalence ratio. Therefore, a ceramic reactor was selected to achieve the most adiabatic

33
 blowout data possible. The first design challenge arising from the use of a ceramic

reactor is the issue of cracking. As previously discussed, the low thermal conductivity

and tensile strength of the ceramic make it prone to cracking. When operating at

atmospheric pressure, the reactor operates at less than 1 psig over the ambient pressure.

With a very slight overpressure, leakage from the reactor due to cracks is minimal. When

adapting the hardware to operate at high or low pressure, however, the potential leakage

from even a small crack is dramatically increased due to the higher pressure differential.

Providing a positive seal to the reactor is critical to successful operation. Two sealing

rings were designed to address the problem of reactor containment. The two sealing rings

are machined to accommodate gaskets as well as seat the reactor. Mica gaskets with a

stainless steel core were used for testing. The sealing ring configuration is shown in

Figure 3.2.

Sealing Rings

Figure 3.2: Reactor stack up with sealing rings

Figure 3.2 shows a cross-section of the HPWSR assembly. The toroidal reactor

section is shown in purple, and is contained by the upstream water cooled plate, yellow,

34
and the air cooled exhaust housing, brown. Clamping studs, blue, provide the force to

seal the gaskets. The sealing rings are shown in light yellow and sandwich the jet ring.

The sealing rings provide a positive seal outside of the reactor and do not depend upon

the reactor walls to seal the combustion volume.

In order to achieve high pressure operation, the reactor assembly from Figure 3.1

must be fit into an appropriate flange to interface with a pressure vessel. It would be

possible to run at elevated pressure without the need for a pressure vessel, but the

pressure differential across the reactor, and the possible leak paths are comparatively

greater. Enclosing the reactor in a pressure vessel allows the surrounding pressure to be

increased, mitigating the influence of leaks. Additionally, enclosing the reactor in a

pressure vessel allows the reaction pressure to be precisely controlled by a back pressure

valve, downstream of the assembly. An external pressure vessel allows a lighter

construction of the well stirred reactor housing, as the vessel environment can be

maintained at a slight overpressure to the reactor pressure, minimizing stresses.

All instrumentation lines as well as combustion air, fuel and cooling flows were

brought into the pressure vessel to supply the reactor. Previous combustion rigs used to

mount the hardware in a flange with threaded holes to allow pass through the required

plumbing while fittings provided positive seal. This “instrumentation flange” approach

was used for the HPWSR, and allows interface with existing facilities. The supply

plumbing was designed to fit within the 13.5-inch diameter vessel that would be used for

testing. Figure 3.3 shows the HPWSR assembly inside the pressure vessel.

35
 Pressure Vessel

Air
Manifold

Reactor
Assembly

Figure 3.3: HPWSR hardware with pressure vessel

In Figure 3.3, the instrumentation flange is shown in orange. All of the plumbing

must pass through the holes on the flange, and make a 90° turn upstream to reach the

reactor. Custom made fittings are used to seal tubing and instrumentation passing through

the flange. Air is fed through the instrumentation flange at two locations, 180° out from

each other. The air flows into an arc-section manifold upstream of the reactor. The

manifolds distribute the air to two feed tubes, which flow to the jet ring. Two manifolds

feed four inlet tubes. Previous WSR designs used either a single or double feed to the jet

ring. Less feed tubes tend to prioritize the inlet region for flow, and starve the region on

the opposite side. In order to promote even and consistent flow and mixing, four inlet

tubes were used. The inlet tubes were welded to the jet ring to eliminate compression

fittings, a possible source of leaks. A detailed view of the inlet tubes entering the jet ring

is shown in Figure 3.4.

36
 Figure 3.4: Jet ring inlet design

As shown in Figure 3.4, caps are welded on the ends of the tubes and a 0.062-inch

hole is drilled in the center. Additionally, two slots are cut in the side wall of the tubes.

The cap and slot design helps to promote bulk flow circumferentially around the jet ring,

while the hole in the cap prevents the jet directly under the inlet from being starved of

flow.

3.2.1. Autoignition Issues

Auto-ignition is a combustion reaction initiated in a fuel-oxidizer mixture with the

absence of an external ignition source. In order for a fuel to auto-ignite, it must be mixed

with an oxidizer within the flammability limits of the fuel. A typical hydro-carbon fuel

cannot auto-ignite in the absence of oxygen. Auto-ignition events are caused when a fuel-

air mixture is heated to a high temperature. For example, the auto-ignition temperature of

Jet-A is 410 °F while the auto-ignition temperature of hydrogen is 997 °F. [21] The time in

which the auto-ignition reaction requires to begin is called the ignition delay time, and is

37
a function of temperature, pressure, fuel used, and equivalence ratio. When operating in

premixed systems, such as the WSR, it is possible for auto-ignition to occur at any point

after the fuel and air are mixed given the proper conditions. Therefore, it is important to

understand the auto-ignition characteristics and limits at the operating conditions.

A study by Lefevbre et al[22] showed that reaction inlet temperature had the

strongest influence on ignition time delay in lean propane-air mixtures. Lefevbre also

showed that the effect of varying equivalence ratio on ignition delay was weaker than the

effect of temperature. At a given fuel-air ratio, less ignition energy is required to initiate

the reaction with higher inlet temperatures.

The autoignition characteristics of fuels are important in gas turbine and reactor

design as the ignition delay and autoignition temperature represent the upper limit of

residence time and fuel-air mixture temperature that is acceptable in the system [22]. The

effects of ignition delay are more pronounced in applications where the primary heat sink

for cooled components is the fuel. In systems that utilize premixed combustion, such as

the WSR, short ignition delay presents a problem for large reactor volumes or long

residence times in the feed lines. This is because the reactants can auto-ignite in the feed

lines before entering the combustor. The issue is magnified with the use of heavy

hydrocarbons, such as JP-8, that require higher temperature to maintain the vapor state.
[21]

38
 3.2.2. Fuel Vaporization

The atmospheric WSR used an external vaporizer is shown in Figure 3.5. The

large, heated volume of the vaporizer provides a long residence time for mixing to occur.

Fuel vaporization is achieved with a Delavan air assist nozzle, shown in Figure

Figure 3.5: Air assist atomization nozzle [2]

The air assist nozzle shown in Figure 3.5 uses an air stream and swirler to break

up the fuel flow into small droplets. Smaller droplets evaporate more rapidly, and,

therefore, it is advantageous to obtain the smallest droplet size possible. When evaluating

the evaporation of droplets in a hot flow, the evaporation constant [21] is used. The

evaporation constant, as proposed by Lefevbre, takes into account the fact that the heat

the droplet absorbs during the heat-up period reduces the available heat for the

vaporization of the droplet, lowering the evaporation rate over a steady state

approximation. In other words, the energy expended by the environment on a droplet

during phase change lowers the overall temperature of the environment, reducing the

energy available for evaporation of additional droplets. The effect of droplet cooling will

lead to longer evaporation times then when a steady state approximation is used.

Lefevbre defines the evaporation constant as . [21] Figure 6 shows the effect of

39
droplet size on evaporation time. Multiple evaporation constants are also represented to

illustrate the impact of evaporation constant on the evaporation process.

1000

Evaporation Constant = .005 mm2/s

900 Evaporation Constant = .01mm2/s
Evaporation Constant = .02 mm2/s

800

700
Evaporation Time (ms)

600

500

400

300

200

100

0
0 20 40 60 80 100 120
Droplet Size (um)

Figure 3.6: Impact of droplet size on vaporization [36]

In Figure 3.6, the evaporation time as a function of droplet size is shown for three

different evaporation constants. The clear trend is that as droplet size increases, the

evaporation size increases exponentially due to the increase in droplet volume.

The internal passages of the nozzle are sized such that there is a small pressure

drop from inlet to exit, allowing the pressure to be driven by downstream conditions. The

coupling of upstream and downstream conditions was mitigated by the use of a needle

valve immediately upstream of the nozzle to produce a large pressure drop in the system.

The air flow rate through the nozzle was maintained constant over the majority of testing

to ensure repeatability. In atmospheric testing, as much as one-quarter of the total air

flow rate was passed through the nozzle.

40
 The external vaporization technique, detailed above, has been used extensively

during atmospheric testing, and has been used for over several hundred hours of stable

operation. The main disadvantage of this vaporization scheme is the long residence time.

As the reactor pressure increases, the temperature required to maintain a fuel vapor also

increases. Increasing mixture temperature brings the mixture closer to autoignition. Long

vaporizer residence times further increase the risk of an autoignition event when

operating at high pressure. Therefore, the ability of the vaporizer to safely contain

pressure was investigated.

3.2.3. Vaporizer Safety

The vaporization of JP8 poses a greater challenge than other lighter liquid fuels

since more heat is required to completely vaporize JP8, coupled with a lower autoignition

temperature. The current vaporizer design has been successfully and safely used to

vaporize a variety of liquid hydrocarbon fuels, including JP8, during atmospheric testing.

The WSR system is designed such that the pressure drop across the nozzles in the jet ring

is the largest in the system. Having the largest pressure drop located at the nozzles means

that as the reactor pressure is increased, the vaporizer pressure must be increased by a

larger factor. As the vaporizer pressure is increased, higher fuel temperatures are required

to achieve complete vaporization. The temperature required to maintain a vapor is called

the dewpoint. The dependence of pressure on the dewpoint of JP-8 was presented by

Stouffer et al.[36] and is shown in Figure 3.7. A JP-8 surrogate was developed by Stouffer

et al.[36] to allow the fuel dewpoint to be calculated by the program SUPERTRAPP.[36]

SUPERTRAPP is a program released by NIST to calculate the thermophysical properties

41
of hydrocarbon mixtures. It is clear that as the pressure increases, more energy is required

to vaporize the fuel. Additionally, as equivalence ratio is increased, more fuel in present

in the mixture, driving the dewpoint even higher. The dewpoint dependence on

equivalence ratio becomes significant when investigating soot formation and rich-

emissions with a WSR.

