93w '
gk

. wnm
‘ {v I

331$
" v

St.-

‘ngzlfik‘? ‘-
(a.

. N“;

1.

.1. - .-
121 w «3’5

 

V "a?
3. N {Na
1619““. .f
,
x“
‘ ff‘ b
1311‘“-

‘1
‘i'f’iflaifi‘ r

 

'1’ E .v-
“ff-tflfphr. .
r'fiw‘y'zz .
« .5 .1,
w. .A ,
. ~ , n ‘ ‘3; - 4
fl .: A . "‘ a (:3.
l

.nfij ”,1 ‘ ,5
y'rtih‘hl. .01 4 " . '3 '4‘ ., n 5%! ‘ t ngg
.‘ " “ "M ~ . ; k»

7‘ “L 4 .

f . 3&1“ «*5; Jim?
WM 3 ~37 $5“?
1w wa

J.
I‘

, 3v
#3:.

 

   

This is to certify that the
thesis entitled

Examination of a Simulated Micro-Gravity Device For
Evaluating Flame Instability Transitions And Flame
Spread Over Thin Cellulosic Fuels

presented by

Stefanus A. Tanaya

has been accepted towards fulfillment
of the requirements for the

MS. degree in Mechanical EngineerigL

 

 

 

 

‘ Major Professor !
8/26/04
Date

MSU Is an Affinnative Action/Equal Opportunity Institution

 

 

 

LIBRARY |
Min" E5 ‘7 Qtate
University ]

 

 

+4 4 A‘s—«a.- —¢

p

——o. -' ~—* -0 «4—-

v- ‘u‘ v. fl....-——---.—.4

L- A. .‘m .- ..A. ~-..

_ -J- A

A A

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

95
MA

019.7,
Ynyzngz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 czrcTncroammdd-ms

 

EXAMINATION OF A SIMULATED MICRO-GRAVITY DEVICE FOR
EVALUATING FLAME INSTABILITY TRANSITIONS AND FLAME SPREAD
OVER THIN CELLULOSIC FUELS

By

Stefanus A. Tanaya

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

2004

Abstract
EXAMINATION OF A SIMULATED MICRO-GRAVITY DEVICE FOR
EVALUATING FLAME INSTABILITY TRAN SITIONS AND FLAME SPREAD

OVER THIN CELLULOSIC FUELS
By

Stefanus A. Tanaya

The National Aeronautical And Space Administration (NASA) Glenn Research
Center at Lewis Field in Cleveland, Ohio, supports a research program called the
Analysis of Theme-diffusive and Hydrodynamic Instabilities in Near-extinction
Atmospheres (ATHINA), which conducts studies of micro-gravity flame spread. This is
an important research interest for NASA since it is related to NASA space mission safety.
All of the NASA combustion and flame tests are performed in a drop tower, which
creates a micro-gravity environment. However, the drop tower tests are expensive and
there are severe time limitations on the resulting unstable flame from formation, or what
is commonly known as flamelets formation.

To overcome the above limitations on the drop tower, a simulated micro-gravity
device was constructed to perform flame instability transition and flame spread tests over
thin cellulosic fuels. This device, called the Simulated Micro-gravity Flame Tunnel
(SMFT), successfully simulates micro-gravity effects on combustion that are similar to
those in the drop tower. The experimental results demonstrate that the process of flame
change from the stable to the instable form is reversible. Also, for oxidizer flow velocities
near the range that will produce flame instability, the flame spread velocity is highly
influenced by the flame stability. This is not the case when the oxidizer flow velocities

are in the higher values of their range.

I dedicate this thesis to :

My parents, my sister, and my grandparents for their endless support and encouragement

for me in pursuing my goals.

iii

Acknowledgements

I would like to thank Dr. Indrek S. Wichman, my academic advisor, for his
assistance, support, guidance, and encouragement during my course as a Mechanical
Engineering graduate student at Michigan State University. It is truly an honor to work
on such a significant project with The National Aeronautical And Space Administration
(NASA) Glenn Research Center at Lewis Field in Cleveland, Ohio. I would like to thank
the NASA Glenn Research Center for sponsoring and funding this research project under
NASA contract NCC3-662.

I would also like to thank Dr. Fletcher Miller and Dr. Sandra Olson of NASA
Glenn Research Center for sharing their expertise and insight in performing this research
project. Thank you also to the rest of ATHINA team members who taught me many
valuable lessons in engineering research.

Finally, I would also like to thank my fellow graduate students for all their
assistance and expertise they shared with me during this research project. Last but not
least, I would like to thank my family for always encouraging me to get the best

education for myself.

iv

Table of Contents

List of Tables .................................................................................. vii
List of Figures ................................................................................. viii
Nomenclature .................................................................................. xix
Chapter 1 — Introduction and Background ................................................. l
1.1 Introduction ........................................................................ 1
1.2 Literature Review ................................................................ 10
Chapter 2 — Experimental Setup ............................................................ 12
Chapter 3 - Simulated Micro—gravity Flamelet Tunnel (SMFT)
Design and Development ................................................... 21
Chapter 4 — Experimental Procedure ....................................................... 36
4.1 — Pretest ........................................................................... 36
4.2 — Flow System ................................................................... 42
4.3 - Test .............................................................................. 45
4.4 — Post-Test ........................................................................ 48
Chapter 5 — Data Collection Analysis ...................................................... 49
Chapter 6 — Flammability Map .............................................................. 69
Chapter 7 — Hot Wire Probe .................................................................. 73
Chapter 8 — Ramp Down Test ............................................................... 85
8.1 — Raw Data ....................................................................... 88
8.2 — Weighted Average ............................................................ 93
8.3 —- Normalized Average .......................................................... 97
8.4 - Percent Change From Initial Steady to Steady State Flamelet Count 103
Chapter 9 — Ramp Up Test ................................................................... 106
9.1 — Flame Spread Velocity ........................................................ 113
9.2 — Velocity Ratio .................................................................. 120
9.3 - Transition Time ................................................................ 126
9.4 - Thick Fuel Sample Flame Spread Velocity ................................. 134
9.5 — Thick Fuel Sample Velocity Ratio .......................................... 138
9.6 - Velocity Ratio Comparison for All Fuel Samples ......................... 142

Chapter 10 — Miscellaneous Analysis ...................................................... 148

10.1 — Flame Position vs. Time .................................................... 148
10.2 — Flame Spread Velocity vs. Oxidizer F low Velocity ..................... 153
Chapter 11 — Smoke Wire .................................................................... 161
Chapter 12 — Alternative Fuel Test ......................................................... 165
Chapter 13 — Conclusion and Suggestions for Future Work ............................. 168
Appendices ..................................................................................... 172
Appendix A ........................................................................... 173
Appendix B ........................................................................... 176
Appendix C ............................................................................ 179
Appendix D ........................................................................... 182
Appendix E ............................................................................ 187
References ....................................................................................... 192

vi

List of Tables

1. Raleigh number for different gap spacing .......................................... 50
2. Iterative process for the normal position ............................................ 66
3. Iterative process for the inverted position .......................................... 66
4. Thermal resistance model analysis .................................................. 67
5. Hot wire probe measurements comparison ......................................... 77
6. Theoretical volumetric flow rate .................................................... 80
7. Experimental volumetric flow rate .................................................. 81
8. Volumetric flow rate comparison .................................................... 81
9. Boundaries corrected experimental volumetric flow rate ......................... 84
10. Volumetric flow rate comparison ................................................... 84
ll. Velocity Ratio Summary ............................................................ 125
12. Time Constants for 3 millimeter Gap Spacing ................................... 131
13. Time Constants for 4 millimeter Gap Spacing ................................... 132
14. Time Constants for 5 millimeter Gap Spacing ................................... 133
15. Cellulose Fuel Sample Thickness and Area Density ............................. 134
16. Average Flame Spread Velocity for The Ramp Up Tests ...................... 152

vii

List of Figures

Note: Images in this thesis / dissertation are presented in color.

1. NASA S-second “Zero-g” facility
produces ~10'5 g over the duration of the drop .......................................... 3
2. NASA drop rig experimental setup diagram ............................................. 4
3. F larnelets observed during micro gravity combustion process. ...................... 5
4. F lamelet formation in NASA Zero-g facility. .......................................... 6
5. Micro-gravity Flamelet Tunnel (SMFT) ................................................. 7
6. Flamelets traveling along a solid thin fuel (cellulose) .................................. 9
7. A close-up view of flamelets .............................................................. 9
8. The Micro-gravity Flamelet Tunnel (SMFT) — normal (top) and inverted (bottom)
positions ..................................................................................... 12
9. Main sections of the SMFT ............................................................... 13
10. Contraction chamber diagram showing the air inflow into the test section through
the porous plate at the topmost side of the contraction ................................ 14
11. Interface between contraction chamber and test section for large gap spacings
(left) and small gap spacings (right) ..................................................... 15
12. Test section cross-section view showing the lefi-to-right flow across the fuel
sample in the test section .................................................................. 15
13. Test section diagram showing the position of the igniter wire at the downstream

edge of the sample ......................................................................... 16

viii

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Test section photograph view toward the contraction

chamber from the exhaust ............................................................... l7
SMFT flow and flow metering system ................................................. 19
The MSU Flame Rig on which the SMF T was based ................................ 21
The MSU Flame Rig test section ........................................................ 22
The SMFT test section .................................................................... 23
The MSU Flame Rig original design drawn in AutoCAD ........................... 24
Design concept #1, shown in the normal (top) and in the inverted (bottom)
positions ..................................................................................... 25
Design concept #2, shown at normal (top) and inverted (bottom) positions ....... 26
Design concept #3 (the rotisserie), shown at normal (top) and inverted (bottom)
positions ..................................................................................... 27
The stand and shaft mounting plate interface ........................................... 28
The final SMFT design .................................................................... 29
The final SMFT design. The apparatus is shown in the bottom View, both with
(top) and without (bottom) the “rotisserie” shaft ...................................... 30
Side view of SMF T final design in both normal (top) and inverted (bottom)
positions ..................................................................................... 31
SMF T General dimensions. All units are in millimeters ............................. 32
Camera stand ................................................................................ 33
SMF T - Normal position .................................................................. 34
SMFT —- Inverted position ................................................................ 34
SMFT — Inverted position (2) ............................................................ 35

ix

32. The gap spacing adjustment handle (bottom view) .................................... 36

33. The depth gauge ............................................................................ 37
34. The flexible ramp at the upstream location ............................................. 37
35. The sample specification (top view) ..................................................... 38
36. The sample installed on the sample holder using tape ................................. 39
37. The igniter wire ............................................................................. 39
38. The quartz glass installed above the test section ....................................... 40
39. Mounting the camcorder at the top of the camera stand ............................... 41

40. The PC (with LabView), the main power supply, the programmable power supply

(PPS), and the voltmeter .................................................................. 43
41. The LabView interface .................................................................... 44
42. Flat flame front is observed after ignition .............................................. 45
43. Initial flarnelet formation (top) and flamelets fully formed (bottom) ............... 46
44. F lamelet paths of travel afier the LED lights are turned back on .................... 47
45. Flame fi'ont reformation (for ramp up tests) ............................................ 47

46. Velocity profile in F LUENT for the normal position (g downward) at the
sample’s leading edge ..................................................................... 51
47. Velocity profile in F LUENT for the inverted position (g upward) at the test
section outlet ............................................................................... 51
48. Velocity profile in F LUENT for the normal position (g downward) at the
sample’s trailing edge .................................................................... 52
49. Velocity profile in FLUENT for the inverted position (g upward) at the sample’s

trailing edge ............................................................................... 52

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Velocity profile in FLUENT for the normal position (g downward) at the test
section outlet .............................................................................. 53
Velocity profile in FLUENT for the inverted position (g upward) at the test
section outlet ............................................................................... 53
Bitmap analysis for the F LUENT velocity profile results ............................ 55
The SMF T thermal resistance model for the normal position of the test section..56
The SMF T thermal resistance model for the inverted position ...................... 63
F lamelets consuming the sample. This photograph was taken with the room lights
on ............................................................................................. 67
Flamelets consuming the sample. This photograph was taken with the LED
ambient lights on ........................................................................... 68
Flammability map for the Simulated Micro-gravity Flame Tunnel (SMF T) ...... 70

Fully developed Poiseuille flow with the parabolic profile U = Umax (l-yZ/(h’l2)2),

where Umax is the centerline velocity .................................................... 73
The hot wire probe calibrator ............................................................. 76
Photograph of the hot wire probe calibrator ............................................ 76
Velocity profile of test section, 7 millimeter gap spacing ............................ 78
Velocity profile of test section, 5 millimeter gap spacing ............................ 79

Velocity profile of test section, 7 millimeter gap spacing. The second data points
fi'om the boundaries were linearly interpolated between zero (at the wall) and the

third data points to offset the wall effects. For the 5 cm/s case, the third data point

at the lower boundary is interpolated as well ........................................... 82

xi

65.
66.
67.
68.
69.
70.
71.
72.
73.
74.

75.
l 76.
77.

78.

79.

80.

81.

82.

. Velocity profile of test section, 5 millimeter gap spacing. The second data points

from the boundaries were linearly interpolated to offset the wall effects ........... 83
The initial flamelet breakup ............................................................... 86
The formed flamelet count ................................................................ 86
The steady state flamelet .................................................................. 87
Overall initial flamelet breakup vs. deceleration rate ................................. 88
Overall formed flamelet count vs. deceleration rate ................................... 89
Overall steady state flamelet count vs. deceleration rate .............................. 89
Four millimeter gap spacing initial flamelet breakup vs. deceleration rate. . . . . ....90
Four millimeter gap spacing formed flamelet count vs. deceleration rate .......... 90

Four millimeter gap spacing steady state flamelet count vs. deceleration rate. . . ..91

Steady state data bandwidth ............................................................... 92
Weighted average overall initial flamelet breakup vs. deceleration rate ............ 93
Overall initial formed flamelet count vs. deceleration rate ........................... 94
Weighted average overall steady state flamelet count vs. deceleration rate ........ 94
Weighted average 4 millimeter gap spacing initial flamelet breakup vs.

deceleration rate ............................................................................ 95
Weighted average 4 millimeter gap spacing formed flamelet count vs.
deceleration rate ............................................................................ 95

Weighted average 4 millimeter gap spacing steady state flamelet count vs.

deceleration rate ............................................................................ 96
Normalized average overall initial flamelet breakup vs. deceleration rate... ......97
Normalized average overall formed flamelet count vs. deceleration rate .......... 98

xii

83. Normalized average overall steady state flamelet count vs. deceleration rate. . . ..98

84. Normalized average 4 millimeter gap spacing initial flamelet breakup vs.
deceleration rate ............................................................................ 99

85. Normalized average 4 millimeter gap spacing formed flamelet count vs.
deceleration rate ............................................................................ 99

86. Normalized average 4 millimeter gap spacing steady state flamelet count vs.

deceleration rate ........................................................................... 100
87. Normalized average steady state bandwidth .......................................... 101
88. Normalized average initial steady bandwidth ......................................... 102

89. Three millimeter gap spacing percent change from formed flamelet count to
steady state flamelet count ............................................................... 103
90. Four millimeter gap spacing percent change from formed flamelet count to steady
state flamelet count ...................................................................... 104
91. Five millimeter gap spacing percent change from formed flamelet count to steady
state flamelet count ...................................................................... 104
92. Flame Front Reformation in Solid Thin Fuel ........................................ 106
93. F lamelets formed from the original coherent flame front (which consumes the

entire sample). Note the black region to the left before the flamelet “tracks” are

formed .................................................................................... 107
94. Initial flame front reformation ......................................................... 108
95. Flame front reformation in progress .................................................. 108
96. Flat flame front fully reformed ........................................................ 109
97. Post test picture .......................................................................... 109

xiii

98. On-screen tracing ........................................................................ 11 l

99. Flame travel trace on a transparency .................................................. 1 1 1

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

Flame spread velocity graph - 4 millimeter gap spacing, 2.16 cm/s2

acceleration rate ........................................................................ 1 13
Flame spread velocity graph — 4 millimeter gap spacing, 5.05 cm/s2 acceleration
rate ....................................................................................... 114
Flame spread velocity graph - 4 millimeter gap spacing, 7.58 cm/s2 acceleration
rate ....................................................................................... 114
Flame spread velocity graph - 4 millimeter gap spacing, 10.1 crrr/s2 acceleration
rate ....................................................................................... 115
Flame spread velocity graph - 4 millimeter gap spacing, 15.15 cm/s2
acceleration rate ........................................................................ 1 15
Mollified flame spread velocity graph — 4 millimeter gap spacing, 2.16 cm/s2
acceleration rate ........................................................................ 1 17
Mollified flame spread velocity graph — 4 millimeter gap spacing, 5.05 crn/s2
acceleration rate ........................................................................ 1 17
Mollified flame spread velocity graph —- 4 millimeter gap spacing, 7.58 cm/s2
acceleration rate ........................................................................ 1 l8
Mollified flame spread velocity graph — 4 millimeter gap spacing, 10.1 cm/s2
acceleration rate ....................................................................... 118
Mollified flame spread velocity graph — 4 millimeter gap spacing, 15.15 cm/s2

acceleration rate ....................................................................... 1 19

xiv

110. Velocity ratio graph — 4 millimeter gap spacing, 2.16 cm/s2 acceleration

rate ...................................................................................... 120
111. Velocity ratio graph — 4 millimeter gap spacing, 5.05 crn/s2 acceleration

rate ..................................................................................... 121
112. Velocity ratio graph — 4 millimeter gap spacing, 7.58 cm/s2 acceleration

rate ..................................................................................... 121
113. Velocity ratio graph — 4 millimeter gap spacing, 10.1 cm/s2 acceleration

rate ..................................................................................... 122

114. Velocity ratio graph —- 4 millimeter gap spacing, 15.15 cm/s2 acceleration

rate ..................................................................................... 122
115. Velocity ratio summary for all gap spacings ..................................... 124
116. Blue flame front upon flame front reformation .................................. 126
117. Fully reformed flame front ......................................................... 127
118. Time constants for 3 millimeter gap spacing ..................................... 129
119. Time constants for 4 millimeter gap spacing .................................... 129
120. Time constants for 5 millimeter gap spacing .................................... 130

121. Flame spread velocity graph - 4 millimeter gap spacing, 3CHR fiiel sample -
2.16 cm/s2 acceleration rate ........................................................ 135

122. Flame spread velocity graph - 4 millimeter gap spacing, 31ETCHR fuel sample
- 2.16 cm/s2 acceleration rate ..................................................... 135

123. Flame spread velocity graph - 4 millimeter gap spacing, 17 CHR fuel sample -

2.16 crn/s2 acceleration rate ....................................................... 136

XV

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

Flame spread velocity graph - 5 millimeter gap spacing, 3 CHR fuel sample -
1.74 crn/s2 acceleration rate ........................................................... 136
Flame spread velocity graph - 5 millimeter gap spacing, 31ETCHR fuel sample
- 1.74 cm/s2 acceleration rate ......................................................... 137
Flame spread velocity graph - 5 millimeter gap spacing, l7 CHR fuel sample -
1.74 cm/s2 acceleration rate ........................................................... 137
Velocity ratio graph - 4 millimeter gap spacing, 3CHR fuel sample - 2.16 cm/s2
acceleration rate ........................................................................ 138
Velocity ratio graph - 4 millimeter gap spacing, 31ETCHRR fuel sample - 2.16
cm/s2 acceleration rate ................................................................. 139
Velocity ratio graph - 4 millimeter gap spacing, 17CHR fuel sample - 2.16
cm/s2 acceleration rate ................................................................. 139
Velocity ratio graph - 5 millimeter gap spacing, 3CHR fuel sample - 2.16 cm/s2
acceleration rate ......................................................................... 140
Velocity ratio graph - 5 millimeter gap spacing, 31ETCHR fuel sample - 2.16
cm/s2 acceleration rate ................................................................. 140
Velocity ratio graph - 5 millimeter gap spacing, 17CHR fuel sample - 2.16
cm/s2 acceleration rate ................................................................. 141
Flame front in 17CHR fuel

sample .................................................................................... 143
Flamelets in 17CHR fuel sample ..................................................... 143
Velocity ratio comparison for all fuel samples for 4 millimeter gap spacing...144