Effect of Pressure on JP8 Dewpoint

500

450
Dewpoint Temperature (°F)

400

350

300

250 P=1 atm

P=5 atm
P = 17 atm
200
P=25 atm
P=50 atm
150
0 0.5 1 1.5 2 2.5 3
Phi

Figure 3.7: Dewpoint pressure dependence for JP-8 [36]

As mixture temperature increases, a point exists where the pressure in the

vaporizer will be so great that the temperature required to vaporize the fuel exceeds the

autoignition temperature. The autoignition temperature of JP-8 is 460 °F. In order to use

the existing vaporizer for high pressure testing, the test matrix will be limited such that

the mixture temperature does not approach that of autoignition.[36]

42
 After analyzing the limitations of the vaporizer hardware available for the

experiments at high pressure, it was determined to limit the reactor pressure to 3 atm.

Based on this desired maximum operating pressure, the flow rates (calculated from

desired residence time and reactor volume), and the jet ring nozzle geometry, the required

upstream pressure could be calculated. It was found that the highest vaporizer pressure

that would be experienced during testing at 44 psia reactor operating pressure was 165

psia, corresponding to τ = 3 ms and ϕ = 0.3. Knowing the dewpoint for all of the test

conditions and the maximum pressure that would be experienced in the vaporizer, fuel

and air temperatures could be determined such that the mixture was a vapor at all

equivalence ratios desired.

Figure 3.8 shows the variation of the mixture temperature with equivalence ratio, for an

air temperature of 450 °F and a fuel temperature of 350 °F.

Figure 3.8: Mixture temperature vs. dewpoint at maximum vaporizer pressure

43
 Maintaining the mixture temperature below the autoignition temperature of JP-8

for all operating conditions is desirable for safe operation. The next step in the analysis

was to ensure that the construction of the vapoirzer would stand up to a un-planned

detonation at the elevated pressures desired.

The pressure rise of a detonation in a JP8-air mixture was calculated assuming

that the pressure wave propagates at the Chapman-Jouguet speed, and thus passes

through the Chapman-Jouguet point on a Rankine–Hugoniot plot. Figure 3.9 shows an

example of a Rankine-Hugoniot plot with the important plot features noted. [20]

Figure 3.9: Rankine-Hugoniot plot with annotated regions [20]

The calculation of the detonation pressure rise requires that the ratio of specific

heats (γ), the reactant pressure and temperature, and the heat release. The ratio of specific

heats for JP8- air mixtures was given through SUPERTRAPP, while the reaction

temperature and pressure was defined through by the desired test conditions. The heat

44
release was determined by stoichiometry knowing the approximate composition of the

JP8 surrogate used in the analysis and assuming complete combustion. Substituting in the

values relevant to the desired test conditions yielded a maximum detonation pressure rise

of 10.14. A detonation pressure rise, P2/P1, of ~10 for stoichiometric JP8-air mixtures is

confirmed in the literature. [20] Based on the maximum pressure rise that could be

experienced during testing and the maximum vaporizer pressure (i.e.,165 psia) the peak

pressure to be taken by the hardware in the event of a detonation is 1650 psia. After the

peak pressure during a detonation was determined, the stresses on the vaporizer could be

calculated. The vaporizer structural analysis presented in Appendix B shows that the

structure of the existing vaporizer is sufficient to handle a detonation event up to an

initial reactor pressure of 3 atm.

3.3. INTERNAL VAPORIZATION

When using an external vaporizer, the residence time of the mixture in the

plumbing is long, increasing both the heat loss from the fluid and the time the mixture

has to react. The high temperature needed to achieve the flash vaporization, required by

the internal vaporization technique, increases the autoignition risk. The system must be

capable of vaporizing and thoroughly mixing the fuel and air without burning upstream

of the jet ring. In order for the HPWSR to achieve the elevated pressures, a different

vaporization scheme had to be developed. The following section discusses the

development of an internal vaporization technique that reduces the mixture residence

time and allows high reactor pressure. Additionally, the implementation of this system for

low pressure operation is discussed.

45
 3.3.1. Experiments with a WSR at Elevated Pressure

Up to a reactor pressure of 3 atm, vaporization was handled using the external

vaporizer. The safety analysis in Section 3.2.3 showed that the vaporizer was structurally

capable of limited high pressure operation. Above 3 atm, however, structural limitation

mandated a different method. The external vaporizer was removed and the fuel lines were

routed directly into the pressure vessel to feed injectors mounted in the WSR air feed

lines. In order to achieve vaporization inside the pressure vessel near the jet ring, the

properties of a super critical fluid were utilized. A supercritical fluid is one where the

temperature and pressure are maintained above the critical point. In this state, there is no

distinct liquid or vapor phase. Additionally, relatively small changes in pressure result in

large changes in density. The internal vaporization technique uses orifice tubes that act as

throttling valves. As the high pressure fuel flows through the orifice into a significantly

larger area, the pressure drop causes flash vaporization. The orientation of the orifice

tubes is shown in Figure 3.10.

46
 Secondary Fuel
Leg

Orifice
Tubes

Fuel Supply
Air Feeds

Figure 3.10: Fuel injection points for internal vaporization

Figure 3.10 shows four of the orifice tubes used for high pressure testing. The

tubes (shown in red) are fed supercritical liquid JP-8, and spray directly into the 4 air-

feed lines (shown in green) leading to the jet ring. The air lines act as the mixing volume.

Utilizing flash vaporization allows the nozzles to be placed closer to the reactor, reducing

the residence time required for mixing. Smaller residence time allows for mixture

temperature potentially to reach levels that would auto-ignite in the higher volume,

external vaporization set-up.

The orifice tubes are closed end stainless steel tubes with a laser drilled orifice in

the end. The main disadvantage of using a fixed geometry nozzle is the limitation on

operating flow rate. Nozzle pressure drop is a function of flow rate, and a higher flow rate

47
will require a large pressure drop across the nozzle. The supercritical fuel must be

maintained at pressure, while the pumps are only capable of operating up to 3000 psi.

These limits define the orifice size that must be used. In order to increase the operating

range of internal vaporization two fuel flow circuits were installed utilizing 50 and 100

micron orifice tubes. During low flow situations, the 50 micron circuit is used. As the

flow rate in increased, the 50 micron circuit is turned off, and the 100 micron turned on.

If it is desired to run the flow rate higher both can be operated for a total of 150 micron

orifice size.

The final aspect of the internal vaporization design was a fuel line heat trace. The

fuel is heated to 650 °F, which is above the critical temperature. At these high

temperatures, the heat loss rate from the fuel between the heater and nozzles can be

significant. If the fuel falls below the critical temperature, it will pass through a two

phase region rather than moving directly to the vapor phase, which may result in

unsteady combustion. A 0.375-inch tube was fit around the fuel lines, with hot nitrogen

flowing through. The nitrogen was heated to 800 °F to allow some heat loss from the

flow without allowing the inner co-flow fluid to drop below the fuel temperature.

3.4. FINAL ASSEMBLY

After the issues of reactor sealing, instrumentation and vaporization were

addressed, the final reactor assembly was installed in the pressure vessel. Figure 1 shows

the HPWSR installed in the instrumentation flange with all of the necessary plumbing.

The air manifolds, reactor stack-up and fuel lines can be seen. Additionally,

48
thermocouple and pressure instrumentation is passed through fittings in the flange

allowing a pressure seal.

Figure 3.11: Final HPWSR hardware installation

49
 In Figure 3.11, flows proceed to the right, and the exhaust section can be seen in the

back on the right side of the figure. The facility controls reactor pressure with a back-

pressure valve. The back-pressure valve used for testing has a maximum temperature

rating. Air is used to quench the exhaust stream and prevent damage to the back-pressure

valve. The 1-inch lines that carry the exhaust quench flow can be seen at the midpoint of

the flange, on the right side in Figure 3.11. For the low air flow rates at which the WSR

operates, air quenching is sufficient to drop the exhaust temperature. For higher flow rate

applications, a water quench spray is used.

3.5. TEST PLAN AND PROCEDURE

Lean blowout experiments were accomplished in three phases, shown below:

1. Atmospheric experiments to provide baseline against previously acquired data

2. Elevated pressure experiments to demonstrate both the external and internal

vaporization technique

3. Experiments at sub-atmospheric pressure to provide a comparison for nitrogen

dilution lean blowout data and to demonstrate the extended capabilities of the

WSR

The following sections details the experimental procedure and test conditions used in

each of the test phases.

3.5.1. Experiments at Atmospheric Pressure

Prior to testing, all the air heaters and heated lines feeding the reactor are turned

on and airflow established. The pre-heating process serves two purposes. First, the hot

50
air heats all the metal plumbing that the fuel/air mixture will contact and heats the

ceramic in the reactor. When using a liquid fuel such as JP-8, all of the plumbing must be

maintained above the dewpoint of the mixture. If there are any cold spots in the system,

fuel vapor may condense and form liquid fuel droplets in the feed lines. Preheating the

ceramic reactor makes ignition easier and helps to reduce the thermal shock experienced

by the ceramic. During atmospheric testing, the liquid fuel is heated to 300 °F by a

Mokon oil heater. The oil is also used as a coolant for the emissions probe.