Velocity ratio comparison for all fuel samples for 5 millimeter gap spacing... 144

xvi

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

Velocity ratio comparison for all fuel samples for 4 millimeter gap spacing —- log
scale ....................................................................................... 145
Velocity ratio comparison for all fuel samples for 5 millimeter gap spacing — log
scale ....................................................................................... 145
Velocity ratio comparison for all fuel samples for 4 millimeter gap spacing — log
scale (exclude 17CHR) ................................................................ 146
Velocity ratio comparison for all fuel samples for 5 millimeter gap spacing — log
scale (exclude 17CHR) ................................................................ 146
Flame position versus time graph — 4 millimeter gap spacing, 2.16 cm/s2
acceleration rate ........................................................................ 149
Flame position versus time graph — 4 millimeter gap spacing, 5.05 cm/s2
acceleration rate ........................................................................ 150
Flame position versus time graph — 4 millimeter gap spacing, 7.58 cm/s2
acceleration rate ........................................................................ 150
Flame position versus time graph — 4 millimeter gap spacing, 10.1 cm/s2
acceleration rate ........................................................................ 151
Flame position versus time graph — 4 millimeter gap spacing, 15.15 cm/s2
acceleration rate ........................................................................ 151
Raw data for the normal position flame spread velocity vs. oxidizer flow
velocity .................................................................................. 154
Raw data for the inverted position flame spread velocity vs. oxidizer flow

velocity .................................................................................. 1 54

Averaged normal position flame spread velocity vs. oxidizer flow velocity... 155

xvii

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

Averaged inverted position flame spread velocity vs. oxidizer flow velocity.. 155
Averaged normal and inverted positions flame spread velocity vs. oxidizer flow
velocity ................................................................................... 157

Averaged normal position flame spread velocity vs. oxidizer flow velocity

where the higher values of oxidizer flow are included ............................ 158
The smoke wire setup .................................................................. 161
The smoke lines when the flow is initiated ......................................... 162
The smoke lines when the flow is in a nearly steady state ........................ 163
The smoke lines when the flow reaches a nearly steady state .................... 163
Flame front after ignition for the SIBAL fuel sample ............................. 166
Flame front consumes the cotton ..................................................... 167
Smolder front consumes the fiberglass after all cotton is consumed ............ 167

xviii

Arabic

A

ar

As
AVG(ti)
Dh

dr

g

h’

hamb
kcoppcr

kglass

hox

mMI-‘(‘

NUD

Nomenclature

Cross-sectional area (m2)

Acceleration rate (m/s2 or cm/sz)

Surface area (m2)

Average velocity from discreet data points

Hydraulic diameter (m)

Deceleration rate (In/s2 or cm/52)

Gravitational constant (9.81 m/sz)

Test section gap spacing (m or cm or m)

Ambient air convective heat transfer coefficient (W/(m2.K)
Copper plate heat sink glass thermal conductivity (W/(m.K))
Quartz glass plate thermal conductivity (W/(m.K))

Oxidizer convective heat transfer coefficient (W/(mZK)
Material thickness (m or cm)

Mass flow rate (kg/s or gr/s)

Mass flow rate of the mass flow controller (kg/s or gr/min or gr/s).
Number of data points
Nusselt number

Pressure (Pa)

Perimeter (m)

xix

Q
qbottom

Chen

R

B

RaL
R-avg
Rbottom
ReD
Ri
Rtop

T
Tamb
Tcoppcr
Tgiass
u

Uavg
Vfinal
Vi
Vinitial

Umax

Volumetric flow rate (m3/s or SCFH)

Heat loss through the bottom part of the test section (W)
Heat loss through the top part of the test section (W)
Velocity ratio

Universal gas constant (8.314 J/(mol.K))

Raleigh number

Averaged velocity ratio

Thermal resistance of the bottom part of the test section (K/W)
Reynolds number

Thermal resistances (K/W)

Thermal resistance of the top part of the test section (K/W)
Temperature (K)

Ambient temperature (K or °C)

Copper plate heat sink temperature (K or °C)

Quartz glass plate temperature (K or °C)

Velocity (m/s or cm/s)

Average velocity (m/s or cm/s)

Final oxidizer flow velocity (m/s or cm/s)

Velocity data point

Initial oxidizer flow velocity (m/s or cm/s)

Maximum velocity (m/s or cm/s)

XX

Voxidizer

Vreq

v(t)
xavg

Xi

xNORM

x(t)

Greek

At

AT

pair

Oxidizer flow velocity

Required velocity supplied by the mass flow controller to the test section

(m/s or cm/s)

Flame spread velocity as a function of time (m/s or cm/s)

Average value of data points
Flame position data point

Data point

Normalized value of data points
Flame position (m or cm)

Vertical position (m or cm or mm)

Thermal diffusivity (mZ/s)

Volume expansion coefficient (K")
Time difference (5)

Temperature difference (K or °C)
Density (kg/m3)

Density of air (kg/s or gr/s)

Kinematic viscosity (mz/s)

xxi

Chapter 1

Introduction and Background

1.1 Introduction

This project is sponsored by the National Aeronautical and Space Administration
(NASA) Glenn Research Center at Lewis Field in Cleveland, Ohio. Within the NASA
Glenn Research Center is a division called the Micro-gravity Science Division (MSD).
The main objective of MSD is to use low-gravity analysis and experimentation to better
understand fundamental low-gravity physical processes that are important to NASA and
its various space mission programs. One of these areas is micro-gravity combustion
science, which addresses the primary issue of spacecraft fire safety. One of the MSD
projects is called ATHINA, an acronym for Analysis of Thennmdiffirsive and
Hydrodynamic Instabilities in Near-extinction Atmospheres. ATHINA’s main objective
is to study flame spread over combustible materials in the near-extinction limit under
Micro-gravity conditions, whether simulated or actual.

Previous work by L.M. Oravecz (Oravecz (2001)) results in the construction of a
facility for examining earth-bound simulated zero-g flame spread, called the MSU Flame
Rig. Although this facility is very simple, micro-gravity combustions over thin solid fuel
sample was successfully simulated and many valuable experimental results were
obtained. Improvements can be made to this facility. The purpose of this research is to
extend and improve the existing apparatus. The current improved facility is called the

Simulated Micro-gravity Flame Tunnel (SMFT) and it is capable of performing tests

under wider range of conditions. Controls over oxidizer flow and test section geometry
are improved in the SMFT. In addition to the physical improvements, several additional
tests were also performed.

There are several ways to create a simulated Micro-gravity environment on earth.
One classical method is to utilize a “drop” facility. ATHINA has used this technique to
create experiments under micro-gravity conditions. In carrying out these experiments, the
experiment setup and all of the necessary instrumentation for the drop is enclosed in a
test chamber. The test chamber is placed into a droppable rig which undergoes free-fall in
a 15 S-meter depth well, producing a 5-second micro Gravity condition inside it. This
facility is called the NASA “Zero-g” facility because the g-level attained is a very good
10'5-g, where l-g =980 cm/sz. Another NASA facility, used at the very beginning of the
ATHINA in the late 1990’s is the 2.2-second drop tower. Here the g-level is close to 103-
g, approximately two orders of magnitude larger than the “Zero-g” facility. Hence, not
only is the drop time less than half, the g-level is higher by nearly a factor of 102.

The drop facility experimental setup contains a test section, a fuel sample (thin
cellulose), and an ignition system. The ignition system consists of an igniter wire that is
laid across the entire downstream width of the sample. Once the ignition is initiated a flat
flame fiont is formed that spreads upstream (against the oxidizer inflow) in the
longitudinal direction of the sample holder. For the ATHINA experiments, the test
section uses an opposed flow oxidizer (in this case air). In other words, the flow direction

of the air is opposite to the direction of flame spread (See Figure 2).

     

TO WIND TUNNEL
EXHAUSTERS

MEZZANINE
ASSEMBLY EAN ROOM
CL VACUUM PUMPS
CONTROL ROOM
EXPERIMENTAL
~. PACKAGE
STEEL VACUUM CHAMBER
CONCRETE-LINED
SHAFT
DECEIERATOR
DECELERATOR CART STOP
ACCELERATOR
ACCUMULATOR
io-Fr-THiCK CONCRETE BASE

~w—n

BOTTOM OF SHAFT 510 FT BELOW GFDUND LEVEL

Figure 1. NASA 5-second “Zero-g” facility produces ~10‘5 g over the

duration of the drop. The tunnel is over 150m deep.

.

   
 
 

 

 

 

 

 

 

Test Section
...,
Sample
Holder
Thin Solid
Fuel Sample

 

 

 

 

I Opposed Flow Oxidizer

 

 

Figure 2. NASA drop rig experimental setup diagram. Note that the flow direction
opposes the flame spread direction. Typical samples are ~15cm in width, ~15 cm in
length, with a thickness measured in microns and a surface density of order 102

mg/cmz.

As soon as a flat flame front is visually observed by on-board camera, the test
chamber in the drop facility is dropped into the well producing a micro-gravity condition
inside the test chamber. The transition to the micro-gravity condition is visually
observable by the color change of the flame (from yellow to blue) Here, the lack of
buoyant motion decreases the flux or supply of oxidizer to the flame, decreasing its
temperature and preventing the formation (and oxidation) of radiating soot (the yellow
regions). Under this weakened combustion condition, the flame front will sometimes
experience a greater susceptibility to instability and therefore a “near extinction limit” is

reached. As a result, the flame will “break up” from its 2-D coherent flame front into

what is commonly referred to as flamelets or a “flamelet front” consisting of numerous
individual 3-D cap-like flame structures. A “flamelet” is thus a form of flame either in its
near extinction limit or just a flame that for one reason or another has become very small.
We may, for the sake of our discussion, consider a hill flame front as the most stable
form of flame spread over a thin fuel sample, a smoldering front as the least stable form,
and a flamelet fi'ont as the next least stable (is. intermediate) form. The least stable form
of burning is of course when the flame is either totally extinguished or will momentarily
be extinguished. The ATHINA group at Michigan State University (MSU) studies this
phenomenon of flamelet formation, propagation (spread), and extinction, in order to
determine certain flame behaviors when subjected to a micro-gravity or simulated micro-

gravity condition.

 

Figure 3. Flamelets observed during micro gravity combustion process. Note
first that the flamelets are three dimensional, that the flame front appears most
vigorous at the sides (where it is brightest), and that the flamelet cuts a “track” into

the sample (the burned out black regions).

There are two major shortcomings of the NASA drop tower experiments. The first
is the cost Each experiment must be performed full-staffed by a large group of engineers
and technicians because of the extensive pre-test preparations and post-test recovery
procedures (the test chamber must be winched back up from the bottom of the 155-meter
well). As a result, usually no more than one test (at most two) can be performed each day.
The second shortcoming of the drop tower is the severe limitation on the micro-gravity
time. The longest time that the experimental setup inside the test chamber endures the
micro-gravity condition is 5.2 seconds. This time period is sufficient only to observe the
initial flamelet formation behavior. The test chamber drops to the bottom of the well and
flamelets are extinguished before any significant flame spread can be observed. In fact,

over the 5.2 seconds of drop time the flamelets spread at most a few millimeters.

 

t=.63sec

 
   

  
 

t=l43sec

   

Figure 4. Flamelet formation in NASA Zero-g facility. The flamelets spread
only a few millimeters (at most) over the course of the drop, but their morphology

undergoes considerable evolution, as seen.

Under a grant for fimded research from the NASA Glenn Research Center on
ATHINA, the Department of Mechanical Engineering at the Michigan State University
undertook to design, fabricate, and test a device that will simulate a micro-gravity
combustion for lengthy time intervals up to 103 times longer than those permitted by the
NASA drop facilities. In the initial research, an experimental apparatus was the designed
and fabricated. This facility was later improved by adding a capability to run a test in an
inverted position (upside-down). This device is called the Simulated Micro-gravity

Flamelet Tunnel or abbreviated to SMFT.

 

Figure 5. Micro-gravity Flamelet Tunnel (SMFT).

The SMFT overcomes some of the shortcomings of the NASA drop facility since
each experiments is staffed by one person and many experiments can be performed each
day. More importantly, there are virtually no time limitations in which the experiment
will undergo a simulated micro-gravity condition. The only limit to the experiment time
is the length of the fuel sample. It is important to clarify that the SMFT simulates micro-
gravity by suppressing the buoyant flow (by examining the flame-spread process in a
narrow wind tunnel) and by manually decreasing the oxidizer flow velocity to a range in
which flame fionts experience flame front to flamelet transitional instabilities. Flamelets
are formed as shown in Figure 6, where their photograph (still image) was taken
moments after the flame front underwent transition to ten individual flamelets. The rate in
which the oxidizer velocity is decreased is referred as the “Flow Deceleration Rate”. The
deceleration rate is a measurement of negative acceleration (deceleration) according by

Equation (1):

Vial—V“,
d = mm fi 1
. At ()

 

Oxidizer Flow

Flamelet Travel

Figure 6. Flamelets traveling along a solid thin fuel (cellulose). The dark

 

region, which shows fuel sample completely consumed, experienced only flame
spread. The line of flamelet breakup can clearly be seen, followed by flamelet

spread, flamelet bifurcation and extinction and finally sample burnout.

 

Figure 7. A close-up view of flamelets.

1.2 Literature Review

Several works on the flame spread analysis was done in the past by various
researchers. Most of this project however, do not deal with low speed, low g research in
the instability regime. Some of the works that involve low speed, low g conditions were
reviewed as the basis of this project.

This project is based on the apparatus and research work by L. Oravecz (Oravecz
(2001)). The work done by Oravecz was mainly the development and calculations of a
narrow-gap cell or what is often referred as ‘Hele-Shaw’ apparatus. The theory of
buoyancy suppression is the based of Oravecz work in constructing the MSU Flame Rig.

The works by I. Wichman and F. Williams (W ichman (1983)) deals with a
simplified model of flame spread in an opposed flow oxidizer. The purpose of this work
was to analyze a flame spread model in where the heat release occurs at the interface
between two different media. Each media was described as moving at different but
constant velocities. As a result, a flame spread formula was created for an opposed flow
oxidizer condition.

Another related research was done by O. Zik (Zik (1997)), which investigate the
fingering instability in combustion. The spacings between the flame fingerings was said
to be determined by the Peclet number. Zik’s work has a very similar setup to the MSU
apparatus, where a thin fuel sample is subjected to an opposed flow oxidizer. The
buoyancy effect was suppressed as well.

8. Olson (Olson (1991)) performed the study of micro-gravity flame spread over a
thin solid fuel. The experimental setup was also an opposed flow oxidizer. It was studied

that at the near-quenching regime (low oxidizer flow rate), the oxidizer transport-limited

10

chemical reaction is the governing mechanism for the flame spread. At the high oxidizer
flow rate, residence time limitation is imposed on the flame spread.

A similar work was done by A. Femandez-Pello (F emandez-Pello (1981)).
Although the main subject of Fernandez-Pello’s work was to investigate the ambient
oxygen concentration, some of the experimental results regarding the flame spread
velocity were used as a reference for this research. Femandez-Pello’s study did not
involve a low oxidizer flow case (i.e. the instability regime). For air (which is the
oxidizer at the MSU apparatus), the flame spread velocity for oxidizer flow higher than
10 cm/s is a constant value. For a very high oxidizer flow (above 100 cm/s), the flame

spread velocity starts to decrease as the oxidizer flow is increased.

11

Chapter 2

Experimental Setup

All experiments for this project will be carried out in the above-mentioned SMFT.

Figure 8 shows CAD drawings of the facility.

 

Figure 8. The Micro-gravity Flamelet Tunnel (SMFT) — normal (top) and inverted

(bottom) positions.

As shown, the SMFT has the capability of running test both in normal (right side up) or

inverted (upside down) positions. This capability was developed by mounting the facility

to an aluminum shaft (or ‘spine”) that has full-rotational degree-of-fieedom mountings at
both ends.
There are three main sections of the SMFT: The plenum chamber, the contraction

chamber, and the test section (See Figure 9).

 

 

’/
é
[I
é!
.4
fit
4:
fl
fl;
fr
x5.
fr
4?
fit
/
/
i
I;

 

Figure 9. Main sections of the SMFT.

The SMFT is a pressure-driven flow system. Pressurized air is injected at the inlet while
the outlet remains at the atmospheric pressure condition (p = l atrn). In sequence, the air
flows from the plenum to the contraction chamber, and finally will reach the test section
before exiting from the outlet (or exhaust).

The oxidizer (air) is injected via a standard l/4-inch hose fitting that is connected
to a hollow aluminum cylinder mounted laterally inside the plenum. Referring to Figure
9, the cylinder is extruded to the out-of-the-page direction through a circular tube
adjacent to the duct with the ‘INLET’ label. The cylinder has holes in its surface that
disperse air inside the plenum chamber. Most of the inlet air flow is directed toward the
rear wall in order to eliminate directional momentum and to produce a more uniform pre-

contraction flow.

13

From the plenum, air is directed to the contraction chamber. The main purpose of
the contraction is to regulate the flow or velocity profile once air exits the contraction
chamber and enters the test section. The contraction chamber consists mainly of a ramp
and a porous plate. The ramp directs the air to a jet slit opening on the topmost side of the
contraction chamber outlet. In fi'ont (to the lefi) of this opening, a porous plate is placed

to straighten out the airflow. Figure 10 illustrates the arrangement.

Sink Edda——

 

 

 

 

    
 
  
   
      
 

   

\\\\

\\\\\\\\\\\\\\

. s\\\\

 

     

Seal
(DrIwnAsSolirlHldl)

Paous Plate
(Drawn As Dotted Hatch)

Figure 10. Contraction chamber diagram showing the air inflow into the test

section through the porous plate at the topmost side of the contraction.

The flow exiting the contraction chamber enters the test section. It is important to
note that the sample holder is height-adjustable and is sealed in such way so that all of the
airflow from the contraction chamber (Figure 10) will be directed above it. While a test is
in progress, the top of the test section is covered by a sheet of quartz glass. The vertical

distance between the quartz glass and the sample holder is the gap spacing. At the

14

minimum gap spacing, the seals will prevent air from going under the sample holder and
at the maximum gap spacing the top surface on the sample holder is aligned with the

bottom of the porous plate. Figure 11 shows the above-mentioned setup.

Porous Plate

From :fi

Plenum

     
      
 

 

 

 

 

M V
Figure 11. Interface between contraction chamber and test section for large gap

spacings (left) and small gap spacings (right).

 

 

 

 

 

 

——Quartz Glass e—Fuel sample
—Sample Holder —Copper Heat Sink
wfi -o -o -o -t> -l> -t> -(> -o A -(> -(> 94>
4> =9 w w'fl

 

 

Aluminum Base

Figure 12. Test section cross-section view showing the left-to-right flow across the
fuel sample in the test section. Note that the distance from sample to copper heat

sink is identical to the quartz-to-sample spacing, herein called the gap spacing.

The fuel sample itself is suspended above a copper heat sink or substrate. The
vertical distance between the fuel sample and the copper heat sink is equal to the above-
mentioned gap spacing between sample and quartz glass. The space between the quartz
glass and the sample holder forms the wind tunnel. If the gap spacing is small enough, the
buoyancy motion will be suppressed" so that the experiment inside this tunnel will
simulate a micro-gravity condition.

Similar to the system at NASA drop facility, an igniter wire is employed to ignite
the fuel sample on its downstream side. The current passing through the wire is regulated

by a power supply. Figures 13 and 14 shows the test section in iso-plane views.

Fuel sample
—Sample Holder /I: I—Copper Heat Sink

 

 

 

 

 

——lgniter Wire

Figure 13. Test section diagram showing the position of the igniter wire at the

downstream edge of the sample. In this figure the flow is Ieft-to-right.

" This is the SMFT hypothesis.

 

Figure 14. Test section photograph view toward the contraction chamber from the

exhaust.

Initially, the flow velocity entering the test section of the SMFT was determined
using the relations between the volumetric flow rate (Q) and the average velocity (Um)
of air entering the test section. Here Q is represented in SCFH and is monitored by a
series of rotameters. The relationship between Q and Uavg was created as a table. A valve
placed between the pressure regulator and the rotameters then adjusts the velocity

manually. To relate Q and Um , the following relations were used:

m: '. (3)
1):; - (IdealGaslaw), (4)
rh=pUmA, (5)

l7

 

OJ. (6)
p
._pUmgA
Q— p/y’ (7)
Um=——}gp . (8)
_pA

where p is pressure, 3 is the universal gas constant, T is the temperature, and A is the
inlet jet area to the test section. From Equation (7), it is clear that if the pressure of the
system is changed, as by introducing new parts/accessories to the system, then the
relation between Q and Uavg is changed. It was later determined that the method of
ramping down the flow velocity to obtain instabilities of the flame by manual adjustment
of the valve is not a very accurate way to conduct a test. One of the biggest problems is
the variation in the deceleration rates from test-to—test or from one person to another since
there will be no two experimentalists that turn down the valve at exactly the same rate.
To overcome this problem, a Mass Flow Controller (MFC) model MKS - M1008 was

purchased in place of the valve. Figure 15 illustrates the system.

Replacing the valve by the MFC means that a new table for Q and Uavg relation is
newed. Instead of using the Q to determine velocity, the relationship between the Mass
Flow Rate (:51 ) and Ualvg will be used (Equation 5). Based on the calibration curve issued
by the manufacturer, the relationship between the input voltage and the Q on the MFC is

linear, where 0 Volts produced 0 liter/min and 5 Volts produced 30 liter/min.

18

PROGRArhAABLE
l l POWER SUPPLY

VOLT- /

METER VOLTREADINGS " " *

SHOULD 'MATCH' I
\x l

 

 

 

 

 

 

 

BLDGAIR

7 SUPPLY
PRESSURE PRESSURE MASS FLOW
REGUIATOR m CONTROLLER
WI GAUGE VALVE

Figure 15. SMFT flow and flow metering system.

Based on Equations (4) and (6), Equation (7) becomes:

(7)

Mass flow rate of the MF C will be determined based on the Q of the flow exiting the
atmospheric pressure and temperature. A sample calculation for a gap spacing of 5 mm
with U..,s of 20 cm/sec is shown below:

pm, = 1.23kg/ m3 or 1.23gr/ liter ,

MFC Max Flow = 301iter/min " 1.23 gr/liter

= 36.9 gr/min

= 0.615 gr/sec,
n'rMFC = 0.615gr/sec,
ti! = pUmA , (5)
Uavg = 0.2 m/s,
A = (Gap Spacing) x (Sample Holder Width)
= (0.005 m) x (0.33 m) = 0.00165 m2,

"'2 = (1.23kg/ m3 ).(O.2m/ s).(0.00165m2)

= 0.0004059 kg/s

= 0.4059 gr/sec.
MFC: 5 Volts for 0.615 gr/sec (Max Flow)
Since the calibration curve is linear, then the voltage required to produce 20 crn/s

at the 5mm gap spacing (or 0.4059 gr/sec) is determined by:

V,” = (5V)[—,"’—-I= 5V.(0.4059/0.615) = 3.3 mus

mm

In order to produce an average velocity of 20 cm/s on the test with 5 mm gap spacing, a
3.3 V input voltage to the MF C is required. Based on this calculation, a spreadsheet was
created. The MFC is controlled by a programmable power supply that is connected to a
personal computer (PC). The software LabView 6.1 is used to control the programmable
power supply.

Also, since the mass flow controller can only handle 40 psi of inlet pressure, the
test with the largest gap spacing is only capable of maximum velocities less than 40 cm/s

(Unlike the ‘old’ system which was capable of velocities over 40 cm/s).

20

Chapter 3

Simulated Micro-gravity Flamelet Tunnel (SMFT) Design and Development

The Simulated Micro-gravity F lamelet Tunnel (SMFT) was developed from what
originally was a device called the MSU Flame Rig, refer to Oravecz (2001). The MSU
Flame Rig and the SMFT are very similar systems. Both have the same plenum and
contraction chambers. The test sections, although identical in principal, are structured

differently.

 

Figure 16. The MSU Flame Rig on which the SMFT was based. Note that this
apparatus does not have the capability to run experiments in the inverted (upside

down) position.

21

Since the MSU Flame Rig does not have the capability to run experiments in the
inverted (upside down) position, the sample holder is simply supported by a series of
shims. Adding or reducing the number of shims adjusts the gap spacing between the
sample and the quartz glass. The copper heat sink of the MSU Flame Rig was bolted to
the slotted hole on the sample holder. To adjust the gap spacing between the sample and
heat sink, the heat sink is traversed along the above-mentioned slotted holes (refer to

Figure 17).

    
   
  

Fuel Sample

Sample Holder

 

Gap Spacing \
§‘\

    

 

 

_ \msmm
‘1‘ " Aluminum
Block
Ezfi @— _
, ~ ~ -~ Shims
\ Aluminum
Base

 

 

 

Figure 17. The MSU Flame Rig test section. Note the use of shims to adjust

the gap spacing. The air flow is from left to right.

When designing the SMFT, one needs to consider a new method of securing the
sample holder and the heat sink since the SMFT has the capability to run the experiments
in the inverted position. Instead of using shims, four threaded holes were made in the

aluminum base. Carriage bolts were inserted to those threaded holes from the bottom of

22

the device. The other end of those carriage bolts were drilled and tapped to accept four
smaller shoulder bolts that support the sample holder (refer to Figure 18). A similar
system is also employed for the heat sink so that is can be adjusted independently of the

sample holder. In order to seal gaps between components, rubber seals were used.

 

Fuel Sample

 

 

 

 

/___4 Sample Holder

 

 

 

 

 

Gap Spacingr\
Quartz Glass

 

 

\

 

 

Shoulder i\ \ ....

 

 

 

 

 

BOIIS I. 1“ .. \H:-:
.1 .
I E: E: EI $3.2“

 

\

//

 

is: .2; m...

Base

 

 

 

 

Carriage Bolt

 

 

 

Figure 18. The SMFT test section. Note the use the bolts and threaded holes
to adjust gap spacing. The bottom end (upper end in picture above) of the carriage
bolts were drilled and tapped to accept the threads of the shoulder bolts. The
aluminum platforms are drilled so that the shoulder part of the bolt fits. The upper
part of the aluminum platforms are held by the bolt heads, the bottom part is held
by the diameter of the carriage bolts’ thread. By turning the carriage bolts, the
aluminum platforms (which are attached to the sample holder and heat sink) move

up and down. The air flow is from left to right.

23

The next step was to determine a method that would enable the SMFT to be
rotated 180°. Several designs were proposed and evolved into the current SMF T design.
The first step was to transfer the original design to AutoCAD in 1:1 scale. These results
are shown in Figure 19. From this AutoCAD file, several design concepts were

performed.

 

Figure 19. The MSU Flame Rig original design drawn in AutoCAD.

24

The first concept was to build a stand that could hold the apparatus in an inverted
position with all the non-welded legs removed. When running experiments in the normal
position, the legs were to hold the apparatus. To invert the apparatus, two persons would
manually lift the apparatus and plac it in the stand upside down. This method is
economical and easy, but due to the heavy weight of the apparatus (~80 kilograms), is not

safe.

 

Figure 20. Design concept #1, shown in the normal (top) and in the inverted

(bottom) positions.

25

The next concept was to attach a circular-cross-section shaft on one side of the

apparatus, mmmt both of its ends to allow full

A " ‘ Anni-An n? 1‘ J
“vbnvv v.

This shaft
will perform similar to a door hinge (Figure 21). A stand would be fabricated to “catch”

the apparatus when it is inverted. This arrangement is effective but will take much of the
needed lab space, therefore this concept were not approved.

 

Figure 21. Design concept #2, shown at normal (top) and inverted (bottom)

positions.

26

The next concept was based on concept #2, but instead of attaching the shaft at
one side of the apparatus, the shaft would be attached below the apparatus while

.I

mounting its ends to allow the full ‘ " ' degree of A All legs were removed
and a new stand will support the apparatus (which is attached to the shaft) in both the
normal and inverted positions. This system is referred as the “rotisserie” system, as

illustrated by Figure 22. This design is effective, compact, and safe.

 

Figure 22. Design concept #3 (the rotisserie), shown at normal (top) and

inverted (bottom) positions.

27

The final design of the SMFT is based on concept #3 with several improvements
included. The stand is built using unistruts, which are slotted steel U-channels. This
material is low-priced and the slotted holes on the channels (along with the supplied
accessories) ease the process of assembly. The first step alter the stand was built was to
make it height-adjustable. L-brackets were used to connect the stand to the shaft
mounting plates. These L-brackets were bolted both to the slotted holes at the stand and
to the fixed holes at the mounting plates. To adjust the height, the bolts holding the L-
brackets can be taversed along the slotted holes or moved to the next holes. Figure 23

illustrates this system.

Shafi mounting Plate is
height adjustable

   
   
  

Locking Bolt

 

 

 

 

 

Stand
(unistrut)

 

 

 

Figure 23. The stand and shaft mounting plate interface.

28

In addition to the above, wheels were added to make the apparatus moveable
around the lab. Also, a locking bolt was installed to secure the shafi position at the

mounting plates. Figures 24 to 27 show the final design in AutoCAD.

 

Figure 24. The final SMFT design. The SMFT is shown in the normal

position (top) and with the flow tunnel removed (bottom).

29

 

Figure 25. The final SMFT design. The apparatus is shown in the bottom

view, both with (top) and without (bottom) the “rotisserie” shaft.

 

 

 

 

Figure 26. Side view of SMFT final design in both normal (top) and inverted

(bottom) positions.

31

 

 

. i .3 31...]
General: Front View

0.6625 ...j 130
30.44-30.o~j —~—82.4

Side view

-. 66.040 a1
170508

I
l
,,__r,

15.000

88.900

Stand: Front view

Figure 27. SMFT General dimensions. All units are in millimeters

93.900

l

 

r“:_________ ___- __ _

Front view

 

.".;_"_-"_r\;_'. '-- .
. ‘ n

  

-|—-

Side view

*i rfilm 7

Rear view

..——~.__. 170.180 *H n.~_.1

 

32

Side view

{"1 _

 

 

Tests are recorded using 3 SONY TRV-9OO Digital Camcorder. To mount this
camera to the apparatus, a camera stand was built on the top of the test section, as shown

by Figure 28. This entire assembly rotates with the SMFT.

 

Figure 28. Camera stand.

Figures 29 to 31 show photographs of the SMFT after completion of the design

changes.

33

 

Figure 30. SMFT - Inverted position.

34

 

Figure 31. SMFT — Inverted position. Note the minimal clearance at the bottom of

the photograph between the camera stand and the structural braces.

35

Chapter 4

Experimental Procedure

The experimental procedure of the Simulated Micro-gravity F lamelet Tunnel

(SMF T) will be explained in this section. Refer to Figure 15 for the flow system diagram.

4.1. Pretest

1. Adjust gap spacing by adjusting the handle (mounted on the carriage bolt’s heads)
for the sample holder and the heat sink (Figure 32). Measure distance using the

depth gauge (Figure 33).

 

Figure 32. The gap spacing adjustment handle (bottom view). Both the sample

holder and heat sink have four adjustments for each corner.

36

 

Figure 33. The depth gauge. The digital caliper measures the gap between the
bottom of the quartz glass to the sample holder’s surface. The same procedure is

used for the heat sink.

2. Install a ramp fi'om a flexible material (such as tape) that connects the sample

holder edge to the heat sink edge at the upstream location (Figure 34).

 

Figure 34. The flexible ramp at the upstream location.

37

3. Cut the fuel sample (250 x 220 millimeters) to the specifications shown in Figure

35.

 

 

 

 

 

 

 

 

 

 

Igniter Wire
250
303
l
i Sample
i ,
v ~ ““
7 1. .. Mr 7 ‘;:\- HeatSink
\\ Mica
l-‘——— 220—.-

 

 

 

Figure 35. The sample specification (top view). Note that the sample is attached to

the mica surrounding the heat sink. All dimensions are in millimeter.

4. Install sample above the heat sink, secure using tape (preferably duct tape due to
its strength). Make sure the sample is level and does not have any wrinkles in it

(Figure 36).

38

 

Figure 36. The sample installed on the sample holder using tape.

5. Install the igniter wire. Make sure igniter wire is straight and laid flat above the
sample (Figure 37). Use care when installing the igniter wire since this item is

very fragile.

 

Figure 37. The igniter wire. Note that the igniter wire needs to be placed flat above

the sample, or a flat flame front will not be observed once the test is initiated.

39

6. Cover the test section with the quartz glass. Slide the quartz glass into place and
press firmly (horizontally) against the seal above the slit-jet opening of the inlet
flow channel. Carefully lower the clamps to secure the glass (Figure 38). Do not

use excessive force when handling the quartz glass and make sure that the glass is

centered and rests above the rubber seal (Figure 38).

 

Figure 38. The quartz glass installed above the test section. Note the use of clamps

(right) and the position of glass relative to the seal.

7. Level the test section using the level-gauge.
8. Install digital camcorder on the stand (Figure 39). Set the cassette on the desired

position. Close the drapes and turn on the LED ambient lights.

40

 

Figure 39. Mounting the camcorder at the top of the camera stand.

41

4.2. Flow System

Check all connections of hoses and fittings. Ensure that there are no leaks.

fl
0

2. Turn on the shop air.

3. Turn on the main power supply (make sure the igniter wire switch is ofi), the
Programmable Power Supply (PPS), the volt meter, and the video monitor (Figure
40). Press the ‘Output On/Ofi’ button on the PPS until the output is indicated as
‘On’.

4. Using the supplied table, set the proper parameters on the PPS using the LabView

interface as shown in Figure 41 .

42

 

Voltmeter

Figure 40. The PC (with LabView), the main power supply, the programmable
power supply (PPS), and the voltmeter. This system controls the mass flow

controller (MFC).

43

 

Figure 41. The LabView interface. The ‘Initial Voltage’ value will set the voltage

supplied to the PPS at the beginning of the test. The ‘Intermediatc Voltage’ will set
the next voltage, and the ‘Final Voltage’ will set the final voltage value. For a
regular test, where ramping up after flamelets are formed is not necessary, the
‘Intermediate Voltage’ and ‘Final Voltage’ can be set at the same value. The ‘Initial
t’ is the time between the initiation (Run) and the first ramp down. The ‘Ramp
Down t’ is the desired time for ramp down. The ‘Inter Hold t’ is the time over which

the intermediate voltage is held, and the ‘Ramp Up t’ is the ramp up time.

5. To initiate the flow, press ‘RUN’ and immediately ‘STOP’. This will set the flow

at the initial value.

4.3. Test
1. Press the ‘RECORD’ button on the camcorder.

2. Turn on the igniter wire and wait until a flat flame front is observed (Figure 42).

 

Figure 42. Flat flame front is observed after ignition.

3. Press ‘RUN’ at the LabView interface and turn the LED’s off at the same time.
This will mark the beginning of the program run for post-test analysis, 3 seconds
before ramp down starts.

4. Once flamelets are formed (Figure 43), turn the LED’s on so that the flame travel

pattern can be observed (Figure 44).

45

 

Figure 43. Initial flamelet formation (top) and flamelets fully formed (bottom).

46

 

Figure 44. Flamelet paths of travel after the LED lights are turned back on.

5. For mug-up tests, the flow is automatically ramped back up to observe the flame

front reformation (Figure 45).

 

Figure 45. Flame front reformation (for ramp up tests).

6. Once the flame travels to the end of sample and extinguishes, press the ‘STOP’

button at the camcorder.

47

4.4. Post-test
1. Carefully remove the quartz glass from the test section and set it aside.
2. Remove the unburned sample from the test section.
3. Clean all residue and debris.

4. Download the test file into the PC using Adobe Premiere.

48

Chapter 5

Data Collection Analysis

As explained above, the Simulated Micro-gravity Flame Tunnel (SMFT)
hypothesis states that if a combustion process over a solid thin fuel is conducted in a
narrow gap tunnel, the buoyancy forces will be significantly suppressed so that there will
be no effect on the experimental results. The SMFT’s capability to run experiments in
both normal (right side up) and inverted (upside down) positions is based on this
hypothesis.

If buoyancy has no effect on the experiment results (i.e., it is suppressed by the
SMFT), experiments conducted in the normal and inverted positions should yield
identical results. It is important to note that in reality the experimental results between the
normal and inverted positions are not identical. There are two significant observed
differences between the normal and inverted position test results. The first is the flamelet
size. For identical tests, inverted tests produce smaller flamelet diameters. The work on
flamelet diameters is currently in progress at MSU, so all flamelet size comparisons in
this thesis were made qualitatively. The second difference is the degree of flamelet
oscillation. For identical tests, flamelets in inverted tests tend to oscillate more than in
normal tests. As with the flamelet size, the work of quantifying the oscillation is in
progress.

It is necessary to check if buoyancy effect can really be neglected. Oravecz (2001)
calculated the Raleigh ntunber to determine the effect of buoyancy inside the test section.

Buoyant motion develops when the Raleigh number exceeds 1708. Table 1 shows the

49

Raleigh number values versus test section height, assuming the temperature difference

between the flame and the ambient is 700K.

Table 1. Raleigh number for different gap spacing

 

Table 1 shows that even at the gap spacing exceeding the normal maximum value used
for experiments (7 millimeter), the Raleigh number is below 1708. According to this
calculation, it is safe to assume that buoyancy inside the SMFT test section can be
neglected.

To support this simple calculation, two FLUENT models were constructed. They
are both identical except that for the normal position, gravity is ——9.81 m/s2 (downward)
and for the inverted position, it is +9.81 rn/s2 (upward). The fuel sample temperature is
1200K and the ambient and inflow gas temperature is 300K. The sample length is
reduced to 1/3 of its original length to simulate a partially consumed sample undergoing
combustion. Figures 46 to 51 show the results. Note that for Poiseuille flow, the
maximum velocity (Umax) is 1.5 times the average velocity (U avg). The F LUENT model
was constructed with the average flow velocity of 6 cm/s, and as Figures 46 to 51 show,

the maximum velocity is reasonably close to 9 cm/s (9 cm/s = 1.5 x 6 cm/s).

50

 

 

Figure 46. Velocity profile in FLUENT for the normal position (g downward) at the

sample’s leading edge.

 

Figure 47. Velocity profile in FLUENT for the inverted position (g upward) at the

test section outlet.

51

 

 

Figure 48. Velocity profile in FLUENT for the normal position (g downward) at the

sample’s trailing edge.

 

 

Figure 49. Velocity profile in FLUENT for the inverted position (g upward) at the

sample’s trailing edge.

52

 

Figure 50. Velocity profile in FLUENT for the normal position (g downward) at the

test section outlet.

 

Figure 51. Velocity profile in FLUENT for the inverted position (g upward) at the

test section outlet.