After the hardware preheating is complete, the air flow rate is decreased to 100-

150 SLPM. The igniter is activated and gaseous fuel incrementally introduced until light-

off is achieved. Successful ignition is verified with the type-B thermocouple in the

reactor volume. Ethylene is used as the fuel for light-off due to its wide flammability

limits. The reactivity of ethylene allows combustion to be established at a lower

equivalence ratio than JP-8 at the same conditions. During the light-off procedure, the

equivalence ratio is maintained below 0.6 to avoid unnecessary heat stress on the

ceramic. As the reactor temperature increases, air is added to maintain the temperature

below the melting point of the ceramic. The next step is to switch to liquid fuel. The

liquid fuel flow rate is slowly increased while simultaneously decreasing the gaseous fuel

flow rate. Changing both flow rates will maintain an approximately constant reactor

temperature. After the reactor is running entirely on liquid fuel, it is allowed to reach a

steady state temperature. Temperature traces in the data acquisition software are used to

indicate when the reactor temperature has stopped changing and reached equilibrium with

the fluid temperature in the reactor. The fuel and air flow rates are adjusted to the first

data point in the test series.

51
 The air flow rate is metered using a thermal mass flow controller manufactured by

Brooks. Blowout is achieved by reducing the fuel flow rate while maintaining a constant

air flow rate. The timing of the fuel flow rate reductions is very important in the accuracy

of lean blowout data. If the fuel flow rate is decreased too fast the reactor will not be able

to equilibrate, and the walls will be at a higher temperature for a given equivalence ratio.

Hotter walls will tend to stabilize the flame, resulting in a blowout equivalence ratio that

is lower than the true adiabatic blowout. The goal of the blowout study is to begin to

determine the lean stability margins for JP-8 in a WSR. Several different air flow rates

were tested. The air flow rate changes the residence time in the reactor, and an increase in

the former results in an increased air loading parameter. Table 3.1 shows the test matrix

used for the atmospheric portion of this study. Each of the blowout conditions was

repeated 3 times to ensure that each result is statistically significant.

Table 3.1 Atmospheric Blowout Test Matrix

Air Flow
Fuel LP Starting
Rate (SLPM)
275 1.44
300 1.41
JP-8 0.6
500 2.23
650 2.84

3.5.2. Alternative Fuel Testing

The properties of new fuels cannot be guaranteed to be the same as those of JP-8.

Atmospheric pressure testing was conducted to evaluate the impact of fuel Cetane

number on combustion stability. Derived Cetane number (DCN) is a representation of the

ignition delay of certain fuel. DCN is typically applied to diesel and heavy aviation field,

52
such a JP-8 and its alternative variants. High DCN fuels have shortened ignition delay

times, and therefore are easier to ignite. A lower DCN results in a fuel that requires more

energy to ignite. In this manner, DCN is the opposite of the octane rating used for

automotive gasoline.

While the majority of properties of JP-8 are specified, there is no specification for

the DCN. The same is true for the alternative fuels that are gaining attention. If a specific

alternative fuel is shown to meet the density, flashpoint and lubricity requirements

(among others) as specified in MIL-DTL-83133, [7] it cannot be guaranteed to possess a

similar DCN. The possibility of a widely varying DCN could mean that an alternative

fuel is more difficult to ignite in a gas turbine application, reducing effectiveness. To

address the impact of DCN on combustion, alternative fuels were selected that bracketed

the DCN of JP-8, one higher and one lower. The fuels selected for the study are shown in

Table 3.2. The goal of the study is to quantify what impact, if any, DCN has on the

combustion stability in a WSR.

Table 3.2: Selected Alternative Fuels

C11.4H22.3

C10.8H23.2

C11.9H25.6

53
 Table 3.2 shows the properties of the two alternative fuels, along with the

properties of JP-8 for comparison. The alternative fuels are iso-paraffinic kerosene (IPK)

manufactured by Sasol, and Hydro-Treated Renewable Jet fuel R-8 (HRJ). The average

fuel composition from GCxGC analysis is shown, along with the DCN.

Test conditions were chosen to achieve the largest loading parameter range

possible with the atmospheric reactor. The low flow rate was chosen based on initial

exploratory work that showed inconsistent blowout data with flow rates below 275

SLPM. The high end of the matrix was facility-limited to 675 SLPM. Three blowout

points were taken at each condition to ensure the results are statistically significant.

Table 3.3: Test Conditions for Alternative Fuels

Air Flow
Fuel DCN LP Starting
Rate (SLPM)
275 1.44
325 1.71
375 1.97
JP-8 47.7 0.6
500 2.62
600 3.13
675 3.51
275 1.44
325 1.71
375 1.97
IPK 31.5 0.6
500 2.62
600 3.13
675 3.51
275 1.44
325 1.71
375 1.97
HRJ R-8 59.1 0.6
500 2.62
600 3.13
675 3.51

54
 Lean blowout experiments were conducted using the metal reactor assembly. The

metal reactor assembly was installed in the Atmospheric Combustion Research Facility,

and, therefore, it uses all of the same equipment as the previous atmospheric WSR tests.

The procedure for warm-up and reactor ignition is also the same as in previous testing.

For each condition, the equivalence ratio was set at 0.6 and the reactor allowed to reach a

steady temperature. A constant measured temperature in the reactor indicated that the

ceramic walls were near the same temperature as the fluid, and heat loss was in

equilibrium with the heat release of the combustion. Fuel flow rate was then reduced,

with temperature equilibrium reached between each step.

3.5.3. Nitrogen Dilution

Nitrogen dilution data was collected during the atmospheric test campaign. As

previously discussed, the excess nitrogen in the air will lower the combustion

temperatures. The lower combustion temperature can be correlated to a decrease in

pressure through kinetic theory. From the definition of loading parameter, there is an

inverse exponential dependence on reaction pressure. Therefore, a decrease in reaction

pressure will yield large increases in air loading parameter, extending the stability curve

for the reactor. The literature [10] shows that nitrogen dilution can increase the air loading

parameter over 2 orders of magnitude without the need for pressure vessels or new

hardware. In order for the technique to be validated, nitrogen dilution data must be

compared to true low pressure data collected in the same reactor.

The HPWSR hardware, with fused silica reactor was used to collect nitrogen

dilution data. The hardware was installed in the Atmospheric Combustion Test Facility.

55
The facility hardware and instrumentation is the same as with previous atmospheric

testing. Nitrogen was added to the air flow upstream of the heater. Mixing occurred over

the length of plumbing leading to the reactor.

Two different nitrogen mixing schemes were used in testing. The first was

nitrogen addition. When testing with nitrogen addition, the air flow rate was set to obtain

the desired residence time, and then dilution nitrogen was added. This method increased

the overall flow rate of the mixture into the reactor. The second method used was

nitrogen replacement. When using the nitrogen replacement method, the air flow rate was

decreased as nitrogen flow rate was increased to maintain the same combined overall

mass flow rate of nitrogen and air. The mass flow rate calculation accounted for the

difference in molecular weight between air and nitrogen. Table 3.4 shows the test matrix

for nitrogen dilution testing.

56
 Table 3.4: Conditions for Nitrogen Dilution Testing

Nitrogen
Air Flow
Test Flowrate
Rate (SLPM)
(SLPM)
500 5
500 25
Nitrogen 500 35
Addition 500 50
500 75
500 100
475 25
450 50
400 100
350 150
Nitrogen
275 267
Replacement
275 225
262 238
258 242
254 246

The conditions shown in Table 3.4 represent both the nitrogen addition and

nitrogen replacement. Higher blowout equivalence ratios and, therefore, lower simulated

pressures are achieved with the lowest air-to-nitrogen ratio.

3.6. EXPERIMENTS WITH AN EXTERNAL VAPORIZER

Fuel was heated with a 3kW sand bath for all conditions. The sand bath is

fluidized by nitrogen. The fluidization process greatly increases the heat transfer between

the medium and the fuel, while still allowing high temperatures to be reached. The heat

transfer to the fuel was further maximized by using a 20-foot coil of 1/8 inch tubing. The

small diameter of the tubing resulted in high velocity, while the length provided a large

surface area over which heat transfer would occur. Prior to heating, the fuel was sparged

57
with nitrogen for a day. The bubbling of nitrogen helps to remove dissolved oxygen from

the fuel, helping to prevent oxidative coking that may clog the fuel lines at high

temperature. When testing with the external vaporizer, the fuel was heated to 350 °F.

The fuel was delivered with ISCO 1000D continuous flow syringe pumps. The

ISCO pump system is capable of delivering 2500 psi fuel pressure, and delivering a flow

rate of .01-204 ml/min. The pumps meter the flow, so no additional flow meters were

required. ISCO published a LabView interface, allowing the pumps to be controlled by

the main LabView program. This also permits the fuel flow rate and pressure data to be

logged in the same spreadsheet as the other pertinent test data.

The combustion air for experiments at elevated pressure was heated with a 15kW

circulation heater, while the combustion air for early atmospheric experiments was heated

with a 3 kW circulation heater. Flow was provided using the shop air at the facility and

had a maximum pressure of 100 psi. The air flow was metered using a CFM025 Coriolis

meter. The Coriolis meter has a large turndown ratio, thus allowing the same flow meter

to be used over a large range of test conditions. The general flow path used for testing

with the external vaporizer is shown in Figure 3.12.

58
 Figure 3.12: Flow path schematic for experiments with an external vaporizer

One of the main concerns when using a Coriolis flow meter is the back pressure.