53

Note that the scaling is almost identical between the normal and inverted positions
and although that there is no significantly large difference in velocity profile between the
two positions, on the fluent computations, the velocity profile for the normal position -on
the quartz glass side- shows a slightly larger area of higher velocity (4.89 x 10'2 — 5.43 x
10'2 m/s) than that of the inverted position (especially noticeable in Figures 46 and 47).
This comparison is shown in Figure 52, which is a zoomed view of FLUENT velocity
profile results for both positions. To compare the cross-sectional area of the 4.89 x 10'2 —
5 .43 x 10'2 m/s velocity range, a bitmap analysis was performed. The images from fluent
are converted into bitmaps. There are 88 pixels for the 4 millimeter gap, which means 1
pixel represents ~ 0.0455 millimeters (4 millimeter/ 88 pixels = 0.0455 millimeter/
pixel). The cross-sectional area difference between the normal and inverted positions is
represented by 8 pixels (the normal position has more area), which means in the real
apparatus it is ~ 0.364 millimeter (8 pixels x 0.0455 millimeter / pixel = 0.364
millimeter), ~ 9 percent of the gap spacing (100 % x (9 millimeter / 4 millimeter) = 9.125
%).

This suggests that there is a difference in the convection rate on the quartz glass
side between the normal and inverted positions. The normal position seems to have a
higher convection rate that will certainly affect the flame behavior. A higher convection
rate means there is more oxygen supplied to the flamelets. The lower convection rate in
the inverted position case will weaken the flamelets. Since they are less stable, they will

have a greater tendency to bifurcate (split), as observed in the experimental results.

54

88 Pixels for
4mm gap

 

Figure 52. Bitmap analysis for the FLUENT velocity profile results. Note that the

normal position has a larger cross-sectional area of the higher velocities range.

Note that the FLUENT analysis simply models the oxidizer (air) flow inside the
test section to observe the buoyancy of the oxidizer, not the flame itself, which is another
gas phase not modeled here due to its complexity. Although the oxidizer flow is not
influenced by buoyancy, the flame itself might. Ifthe flame is affected by buoyancy, then
it will have a tendency to flow upward and actually ‘touch’ the upper boundary of the
tunnel. At the normal position, the upper boundary is the glass. At the inverted position,
the upper boundary is copper. Since copper is more conductive than glass, the flame will
release more heat to the copper. Since this is the case for inverted position, at that
position more heat is dissipated from the flame and therefore weakening it. This explains
the smaller flamelet sizes, the higher degree of flamelet oscillation, and a greater
tendency to bifurcate for the inverted position.

A heat transfer analysis was performed to determine the difference in heat losses
from the flame to the atmosphere in the normal or inverted positions to observe whether a

larger scale heat transfer than the convection rate difference previously explained is

55

causing the difference in flame behavior between the two SMFT’s positions. For this, a

thermal resistance model is constructed. Figure 53 shows this model for the normal

position.

 

Tamb

 

 

 

 

Quartz Glass

 

 

WWW R2 Rtep

 

 

 

 

 

Fuel Sample

 

 

 

 

 

 

 

 

 

Copper Plate Rbouom
.. -2... -..‘:...-.e.._v....- .7 ' . ...-.n';..... ...... .1 ”run-..- -... ....-. :':.',..._'e_ ._ <_ .._'
maxi—149%; *2. 3‘ 42:11; 'Ii-"i-‘i'u HF.- :*~ir :-:-: a“: s; ”aces-'1"; arm-t "15.-s: ...,

 

 

 

 

 

 

 

 

Tamb “-

 

 

 

Figure S3. The SMFT thermal resistance model for the normal position of the test

section.

The objective of this analysis is to calculate how much heat is transferred from the flame
to the ambient air across the top (qtop) and the bottom (qbomm) of the channel. An iterative
method will be used to solve this heat transfer problem. Referring to Figure 53, the

following resistances were determined:

 

R1 = h (9)

56

 

 

 

122-.- (10)
k8,“.A
1
R = 11
3 had, ( )
l
R = 12
. hm“, ( )
L
R5: (13)
kW,.A
1
R = 14
. hWA ( )

Here, R1, R3, R4, and R6 are determined from a convective heat transfer model, while R2
and R5 are determined from a conductive heat transfer model. Also, Rmp is R] + R2 + R3
and Rhona“, is R4 + R5 + R6. The L’s are the material thicknesses, which are 0.00635
meters (1/4 inch) and 0.00157 meters (0.062 inch) for the quart glass and copper plate,
respectively. The thermal conductivity, k, is 1.4 W/(m.K) for the quartz glass, 391.1
W/(m.K) for the copper plate (Alloy 110), and 2.56 x 10-2 W/m.K for ambient air. The
kinematic viscosity, 1), for air is 1.62 x 10'6 m2/s. The thermal diffusivity, a, for air is
2.29 x 10'6 m2/s. The volume expansion coefficient, [3, for air is 3.3x10'3 K". The unit
area is set to 1 m2 (1m x 1m sections). The gravitational constant, g, is 9.81 m/sz. For the

normal position:

57

Determmmg RI:

_ gfi(Tamb - Tglass )L3
00

Ra 1.

 

(15)

Here, L is the characteristic length, defined by the surface area per perimeter as:

-115
P

L (16)

Thus,

1»:2

L:
(l+1+1+1)m

 

=1/4m=0.25m,

and

(9.81m/s’)(.0033K")(T,,,, —Tg,m)0.25m3
— (lts.2t10"’m2 /s)(22.9x10'5m2 ls)

 

0L

To calculate (T a.,,i,-ng,,s.,.), an iterative process was used (see the next section for

explanation).

NuL = 0.151%,“ (10’<Rm<lo' ‘) - For the upper surface ofthe heated plate. (17)

_ NuLk

 

(18)

R. = (9)

58

R2 =———— (10)

0.00635»:

= = 0.0045K/W
(1.4W/(m.K))(1m2)

 

2

Determining R3 and R4:
R3 and R4 are identical. Since this is an internal flow problem, the first criterion that

needs to be determined was the flow turbulence. To do this, a Reynolds number
calculation was performed. Assuming the velocity, u, is 0.06 m/s (6 cm/s) and the gap

spacing is 0.004 meter (4 millimeter):

 

ReD = “D" (19)
v
where Dh is the hydraulic diameter, defined as
4A
Dh = T (20)

where A is the cross sectional area and P is the cross sectional perimeter. Thus,

_ 4(0.36mx0.004m)

h — = 0.00791»: .
2(0.36m + 0.0041»)

 

therefore,

59

_ (0.06m/s)(0.0079lm)

=29.034,
D (16.2x10'6m2 Is)

 

sothat

Rep = 29.034 << 2300, or Rep / RerURBULENT = 29.034 / 2300 = 0.013 << 1.

This means the flow inside the test section is definitely laminar. Oravecz (2001)
determined that the flow above the fuel sample is fully developed, therefore Table 8.1 in
Incropera (1996) can be used to determine the Nusselt number. First, the base / height
ratio must be determined. In this case, the base is 0.36 meters wide and the height (gap
spacing) is 0.004 meters. The means the base / height ratio is 90. The largest finite
number given in the table for the ratio is 8, so 90, for engineering purposes, can be
regarded as infinity. For the ratio infinity, the Nusselt numbers for the uniform heat flux
and uniform temperature cases are 8.23 and 7.54, respectively. The average of the two

values yields 7.885, which is the Nusselt number value used in this calculation.

 

 

ho, = M“ (21)
Db
... = (7.885)(0.0265W/m.K) = 25.539", «”810.
(0.00791m)
R _ 1 (11)
3 h .A

l

R, = 2 2 =0.0392K/W.
(15.539W/m .K)(lm )

 

60

Since
R3 = R4 ,
we also have

R, = 0.0392K/W.

Determining R5:

L
k .A

W

R5:

 

_ 0.001651m
’ (391.1W/(m.K))(1m2)

 

= 4.027x10‘6K/W

Determining RQ:

__ gfl(Tamb - Tgias: )L3
— Dd

 

Ra L

where

Here (as before)

L- 1m
(1+1+1+1)m

 

=1/4m =0.25m,

so that

61

(13)

(15)

(16)

a = (9.81m/s2x.0033K-‘)(T,,,, -TW,)0.25m3
" (16.2x10’5 m2 /s)(22.9x10‘5m2 ls)

 

To figure the (Tam-Twp”), an iterative process was used (see the next section for
explanation).

Nu L = 0.27RaLm (105<RaL<10'0) - For the lower surface of the heated plate. (22)

 

 

Nuk
h— " 18
L ( )

1
R = . 14
‘ hwy/1 ( )

For the iterative process, Rm and Rheum. were calculated separately, as well as qmp
and qmm. For Rmp, hr is the only unknown heat transfer coefficient, so a guessed value
was used to start the iteration. From the guessed hr value, R. was determined. Rio, is R1 +

R2 + R3. From this Rmp, the qmp value was calculated. Then (Tamb-Tglass) is determined by:

AT = q.R (23)

From (Tmb-Tglm), the new Raleigh number was calculated - this is the beginning of the
next iteration step. From here, the new Nusselt number and h; were determined. From the
new h. value, new R. and q; values were calculated. From the new qr value, (T ,mb-Tgiass)

was again determined using Equation (23). This process was repeated until there were no

significant differences in the q values, with the difference less than 0(10‘4). Rhona“ is

62

calculated using the same procedure, with the h5 value as initial guess. Table 2 shows the
iterative process.

For the inverted position, the process was generally the same. The values of R.
and R5 will be different between the normal and inverted positions due to the orientation
of the hot surface (the quartz glass and the copper). Figure 54 shows the thermal

resistance model for the inverted position.

Tamb

 

 

 

Copper Plate

 

 

 

 

 

 

 

 

Fuel Sample

 

 

 

 

 

 

Qw‘zmass \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ R2 mm“

 

 

 

 

 

 

 

 

Tamb L

 

 

 

Figure 54. The SMFT thermal resistance model for the inverted position.

Note that for this position, Rm, is R4 + R5 + R5 and Rbonom is R. + R2 + R3. For the

inverted position:

63

Recalculating R1:

= gfl(Tamb — Tglass )L3

RaL (15)
va

 

L is the characteristic length, defined by the surface area per perimeter as:

P (16)

 

L = =1/4m=0.25m
(1+l+1+1)m

(9.81m/s2 )(.0033K" )(Tm - Tw)0.25m3
a =
L (16.2x10” m2 /s)(22.9x10'5m2 Is)

To figure the (Tam-Tease), an iterative process was used.

NuL = 0.271%,“ (105<RaL<10'°) For the lower surface of heated plate (22)

 

 

Nuk
h— L 18
L ( )

1
R = 9

Recalculating 3,5:

= gm... - T... )L’

RaL (15)
00

 

P (16)

 

L: =1/4m=0.25m
(l+1+1+1)m

= (9.81m/s2)(.0033K")(Tm —rwp,,,,)o.25m3
(16.2x10'5m2 /s)(22.9ir10‘5m2 Is)

a].

To figure the (Tam-Team), an iterative process was used.

NuL = 0.15RaL”3 (107<Ra1,<10I ‘) For the upper surface ofheated plate (17)

 

Nu k
h= L 18
L ( )

l
R = 14
6 hamb’A ( )

Using the recalculated R; and R5, the same iterative process as previously explained was

used to determine qbomm, as shown in Table 3.

65

Table 2. Iterative process for the normal msition.

 

Table 3. Iterative process for the inverted msition.

 

66

Table 4. Thermal resistance model analysis.

NORMAL INVERTED

   

Difference in heat loss: Inverted is 1.17 % greater.

This analysis shows that there is no significant difference in heat loss between the normal
and inverted positions (< 2%). This means that the difference observed in flamelet
behavior during the inverted test is likely not caused by the difference in heat loss.

Figure 55 and 56 show flamelets consume the fuel sample. They travel in the
longitudinal direction of the sample. Note the brighter ‘spots’ along the sides of the
flamelets. These are the areas where the maximum temperature of the flamelet is

observed.

 

Figure 55. Flamelets consuming the sample. This photograph was taken with the

room lights on.

67

 

Figure 56. Flamelets consuming the sample. This photograph was taken with the
LED ambient lights on. Note that one of the flamelets was just bifurcated into two

individual flamelets.

68

Chapter 6

Flammability Map

All of the experimental data were collected and stored as Audio Video Interleave
(avi) files on the lab PC. The experimental data were analyzed to produce the
flammability map. The purpose of constructing this map was to quantitatively summarize
the micro-gravity combustion experiment results since each test condition has a unique
result. This map is also useful for future researches because expectation for each test
condition has been determined.

The results were categorized into different regimes. These regimes are:
“Extinguished”, “F lamelet”, and “Flat Front”. “Extinguished” means that all flames were
quenched after the oxidizer flow was turned down. “Flamelet” means that turning down
the oxidizer flow resulted in the formation of flamelets. “F lat Front” means the test
condition where a flat flame front was formed after the oxidizer flow is turned down.
This means the reduced oxidizer flow rate didn’t produce the instability required for the
flamelet front to form.

Some of the test results were in-between these categories. One case is a test where
blue flame cusps were formed. These cusps are apparently a transition phase between the
flat flame front and flamelets. Another case is a test where flamelets were formed for a
brief moment and then extinguished. This test will be categorized as in-between flamelet
and extinguishment.

As explained in the previous section, there were observed (not calculated)

discrepancies between the normal and inverted SMFT tests. The flamelet has a different

69

behavior and a difference physical appearance. The flammability maps between the
normal and inverted positions are also different. These maps are plotted in Figure 57.
Note also that the 1 / gap spacing scale on the y-axis is a total gap spacing, that is, the

total gap between the copper plate and the quartz glass, or twice the normally-discussed

gap Spacing-

 

 

SMFT Flammability Map
for Total Gap Spacings (Fuel-to-Glass + Fuel-to-Heat-Sink)

 

 

 

 

I NORMAL Lower
Limit Flow (cmls)

A NORMAL Upper
Limit Flow (cm/s)

e INVERTED Lower
Limit Flow (cmls)

o INVERTED Upper
Limit Flow (cm/s)

 

 

 

1/ Gap Spacings (mm'l)

 

 

 

 

0 5 10 15

Flow (cm/s)

Figure 57. Flammability map for the Simulated Micro-gravity Flame Tunnel

 

 

 

(SMFT). The shaded areas between the curves are the “Flamelet” regions. Note that
the inverted position tests have a narrower flamelet band. The area to the left of the
flamelet region (lower limits) is the ‘Extinguished” region, and the area to the right

of the flamelet region (upper limits) is the ‘Flat Front” region.

70

Trend curves were fitted to the raw data. For the normal position, the second order
polynomial curve equation for the lower limit is y = 0.0017x2 + 0.0083x + 0.0859. For
the upper limit, due to the shape of the raw data which has multiple x values for a single y
value, or x(y) instead of y(x), a second order polynomial curve was fitted in the x-
direction as x=-901.54y2 + 340.77y - 17.163. For the inverted position, the second order
polynomial curve equation for the lower limit is y = 0.0022x2 - 0.0048x + 0.0898. For the
upper limit, a second order polynomial curve was also used giving y = 0.0021x2 -
0.0166x + 0.1075.

For the normal position, as the l / gap spacing increases (i.e., the gap spacing
decreases), the lower limit, or the minimum oxidizer flow rate required to produce
flamelets increases. This is expected since a smaller gap spacing naturally causes the
flame to consume a higher percentage of the available oxidizer, therefore as a counter, a
higher forced oxidizer flow is needed. Also for the normal position, it is interesting to
observe that as the gap spacing decreases, the upper limit, or the maximum oxidizer flow
allowed before flamelets turn back into a flat flame front, increases until 5.5 millimeter (1
/ gap spacing = 0.18 m") then decreases until the minimum gap spacing of4 millimeter
(1 / gap spacing = 0.25 mm"). Recall that these are total gap spacings, so 5.5 and 4
millimeters corresponding to normally-discussed gap spacings of 2.75 and 2 millimeters,
respectively. In reality, these are smaller gaps than normally tested. A theory was
developed that the minimal distance between the flame and the physical boundaries of the
test section might interfere with the flame behavior. At gap spacings 3 millimeters and
above, there is enough clearance between the fuel sample and the quartz glass and copper

plate so that the flame thickness is not affected. This theory is further shown by the fact

71

that during test with gap spacing less than 3 millimeters, it is difficult to start the flat
flame front using the igniter wire.

For the inverted position, both the lower limit and upper limit behaved as
expected, without the anomaly observed in the normal position upper limit. It is possible
if the gap spacing was decreased further more that this anomaly will be observed in the
inverted position tests. There are physical limitations (such as igniter wire clearance,
paper flexibility) that prevent nmning the SMFT with gap spacings smaller than 2

millimeters.

72

Chapter 7

Hot Wire Probe

The hot wire probe measurement investigated the velocity profiles inside the test
section, and was used to observe any significant discrepancies with the theoretical
profiles.

For analysis, the flow in the test section will be treated as a Poiseuille flow, which

is a fully developed flow through a straight channel of constant cross sections. Figure 58

 

 

 

 

 

illustrates the Poiseuille flow.
y=h' / 2
Y
x 33
_ _ __ _ ._ ...... *7 _ _ _ F0
J
y= -h' / 2

Figure 58. Fully developed Poiseuille flow with the parabolic profile U = Um (I-

y2/(h’l2)z), where U.“ is the centerline velocity.

This assumption holds valid especially in the region where the cross section is constant
(downstream of the porous inlet (Figure 11) and upstream of the fuel sample). The
hotwire probe measurement that will be used for comparison was made upstream of the
fuel sample. A sheet of Plexiglas with a cavity to insert the probe stem replaced the
quartz glass. The probe was located 5 centimeters upstream of the fuel sample leading

edge (and 19 centimeters downstream fi'orn the porous plate inlet of Figure l 1). The

73

result of the Poiseuille Flow calculation will be used to construct the theoretical velocity
profile equation. Equation 14.49 of Schlichting (1960) states that a parabolic vertical

velocity distribution has the form

2

_ _ y
140’) - Umfl (h'/2)2) (24)

 

where h’/2 is the channel half-width (see Figure 51). Equation (24) will be used to

construct the theoretical velocity profile. To determine the theoretical average velocity,

 

 

U =—‘-Wf(1- y w dy=hlf(1— ’2 )U 1’— (25)
“’8 h'_,,,,, (h'/2)2 ...... , (h'/2)2 “I172

Next, the relationship between the theoretical average velocity and the theoretical

maximum (or centerline) velocity will be determined. To simplify, let:

 

y _
(h'l2) — 5' (26)
Therefore,
Um =Umax j(1-;2)d;=Umax[g-‘5—33] =Umax(l-§), (27)

74

avg

U =§Umax. (28)

To determine the experimental velocity profile the hot wire probe was calibrated
using a nozzle designed in FLUENT to give an exit-section plug flow. Connecting the
nozzle to a mass flow controller results in a known value of volumetric flow rate. Using

Equations (3) to (5), and (25) below, the exit velocity from the nozzle outlet can be

calculated as follows:
:5: = pr}, (3)
p = 1%- - (Ideal Gas law) , (4)
ti: = pUmA , (5)
ti:
U = __ 29
..., pA ( )

The probe was placed at the center of the nozzle outlet. The exit velocity of the nozzle
was plotted versus the probe voltage reading, and a calibration curve was created. The
calibration process took place before each gap spacing measurement was taken. Figure 59
shows the AutoCAD design of the nozzle and its flow alignment system, which was built

from straws and metal screens.

75

 

Figure 59. The hot wire probe calibrator. Note the presence of straws and screens to
align the flow in the axial direction. The flow from the mass flow controller is

supplied from the bottom of cylinder.

 

Figure 60. Photograph of the hot wire probe calibrator.

76

The measurements were taken at the 7 millimeters and 5 millimeters gap
spacings. The data show that there are discrepancies between the theoretical and

experimental velocity profiles, as shown in Table 5.

Table 5. Hot wirgirobe measurements compafl'son

 

fl

 

 

 

 

 

 

 

Centerllne
Gap Spacings ©3323: Velocity:

(mm) ( cm ls) Measured I
Theoretical

7 22.0 0.94

7 12.0 1.01

7 5.0 1.12

5 30.0 0.97

5 15.0 1.08

5 5.0 1.19

 

 

 

 

This is mainly caused by the imperfections of the test section shape, such as the presence
of the ramp downstream of the probe. Also, there might be some discrepancies between
the dialed-in and output values of the mass flow rate fiom the mass flow controller.

Figures 61 and 62 show the velocity profiles, both theoretical and experimental.
Note that in the 5 millimeter case, average velocity 30 cm/s, the experimental centerline
velocity is not located at the centerline, instead it is shifted upward approximately 0.2
millimeters. This may be caused by operator error when initializing the probe travel
measurement from the base of the test section (i.e., the sample holder). Note also that the
top and bottom boundaries of the test section, the experimental velocity is not zero. At
the top boundary, the probe was not measuring velocity near a wall; instead it is
measuring velocity at the center of a cylindrical tube base (i.e., the probe stem). This
means the no-slip assumption does not hold valid for this measurement and a zero

velocity is not expected. At the bottom boundary, which is the sample holder’s surface,

77

the probe detects a voltage change higher than it should since the aluminum surface of
the sample holder acts as a heat sink to the probe. Therefore when the probe is located

close to the surface, a radiation effect might take place.

‘ 7mm Gap Spacing Velocity Profile

Velocity (cm/s)
40.00

    

 

 

l
I
‘ l
l 0.00 10.00 20.00 30.00 "
l 0.0 y... __T__--_ --.,» «—
. l l I
[ l I ‘ l
r 1.0 9—9— i
r l
l J
l 2.0 r, l — 7-, n ,_ (l
A f—o—zz cm/s H
. E . i:
g E 3.0 ‘H_ a? _ _ l—l—12cmls 3|
l 0 ‘ 1+5 cm/s 1‘
g E 40 l : . 5 g ------ Theoretical 22 cm/s l
! g ' ‘ I 7: 7 W h . 1,-"im ’7‘ l ------ Theoretical 12 cm/s I)
l ..1’ L ------ Theoretical 5 cm/s j
i 50 I ' P ’ ’ "i M" j
E l l
l l 2 l
, l 4
‘ 7.0 1.:- - . 7 fl , - “if. ,. I

Figure 61. Velocity profile of test section, 7 millimeter gap spacing. Note the
discrepancy between the theoretical and experimental profiles at the top and bottom

of the duct.

78

 

 

 

5mm Gap Spacing Velocity Profile

 

 

 

 

 

 

Velocity (cmls)
0.00 10.00 20.00 30.00 40.00
0.0 f A
. . . 'l '
0.5 ,L
i‘.
1.0 - . l —
i A- _
A1'5‘ l“— " “; +30crfls 7
E 2.0 1 %~—~ —~—i4»-- +15cmls
3' + 5 cmls
o 2.5 .
g vvvvvvv Theoretical 30 cmls
g 3.0 vvvvvvv Theoretical 15 We

:59.“
moan

 

 

 

 

------ Theoretical 5 cmls

 

 

 

 

 

 

9"
0

ti ____-______‘_ _ —_____v

 

 

 

Figure 62. Velocity profile of test section, 5 millimeter gap spacing.

Table 5 shows that the maximum discrepancy observed between the theoretical

and experimental centerline velocity is 12%. For the purpose of this research, this is a

sufficiently small discrepancy and therefore the measured velocity profile can be used in

the experiments.

More further, a volumetric flow rate per unit width produced from the theoretical

formula and the experimental measurements can be compared. Integrating the velocity

profile with respect to area produces the volumetric flow rate: in other words, the

volumetric flow rate is the area under the velocity profile curve.

79

h

Q= lama. (30)
-h

Since the volumetric rate sought is per unit width, dA is dy, hence
I:

Q= Iu(y)dy . (31)
4.

where

2

_ _ y
u(y)—Umax(l (h’/2)2

 

) . (24)

Therefore,

 

 

 

h'l2 h/2 3 h'l2
_ _ _ _ ...1. y
Q_-»!/2Uma- (1172) ”My Um J20 (if/2) “My UMP 3(h'/ 2) La

(32)

For both the 7 millimeter and 5 millimeter cases, Table 6 shows the unit conversion to SI

as well as the theoretical volumetric flow rate (m3/s).

T_able 6. Theoretical volumetric flow rz_1_t§_._

 

 

 

 

 

 

 

 

 

 

 

 

. Avera e Centerline Volumetric Flow
Gap Spacrng (mm) Velocity %m/s) Velocity (mls) “’2 (m) Rate (male)
7 0.22 0.33 0.0035 0.00154
7 0.12 0.18 0.0035 0.00084
7 0.05 0.08 0.0035 0.00035
5 0.30 0.45 0.0025 0.00150
5 0.15 0.23 0.0025 0.00075
5 0.05 0.08 0.0025 0.00025

 

 

For the measured volumetric flow rate, Microsoft Excel computes the area under the

curve. Table 7 shows the results.

80

Table 7. Experimental volumetric flow ra_te_.

 

 

 

 

 

 

 

 

 

 

 

 

. Aver r ' '
Gap Spams wwwmfii’n‘ils) V3321: 1'23.) m2 <m> V°é”.’?§2£l$s£'§’w
7 0.208 0.312 0.0035 0.00165
7 0.121 0.182 0.0035 0.00100
7 0.056 0.084 0.0035 0.00050
5 0.290 0.436 0.0025 0.00175
5 0.162 0.243 0.0025 0.00098
5 0.060 0.090 0.0025 0.00038

 

 

Table 8 shows the volumetric flow rate comparison between the theoretical and analytical

results.

Table 8. Volu_rnetric flow gate comparison.

 

 

 

 

 

 

 

 

 

 

 

 

Desired Theoretical Experimental
Gap Spacing (mm) Average Volumetric Flow Volumetric Flow % Difference
Velocity (cm/s) Rate (ma/s) Rate (m3/s)
7 22 0.00154 0.00164 7.05
7 12 0.00084 0.00100 19.19
7 5 0.00035 0.00049 42.70
5 30 0.00150 0.00174 16.45
5 15 0.00075 0.00097 30.44
5 5 0.00025 0.00037 51.40

 

 

The large discrepancies (up to 51.4%) are mainly due to the fact that at the boundaries,
the probe didn’t measure the true velocity (See Figures 54 and 55). To overcome this
problem, correction curves were made for the 7 millimeter and 5 millimeter cases by
forcing the boundary velocities to be zero, and linearly interpolating the next data points

fi'om the boundaries, as shown in Figures 63 and 64.

81

9° .N
o o
l

A

Dlstance (mm)
'o

9‘
o

 

0.00
0.0 I -_ .1" —1'*r 7., f i ‘7,

1.0 -

n
”7' *#r

 

Corrected 7mm Gap Spacing Velocity Profile

10.00

 

Velocity (cm/s)

20.00

 

30.00

 

 

 

 

 

 

40.00

 

17+" “ 52’ Em; ”T T]
i + 12 cm/s .
+ 5 cm/s }
, ------ Theoretical 22 cm/s I
1 ------ Theoretical 12 cm/s '

J: - ;-_T,he9rst£a|§ cmfi J

 

Figure 63. Velocity profile of test section, 7 millimeter gap spacing. The second data

points from the boundaries were linearly interpolated between zero (at the wall) and

the third data points to offset the wall effects. For the 5 cm/s case, the third data

point at the lower boundary is interpolated as well.

82

 

Corrected 5mm Gap Spacing Velocity Profile

[7A- ......_7. . --. V “..., .7 A , 2- -.z. ..- m
E
E

 

 

 

 

Velocity (cm/s) 1
i 0.00 10.00 20.00 30.00 40.00 ‘
j 0.0 I r 7t E j
: ‘7 " . » E E
1 0.5 7 ~ ‘7 7 _; 7 7 7 7- 7 E
“The—'24,- - r
E 1.5 77 E 77 7k} 7 77E7 7777.7 7 77-777 7 77- 7 77.
1 ... l T. 1 'En : —e—— 30 cm/s jl
E E 2.0 7 7 7 l 7 71737 77E—7 E i—a—15cm/s E}
E v ' E '- E, E—*— E'
E 8 2-5 r *r 1‘ m *E ' *1“ "i—P‘ { 5cm/s. LE
3 g E J ; E E I! g ------ Theoretical 30 cm/SEԤ
g 3.0 __ - f- i if —P _ '- ~ ------- Theoretical15cm/s}1
E 3.5 7 - -1 z ,4. _ E m . A- E1:‘1;:TE??°.'?L'¢§'..5..°_"1’§-J1
E 40 7 + ,5 ; , z r‘ l
E E"‘ E 1 l
4.5 7 77 , 7 77, 77 77
E l L {E E
E 5.0 E E l

L,, __,. W!” n m .. 7.. ... . ,fiv . .. .#_ . ,7, .7 .. ,7 in
Figure 64. Velocity profile of test section, 5 millimeter gap spacing. The second data

points from the boundaries were linearly interpolated to offset the wall effects.

With these corrections, the volumetric flow rates were recalculated and the results
are shown in Table 9. Table 10 compares the theoretical volumetric flow rates and the
boundary corrected value. With the corrected curves, the largest discrepancy decreases
from 51.4 % to 30.19 %. The smallest discrepancy decreases from 7.05 % to 0.73 %. It is
believed that the corrected curves are closer to reality than the non-corrected ones due to
the wall effects explained above, therefore the flow system with the mass flow controller
can be used for this experiment with a reasonable amount of confidence. Note that the

discrepancies between theoretical and measured flow rates increases as the flow velocity

83

decreases, suggesting that hot-wire measurements at low flow velocities are less accurate

than at higher flow velocities.

Table 9. Boundaries corrected experimen_tal volumetric flow ra__t_e_._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Average Centerline Volumetric Flow

Gap Spacnng (mm) Velocity (mls) Velocity (m/s) hi2 (m) Rate (ms/s)

7 0.208 0.312 0.0035 0.00153

7 0.121 0.182 0.0035 0.00092

7 0.056 0.084 0.0035 0.00043

5 0.29 0.436 0.0025 0.00155

5 0.162 0.243 0.0025 0.00086

5 0.06 0.09 0.0025 0.00033
mic 10. Volu_rnetric flow rate comparison.
(3 S 'n mm) Desired Average Vo-lruhrfigffigclgiiw Efsérrienfiteerga' °/ Difference

ap pacr g( Velocity (cm/s) Rate (m3/s) Volumetric Flow °
Rate (ma/s)

7 22 0.00154 0.00153 0.73

7 12 0.00084 0.00092 9.31

7 5 0.00035 0.00043 23.63

5 30 0.00150 0.00155 3.28

5 15 0.00075 0.00086 14.57

5 5 0.00025 0.00033 30.19

 

 

 

 

 

 

84

 

 

Chapter 8

Ramp Down Test

The objective of this section is to investigate the influence of the deceleration rate
on flamelet formation. As previously explained, besides locating the experiments in a
narrow buoyancy-force-suppressing tunnel, the SMFT simulates micro-gravity gravity
also by decreasing the oxidizer velocity to a range which a flame front will experience
instability. This simulates what happens to the oxidizer flow in the NASA “Zero—g”
facility. The rate in which the oxidizer velocity is decreased is referred as the
“deceleration rate”. The deceleration rate is then a measurement of (negative)

acceleration according to Equation (29).

d __ Knitter! — V/inal
, - T

(29)
The experimental procedure was explained in the “Experimental Procedure”
section above. It is important to observe the flamelet counts as they are formed. Figure 65
shows the initial flamelet breakup, Figure 66 shows the formed flamelet count, and
Figure 67 shows the steady state flamelets traveling on the fuel sample. Due to the
variable nature of this experiment, a certain level of subjectivity is involved when judging

the flamelet counts.

85

 

Figure 65. The initial flamelet breakup. Note that in this stage right after the
oxidizer ramp down is completed, some f'lamelets travel in the lateral (sideways)

direction. They may split or join with the other flamelets.

 

Figure 66. The formed flamelet count. All flamelets travel in the longitudinal
direction (rightward in this photograph). All splitting and merging occur along their

travel path.

86

 

Figure 67. The steady state flamelet. This type of image normally occurs about ’/4 of
the flamelet travel, where the flamelet count does not significantly change.

Subjectivity is involved when determining the steady state count.

Most of the results shown in this section are the ones fi'om the 4 millimeter gap
spacing, except where comparisons between all gap spacing results are needed. All other
results are included in the Appendix. The results of this section are divided into four sub-
sections. Each section contains graphs from the overall data (mixed between 3 millimeter,
4 millimeter, and 5 millimeter gap spacings) and individual gap spacing data. 4211
extingaished test and data anomalies will not be accounted tor the result. 6 millimeter

and 7 millimeter gap spacing cases were excluded since the flamelets produced in those

87

tests tend to form toward a more stable state, such as flame cusps that are very close to a
full flame fiont. Note that the plot legends labeled ‘New’ are not different types of tests,
instead this notation denotes different tests that were done under the same conditions at a

later time.

8.1.Raw Data

The following figures present the raw data. The test data were not processed or

altered in any way. The data are displayed ‘as-is’.

 

.-. -__._._._ . W_._.1

 

 

 

 

 

 

 

 

 

 

Initial Flamemet Breakup vs. Deceleration Rate
16

o. 14 1» s
a A —-—-—- —~~————»,u
g 12 77 7 7 a 77 7 77777.77 7 a 77 77 7 77 77 7777 0 4mm ‘*

A A
g 10 ”T“ “T— . I‘- * ..._ _ _._.. H - 5mm
in! A I I
g 8 77 —7————— .7‘7 7— e7 777— 777— 77 7 ‘Smm
E I I A 3mm new
3 6 7—aa a7 7n 777 7 777977 77
ll. 0 A A I O . 4mm new
3 4 777— 7-7- a 7477-77 797 I —— 4 77 —— 7 .5mmnew
,4; e I e _
E 27 7 7 70— 7 -I~- 797 7n 7 7 77 77 777

0 ,7 I
0 5 10 15 20
Deceleration (chsZ)
Wham“- ~__ 2 _ w ___ 5 mm‘_ --.nF

 

 

igure 68. Overall initial flamelet breakup vs. deceleration rate.

88

 

 

 

 

 

#75 A H7"?

 

 

 

 

 

 

 

04mm
I5mm
A3mm
a 3mm new
0 4mm new

 

 

 

 

Formed Flamemet Count vs. Deceleration Rate
16
g 14 ‘ e
0 12 a fi
*6 1o 7 i . i
3 A I I A
E 8 +7 7—7—77 7—77-9 —+— 7 7 e
O I D
E 6 7777—a-7t7—7—n—7—7—47I a -I 7
In C A I I O O
O 4 ~ --——————-7I»—-I 7‘77- «I 7—7 707-77 I777- 770-77— 7—7—77
E 2 at...____ _..__...__2 ...-._:_-_.___- _-______,____._ -_.. ..__
O l T l
0 5 1O 1 5
Deceleration (cmls2)

20

 

I 5mm new

 

.___._.._2 -J

 

Figure 69. Dverall formed flamelet count vs. deceleration rate.

 

16
14
12

SS Flamelet Count

ores-ores

 

L .-.”--.

Steady-State Flamelet Count vs. Deceleration

 

10*

 

 

 

 

 

 

 

 

 

 

 

 

 

-.- .. _____ -___.. a, --. __- ,_ g __ _
I A I
7E- t~-~O-----~— a e e
O A O A O
— ~7——-—--a m-H—7—en- e
A O O O I I
7...— ..... 2. _. _ ”n . _ _ -
I I I I A
I I
0 5 10 15
Deceleration (cmls2)

A 3mm

a 3mm new
0 4mm new
I 5mm new

 

 

 

 

 

 

 

 

Figure 70. Overall steady state flamelet count vs. deceleration rate.

89

 

 

 

 

 

 

 

 

 

 

 

 

Initial Flamemet Breakup vs. Deceleration Rate T
l
16 g
a. 14 77 77 7 7 7— 77 — 7777—7
:3
4.: 12 7 7
2
m1o7E77 7— 7— 7 77777777777
15’ [0 4mm 1
a a 77 7 7 7.77777 7+7 7 7 77777777 7 .
E 0 4mm new i
2 6 — o 7 797 77777 7 7 7 797 777 7 _m__-_.
E 4 e e
a 77.7 7e— 7—77 7477777- 77
a . .
- 2 777 7 797 77777777
0 Y . 7
0 5 10 15 20
Deceleration(cml32)

 

 

 

 

 

Figure 71. Four millimeter gap spacing initial flamelet breakup vs. deceleration

 

 

 

 

 

 

 

 

 

 

 

 

rate.
Formed Flamemet Count vs. Deceleration Rate
16
14.777 777-7 777— 7777 77 77 777- 7
‘5
o 12
0
g 10 77 7 7 777 7 77 — 7— — — 777 777 777—
3 0 4mm
E 8 ..-. _. 77 _ ...77_7 _ #7 . .____7§ .--.-_..---_---..- - 5..- _ «7.. 7-.....4
.2 e 0 4mm new
“. 6 4... § _ __+-__ Q ..-—.2_ _- .. A- __ ._.__._ -.--.-_.__.
8 . . .
E 4 77 77 77777 o 777+ — —7 7 7o 777 7777
O O |
IL 2 -.-- .. ._ _._ . ___ ._ __ .‘_______ --.--. .... ;
. . I 7 I
0 5 10 15 20
Deceleration (chsZ)

 

 

Figure 72. Four millimeter gap spacing formed flamelet count vs. deceleration rate.

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F"””’””"”“"’ ...__,_.....,..__.2_
Steady-State Flamelet Count vs. Deceleration
16
14
a O 0
g 12 9 9777—7777. 777 77
o e e e
o 10 ~ 77707—70774—7- *~ ...... ...__
3 7 7 . E 4mm 1
a 8 777 777—--———7- ~ .:
E 0 4mm new i
.2 6
II.
a 4 7777 7 -
2
o T i l
0 5 10 15 20
E Deceleration (cmls2)

 

 

Figure 73. Four millimeter gap spacing steady state flamelet count vs. deceleration

rate.

As observed in Figures 68 to 73, the deceleration rate does not seem to have a
particular influence on the initial flamelet breakup, formed flamelet count, and the steady
state flamelet count. While the initial flamelet count and the initial steady flamelet count
seem to be scattered, the overall steady state flamelet count is handed between 6 and 14.
For the 3 millimeter gap spacing, the bandwidth is between 7 and 14. For the 4 millimeter
gap spacing, the bandwidth is between 7 and 13 (with exception of extinguished case,
where it goes to zero). For the 5 millimeter gap spacing, the bandwidth is between 6 and
10. Figure 74 illustrates this phenomenon. This is an indication that an overall steady

state phenomenon can be observed in the experiments regardless of the deceleration rates.

91

This also confirms the experimental trend observed in the lab, that the 3 millimeter gap
spacing test results are the most varied / random ones (shown by largest bandwidth). To
better understand the data collected, several different ways of presentation will be carried

out in the next sub-sections.

Steady State Raw Data Bandwidth

'16
14 e 7

12 7777-77\.\
10 7 \777 +Upper Limit

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 __ » ___ --. + Lower Limit

./ --0verall Upper Limit
6 " " —_ Overall Lower Limit
4 . ._._ _ .______._-_.__-___-_.__ ___ __ _ __---.__
2 7 __ ,__.__ _.__ ___ _
0

3mm 4mm 5mm
Gap Spacing

 

 

’ Figure 74. Steady state data bandwidth.

92

8.2.Weighted Average
The following figures present the data after they are averaged for each value of
the deceleration rate value. The averaging scheme will be weighted average according to

the formula:

_ X, +X2 +...+XN

30
m N ()

X

where Xavg is the average value, X1, X2, etc. are the data points fi'om each experiment and
N is the number of experiments performed. The averaged values are connected by a

(fictitious) smoothed curve.