Due to the construction of the meter, a large pressure drop can cause the internal passages

to choke, potentially damaging the meter. In isentropic air, the pressure drop required to

choke the fluid is 52.5%. For the safety of the hardware, the maximum allowable

pressure drop across the Coriolis meter was set to 25%. The back pressure and overall air

flow rate was regulated with two pressure operated, fail closed, Badger Research control

valves. Two different valve trims, with different valve flow coefficients (Cv), were used

to allow fine and coarse flow control. The coarse control valve had a Cv of 2, as was used

to set the flow rate and pressure in the general area desired for the specific test condition.

The coarse valve was manually set to the same setting until a large change in the air flow

rate was needed. The fine control valve had a Cv of 0.5. This valve was controlled with a

PID loop within a controller. After the desired flow and meter backpressure was dialed in

59
the fine control valve was set to automatic mode, and the flow rate was maintained for

the test.

3.6.1. Sub-atmospheric

After many hours of successful operation at atmospheric pressure with the

external vaporizer, the same setup was used for low pressure testing. The main

consideration in adapting the atmospheric vaporizer hardware to low pressure testing was

to manage heat loss. In order to maintain a constant reactor residence time, as the reactor

pressure decreases, the air flow rate must be decreased as well. The low flow allows more

time for the fluid to lose heat through the plumbing. A heat trace was applied to the fuel

tubing to minimize heat loss. The fuel line was passed through a 0.375-inch line, which

was attached to a 1200 W heater. Nitrogen was heated and flowed along the fuel lines.

Extra ceramic fabric insulation was added on the exterior of the plumbing throughout the

entire system to further combat heat loss from both the air and fuel. Additionally, tape

heaters were wrapped around the vaporizer body, the combustion air inlet and the

pressure vessel. These heaters were controlled to 300 °F and served as system preheat.

The final aspect of thermal management was heated nitrogen that was flowed into the

pressure vessel. The nitrogen was heated to 350 °F and served three purposes. The first

was to preheat the reactor assembly and internal supply lines, while the second purpose

of the nitrogen was to provide an inert environment inside the pressure vessel as an extra

safe guard against leaking fuel/air mixture. The final purpose of the nitrogen was that the

pressure difference between the reactor and the vessel could be minimized, theoretically

eliminating the majority of possible leak paths.

60
 Blowouts were achieved at three (3) different sub-atmospheric pressures. The

lowest pressure the facility was capable of achieving was 4 psia. In order to further

increase the loading parameter the air flow rate was increased at the lowest pressure. The

test conditions used are shown in Table 3.5.

Table 3.5: Test Conditions for Sub-Atmospheric Blowout with External

Vaporization
Air Flow Reactor
Fuel Rate LP Pressure Starting
(lbm/min) (psia)
1.11 8.82 10.02
0.71 14.03 6.07
JP-8 0.6
0.52 21.00 4.07
0.80 31.47 4.14

3.6.2. High Pressure Operation

The first set of high pressure lean blowout data was conducted with the external

vaporizer, allowing more direct comparison between atmospheric and low pressure

blowout data taken with similar hardware. As discussed in Section 3.3.1 the external

vaporizer could only safely support testing up to 3 atm. The initial testing, then, was used

to validate the hardware and reactor stack-up for its suitability at high pressures. The

ability of the reactor to maintain a positive seal, the ability of the cooling flows to

maintain acceptable operating temperatures and the controllability of the air and fuel

flows at the higher pressures were of specific interest. The same flow set-up was used for

the high pressure testing as was used for the low pressure testing. The 15kW heater was

sufficient to provide the required air temperatures for high pressure operation. The test

conditions used for reactor checkout and initial high pressure blowout data are shown in

Table 3.6.

61
 Table 3.6: Test Conditions for High Pressure Blowout with External Vaporization
Air Flow Reactor
Fuel Rate LP Pressure Starting
(lbm/min) (psia)
2.84 3.25 29.45
JP-8 4.24 3.0 38.18 0.6
3.90 2.0 45.50

3.7. EXPERIMENTS USING THE INTERNAL VAPORIZATION

TECHNIQUE

While high pressure testing with the external vaporizer served as a proof of

concept and afforded time to validate the hardware and operating procedure, the safety

limitation of the vaporizer prevented the reactor pressures from reaching the target of 5

atm. The next phase of testing was to evaluate the internal vaporization technique, and

utilize the lower residence time to achieve ever higher reactor operating pressures.

3.7.1. High Pressure Operation

With the internal vaporization technique, the fuel is heated with the fluidized sand

bath to 750 °F in order to facilitate supercritical flash vaporization. For the high pressure

testing, 100 and 50 micron orifice tubes were installed on the two fuel legs. This

combination of orifices provide adequate pressure drop to achieve vaporization while still

permitting the required flow over a range of pressures.

The temperature of the nitrogen heat trace over the lines was increased to 800 °F.

Now that there was no longer a vaporizer body and fuel nozzle, the air from the 15kW

heater was routed directly into the air feed lines. The same thermal management

techniques (i.e., tape heaters and ceramic insulation) used in the external vaporization

tests are utilized here.

62
 The goal of the second phase of high pressure experiments is twofold. The first is

to demonstrate the feasibility of the direct vaporization method and to establish some

baseline operability data for the reactor. The second is to use the new technique to

achieve blowout data at a reactor pressure of 5 atm. Additionally, several blowout points

previously tested with the external vaporizer were tested to evaluate the repeatability of

the new vaporization technique. The test conditions used are shown in Table 3.7Table

3.7.

Table 3.7: Test Conditions for High Pressure Blowout with Internal Vaporization

Air Flow Reactor

Fuel Rate LP Pressure Starting
(lbm/min) (psia)
2.20 7.73 15.80
2.20 2.53 29.40
JP-8 0.6
3.90 2.04 45.50
6.67 1.49 73.00

3.7.2. Sub-atmospheric

After the internal vaporization technique was tested at high pressure, it was

decided to evaluate the performance of the system at low pressure. The low pressure

vaporization is the difficult to maintain due to the high heat loss associated with low air

and fuel flow rates. The negative impact of heat loss is partially offset by the reduced

dewpoints caused by sub atmospheric pressures. If any of the orifice tubes is not

producing complete vaporization, but rather producing fuel droplets, the impact would be

more evident in the reactor at low pressure. The 50 micron orifice tubes were used, while

all of the other systems remained the same. Several of the low pressure conditions tested

with the external vaporizer were repeated with the internal vaporizer. Additionally, a final

63
high flow rate, low pressure condition was conducted to maximize the loading parameter.

The test conditions are shown in Table 3.8.

Table 3.8: Test Conditions for Sub-Atmospheric Blowout with Internal

Vaporization
Air Flow Reactor
Fuel Rate LP Pressure Starting
(lbm/min) (psia)
0.98 6.57 11.01
0.98 7.10 10.56
JP-8 1.48 11.10 10.35 0.6
1.97 17.03 9.59
1.31 39.09 4.81

3.8. TOTAL TEMPERATURE PROBE

The design of the temperature probe is discussed in Section 2.5.2. In practice, the

choked conditions required to measure the mass flow are achieved with orifices. The

ceramic probe is constructed from a 0.375-inch diameter extruded alumina tube with a

closed end. A 0.062-inch hole is drilled into the side of the tube with a diamond-burr bit.

The orifice in the side of the probe creates a hole in cross-flow condition. The orifice in

the ceramic is inserted into the flame. Downstream of the probe, a 0.062-inch orifice

fitting is inserted into the system. The downstream orifice meters the system and

provided the temperature and pressure measurements required for the temperature

calculation.

The ceramic portion is inserted into the flame, where the pressure is known.

Sample is drawn through the system to a second orifice, there the temperature and

pressure are measured. The sample is collected using a Venturi vacuum pump. The pump

64
is connected to 80 psi shop air, and it is capable of sustaining a vacuum down to 3 psia. A

low downstream pressure is essential to ensure that both of the orifices are choked.

Starting at atmospheric pressure, the first orifice must be bought down to 7.35 psia, while

the second orifice will need to be at 3.675 psia for the system to be choked, assuming

isentropic, ideal flow. The differential pressure across the second orifice is measured

using a pressure transducer, while the temperature is measured using a type-K

thermocouple inserted downstream of the orifice fitting. Figure 3.13 shows the set-up that

was used for testing, with important components called out.

Figure 3.13: Laboratory set-up of the ceramic total temperature probe

In Figure 3.13, the pressure tap nearest to the vacuum pump was used to ensure

that the system was choked. The probe static pressure tap, indicated in the figure, is used

65
in the calculations, and is a function of the gas temperature. The thermocouple indicated

is a K-type and is used to measure the fluid temperature immediately after the second

orifice, providing in the calculation.

When the system is used, the probe was set to the desired location within the

flame, and the vacuum pump activated. The vacuum pump was controlled by changing

the pressure of the shop air. After the vacuum was established, the system was allowed to

preheat. In most cases, it was not possible to choke the orifice system at ambient

temperature. The second orifice had to exceed 100 °F before choked flow could be

established in both orifices. After it was verified that the system was choked, the

temperature and pressure measurements were collected.

3.8.1. Thermocouple Comparisons

The first test of the temperature probe was to compare the temperature calculated

with the probe with thermocouple readings taken in close proximity to the probe orifice.

Comparison to thermocouple readings allowed a rapid evaluation of the basic operation

and accuracy of the probe upon which to base further testing efforts. The probe was

evaluated over two different burners. The first was in the exhaust of the WSR. The

exhaust of the WSR provides a high flow rate and a high heat release environment that

the probe is unlikely to perturb. The second flame tested was the flat flame of the

McKenna burner. This burner has a smaller flame and has lower flow rate. However, the

flame is very steady, optically accessible, and well characterized. The McKenna burner

represents a case when the presence of the probe may affect the flame. The conditions

tested for each of the burners are shown in Table 3.9.