E Initial Flamemets Breakup vs. Deceleration Rate 3

 

 

 

I E
E 167 7 7 77 77 7- 7 7 —7 —— 7
$.14 7 7 7 7.7 7777777777 7
7 3 ‘ 9 4mm EE
Eg127777777777777777777777 I5mn E
E I- A A E

E '3 10 77 7 77.7777 7r 7 77777 7777 77 777 7 7 E A 3mm 1}
j 7: X‘ ' 7X ' E A 3mm newEE
l __ _* *"‘.’f ’g" ‘ ’ “' T T “ u ‘ i t,
E g ' X ' X ‘ 0 4mm newEE
E E E 77. 377:7 7>6 7 7 7 7 e 7 7 E EE
l — e a x a . I e E l 5mm newE
; g2; - i 7|» a7 e777 + 7 x 7 e 7 E El
E 'E x I e Lgiéyez v- ,
E _ — 7 7+77777a7 «77 I 7 77 7 7 7 E
l

I o , i

. 0 5 10 15 20

Deceleration (chsZ) E

Figure 75. Weighted average overall initial flamelet breakup vs. deceleration rate.

93

 

Formed Flamamets Count vs. Deceleration Rate

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16 ~ .___. — —- -~7 77 ---—~ ~77 7-77 7* E
*a 14 ...--.“ -- - W _- -_ - - m-” _ l . 4mm 9:
3 A ii
0 12 777777;- 77 77 . 7.77 77 I 5mm E‘
3.; 1o ..---- -5--- _ _ _ . 3mm
X A I I A
g 8 7777 7.7-7X77;--—7t~——-—77—7-777~>o<-- . x 7777 ‘ 3mm new
‘L 6 45...... .-.*___....M__._--_.2$ .-.___--_‘._-a...x_,.7__'_rv-__.. -____~__--__a_____. . 4mm new
g 4 .... 9.2;- . z - 5mm new.
" 7 7 AVG ‘z
.2 2 --.. ___--.“ _r. .. X E
E o l f l
E 0 5 10 15 20
E Deceleration (chsZ)
Figure 76. Overall initial formed flamelet count vs. deceleration rate.
Steady-State Flamelets Count vs. Doceleratlon
16
14 A7777; 777—7777;
A a
7.- 12 5-24-... >.< __. . 4mm
3 10 O A o A , X , o I 5mm
0 777—77—- ~O7-t-—Ofl 77o-
: 1X , F" M a 3mm
'5 8 +— I-77‘~—lx' -- X77777— 7777 — a 3mm new
E I I I' I A
g 6 _t_..._._ x .W .__#__._.___,__ 0 4mm new
3; I 5mm new
4 -E--- 7
to x 7AVG
2 _____
0 E . .
O 5 10 15 20
Deceleration (cm/32) E

 

 

 

 

Figure 77. Weighted average overall steady state flamelet count vs. deceleration

rate.

94

 

 

 

 

 

 

 

 

 

 

 

 

 

E Inltial Flamemeta Breakup va Deceleration Rate
16 7E ___—___- _ —7___7___ _77_7
n.
3, 14 7 7
I
2 12 777— 777 77777—777777 77 77
m l.
s 10 W__.._ W ___ _- W W. W. W --.-WW WW ._ W W WW... 0 4mm
'cE'i 3 W-W-W WW WW -WWW . WWW-W ___W W .. 0 4mm new
.2 6 .1.__.___.W_ ‘ % +AVG E
‘i . .. .. .0
g 4 ME7 7777 77 7777 £23... v in -W.._W ....WWW
5 2 1» 777 777 9 7777-77-79 7777 —77--77 7-7
O E E 1
0 5 10 15 20
Deceleration (cmlsZ)

 

 

Figure 78. Weighted average 4 millimeter gap spacing initial flamelet breakup vs.

deceleration rate.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Formed Flamemets Count vs. Deceleration Rate
3 16 --~-—
3 14 777—7777777—7—77 7
9.. 12 w
% 10 “W ___ E 0 4mm E
E 8 -—7——7—-—-77+-77777-777o-7 e 777—7-77 E ° 4mmnewE
I; 6 ‘5‘_$--"“‘_ ‘ $‘""—_'j7?‘m E+AVG E
E 4 .. --.—\fi /.“‘“".- ._. 77.7..
o 2 7777—7 .7777 o7 77—7
‘L o E E
0 5 10 15 20
Deceleration (chsZ)

 

Figure 79. Weighted average 4 millimeter gap spacing formed flamelet count vs.

deceleration rate.

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Steady-State Flamelets Count vs. Deceleration
14
O O
12 77 77) 7777.777 77 o7 — 7 — o 7 7777777 7
g 10 777 777. 7.7-
o 5‘
3 8 7 77777 . ’ 4mm E
.3 o 4mmnew
E 6 +AVG
E
m 4
m
2
o T T f
O 5 10 15 20
Deceleration (emIsZ)

 

Figure 80. Weighted average 4 millimeter gap spacing steady state flamelet count vs.

deceleration rate.

Figures 75 to 80 demonstrate once again that the deceleration rate has no
particular (non-random or systematic) influence on the initial flamelet breakup, the
formed flamelet count, and the steady state flamelet count. But these figures show that
the weighted average trend (the shape of the connecting curve) of the initial flamelet
breakup and the formed flamelet count is generally duplicated by the steady state flamelet
count. In other words, the connecting curves have similar shapes over time between the

initial flamelet breakup, the formed flamelet count, and the steady state flamelet count.

96

8.3.Normalized Average
One way to determine the amount of data variation is by subjecting the data points
to normalization. The normalization procedure is performed by first averaging all data

points, then dividing each of the data points by the averaged value determined earlier,

 

according to the formula:
X... = X’ (31)
X”

where Xnom. is the normalized average value and Xi is data from each individual
experiment. So when the data point value is equal to the overall average (Xi = Xavg), the

value of Xnonm is one.

 

 

 

 

 

 

 

 

 

 

 

 

Initial Flamemets Breakup vs. Deceleration Rate
2.5
A
g- ‘ WWWW--- __W
a; 2 ‘ "7* 2‘ “fl . 4mm
2 A A A I 5mm
8 1 5 ”HMHH 5 "WWI” " ‘ W" " I" ’ ”' _'_“”"““
.2 o A o ‘ 3mm
g " ' A 3mm new
a 1 re - —I a—— n ' --*-—Ir"“ -' —— ‘ ”— err-fl' ”—rr
E o A A I o 9 4mm new
-' I I A O O I 0
Z?- 05 ___...WW--.-.__.. .-..W.W___._W__-._...WWWW ' 5mm new
5 o I o I I
O T i r
O 5 10 15 20
Deceleration (chsZ)

 

 

L... _-____ ___..W. ___... W. _ -___.__ __ __ ___- .... “___--W .W___.._..._ -.... ___- ____ _. . _ - _

Figure 81. Normalized average overall initial flamelet breakup vs. deceleration rate.

97

 

 

 

 

 

 

 

 

 

 

 

 

Formed Flamemets Count vs. Deceleration Rate
2.5
E A
g 2 “E ‘- 7‘77— — ~777— —7 7777—”.— W-..__ -..
3 4 A 0 4mm
A A

g 1.5 747777 _. A _W _;W_ __ ___,-_____7 ___E_____#___ ______V_< I 5mm
0 ‘ ' I A A 3mm
E a» A o o
2 1 W W ,! -W_._Ws»_ __,rdg_____‘m._m______ ___ 777-1 A3mm new
"- Ia A are 97.-I A I o . 4mm new
8 I A A I O O
E I I A H e I e _l 5mm new
0 0.5 E77777‘7-77I7
ll. 0 O

o T F l

0 5 1O 15 20
Deceleration (cm/s2)

 

 

 

Figure 82. Normalized average overall formed flamelet count vs. deceleration rate.

 

 

 

 

 

 

 

 

 

 

 

 

 

Steady State Flamelets Count vs. Deceleration Rate
2.5
E 2 0 4mm
8 I 5mm
_‘é 1'5 . ‘ A ‘i' “i" A “F A 3mm
0 A A 3
E 1 WW9. t ._Wé‘ng--WW,_WWWW__WW_W. A 3mm new
'-' u "' E
u. E I . A I 0 4mm new
3 0-5 I 5mm new
0 l l l
O 5 10 15 20
Deceleration (emls2)

 

Figure 83. Normalized average overall steady state flamelet count vs. deceleration

rate.

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20

Initial Flamemets Breakup vs. Deceleration Rate
2
1-3 o o
5- 1.6
E 1.4 777.7777; 7 7 7 7 777 7 7 7 "6‘” 777 77
g 1'? 0 Q” o
E 0.8 77 O O O
E 0.6 ’ ’
g 0.4 9— °
5 0.2 7—777—77 7 7—7 7
O I . ,
0 5 10 15
Deceleration (chs2)

 

E°4"!P_J

 

 

Figure 84. Normalized average 4 millimeter gap spacing initial flamelet breakup vs.

deceleration rate.

 

Formed Flamemets Count vs. Deceleration Rate

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2
‘5 1.8
3
o 1.6 77777-77
9.. 1.4 7777 7 7 7: 7777 _’_....-.___’ WW _ _-
E 1.2 . . . o
3 1 9 9 o
u. 0.8 e . .
I: 0.6 “~— # “— '— .—** -- ........ WWWW,
§m777, -W__W-_ W
IL 0.2

0 . . 7

0 5 10 15
Deceleration (cm/32)

20

 

 

J7

BE]

 

__ ___.E

Figure 85. Normalized average 4 millimeter gap spacing formed flamelet count vs.

deceleration rate.

99

 

 

 

 

.__.____ PETE

 

 

 

SS Flamelet Count
4) 11

 

 

 

 

 

O 5 10 15 20
Deceleration (cm/32)

 

 

 

Figure 86. Normalized average 4 millimeter gap spacing steady state flamelet count

vs. deceleration rate.

Figures 81 to 86 show that the data variation does not follow any discernible trend
for the initial flamelet breakup, the formed flamelet count, and the steady state flamelet
count. However, these figures show that the steady state normalized average data are
banded in a consistent range. The overall normalized average steady state data is handed
from 0.62 to 1.44 (about 1 +/- 0.4). The 3 millimeter gap spacing steady state data are
handed between 0.62 and 1.24. The 4 millimeter gap spacing steady state data are handed
between 0.91 and 1.31. The 5 millimeter gap spacing steady state data are banded

between 0.74 and 1.23. The summary of this bandwidth is presented in Figure 87.

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Normalized Steady-State Data Bandwidth

1.6
1,4 ._____-__ .1 -f .7 7 _ m. _- .-.-._-_ ___-
1.2 */.-\0 W.

1 “r“ * —--——-—- — ——-‘ -— +UpperLimit
08 /‘—\ +LowerLimit

' / \- -—Overall Upper Limit
0.6 - -7-- 77 . 777 ~77 -—7—-7 77 — Overall Lower Limit
0.4
0.2 -— ________-._-___

0 r

3mm 4mm 5mm
Gap Spacing

 

 

 

 

Figure 87 . Normalized average steady state bandwidth.

Additionally, Figure 88 shows the normalized data bandwidth for the initial steady

flamelet count.

101

 

Normalized Initial STEADY Data Bandwidth

 

 

 

 

 

 

 

 

 

 

 

1.8

1'6 4‘ \/’Q

1.4

1.2 «7 -——- 7— -7——— 7- - .- -1 __
+ Upper Limit

1 v ' ’ ‘W‘M " +Lower Limit

0.8 77777777 777 777» - - ....... m Overall Upper Limit

—0verall Lower Limit

 

 

 

 

 

 

0.2

 

 

 

 

3mm 4mm 5mm
Gap Spacing

 

 

 

Figure 88. Normalized average initial steady bandwidth.

The overall normalized average steady state data are banded from 0.37 to 1.61 (about I
+/- 0.6). The 3 millimeter gap spacing steady state data are banded between 0.46 and
1.61 . The 4 millimeter gap spacing steady state data are handed between 0.37 and 1.49.

The 5 millimeter gap spacing steady state data are handed between 0.52 and 1.57.

102

8.4. Percent Change From Initial Steady to Steady State Flamelet Count

To determine weather there is a steady correlation between the initial steady and
steady state flamelet count (ie. steady change), the data are presented in the form of

percent change from initial steady count to steady state flamelet count according to the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

formula:
%change = (Steady _ State) 7 (Initial _ Steady) x1 00% (32)
Initial _ Steady
I
Percent Change in Flamelet Count From Formed Flamemet
Count To Steady State Flamelet Count vs.
Deceleration Rate
600
500 -~,_ _ -_7- #2‘6 . W
:i 400 77- 777777777- -» » ~~ —- ------
g 300
‘5 200 ‘
‘2’ .
g 100 g ‘ A ‘
0 -..—w;-.a...-a#_---‘__*+__-§ m---
'1 00 I r r
0 5 10 1 5 20
L Deceleration (chZ)
J

 

 

 

Figure 89. Three millimeter gap spacing percent change from formed flamelet count

to steady state flamelet count.

103

 

 

 

 

 

 

 

 

 

 

 

1 Percent Change in Flamelet Count Prom lnitlal Steady
Flamemet Count To Steady State F lamelet Count vs.
I Deceleration Rate
600 7 .
. 50077—777 7 e777-
§4oo 77 — 77» 77» 777—777
5 300 777 777 77» 77—777 » 77777 7 777 77»—
5 200 77 § 7 e 7 .7— . : 77
g 100 . —- 777777 :777777
e. o .. ___ _ __ _ Q ______ on-“ _. o- _
-100 . r
0 5 10 15 20
i Deceleration (ch32)
L- _ __ _ _ _ __ _ _ -_. _____,________.

 

 

Figure 90. Four millimeter gap spacing percent change from formed flamelet count

to steady state flamelet count.

 

(___. ,,_ ___. _ _

 

 

 

 

 

 

 

 

 

 

 

 

 

Percent Change in Flamelet Count From lnltlal Steady j
Flamemet Count To Steady State Flamelet Count vs. i
Deceleration Rate ‘
600
a 500 7— »--——-
g 400 SM...“ _ _____-_._______s_,- _______-_n
8 300 --7
E 200 77777-7» '77 7—
§ 100 - .. .- - -7
l I
0.. o -- _ - a.___w___ : _-_-_ _ _ _ ._-F.-..__ _ _ l __
-100 .
0 2 4 6 8 1O 12 14
Deceleration (cmle2)
l__-

 

 

 

Figure 91. Five millimeter gap spacing percent change from formed flamelet count

to steady state flamelet count.

104

Note that in Figures 89 to 91, the ‘slow’ and ‘rapid’ deceleration rates produce a
(relatively) low difference in the percent change. However, the ‘intermediate’
deceleration values produce the largest differences in the percent change. This suggests
that tests with slower and rapid deceleration rates are more stable in the change of
flamelet count. Tests with ‘intermediate’ (~ 4 cm/s2 to 10 cm/sz) deceleration rates appear
to have more variability.

This phenomenon can be explained by the fact that in slower rate test, the change
of flow is minimal therefore disturbance to the flamelet formation is small. This leads to
the consistency of flamelet counts. In faster rate tests, although the flow change is large,
they occur in such a short time period so that the flame formation changes from flame
front to flamelet in a uniform manner from test to test. In the intermediate rate tests, the
change of flow is neither too slow or too fast which gives the flamelet formation process

a higher degree of variability during the change process form flame front to flamelet.

105

Chapter 9

Ramp Up Test

A study in how the deceleration rate affects the flamelets formation was carried
out in the previous section. As previously explained, as the oxidizer flow rate is reduced,
the flame front will experience an instability that causes it to form flamelets. The
objective of this section is to investigate whether the above process is reversible.

To obtain the reformation of flame front from the state of near extinction (as
shown in Figure 92), the oxidizer flow is increased back to its initial velocity. The rate in
which the oxidizer velocity is increased is referred as the “acceleration rate” according to
Equation (3 3):

V — V
= final Initial . (33)

r A!
Oxidizer Flow

 

’ F lamelet Travel

Figure 92. Flame Front Reformation in Solid Thin Fuel.

106

The experimental procedure was identical to that explained previously up until
(and including) the flamelet formation stage. After the flamelets are formed (Figure 93),
the flow rate was held at this lower value for 90 seconds (one test at 45 seconds), and
then the flow was ramped up to the initial value. During the flow ramp up, the flame front
was reformed, as shown in Figures 94 to 96. After a given test is completed, a snap shot

of the fuel sample was taken for further analysis (Figure 97).

 

Figure 93. Flamelets formed from the original coherent flame front (which
consumes the entire sample). Note the black region to the left before the flamelet

“tracks” are formed.

107

e
o
0
0
O
o
J
I

 

Figure 94. Initial flame front reformation.

 

Figure 95. Flame front reformation in progress.

108

 

Figure 97. Post test picture.

109

The next step of this section was tracing the flame travel path. The process is as

followings:

1. Open the file of the desired test to be traced using Adobe Acrobat 6.0 software.

2. Place test movie on the screen and scale it appropriately. Compare the size of the
ruler recorded at the beginning of the test with a real ruler placed directly on the
monitor screen. In this experiment, the image on the screen is 2/3 of the original
size (Scale 1:15).

3. Attach a transparency sheet on the monitor screen (using regular tape), covering
the entire image of the fuel sample on the monitor.

4. Play the test in the screen. Play it frame-by-frame where appropriate to obtain the
exact screenshot of a specific event (Ramp Down Begin, Ramp Down Done,
Ramp Up Begin, and Ramp Up Done). Note the time (in seconds) where each
event occurs. Time t = O is set when the LabView program is rim, or 3 seconds
prior to ramp down.

5. Trace each event on the transparency. Figure 98 shows the result of such
procedure. Draw manually 3 longitudinal lines labeled V1, V2, and V2 as shown

in Figure 99.

110

 

Ramp Down I L Rarnp up
H H

1 11111111
in: WI, IIIIIIII!

 

6. Based on the trace lines and the time recorded, measure manually the distance
over time for V1, V2, and V3 (i.e. velocity) by measuring the distance of different
trace lines where they intersect with V1, V2, and V3 lines. Record measurements
on the spreadsheet. Calculate the average velocity of V1, V2, and V3 velocity
measurements as well. Make sure to convert scale back to the 1:1 scale when
calculating velocity.

7. Plot velocity versus time, oxidizer flow versus time, and other necessary data.

The experimental results are divided into four different sub-sections: The flame
spread velocity, the velocity ratio, the transition time, the thick fuel sample flame spread
velocity, the thick sample velocity ratio, and the velocity ratio comparison for all fuel
samples. As in the previous section, most of the results shown are for the four millimeter
gap spacing, except where comparisons between different gap spacing results are needed.

All of the other results are listed in Appendixes.

112

9.1. Flame Spread Velocity

Whatman 44 is the fuel sample type used in the MSU Combustion Lab as a
representative cellulosic fiiel sample. This section presents the results of the velocity
measurement versus time during the tests. The acceleration rate is displayed at each
graph’s heading. For example, “4 millimeter Ramp Up 2.16 cm/sz" means that the
experiment is done at 4 millimeter gap spacing, using an acceleration rate of 2.16 cm/sz.
The oxidizer flow velocity will be plotted as well using a separate velocity scale. The

data labeled AVG is the average flame and flamelet spread velocity of each discrete time,

 

 

 

 

 

 

 

asdescribed by:
V11 +V2 t +V3 1.

mm __(l_%)__fl (34)
' 4mm Ramp Up 216cm:

\
[ 0.4000 40
. i
' 0.3500 7 » 35
i 0.3000 30 -_ y
‘ g j-O—V1 ;
i: 3 0.2500 i 7 25 i V2 -
‘ gozooo ». ‘20 j+v3 .
1 m ‘—
i £01500 » ..,15 l AVG“
\ E I—Flow .1
y 0.1000 10 ' ' * -'
‘ i
L 0.0500 - 5 ‘
; 0.0000 . . , o
1 0 10 20 30 40 50 60 70 so 90 100 110

i Time (a)

Figure 100. Flame spread velocity graph - 4 millimeter gap spacing,

2.16 cm/s2 acceleration rate.

113

[— 4mm Ramp Up 5.05cmla2

 

 

 

 

 

 

Figure 101. Flame spread velocity graph - 4 millimeter gap spacing,

5.05 cm/s2 acceleration rate.

4mm Ramp Up 7.58cmla2 5
i

i 0.4000

 

0.3500

0.3000

0.2500 .

0.2000

0.1500

Flame Spread (cmla)

0.1000

0.0500 ~

 

 

 

 

 

 

i
l
i
1 00000
J
1

 

Figure 102. Flame spread velocity graph - 4 millimeter gap spacing,

7.58 cm/s2 acceleration rate.

114

i 4mm Ramp Up 10.1chs2

 

Flame Spread (cmla)

 

 

 

 

 

Figure 103. Flame spread velocity graph - 4 millimeter gap spacing,

10.1 cm/s2 acceleration rate.

1 4mm Ramp Up 15.15cmlsz

 

V 0.4000
I 0.3500

i 0.3000

 

Flame Spread (cmls)
g .

 

 

 

 

 

 

Figure 104. Flame spread velocity graph - 4 millimeter gap spacing,

15.15 cm/s2 acceleration rate.

115

The flame front reforms when the oxidizer flow once again reaches its higher
value. F lamelets occur when the oxidizer flow velocity reaches its lower value. As
observed in Figures 100 to 104, the flamelets travel with a much slower velocity than the
flat flame front. Refer to the “Introduction and Background” section where the physics of
micro-gravity combustion is explained with reference to the NASA drop tower. In the
SMFT, when the oxidizer flow reaches certain threshold, the condition inside the test
chamber will closely simulate that of the micro-gravity condition that causes the
formation of flamelets.

Figures 100 to 104 are ‘raw’ data. One way to present such data is by taking each
data point on the averaged (AVG) curve, and creating a new data point such that each of
the new data points is the averaged value of the old data point and its neighbor. Such

procedure is known as mollification, as shown by Equation (35).

N t + N + N
Nm = old ,lef 31d old,rrght (35)

 

where Now,“ is the data point to the left of the old data point (Now) and Noldfight is the data
point to the right of Now. Figures 105 to 109 shows the result of mollification of the four
millimeter data of Figures 100 to 104. Compared to the raw data figures, these mollified
figures provide a much better visualization tools of how the flame spread velocity

responds to the oxidizer flow change.

116

 

 

 

 

 

 

 

 

 

4mRmpUp2.16ch82

0.4 40

0.35 +~—- -— 7 77 7 ~77 7 7 777— 7 ___, a. 35

0.3 7 — 777 __“fi . 30
‘6‘ 3
E 0.25 4 ‘ I - 25 E

a .

g 0.2 r 20 g.
" §
5 0.15 715 g
“- i:

0.1 7 10

0.05 » _. 5

o 7 I I I 0

0 20 60 80 100
Tlme(a)

 

 

7

 

L; » Flow ’

Mollification 1

 

 

E
i
l

 

Figure 105. Mollifred flame spread velocity graph — 4 millimeter gap spacing,

2.16 cm/s2 acceleration rate.

 

 

 

 

 

 

 

 

 

 

4mm Ramp Up 5.05ch32
0.3 40
. 35
0.25 «77 777 7. 77 - _ _.
‘i? «7 30 A
0.2 7 77.__ _W-
o
g 0.15 477 777—77—777— _#_ # “___ __ 77 20 g
00 . _
E i 7 15 g
0.1 77 7 ___-n g
a . E
' 7 10
0.05 «7 777— 7 . __n
i (F 5
0 . , , 1 . . . 0
0 20 40 60 80 100 120 140
Time (a)

 

I

 

Mollification 1'

 

F'OW , I

 

 

Figure 106. Mollilied flame spread velocity graph - 4 millimeter gap spacing,

5.05 cm/s2 acceleration rate.

117

_. --__-,s, -_....k .;__an- "T"?

 

 

 

 

 

 

 

 

 

0.3 40
35
0.25 m7 7 7 7 77 .7 _7 .- .._-.-.-- .....
' 77 30
7‘ 02 ___ _ 1:
E. 1 25E 7 .
g g Mollification !
0.15 ._..-- .m- . 20 |
g """ How i
. 15 — 777- -—7- 777— — 7 ———-—--
5 0.1 -.-..-__-_.._----_- E
II. II.
10
0.05 7771 _ ._ s __
' 5
0 0

 

0 20 40 60 80100120140
Tlme(a)

 

 

 

1
Figure 107. Mollified flame spread velocity graph - 4 millimeter gap spacing,

7.58 cm/s2 acceleration rate.

 

 

 

 

 

 

 

 

 

F3
.3
0|

7777 ..... Mm“
4mmRampUp10.1cmla2

0.4 35

0.35 777 ___- _ _ ___; W _ . 30

0.3 77 7 __ - . - “___“
7 '_ .. 25 .3
E 0.257 . E. ,,
i 02 ”g Mollification i
"’ 15 E ------ Flow '
E g

_a
O

 

0.1

 

Ul

 

 

 

 

 

 

0 20 40 60 so 100 120 140 5
Tima(a) i
-_.I

 

 

Figure 108. Mollified flame spread velocity graph — 4 millimeter gap spacing,

10.1 cm/s2 acceleration rate.

118

Flame Spread (cm/a)

 

0.4

8

.°
00

0.25

.0
N

.°
a
VI

P
..a

.o
8

O

 

 

 

 

 

 

7

 

 

 

 

 

,;,
01
Flow velocity (cmla)

 

0 20

40 60 80
11me(a)

T

100 120 140

 

 

8

' Mollificationi
. ------ Flow

 

a
0|

..s
O

01

 

Figure 109. Mollified flame spread velocity graph - 4 millimeter gap spacing,

between the completion of ramp up and the full formation of flame front increases. There

is a minimum time interval over which flamelets can be transformed to flame fronts. For

15.15 cmls2 acceleration rate.

Figures 108 and 109 show that at the higher acceleration rates, the time lag

the faster ramp up rate, the flame front reformation process is simply not fast enough.

This phenomenon will be analyzed in the “Transition Time” section.

119

 

9.2. Velocity Ratio

Although section 1 clearly shows the trend of flame spread velocity for the flame
front and flamelets, it is difficult to make a quantitative comparison between the different
gap spacings and oxidizer velocity conditions. To visualize the effect of varying gap
spacing and oxidizer velocity, a ratio of the averaged flamelet spread velocity (Equation
34) and the flow velocity is constructed for each discrete time scale (ti) according to the
formula:

R = AVG(ti) .
V

oxidizer

(36)

This is then plotted for each gap spacing and flow velocities as shown in the following

Figures.

Flame Spread I Flow Velocity lhtlo
4mm Ramp Up 2.16cmlsz

 

-.... .. ..A.....]

0.05

Ratio
0
9

0.02 7 7

 

‘ 0.01 7 7

 

 

 

0 20 40 60 80 100 120 ‘

Time (s)

Figure 110. Velocity ratio graph - 4 millimeter gap spacing,

2.16 cm/s2 acceleration rate.

120

Flame Spread I Flow Velocity Ratio
4mm Ramp Up 5.05crnla2

 

 

 

 

 

Figure 111. Velocity ratio graph — 4 millimeter gap spacing,

5.05 cm/s2 acceleration rate.

 

WV. "77,. . A..__A,i. . , ._.—._. ".7, C...‘

Flame Spread I Flow Velocity Ratio l
4mm Ramp Up 7.58cmls2 I

 

 

 

 

 

O 20 40 60 80 100 120 140

 

Figure 112. Velocity ratio graph - 4 millimeter gap spacing,

7.58 cm/s2 acceleration rate.

121

 

Flame Spread I Flow Velocity Ratio
4mm Ramp Up 10.1cml32

 

 

 

 

 

 

0 20 40 60 80 100 120 140

 

Figure 113. Velocity ratio graph - 4 millimeter gap spacing,

10.] cm/s2 acceleration rate.

{_imm i .7. 7 .g— .A,_.i .#._.,_7_w #__....“i, .4.7,

Flame Spread I Flow Velocity Ratio
4mm Ramp Up 15.15CMI82

 

 

 

 

 

 

 

 

 

 

Figure 114. Velocity ratio graph - 4 millimeter gap spacing,

15.15 cm/s2 acceleration rate.

122

Figures 110 to 114 show that the flames in the “instability” regime (i.e. flamelets)
actually have a larger spread-velocity-to-oxidizer-velocity-ratio (i.e. ‘R’ in Equation (36))
compared to the flame in the stable range (i.e. flame fi'ont). In other words, a flamelet,
relative to the oxidizer velocin. is actually traveling faster than the flame front.

To better study this phenomenon, the velocity ratio for each test is averaged for

the stable (flame front) and near extinction (flamelet) region according to

_ (V1+ V2 + V3) W_ W

 

 

R ’“vgf’m'fm ’ #data (37)
(V1+ V2 + V3) M,“
R — avg/[omelet = #data (38)

The quantity R-avg is thus an average over all tests of the averages for each test for the
flamelet and flame front regimes. Thus, for the flame front and flamelet regimes, the
averages according to Equations (3 7) and (3 8) are once again averaged and presented as
the straight lines of Figure 115. This figure includes all of the gap spacing results. Table

11 represents Figure 115 in a tabular quantitative form.

123

 

 

 

 

0.0400

0.0350 (» 77777 7-.7.-.__.H

0.0300 «7 .77._.__..__ .

0.0250 77.—.7—

Ratio

 

 

,__~.

I Average Flame
Front

 

0.0200 7777—77 __ __

0.0150 __n- .. ,-

.......

 

0.0100 (7

e Average
F lamelet

 

 

0.0050 777777...-

 

 

 

 

 

0.0000

 

3mm
_ 3mm ‘
3mm“
3mm‘
3mm‘
4mm
mm“
mm“
mm-
4mmq
5mm“
5mm:
5mm“

V’ v V
Spacing

5mm

 

 

Figure 115. Velocity ratio summary for all gap spacings.

Table 11. Velocity Ratio Sm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[Gap (mm) Ramp Up Rate Average Flame Average Average of All Average of All

(cm/$2) Front Velocity Flamelet Tests Flame Front Tests Flamelet

Ratio Velocity Ratio Veloclty' Ratio Velocity Ratio
3 2.16 0.0059 0.0162 0.0054 0.0142
3 5.05 0.0044 0.0135 0.0054 0.0142
3 7.58 0.0055 0.0160 0.0054 0.0142
3 10.10 0.0062 0.0119 0.0054 0.0142
3 15.15 0.0053 0.0133 0.0054 0.0142
4 2.16 0.0061 0.0237 0.0061 0.0255
4 5.05 0.0064 0.0241 0.0061 0.0255
4 7.58 0.0061 0.0271 0.0061 0.0255
4 10.10 0.0059 0.0273 0.0061 0.0255
4 15.15 0.0060 0.0254 0.0061 0.0255
5 2.20 0.0120 0.0381 0.0112 0.0369
5 4.84 00113 0.0365 0.0112 0.0369
5 8.07 0.0108 0.0352 0.0112 0.0369
5 12.10 0.0106 0.0376 0.0112 0.0369

 

As previously stated, Table 11 shows quantitatively that the velocity ratio of the

flamelet is actually larger than the velocity ratio of the flame front. As the gap spacing is

increased, the velocity ratios of both flame fiont and flamelet are also increased. This is

due to the increased amount of oxidizer with the larger gap spacing. An increase in the

oxidizer will speed up combustion, therefore increasing the flame spread velocity.

125

 

9.3. Transition Time

The transition from flamelets to flame front will occur in several phases, but the
phase change that is visually observable is the transition from flamelet to blue flame and
the transition from blue to yellow flame. Two time constants are defined. The first time
constant is the transition time between flamelet-to—blue-flarne. This is the time difference
between initiation of the oxidizer flow ramp up and when a blue flame front first appears

on the sample. Figure 116 shows when the blue flame front first appears.

 

Figure 116. Blue flame front upon flame front reformation.

126

This time constant is defined as

Albina = thine—flame — tflmnelel (3 9)

The second time constant is the time difference between when the blue flame front first

appears and when the flame front is fully reformed. Figure 117 shows a fully reformed

 

flame front
Figure 117. Fully reformed flame front.
This time constant is defined as
Atyellow = tyellaw—flane — thine—flame (40)

127

Note that there is no clear definition of these transitions and they are sometimes
hard to observe. Blue fronts sometimes don’t occur, instead the flamelets are suddenly
converted into a fully developed flame front. Therefore a certain subjectivity combined
with good judgment is needed for this part of the research.

It is expected that as the acceleration rate is increased, the time constants
according to Equations (39) and (40) will decrease. This is due to the fact that increased
acceleration time will bring the oxidizer flow back to the higher velocity faster, therefore
transitional events will take place in quicker succession. This part of the research was
conducted in order to determine whether the time constant change is in fact inversely
proportional to the acceleration rate change. To better study this phenomenon, the
following figures show graphs of time constants versus acceleration rates for different
gap spacings. Figures 118 to 120 show that both time constants are not conclusively
inversely proportional to the acceleration rate. For the higher acceleration rate, the time
constant needed to form a blue fi'ont (or a full flame fi'ont as well) is actually increasing
as the acceleration rate is increased. Tables 12 to 14 represent Figures 118 to 120 in a

tabular quantitative form.

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 3mm Flame Front Reformation Time Constant
7
.. 6 77%17 7—7 77777777777
3 +Blue Front
3. 5 ____- - - ___. Transition
Q
§r__-_--w_-
'3 O
. r"
.. : :1
S ..., 53‘ Yellow
a 2 “—h‘-4‘ 7*. -——_-_— ant
5 Transition
~ 1
O . 1 r
0 5 10 15
Acceleration (m2) i

Figure 118. Time constants for 3 millimeter gap spacing.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1i 4mm Flame Front Retormation Time Constant
I. _.
10
9 o._.- +Blue Front
A 2 Transition
g a
K O
:l 7 ”l
a.
E 6
5
g 5 _‘ ._--.. .. _-_- _ - "" '1 + Yell”
.. m Front
0 3 a
2 —
1 _ . ___._.___.O_.
0 r
o 5 10 15
Acceleration (601,82) 1
__ __1

 

 

Figure 119. Time constants for 4 millimeter gap spacing.

129

 

r‘* + W— . hifl
SmmFlemeFrontRetormetlonTlmeConetam i
_i
F 8
+Biue Front
A 7 ——~—a- Transition
3' 6 J- —- — *— —
o.
3
a
E 5 J—~—~ -- ‘
g Q
.34 W _fl _ ____ _ _______~ _- +Yellow
.. ta Front
é 3 4___‘_______ _ _ _g_fl__ ._.. w. Transition
‘2. ’ . ’
2 ° _ —— W
1 w __ _ _
i o 10 15
: AceeieretioMchaZ)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 120. Time constants for 5 millimeter gap spacing.

130

 

_Tgble 12. Time Constants for 3 millimeter Q_ap Spacing

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.15(1)
Bacoiaration Down Down Transition ITransitlon initial Final Transition W
(cmls2) Begin Done (Blue Front) (Flamelet) Velocity Velocity Velocity Velocity
(sec) (sec) (sec) (eec) (cmls) (cmis) (Blue Front) (Yellow Front)
(cmls) (cmls)
5.05 3 0 0 35.8 5.5 5.5
Acceleration Up Begin Up Transition Transition initial Final Transition Transition
(cmis2) (sec) Done (Blue Front)l (Yellow Velocity Velocity Velocity Velocity
(sec) (eec) Front) (cmls) (cniis) (Blue Front) (Yellow Front)
(soc) (calls) (cm)
2.15 90 113 102.57 105 5.5 35.8 13.4272 18.45
5.05(5) _ _
Deceleration Down Down Transition Transition initial Final Transition Transition ‘
(chs2) Begin Done (Blue Front) (Flamelet) Velocity Veloch Velocity Velocity
(sec) (sec) (sec) (eec) (cmis) (cmis) (Blue Front) (Yellow Front)
(cmls) (cmls)
5.05 3 9 8 35.8 5.5 5.5
Acceleration Up Begin Up Transition Transition initial Final Transition Transition
(cmls2) (sec) Done (Blue Front)l (Yellow Velocity Veloclty‘ Velocity Velocity
(eec) (sec) Front) (cmls) (cmis) (Blue Front) (Yellow Front)
(806) (cm) (cm)
5.05 90 105 101.23 5.5 35.8 15.7515
Em) - _______
Deceleration Down Down Transition lTransitlon initial Final Transition Transition
(cmla2) Begin Done (Blue Front) (Flamelet) Velocity Velocity Velocity Velocity
(sec) (sec) (sec) (eec) (cmls) (cmis) (Blue Front) (Yellow Front)
(cmls) (cinls)
5.05 3 0 0 35.8 5.5 5.5
Acceleration Up Begin Up Transition Transition initial Final Transition Transition
(chs2) (eec) Done (Blue Front» (Yellow Velocity Velocity Velocity Velocity
(sec) (sec) Front) (cinle) (cmis) (Blue Front) (Yellow Front)
.. (806) (Q!) (cm)
7.58 00 103 100.83 101.53 5.5 35.8 19.3714 25.4354
10.1(1)
"Bicaiaration Down Burn Hamilton initial Final Transition Transition ‘
(cmls2) Begin Done (Flamelet) Velocity Velocity Velocity Velocity
(eec) (eec) (sec) (cmls) (cmis) (Blue Front) (Yellow Front)
(cmls) (cmls)
5.05 3 9 0.15 35.8 5.5 5.5
Acceleration Up Begin Up Transition Transition Initial Final Transition Transition
(cmis2) (sec) Done (Blue Front” (Yellow Velocity Velocityt Velocity Velocity
(sec) (sec) Front) (cmls) (cmis) (Blue Front) (Yellow Front)
! (80¢) terms) (mm
10.1 80 102 101.7 102.25 5.5 35.8 32.77 35.8
15.1 5(1) t r
Deceleration Down Down Transition initial Final Transition Transition
(cmls2) Begin Done (Flamelet) Velocity Velocity Velocity Velocity
(eec) (sec) (sec) (cmls) (cmls) (Blue Front) (Yellow Front)
(cmls) (cmls)
5.05 3 0 8.333 35.8 5.5 5.5
Acceleration Up Begin Up Transition Transition initial Final Transition Transition
(cmie2) (sec) Done (Blue Front)i (Yellow Velocity Velocity Velocity Velocity
(sec) (sec) Front) (cmis) (cmis) (Blue Front) (Yellow Front)
_ (soc) (en-Isl (cum
15.15 00 101 101.7 102.25 5.5 35.8 35.8 35.8

 

 

 

 

 

 

 

 

 

 

 

131

 

_T_able 13. Time Constants for 4 millimeter G_ap Sw’gg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