66
 Table 3.9: Test Conditions for Ceramic Probe to Thermocouple Comparison

Air Flow Equivalence

Test
Rate (SLPM) Ratio

400 0.60
400 0.70
WSR
400 0.80
400 1.4
12 0.6
12 0.7
McKenna
12 0.8
Burner
12 0.9
12 1

3.8.2. Laser Diagnostics

The second phase of the probe characterization was to compare the results calculated

from probe measurements to temperatures taken with laser diagnostics over well

characterized flames. A study by Hancock et al.[32] used CARS to characterize the

temperature profile of a Hencken burner flame with two different fuels. This study was

repeated and temperature measurements were taken with the ceramic probe. The

temperature results from CARS can be considered the true adiabatic temperatures of the

flame, and are considered the target for the calculated probe temperature. The test

conditions for each fuel are shown in Table 3.10. [32]

67
 Table 3.10: Test Conditions for Hencken Burner Test [32]

Air Flow Equivalence

Fuel
Rate (SLPM) Ratio

12.2 0.5
12.2 0.6
12.2 0.7
12.2 0.8
12.2 0.9
H2 12.2 1
12.2 1.1
12.2 1.2
12.2 1.3
12.2 1.4
12.2 1.5
12.2 0.51
12.2 0.62
12.2 0.71
12.2 0.81
12.2 0.91
C2H4 12.2 0.96
12.2 1.01
12.2 1.06
12.2 1.11
12.2 1.21
12.2 1.3

In addition to comparing the CARS measurements to probe calculations in the same

flame, TDLAS data was taken simultaneously with the probe to characterize the intrusive

impact of the probe in the flame. TDLAS data was taken in connection with another

research group. [17] Data was collected at two different heights above the burner. The

orientation of the probe is shown in Figure 3.14. At each height, data was collected both

while the probe was pulling sample and when the vacuum pump was off and no sample

68
was being collected. Baseline TDLAS temperatures were recorded without the probe in

the flame.

Figure 3.14: Probe orientation for TDLS investigation

Data was collected using ethylene as the fuel. The conditions were a truncated

matrix of the conditions used in the CARS characterization and are shown in Table 3.11.

Table 3.11: Test Conditions for Intrusive Characterization

Air Flow Equivalence

Fuel
Rate (SLPM) Ratio

12.2 0.8
C2H4 12.2 1
12.2 1.2

The experiments conducted with the ceramic temperature probe showed that the

technique has promise as a temperature measurement tool in high temperature

environments. However, due to a tight time constraints in the High Pressure Combustion

Facility, the probe was not used as part of the HPWSR campaign. The results from the

thermocouple and laser diagnostic comparisons are presented in Appendix A.

69
 CHAPTER 4

EXPERIMENTAL RESULTS AND DISCUSSION

4.1. ALTERNATIVE FUEL EVALUATION

The first test conducted with the new HPWSR hardware was a study of the

combustion stability of alternative fuels. The suitability of a cooled metal reactor for

blowout studies was also investigated. The specific focus was to investigate the impact of

cetane number of the blowout characteristics. The cetane number is an indicator of how

easily a fuel will ignite in a compression ignition environment, and is often used as a

measure of diesel fuel quality. Specifically, cetane number is the inverse of octane

number, and inversely proportionate to the fuels ignition delay. A fuel with a higher

cetane number will have a shorter ignition delay time. Currently in the JP-8

specifications, there is no target cetane number, or cetane number range, for new

alternative fuels. As alternative fuels are refined from varying feed stock, there is the

possibility for a wide variation in the cetane number of those fuels relative to JP-8. When

these new fuels are used in systems designed to run on JP-8, the change in the ignition

delay time has to potential to effect the ignition and stability limits of the combustor.

Initial studies in the WSR are intended to determine if there is a noticeable change in the

lean blowout characteristics of alternative fuels from the JP-8 baseline, and quantify the

strength of the correlation between lean blowout and cetane number. The specific test

70
conditions for the study are shown in Table 3.3. While it was intended to compare the

blowout of two fuels to JP-8, limited lab time only allowed one alternative fuel to

becharacterized. The fuel used was iso-paraffinic kerosene (IPK) with a cetane number of

31.5. The comparison between lean blowout of JP-8 and IPK is shown in Figure 4.1

Figure 4.1: Lean blowout of alternative fuels in a metal reactor

The data shown in Figure 4.1 indicates that the equivalence ratio at blowout has a

weak dependence on fuel cetane number. A 34% increase in the fuel cetane number

caused only a 2.8% increase in the equivalence ratio at blowout. IPK has a lower cetane

number than JP-8, meaning that the ignition delay time of IPK is higher than JP-8. As a

result of the increased ignition delay time, IPK is more resistant to ignition and prone to

71
lean blowout at lower velocities of LP than JP-8. However, the dependence of lean

blowout on the cetane number was weak in the current study. Additionally, over the LP

range tested, the blowout equivalence ratio increased by approximately 10%, creating the

beginning of a stability curve. Previous work by Colket et al.[24 investigated the impact of

cetane number in a practical combustor. Seven fuels were tested with Cetane number

ranging from 30 to 80. The results from Colket et al. are shown in Figure 4.2. They

showed that the lean blowout equivalence ratio decreased the derived Cetane number

increased.

Figure 4.2: Sensitivity of lean blowout equivalence ratio to DCN [24]

The results from Colket et al. [24] correspond with the results taken with the metal reactor.

72
 In the practical combustor used, IPK produced a blowout equivalence ratio 3.3%

higher than Jet-A. This agrees well with the results from the metal reactor of a 2.8%

increase in the blowout of IPK over JP-8.

The increase in equivalence ratio at blowout, however, is not consistent with

previous atmospheric data collected. Over a similar LP range with a ceramic reactor, little

variation in blowout is observed. The blowout data collected with the metal reactor is

biased to a higher equivalence ratio by the cooling. The difference in the LBO between

the fused silica ceramic and metal reactor for similar loading parameters is shown in

Figure 4.3.

Figure 4.3: Effect of ceramic and metal reactors on LBO

As LP is increased, more cooling is required to maintain the metal reactor at a safe

operating temperature, increasing the heat loss from the reactor and causing an artificially

73
high blowout point. The ratio of the thermal conductivities of the two materials ( )

is 16. Therefore, for the high pressure test metal reactors are not desirable. However, the

increase in blowout caused by a change in cetane number is still applicable since both

fuels were subject to the same magnitude of heat loss.

4.2. COMBUSTION STABILITY

After studies with the metal reactor showed that heat loss artificially increased the

blowout equivalence ratio, the HPWSR was redesigned to accommodate a ceramic

reactor, as used in previous atmospheric experimental work. Discussion of the design of

the HPWSR can be found in Section 3.2. The following sections present data taken with

the final HPWSR design and a ceramic reactor.

4.2.1. Lean Blowout at Atmospheric Pressure

The first step to characterize the new HPWSR design was to repeat atmospheric

blowout data collected with the previous WSR design. This allowed a comparison of the

data to ensure that the HPWSR design was not artificially impacting the blowout data due

to high heat loss. The atmospheric data was collected over a LP range of 1.5-3, and is

shown in Figure 4.4.

74
 Figure 4.4: Lean blowout in an atmospheric WSR

When operated at atmospheric pressure, there was no significant change in the

blowout equivalence ratio, over the LP range tested. The main limitation with

atmospheric experiments is the maximum mass flow rate attainable for the given reactor

volume. From the definition of loading parameter, the loading can be increased by

increasing the mass flow, lowering the pressure, or lowering the volume. In the initial

experiments, the pressure and volume were fixed, limiting achievable loading parameter.

Operating at low pressure increases the loading parameter, while operation at high

pressure allows more mass to be flowed for a given equivalence ratio. The limited

loading parameter range and minimal change in blowout equivalence ratio correspond

with previous blowout work with a ceramic WSR.[37] The results in Figure 4.5 indicate

that the new HPWSR design is performing as intended and is not skewing the data.

75
Therefore, the ceramic reactor and the HPWSR hardware were ready for testing at low

and elevated pressures.

Figure 4.5: Comparison of previous LBO data in a ceramic WSR[37]

4.2.2. Experiments at Low and Elevated Pressures

Data was collected in the High Pressure Combustion Research Facility at

WPAFB.[6] The pressure in the reactor was varied from 4 psia to 73.5 psia. In addition to

characterizing lean blowout at varying pressures, the internal vaporization technique, as

described in Section 3.3.1, was evaluated. The lean blowout data as a function of LP is

shown in Figure 4.6.

76
 Figure 4.6: Combustion stability curve for JP-8

The maximum LP achievable with the previous atmospheric hardware was

approximately 3 .The pressure variation in the reactor allowed the LP to be

extended to 40 , resulting in over an order of magnitude increase in the

achievable LP in a WSR. Additionally, an increase in lean blowout equivalence ratio of

approximately 28% was observed. The data show that the flame becomes less statically

stable as LP is increased. The LP range of 3-40 achieved with the HPWSR is

significantly larger than the achievable LP range operating at atmospheric pressure. The

increase in the blowout equivalence ratio due to higher air flows and lower pressures (i.e.,

high LP) enabled the plotting of the lean side of the stability curve for a WSR during JP-

8.

77
 Test data across multiple test days is also called out in Figure 4.6. The test data

recorded at near atmospheric pressures on over multiple days of testing correlate very

well. This consistency shows that the hardware is capable of generating repeatable data, a

crucial test for long term operation.

In addition to the day-to-day consistency of the reactor, the blowout points

achieved with both the internal and external vaporization techniques showed close

agreement. The comparison between the two vaporization methods is shown in Figure

4.7.

Figure 4.7: Blowout points taken with internal and external vaporization

78
 A separate way of plotting combustor blowout data is equivalence ratio at

blowout vs. the combustor operating pressure. This type of combustor operating map is

useful in a gas turbine engine to demonstrate the operating pressures required for safe

operation. The lean blowout data as a function of reactor pressure is shown in Figure 4.8.

Figure 4.8: Pressure dependence of JP-8 lean blowout in a WSR

The data in Figure 4.8 clearly shows that as the reactor pressure is decreased, the

blowout equivalence ratio is increased, with the highest blowout point occurring at the

lowest reactor pressure or 4 psia. Low pressure operation represents the most highly

loaded condition and, therefore, illustrates fully the combustion limits of a given fuel.

Conversely, as the reactor pressure is increased, the flame becomes more stable and

blowout equivalence ratio is decreased, with the lowest blowout occurring at the highest

reactor pressure of 73 psia. As evident in the data, it is desirable to operate at the highest

79
pressure possible in order to achieve the most stable reaction. Figure also shows the

reactor operability map. Data sets taken over multiple test days at the same pressure have

very close correlation with each other. Additionally, data taken using external and

internal vaporization at the same condition agree closely, indicating that the fuel is

completely vaporizing and the internal vaporization technique is not biasing results. The

tight correlation between these two data sets shows that the internal vaporization

technique is stable and repeatable even at low operating pressures.

After reactor stability was established for low pressure conditions, the operability

was investigated for elevated pressures. Over a night of testing, the reactor pressure was

maintained at a constant 5 atm and the equivalence ratio was varied between 0.5 and

blowout. During this investigation, emissions data were collected. Figure 4.9 shows

carbon monoxide emission data collected at a constant reactor pressure of 5 atm.

Figure 4.9: CO emissions for JP-8 at 5 atm

80
 When evaluating the emissions of gas turbine engines, it is important to vary the

conditions to discover the local minima, called the bucket. The emissions bucket is the

condition at which the emissions of a given species are the smallest. Emissions of CO

represent the inefficiency of the combustion reaction. If the combustion is more efficient,

more of the CO will be converted to the final product of CO2. As the equivalence ratio

approaches blowout, the combustion temperature decreases, and the conversion from CO

to CO2 is retarded. The equivalence ratio excursion in the WSR at high pressure was

sufficient to show the CO bucket for JP-8. Figure 4.9 shows that the reactor is operating

most efficiently, from CO to CO2 conversion, at an equivalence ratio of 0.46. When the

equivalence ratio is increased, combustion temperature rises, and more complete

combustion results. There exists a point of minimum CO production, if the equivalence

ratio is increased passed this point, the CO production again increases.

Figure 4.10: Minimum CO emissions for JP-8 at 5 atm

81
 Figure 4.10 shows the same CO emissions data, zoomed in on the CO bucket.

For the HPWSR at 5 atm, the minimum CO point occurred around an equivalence ratio of

0.46.

Emissions data were also collected at atmospheric pressure using the same

reactor hardware, allowing the effect of pressure on CO emission to be evaluated. Figure

4.11 shows CO emissions data collected at 1 and 5 atm in the same reactor.

Figure 4.11: CO emissions of JP-8 at 1 and 5 atm

Figure 4.11 shows that significantly less CO was produced at 5 atm reactor

pressure than during atmospheric operation. Higher reactor pressures favor conversion

from CO to CO2, resulting in more efficient combustion and less CO emissions.

Additionally, the equivalence ratio at the minimum CO point is shifted lower at higher

82
pressures. The minimum CO point at 5 atm is at an equivalence ratio of 0.46, while the

minimum CO point at 1 atm is at an equivalence ratio of 0.51.

Figure 4.12: Emissions of CO2 and H2O at 5 atm

Emissions of CO2 and H2O serve as an indication of the combustion efficiency.

These concentrations, on a volume basis, are shown in Figure 4.12. The emissions of both

CO2 and H2O increase linearly with equivalence ratio and have similar magnitudes.

4.2.3. Low Pressure Simulation with Nitrogen at Atmospheric Pressure

The theory of low pressure simulation is discussed in Section 2.1.2. Diluting the

air with nitrogen allows a significantly higher loading parameter than otherwise

83
achievable in the same reactor. Data collected at atmospheric pressure using the nitrogen

dilution technique are shown in Figure 4.13.

Figure 4.13: JP-8 Combustion stability curve using simulated low pressure

The addition of nitrogen allowed the creation of a complete combustion stability

curve for JP-8. Previous atmospheric tests showed minimal change in the blowout

equivalence ratio. The data in Figure is plotted against the nitrogen to fuel mass ratio.

During the nitrogen addition tests, excess nitrogen was added in addition to the

established air flow rate, effectively increasing residence time. In nitrogen replacement

tests, the overall mass flow rate to the reactor was maintained constant. As shown in the

data, nitrogen replacement allowed a significantly higher blowout equivalence ratio to be

achieved.

84
 While the data collected has the appearance and overall trend of a JP-8 stability

curve, LP cannot be calculated until the nitrogen flow rate is correlated to a true low

pressure. Additionally, the peak blowout equivalence ratio in the study is 0.9. Theory

would suggest that the peak blowout equivalence ratio should be unity. The depression of

the peak blowout equivalence ratio could be a result of the incoming diluent causing

areas locally leaner than the bulk flow, creating combustion instability. Additionally, the

high concentrations of nitrogen at the large dilution ratios could be locally reducing

mixing, causing the reactor to blowout below unity. Further experiment and simulation is

needed to determine the cause.

A full calibration of the nitrogen dilution technique was not attempted in the

current study. To fully characterize the method, nitrogen dilution data would be

compared to true low pressure data of the same residence time in the same reactor.

However, the low pressure data points of the current study were sufficient to begin to

visualize a calibration. A comparison of low pressure and nitrogen dilution data is shown

in Figure 4.14, with the true low pressure data plotted on the upper x-axis, and nitrogen

dilution data plotted on the lower x-axis. The equivalence ratio at blowout is shown on

the common y-axis.

85
 Figure 4.14: Correlation of true low pressure to nitrogen dilution data

Data collected at atmospheric pressure and 10 psia correlate very well with

nitrogen dilution data. Past 10 psia, however, the true low pressure data displays a

shallower stability curve slope than the nitrogen dilution data, resulting in nitrogen

dilution over predicting the blowout equivalence ratio at 4 psia. While the data set is not

sufficient to fully anchor the simulation, Figure 4.14 indicates that beyond a nitrogen to

fuel mass ratio of 6, the nitrogen in the system has additional impact of the blowout.

Further study is required to determine how diluent is impacting the system.

86
 CHAPTER 5

CONCLUSION

5.1. PROBLEM SUMMARY

The experimental study presented in this thesis detailed the development and

initial experimental entries using a WSR at high and low pressures. Previous work using

a WSR for gathering emissions data, fuel characterization and stability mapping had been

done at atmospheric pressure. In order to more accurately represent the conditions present

in a gas turbine combustor, elevated pressure operation is necessary. Additionally, the

low pressures encountered at high flight altitudes put more stress on the stability of gas

turbine fuels. Therefore the ability of a WSR to operate at the two pressure extremes seen

in gas turbine aircraft allows the generation of benchmark quality data for emissions and

fuel lean blowout stability.

For the initial experimental studies with the high pressure WSR, fuel was

vaporized and mixed with air externally. The method of external vaporization has been

used successfully in the past for atmospheric WSR work. External vaporization was a

safe and stable method only up to 3 atm, due to autoignition concerns.

87
 To push the WSR to higher pressures, a low residence time vaporization

technique was developed. Fuel was heated above its vaporization temperature, to a

supercritical state, and injected directly into the air lines upstream of the WSR jet ring.

Flash vaporization occurred through the use of orifice tubes, and mixing occurred in the

feed lines. The residence time was maintained shorter than the autoignition time for the

fuel so no burning would occur upstream of the reactor. This method proved stable and

allowed operation up to 5 atm reactor pressure, yielding over an order of magnitude

increase in the loading parameter. The WSR hardware was also used to gather data at

sub-atmospheric pressure, and stable operation was demonstrated down to 4 psia. Low

pressure data applies to aircraft operating at high altitude, where combustion stability is

critical.

While gathering true low pressure data is important to fully understanding the

combustion process, the hardware to do so is expensive and complicated. Another piece

of this study was to determine if a correlation between low pressure combustion stability

and combustion stability with nitrogen dilution can be generated. Using nitrogen dilution

to simulate low pressures in an atmospheric rig can allow quick and simple evaluation of

alternative fuels and new combustor designs.

5.2. CONCLUSIONS

 The WSR was demonstrated to operate stably over a pressure range of 4 psia

to 73 psia.

88
 The loading parameter achieved in the WSR was extended by a factor of 20,

from a maximum value of 2 achieved during atmospheric testing to

40 .

o Previous experiments utilizing the WSR were only able to increase

loading parameter by a factor of 2. The greater range of loading parameter

achieved in the current study allowed the evaluation of blowout

performance at conditions that more closely represent a gas turbine

engine.

 Adiabatic conditions are required to meaningful blowout data. The low thermal

conductivity of fused silica makes it an ideal material for lean blowout studies.

 Nitrogen dilution was shown to simulate the effects of low pressure over a

limited range of pressures.

o The addition of excess nitrogen allowed the construction of a full

combustion stability loop for JP-8, however further study is required to

correlate nitrogen flow rate to pressure. The diluent-pressure correlation

is required to plot nitrogen dilution data against loading parameter.

 Cetane number was shown to have a weak impact of the blowout equivalence

ratio. A decrease in Cetane number of 34% results in a 2.6% increase in

equivalence ratio at blowout.

89
  A ceramic temperature probe using isentropic relations was shown to

accurately measure flame temperature when accounting for conduction and

radiation losses.

5.3. FUTURE WORK

The ultimate goal of the high pressure WSR was to develop a complete stability

loop for JP-8 in conditions similar to a gas turbine engine. In order to continue to develop

the stability curve, higher pressures and mass flows are required. Further study with the

high pressure WSR should continue to adapt the hardware for operations at pressures up

to 20 atm. A 20 atm. capable WSR would allow lean blowout and pollutant emissions

data to be collected for JP-8 that is directly applicable to the conditions seen in gas

turbine engines.

The experiments with nitrogen dilution in the current study showed a promising

correlation with true low pressure data; however, time and facility availability constraints

prevented a full characterization. Future work should focus on taking nitrogen dilution

and low pressure blowout data in the same reactor. The direct comparison between

nitrogen and low pressure blowout data will allow a full correlation that can be used to

better understand future combustion systems and alternative fuels during a preliminary

design phase, without the need for the expensive low pressure combustion facilities.

Further characterization of the ceramic temperature probe is also required. While

the probe showed promising results on both laboratory flames and the WSR exhaust,

further design work is needed to optimize the size and material of the probe, the orifice

90
dimensions and the probe hardware and data collection system. Further, incorporation in

a full scale combustor test would demonstrate the probes ability to measure temperature

in environment too hostile for conventional thermocouple techniques.

91
 BIBLIOGRAPHY

1. A. H. Lefebvre, G. A. Halls. "Simulation of Low Combustion Pressures by Water

Injection." (International Symposium on Combustion) 7, no. 1 (1958).

2. Bohan, B.T.,. ""Analysis of Flow Migration in an Ultra Compact Combustor"."

Master's Thesis, Air Force Institute of Technology, Wright Patterson Air Force Base,

2011.

3. Carlos A. Destefani, Anthony B. Murphy, Elias Siores, Marie-Francoise Elchinger.

"Enthalpy Probe Diagnostics of an Atmospheric Pressure Argon Microwave Induced

Plasma." (Applied Physics) 33 (2000).

4. D. Kretschmer, J. Odgers. "Modeling of Gas Turbine Combustors- A Convenient

Reaction Rate Equation." (Journal of Engineering for Power) 1972.

5. D. R. Ballal, M. D. Vangsness, S. P. Heneghan, G. J. Sturgess. "Studies of Lean

Blowout in a Research Combustor." (Fuels and Combustion Technology for

Advanced Aircraft Engines) 1993.

6. Dale T. Shouse, Jeff Stutrud, Charles Frayne. "High Pressure Combustion Research at

the Air Force Research Labortory." (International Society for Air Breathing Engines)

2001.

7. Edwards, Tim. "Advancements in Gas Turbine Fuels from 1943 to 2005." (Journal of

Engineering for Gas Turbines and Power) 129 (2007).

92
8. G. J. Sturgess, D. Shouse. "Lean Blowout Research in a Generic Gas Turbine

Combustor with High Optical Access." (International Gas Turbine and Aerospace

Congress and Exposition) 1993.

9. G. J. Sturgess, S. P. Heneghan, M. D. Vangsness, D. R. Ballal, A. L. Lesmerises, D.

Shouse. "Effects of Back Pressure in a Lean Blowout Research Combustor." (Journal

of Engineering for Gas Turbines and Power) 1992.

10. G.J Sturgess, S.P Heneghan, M. D. Vangsness, D. R. Ballal, A. L. Lesmerises. "Lean

Blowout in a Research Combustor at Simulated Low Pressures." (Journal of

Engineering for Gas Turbines and Power) 118 (1996).

11. J. W. Blust, D. R. Ballal, G. J. Sturgess. "Emissions Characteristics of iquid

Hydrocarbons in a Well Stirred Reactor." (American Institute of Aeronautics and

Astronautics) 1997.

12. J. W. Blust, D. R. Ballal, G. J. Sturgess. "Fuel Effects on Lean Blowout and

Emissions from a Well-Stirred Reactor." (Journal of Propulsion and Power) 15, no. 2

(1999).

13. J. Zelina, D. R. Ballal. "Combustion and Emission Studies using a Well Stirred

Reactor." (AIAA/ASME/ASEE Joint Propulsion Conference) 30 (1994).

14. J. Zelina, D. R. Ballal. "Combustion Studies in a Well Stirred Reactor." (32nd

Aerospace Sciences Meeting) 1994.

15. James M. Gere, Barry J. Goodno. Mechanics of Materials. Cengage Learning, 2009.

16. Joseph Zelina, Richard C. Striebich, Dilip R. Ballal. "Pollutant Emissions Research

Using a Well Stirred Reactor." (AIAA) 1994.

93
17. Justin T. Gross, Scott D. Stouffer, John Hoke, Drew Caswell, David L. Blunck, Fred

R. Schauer. "Simultaneous Temperature and Species Probe." 8th Annual Dayton

Engineering Sciences Symposium. Dayton, OH, 2012.

18. K. K. Gupta, A. Rehman, R.m. Sarviya. "Bio-Fuels for the Gas Turbine: A Review."

(Renewable and Sustainable Energy Reviews) 14 (2010).

19. L. Rye, C. Wilson. "The Influence of Alternative Fuel Composition on Gas Turbine

Ignition Performance." (Fuel) 2011.

20. Law, C. K. Combustion Physics. New York, NY: Cambridge University Press, 2006.

21. Lefevbre, A. H. Gas Turbine Combustion. Hemisphere Publishing Corporation, 1983.

22. Lefevbre, A., Freeman, W., Cowell, L.,. ""Spontaneous Ignition Delay Characteristics

of Hydrocarbon Fuel/Air Mixtures"." 1986.

23. Mattingly, J.D. Elements of Gas Turbine Propulsion. New York: McGraw-Hill, Inc,

1996.

24. Med Colket, Steve Zeppieri, Zhongtao Dai, Don Hautman. "Fuel Research at UTRC."

5th Annual Fuel Research Meeting, Livermore, CA.: Multi-Agency Coordinating

Council for Combustion Research, 2012.

25. Megan F. Karalus, K. Boyd Fackler, Igor V. Novosselov, John C. Kramlich, Philip C.

Malte. "Characterizing the Mechanism of Lean Blowout for a Recirculation-

Stabilized Premixed Hydrogen Flame." (Proceedings of ASME Turbo Expo) 2012.

26. Moshan S. P. Kahandawala, Matthew J. DeWitt, Edwin Corporan, Sukh S. Sidhu.

"Ignition and Emission Characteristics of Surrogate and Practical Jet Fuels." (Energy

and Fuels) 22 (2008).

94
27. Nenniger, J.E., Kridiotis, A., Chomiak, J., Longwell, J.P., Sarofim, A.F.,.

""Characterization of a Toroidal Well Stirred Reactor"." 20th International

Symposium on Combustion. The Combustion Institute, 1984. 473-479.

28. O'Neil, Alanna Rose. Chemiluminescence and High Speed Imaging of Reacting Film

Cooling Layers. Masters Thesis, University of Dayton, 2011.

29. P. H. Kydd, W. I. Foss. "Combustion of Fuel-Lean Mixtures in Adiabatic Well-

Stirred Reactors." (Tenth Symposium (International) on Combustion) 1965.

30. P. Magre, G. Collin, O. Pin, J. M. Badie, G. Olalde, M. Clement. "Temperature

Measurments by CARS and Intrusive Probe in an Air-Hydrogen Supersonic

Combustion." (International Journal of Heat and Mass Transfer) 44 (2000).

31. R. L Farrow, P. L. Mattern, L. A. Rahn. "Comparison between CARS and Corrected

Thermocouple Temperautre Measurments in a Diffusion Flame." (Applied Optics)

21, no. 17 (1982).

32. Robert D. Hancock, Kenneth E. Bertagnolli, Robert P. Lucht. "Nitrogen and

Hydrogen CARS Temperautre Measurments in a Hydrogen/Air Flame Using a Near-

Adiabatic Flat-Flame Burner." (Combustion and Flame) 109 (1997).

33. Scott D. Stouffer, Benjamin M. Mortimer, David M. Ostdiek, Matthew J. DeWitt,

Robert Pawlik, Richard Reich, Charles W. Frayne, Howard Mayfield. "Soot

Reduction Using a Well-Stirred Reactor." (43rd AIAA Aerospace Sciences Meeting)

2005.

34. Shazib A. Vijlee, John C. Kramlich, Ann M. Mescher, Scott D. Stouffer, Alanna R.

O'Neil-Abels. "Characterizing Combustion of Synthetic and Conventional Fuels in a

Toroidal Well Stirred Reactor." (Proceedings of the ASME Turbo Expo) 2013.

95
35. Simon Blakey, Lucas Rye, Christopher Willam Wilson. "Aviation Gas Turbine

Alternative Fuels: A Review." (Proceedings of the Combustion Institute) 33 (2010).

36. Stouffer, S.D., Ballal, D.R., Zelina, J., Shouse, D.T., Hancock, R.D., Mongia, H.C.,.

""Development and Combustion Performance of a High-Pressure WSR and TAPS

Combustor"." Reno, NV: 43rd AIAA Aerospace Sciences Meeting, 2005.

37. Stouffer, S.D., Pawlik, R., Justinger, G., Heyne, J., Zelina, J., Ballal, D.,.

""Combustion Performance and Emissions Characteristics for a Well-Stirred Reactor

for Low Volatility Hydrocarbon Fuels"." Cincinnati, OH: 43rd Joint Propulsion

Conference, 2007.

38. Terrence R. Meyer, Sukesh Roy, Tohmas N. Anderson, Joseph D. Miller, Viswanath

R. Katta, Robert P. Lucht, James R. Gord. "Measurments of OH Mole Fraction and

Temperature up to 20 kHz by Using a Diode-Lase based UV Absorption Sensor."

(Applies Optics) 44, no. 31 (2005).

39. Turns, S.R.,. An Introduction to Combustion. 2nd Ed. New York: McGraw Hill, Inc.,

1996.

40. W. E. Wesr Jr., J. W. Westwater. "Radiation-Conduction Correction for Temperature

Measurments in Hot Gasses." Industrial and Engineering Chemistry, 1953.

41. Weiss, John P. Longwell and Malcolm A. "High Temperature Reaction Rates in

Hydrocarbon Combustion." (Industrial and Engineering Chemistry) 47, no. 8 (1955).

42. Zelina, J.,. ""Combustion Studies in a Well Stirred Reactor"." PhD Thesis, University

of Dayton, Dayton, OH, 1995.

96
 APPENDIX A

TOTAL TEMPERATURE PROBE RESULTS

A.1 THERMOCOUPLE COMPARISONS

The first test with the temperature probe was a thermocouple comparison in the exhaust

of the WSR. The data for the thermocouple and the probe calculated temperature are

shown in Figure A.1.

Figure A.1: WSR exhaust temperature comparison

97
 The calculated probe temperatures agree within 4% of measurements taken with a

thermocouple at the same height. When used to evaluate a high flow rate, high heat

release flame such as the WSR exhaust, the probe is very capable. The second

comparison for the probe was over a McKenna flat flame burner with an ethylene flame.

The calculated probe temperatures were compared to thermocouple measurements at the

same location. The data are shown in Figure A.2.

Figure A.2: Temperature comparison of McKenna burner flame

Over a McKenna burner, the probe resulted in lower temperature than

thermocouples, with a 9% error or less observed. Comparing the results of the WSR

exhaust and the McKenna burner, it is clear that the intrusive impact of the probe is

biasing the temperature measurements in lower flow rate flames.

98
 While thermocouple comparisons provided a quite diagnostic for probe

characterization, the true test of the accuracy of the measurements is comparison to

CARS data taken over the same flame. Data from Hancock et. al[32] from a CARS

characterization of Hydrogen-Air and Ethylene-Air flames over a Hencken burner was

used to compare the probe calculated temperatures. The data for the two fuels tested are

shown in Figure A.3 and Figure A.4.

Figure A.3: Comparison of temperatures in a hydrogen Hencken burner flame[32]

99
 Figure A.4: Comparison of temperatures in an ethylene Hencken burner flame [32]

When testing over a small, low flow rate flame on the Hencken burner, the heat

loss due to the probe was evident. The calculated probe temperature was 12-18% lower

than the CARS measurements at the same location. The Hencken burner has a flame exit

area that is 1-inch square, while the outer diameter of the probe is 0.375-inch. The most

likely explanation for the low calculated temperature is the conduction losses form the

flame to the probe are so great that the probe temperature is not reaching the full flame

temperature.

A.2 LASER DIAGNOSTICS

The intrusive impact of the probe in the Hencken burner was characterized using

TdLAS.[17] Laser measurements were taken at two different heights over the burner, both

100
while sampling and not sampling. These measurements were compared to a baseline

without the probe present in the flame at all. The results are shown in Figure A.5.

Figure A.5: Intrusive impact of a probe on a Hencken burner flame[17]

The data show that at the lower height, while pulling sample, the flame

temperature, as measured with TdLAS is lower than the baseline, while at the same

height with no sampling results in a temperature higher than the baseline. At the second

height, the test when pulling sample correlated well with the baseline, while with no

sample the temperature was higher than the baseline, although by a lesser amount than at

height 1.

The discrepancies observed with the TdLAS survey can be explained by two

phenomena. Frist, when sample is being pulled through the probe, the movement induces

101
velocity in the surrounding flow. Since the orifice is choked, the flow approaches sonic

velocity as it nears the orifice. This velocity aerodynamically cools the flow, resulting in

additional cooling over the conduction losses, causing the temperature to be lower than

the baseline. When the sample is turned off, the hot ceramic insulates and stagnates the

heat from escaping from the top of the flame, causing an increase in the measured

temperature over the baseline. In the case of the Hencken burner, the probe diameter is

such a large percentage of the burner area that the insulation effect is noticeable. The

same principles apply at height 2, however, since the probe is more removed from the

flame surface, the effect is not as great.

A.3 PROBE MEASUREMENTS WITH RADIATION-CONDUCTION

CORRECTION

After characterizing the probe with laser measurements, it was decided that the

major factor in the temperature discrepancy was the losses due to the probe. The

conduction-radiation correction as proposed by West and Westwater [40] is discussed in

Section 2.5.3. This correction was applied to data collected in the Hencken burner,

McKenna burner and the WSR exhaust. The results are shown in Figure A.6, Figure A.7,

Figure A.8, and Figure A.9.

102
Figure A.6: Comparison of corrected probe hydrogen flame temperatures

103
Figure A.7: Comparison of corrected probe ethylene flame temperatures

104
Figure A.8: WSR exhaust temperature with probe radiation-conduction correction

105
 Figure A.9: McKenna burner ethylene flame temperature with radiation-
conduction correction

The correction improved the correlation to measured data across all conditions. The

Hencken burner with hydrogen flame was improved from 10-15% error to 0-6% error.

The Hencken burner with ethylene flame was improved from 13-18% error to 1-14%

error. Both of the thermocouple comparisons were slightly improved as well, from within

7% for the WSR to within 4%. The McKenna temperature was improved from within 9%

to within 4%. The correction produced a better correlation with the higher temperature

hydrogen flame. In the high temperature flame of hydrogen, the conduction and

radiations losses are greater than with lower temperature flames.

106
 APPENDIX B

VAPORIZER STRESS ANALYSIS

The stresses experienced by the vaporizer during a detonation event can be

calculated based on a peak loading of 1650 psia. The detonation pressure rise is discussed

in Section 3. The first stress calculated was the hoop stress. Classical thin-wall stress

analysis [15] was assumed for the hoop stress on the vaporizer, shown in Equation B.1

Equation (B.1)

The vaporizer is constructed from 304 stainless steel, which as 500 F has an

approximate yield strength of 24 ksi.[15] Knowing the value of the yield strength at the

elevated temperatures that will be present in the vaporizer, and the maximum

107
hoop stress that would be experienced, a factor of safety of 1.7 was calculated for the

current construction. This analysis shows that the vessel is capable of handling detonation

of JP8 under the worst conditions. Note that a pressure rise of 10.14 corresponds to

operation at ϕ = 1. This is a worst case scenario for the test matrix being presented, and

the majority of testing will occur well into the lean regime, below ϕ = 1, thus decreasing

the potential pressure rise resulting from a detonation.

After it was determined that the vaporizer body would contain a detonation, the

bolt stresses were investigated. If a detonation were to occur, the bolts holding the flanges

on the end of exposed to large tensile loads. The conservative the vaporizer would be

assumption that the bolts carry the entire tensile load was used. Bolt pre-loads typically

serve to reduce the stress experienced in a bolt, and in the actual build up would be

utilized.[15] However, to make the estimate conservative, bolt pre-load was not assumed

for the stress calculation. The bolt stress calculation is shown is Equation B.2

Equation (B.2)

The grade 8 steel bolts used to make the vaporizer have a proof strength of 120

ksi.[15] Bases on the stress calculated Equation B.2, the factor of safety in the bolts is 2.65

. As previously mentioned, in the actual construction, a bolt pre-load will be

utilized. A bolt pre-load will reduce the stress on the bolts. Since this will increase the

factor of safety of the system, an appropriate pre-load was calculated. The relationship

between bolt torque and pre-load is defined Equation B.3[15]

Equation (B.3)

108
 Where T is the bolt torque, d is the bolt diameter, K is the torque constant and Fi is

the bolt pre-load. The bolts used in the construction of the vaporizer are 0.3125 inch

bolts. The stress area of the bolt is 0.0581 square inches for a 0.3125-20 bolt. The stress

area for a bolt is different from the geometric cross-sectional area due to the threads.

Torque constant is mainly a function of the friction, and to a lesser extent the thread

geometry. The value quoted for non-plated black finish bolts without lubricant is

K=0.30[15]

The bolt preload is defined in Equation B.3, and is based on the maximum

vaporizer pressure, flange area and number of bolts.[15]

Equation (B.4)

Equation (Equation B.4 yields a bolt preload of 2625 pounds. This bolt force can be

equated to a nut torque. During instillation, it is possible to measure the bolt torque

readily. Using Equation B.3, the required bolt torque is 20.5 ft-lbf. The final step is to

calculate the maximum allowable torque to ensure that 20.5 ft-lbf is within limits. The

maximum bolt torque is determined assuming that the bolt pre-load is 75% of the proof

load [15], and that grade 8 bolts with a proof strength of 120 ksi are used. Using the above

assumptions, the maximum bolt torque is 41 ft-lbf.

In order to hold the vaporizer together in the event of a detonation, the bolt preload must

be greater than the value calculated for the highest pressure case, yet smaller than the

maximum torque value calculated based on the proof load of the bolt. Since the

109
maximum pressure case yielded 20.5 ft-lbf and the maximum torque based on bolt

material constraints is 41 ft-lbf, the current bolt arrangement is sufficient to hold the

vaporizer together in the event of a detonation.

110