215(1)
Dealer-lion Down Begin Down Done Transition Til—mm initial Final Transition Transition
(ends2) (see) (see) (Blue (Flamelet) Veloer Velocity Velocly Velocly
Front) (eec) (eec) (ends) (ends) (Blue Front)i (Yellow Front)
(cmis) (cmisl
5.05 3 9 1079 33.3 3 3
Acceleration Up Begin Up Done Transition Transition initial Final Transition Transition
(cnds2) (eec) (eec) (Blue (Yelow Veioehy Velocity Velocity Velocity
Front) (set-Jr Front) (eec) (ends) (ends) (Blue Front) (Yellow Front)
(cmhL (ell-hi
‘ _2.16 51 65 58.56 59.7 3 33.3 19.3296 21.792
social ,
Decelerflion Dom Begin Down Done Transition Transition initial Final Transition Translion
(crnisZ) (eec) (see) (Blue (Flwneiet) Veloer Veiocly Velocly
Front) (eec)i (see) (ends) (ends) (Blue Front) (Yellow Front)
(cmis) (can)
5.05 3 9 11 33.3 3 3
Acceleration Up Begin Up Done Transition Translion inliai Final Transition Transition
(cndsZ) (eec) (see) (Blue (Yelow Veiocly Velocity Velocity Velocity
Front) (sec) Front) (eec) (ends) (ends) (Blue Front)l (Yellow Front)
(cmis) (alibi
5.05 99 105 101.73 102.49 3 33.3 16.7865 20.6245
753(2)
Deceleration Down Begin Down Done Transition Traiis'lion Inliai Final Transition Transition
(crnisZ) (eec) (eec) (Blue (Fl-neiet) Veiocfly Velocity Velocity Velocity
Front) (eec)) (eec) (ends) (ends) (Bile Front)l (Yellow Front)
(cmis) (cmis)
5.05 3 9 10 33.3 3 3
Acceleration Up Begin Up Done Transition Transition initial Final Transition Transition
(cnds2) (eec) (sec) (Blue (Yelow Veiocly Velocity Velocity Velocity
Front) (aee)i Front) (eec) (ends) (ends) (Blue Front) (Yellow Front)
(ml!) (cu-bi
7.58 99 103 101.5 103 3 33.3 21.95 33.3
10.1(1)
Deceleration Down Begh Down Dorie Transition initial Final Transition Transition
(ends2) (see) (see) (Flamelet) Veloer Velocity Velocity Velocity
(eec) (cmis) (ends) (Blue Front“ (Yellow Front)
(emf!) (ml!)
5.05 3 9 9 33.3 3 3
Acceleration Up Begin Up Done Transition Transition initial Final Transition Transition
(ciids2) (see) (see) (Blue (Yelow Veiocly Velocity Velocity Veloer
Front) (sec)l Front) (eec) (ends) (ends) (Blue Front) (Yellow Front)
(emit) (cull-i
10.1 99 102 99.87 101.63 3 33.3 11.787 29.563
1515(1)
Deceleration Down Begin Down Done Transition initial Final Transition Transition
(crrds2) (sec) (see) (Flamelet) Velocity Velocity Velocity Velocity
(eec) (ends) (ends) (Blue Front) (Yellow Front)
(mls) (cmis)
5.05 3 9 10 33.3 3 3
Acceleration Up Begin Up Done Transition Transition initial Final Transition Transition
(cnds2) (sec) (eec) (Blue (Yelow Veiocly Velocity Velocity Veloer
Front) (eec) Front) (see) (ends) (ends) (Blue Front) (Yellow Front)
(ends) (ends)
15.15 99 101 100.16 101.36 3 33.3 20.574 33.3

 

 

 

 

 

 

 

 

 

 

132

 

Igble 14. Time Constants for 5 millimeter Gap Swing

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

110(5) _—
Deceieration Down Begin [Down Donal Transition Transition initial Final Transition Transition ‘
(m2) (sec) (sec) (Blue Front) (Flamelet) Velocity Velocity Velocity Velocity

(sec) (eec) (crnie) (cmis) (Blue Front) (Yellow
(cmis) Front) (cmis)
111 T j 4133—47—7?“ W 2.5 4.5V 2.5
Acceleration Up Begin UpDone Transition Trarisifion initial Final Transition Transition
(ernls2) (sec) (see) (Blue Front) (Yellow Velocity Velocity Velocity Velocity
(sec) Front) (see) (cmis) (cmis) (Blue Front” (Yellow
(cmis) Front) (crnie)
2.2 95 106 97.5 102 2.5 26.7 6 17.9
434(1) _
Deceleration Down Begin IDown Donal Transition Transition initial Final Transition Transition
(erais2) (eec) (eec) (Blue Front) (Fllnelet) Velocly Velocity Velocity Velocity
(see) (see) (crnie) (crnie) l(Blue Front) (Yellow
(cmis) Front) (crnie)
4.07 3 9 6.3 11.4 26.7 2.6 5.129 2.5
Acceleration Up Begin Up Done Troneillon Trenellon initial Final Transition Transition
(eniIsZ) (sec) (see) (Blue Front) (Yellow Velocity Velocity Velocity Velocity
(see) Front) (sec) (cmis) (cmis) ((Blue Front) (Yellow
(cmis) Front) (crnie)
_ 4.64 99 104 101 103.5 2.5 26.7 12.16 24.26
6.07 (4)
Deceleration Down BeginiDown Donal Transition Transition initial Final Transition Transition
(cmls2) (sec) (see) (Blue Front) (Flamebt) Velocity Velocity Velocity Velocity
(eec) (eec) (cmis) (crnie) “(Blue Front) (Yellow
(cmis) Front) (cmis)
12.1 3 5 7.56 26.7 2.5 2.5
Acceleration Up Begin Up Done Transition Transition initial Final Transition Transition
(crnie2) (see) (sec) (Blue Front) (Yellow Velocity Velocity Velocity Velocity
(we) Front) (see) (crnie) (crnie) “Blue Front) (Yellow
(crnie) Front) (cmis)
6.07 95 96 96.69 96.67 2.5 26.7 16.1363 26.7
1210(1)
Deceleration Down Begin lDown Donel Transition Tramlion initial Final Transition Transition
(cmisZ) (sec) (see) (Blue Front) (FlenIeiet) Velocity Velocity Veloch Velocity
(sec) (see) (cmis) (crnie) ((Blue Front)l (Yellow
(cmis) Front) (crnis
12.1 3 5 3.5 6.33 26.7 2.5 20.65 2.5
Acceleraoon Up Begin Up Done Transition Transition initial Final Transition Transition
(eraisZ) (sec) (sec) (Blue Front) (Yellow Velocity Velocity Velocity Velocity
(see) Front) (sec) (crnie) (crnie) i(Biue Front) (Yellow
(cmis) Front) (cmis)
12.1 96 97 97.33 9 2.5 26.7 26.7 26.7

 

 

 

 

 

 

 

 

 

 

133

 

9.4. Thick Fuel Sample Flame Spread Velocity

So far all experiments were conducted using Whatman 44 fuel (cellulose)
samples. In addition to the Whatman 44, there are several other fuel samples that were
used in this research. These samples, all cellulosic, are: 3CHR, 31ETCHR and 17CHR,
all manufactured by Whatman. Table 15 shows the manufacturer’s specified thickness for

each of this sample.

Iable 15 . Cellulose Fuel Sample Thickness and Area Dens_ity

 

 

 

 

 

 

 

 

 

 

. Thickness ratio
Sam pie Name Thickness Thickness compared with Area Density
(mm) (cm) Whatman 44 (gicm2)
44 0.17 0.017 1.000 0.0074010
3CHR 0.36 0.036 2.118 0.0180615
31 ETCHR 0.50 0.050 2.941 0.0186825
17CHR 0.92 0.092 5.412 0.0410900

 

 

In theory, a thicker fuel sample will affect the flame fiont and flamelet spread
velocity. Using an identical procedures as before, a series of experiments was performed
for the thicker fuel samples. For the thicker fuel samples, the three millimeter gap spacing
was eliminated for two reasons. First, for the thicker fuel samples, the igniter wire has to
be repositioned closer to the quartz glass (refer to Figures 12, 13, and 37). With three
millimeter gap spacing, the wire (or its terminal) will then touch the quartz glass plate.
This will cause ‘sagging’ and imperfect ignition. Second, flamelets cannot be obtained
with thicker fuel samples like the 31ETCHR and 17CHR with a 3 millimeter gap spacing.
Figures 121 to 126 show the flame spread velocity for the thicker fuel sample for four

and five millimeter gap spacings.

134

 

4min Ramp Up 3CHR

 

 

 

 

 

 

 

 

 

 

 

0.40001 - «— -«_ c— 7 7 1 40

0.3500 ‘ . 35
... 03000 1 30
E _______-
0 02500 ... 25 +V1 i
v

02000 i 20
3 —AVG

01500 15
g i ‘ — Flow
3
u. 0.1000 10

0.0500 4 . 5

0.0000 v i i w r T 1 . i . i ---r i s. . 7 0

0 10 20 30 40 50 60 70 80 90 100110120130140150160170
Time (s)

 

 

 

L

 

Figure 121. Flame spread velocity graph - 4 millimeter gap spacing, 3CHR fuel

sample - 2.16 cmls2 acceleration rate.

 

4mm Ramp Up 31 ETCHR

0.4000 - -._,_ - ___“..-Ln 17.-7 «r 40

 

 

 

i

 

 

i
l

 

 

 

 

A A

S
1
!

 

-
‘7

 

 

 

Flame Spread (cmis)
E
l

3
§
l
l
i
r
l
i

 

 

 

p

S
i
(ll

 

 

 

 

0.0000

 

01020304050607080901011121314151517
0 0 0 0 0 0 0

Time (sf
___-_____.-..--f-___- -- ,__J

 

 

 

 

 

Figure 122. Flame spread velocity graph - 4 millimeter gap spacing, 31ETCHR fuel

sample - 2.16 cm/s2 acceleration rate.

135

 

 

 

g

 

i

 

 

 

 

 

 

 

 

F lame Spread (cmis)

 

 

 

 

0.1500 1
0.1000 1
0.0500 1
011900 T 1 T T r l y y 1 1 . Y Q r 0
0 20 40 50 80100120140160180200220240260280
Time (s)

 

 

 

Figure 123. Flame spread velocity graph - 4 millimeter gap spacing, l7 CHR fuel

sample - 2.16 chs2 acceleration rate.

 

5mm Ramp Up 3CHR

§
1
i
\

~40

 

g i
i
1
I

 

 

 

 

 

02500 1». . —

 

 

 

 

.0
é

 

 

Flame Spread (cmis)
i

 

.0
g

 

 

 

 

 

 

010 20 30 4O 50 60 70 80 90100110120130140150
Time(s)

 

 

 

Figure 124. Flame spread velocity graph - 5 millimeter gap spacing, 3 CHR fuel

sample - 1.74 cm/s2 acceleration rate.

136

5mm Ramp Up 31 ETCHR

 

, ___ fl, __._._ ____._ __fi

 

 

 

 

 

 

 

 

 

 

 

1-
l
i 0.4000 T — — __._ i + A- — - 40
5 0.3500 *— - ~- ~ 35

... 0.3000 4 « 30

G

s F” ------ i

0 02500 «» 25 -°- V1

v

'o + V2

3 02000 1 i. 20

3. —AVG

0 0.1500 15 __

E Flow

2

IL 0.1000 1 10

0.0500 4 t 5
0.0000 . r . f i 1 r . . . T . Y . T 0
0 10 20 3O 40 50 60 70 60 90 100 110 120 130 140 150
Time (s)

 

 

 

Figure 125. Flame spread velocity graph - 5 millimeter gap spacing, 31ETCHR fuel

sample - 1.74 cm/s2 acceleration rate.

 

5mm Ramp Up 17CHR

.°
1

.-.-____ .. _ ___. ___"..-h- - ...-..__..- . . ..- 4o

 

 

 

 

g

 

 

 

 

 

 

 

 

 

 

 

 

 

A 0.3000 -

. ——-—-—-—-————-———-r
E 02500 fl -+- V1
V

'o —-— V2

E 02000 i — ———*~~ —— — AVG
a _

to

g °-‘5°° l t ‘ ___“ — Flow
is .

a 0.1000 . -... -.__-

 

 

 

i
/¢
i

 

 

 

p
3

o

8 .

 

‘r T r r r r v ‘r T o
40 60 60 100 120 140 100 180 200

Time (s)

 

 

 

Figure 126. Flame spread velocity graph - 5 millimeter gap spacing, l7 CHR fuel

sample - 1.74 cmls2 acceleration rate.

137

9.5. Thick Fuel Sample Velocity Ratio

As in part 2 of this section, the velocity ratio analysis was made to determine
quantitative comparison between the different gap spacings. To better visualize the effect
of varying gap spacing and oxidizer velocity, a ratio of the flame spread velocity to the
oxidizer flow velocity is plotted for each discrete time scale, as shown in Figures 127 to

132.

 

(if,

Flame Spread I Flow Velocity Ratio
4mm Ramp Up 3CHR

 

 

0.02500

 

0.02000 it

0.01500 r

Ratio

 

0.01000

 

0.00500 ‘J

 

 

0.00000.v.——r..7r.r
0102030405000700090100110120130140150150170

Time (s)

 

 

 

 

Figure 127. Velocity ratio graph - 4 millimeter gap spacing, 3CHR fuel sample —

2.16 cm/s2 acceleration rate.

138

 

“ ‘4 ' "7

Flame Spread i Flow Velocity Ratio
4mm Ramp Up 31 ETCHR

0.016001 - 7 --___- -,__-_ —— -- -, - -. ’1

 

0.01400 1 — ~- - ..--.-“ -. ____

 

0.01200 -

 

0.01000 (~-

 

0.CD800 -

 

Ratio

 

 

 

 

 

0.00200 J»——----——

 

0.00000 r r r r . 1 a
0 20 40 60 80 100 120 140 160

Time (s)

 

 

 

 

 

Figure 128. Velocity ratio graph - 4 millimeter gap spacing, 31ETCHR fuel sample -

2.16 cm/s2 acceleration rate.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F k _1
Flame Spread I Flow Velocity Ratio
4mm Ramp Up 17CHR
0.01000 r — - -—--— ~ ..- 777~—— "—j
0.00900
ammo 1.-
0.00700 «
0.00600
5
. 000500
in
0.00400 .
0.00200 3 i
0.00100 {
i
0.00000 - r T . . A
0 50 100 150 200 250 '1
Time (s) g
__ A. __ J

 

Figure 129. Velocity ratio graph - 4 millimeter gap spacing, 17CHR fuel sample -

2.06 cm/s2 acceleration rate.

139

 

Flame Spread l Flow Velocity Ratio
5mm Ramp Up 3CHR

 

0.02500 1r 1

Ratio

0.02000 <~-» - ---..-

0.01500 . --

 

 

0.01000 1*

 

 

 

000500 ‘J

 

 

o.m ‘L—’—Y_ F r r r v T r ‘l r ‘l T r f T.
O 10 20 3O 40 50 60 70 00 90 100 110 120 130 140 150
Time (s) i

 

 

 

 

Figure 130. Velocity ratio graph - 5 millimeter gap spacing, 3CHR fuel sample -

2.16 chs2 acceleration rate.

 

l'm fl- -___zz__ ___ _ _.____

Flame Spread i Flow Velocity Ratio
5mm Ramp Up 31ETCHR

0.04000 ~— ---

0.03500 " '-

 

 

003000 . ~

 

0.025(1)

0.02000

 

Ratio

0.015(1) 1

 

 

0.010(1)

 

0.00500 « -- ~-

 

 

o.ooooo . . . , - , , ,

Time (s)

 

 

 

 

Figure 131. Velocity ratio graph - 5 millimeter gap spacing, 31ETCHR fuel sample -

2.16 chs2 acceleration rate.

140

 

Flame Spread i Flow Velocity Ratio
5mm Ramp Up 17CHR

0.01200 , 7 ~ ~» , —k Q —_ ,. 7 ,- .7 .. z - --1
001m f . .. .._._...__ ___. ___...v—.. ___._4—- ___.s-Wfi—- -- ~-——- — ~———— —-——~——-————-' "M. -——-———r—-~—«

000000 >- —— ~ —u————~w——~r ”___- ___--._____-_____. — a

 

0.0%00

...... "V \
0,0,0, \ W

 

 

Ratio

 

 

 

 

 

 

 

 

Figure 132. Velocity ratio graph - 5 millimeter gap spacing, 17CHR fuel sample -

2.16 cm/s2 acceleration rate.

Figures 127 to 132 show a similar pattern to the velocity ratio of the Whatman 44 sample,
although the trends are not nearly as “clean”. In the next section, a comparison of this

ratio will be made for all fuel samples.

141

9.6. Velocity Ratio Comparison for All Fuel Samples

A comparison of the velocity ratios for all fuel samples analyzed in section 5 is
now made. For the Whatman 44 test, the velocity ratio for each test is averaged for the

stable (flame front) and near extinction (flamelet) regimes according to:

 

 

(V1 + V2 + V3) flaw M,
R - avgflame—fmni = #data ’ (41)
(V1 + V2 + V3) W,“
R — avg/[omelet = # data . (42)

Then, for the flame front and flamelet regimes, the averages according to Equations (41)
and (42) are again averaged for all tests in each the four and five millimeter gap spacings,
and presented as one composite data point.

For the thicker fuel samples (3CHR, 31ETCHR, and 17CHR), due to the lack of
fuel samples, a single test is done for each thick fuel sample and the various gap spacings.
A total of 6 experiments were carried out (3 thick fuel samples, 2 gap spacings for each
sample thickness). The results will be presented both on linear and logarithmic scales. A
power trend-line is added. The trend-line equations and the R—squared values are shown,
as well.

Note that as the fuel sample becomes thicker, the flamelet and flame front become
much less distinctive. Figures 125 and 126 show an experiment using 17CHR fuel, the
thickest fuel sample available for this research (more than 5 times thicker than the ‘usual’

Whatman 44 fuel). The flame front shown in Figure 133 has a very similar physical

142

appearance to the flamelets shown in Figure 134. Due to this phenomenon, the results in
this section will both include and exclude the 17CHR experiments on the log scale. Note

that the legend “Power” refers to the power law curve fit of the data points.

 

Figure 133. Flame front in 17CHR fuel sample.

 

Figure 134. Flamelets in 17CHR fuel sample.

143

    
 

 

 

 

 

 

 

 

 

 

 

4mm Gap Spacings Velocity Ratio vs. Area Density
0.04 W , ~ , WW WW ~ WW . WWW
: 70 6992
0.035 r .7 X. 7797 9002* ___—_n. ,___~ -7 Aa—_
R‘ = 0 8752
003- ——~ , —- _ W
220-025 -__. * '. 'Fom’emaaro‘" _, r
0 02 _ gmfli , ,, y = 07.0001x'107é: - Flame Front Ratio
‘3 R2 = 09944 Power (Flamelet Ratio)
3 0.015 ,:E9!9r (FlaanMBéfiQ.
0.01
0.005 W «
0 . W . .
0.000 0.010 0.020 0.030 0.040 0.050
‘ Area Density (glam!)
l_,i.__.,, .--_ _K___m__ W, ___

 

 

Figure 135. Velocity ratio comparison for all fuel samples for 4 millimeter gap

spacing.

 

5mm Gap Spachge Velocity Ratio ve. Area Density

   
     

0.04 , W WW ~ - W ,7,

0 035 y : 0 0003x’0 ”9‘

. ._i. _ . .__
R2=O.9994

 

 

 

 

 

 

 

 

 

0.03 ..‘._n_.____,,. .__ ”fir—7 .-.i777w
y = 0.0002x‘1 ”95
0.025- W —~r—~»——~——~~2~ ~~— , Fla R -
R =09901 "‘9'“ 3'”
002 __ I Flame Front Ratio
3 ' . —Power (Flamelet Ratio)
g 0.015 WW , , , him—“l —.,F19‘_”9' GEL“? FWBEWL
0.01 ———————— .
l
i
0.005 ——————- W—W— < — —~W4
. . . , . P
0.000 0.010 0.020 0.030 0.040 0.050
Area Demitymlcmz)

Figure 136. Velocity ratio comparison for all fuel samples for 5 millimeter gap

spacing.

144

 

 

  
 

 

 

 

 

 

 

 

 

If nu, , A? ,,,,_n w --- n-g..__-,_ _A_.__-.n, W
4mm Gap Spacinge Velocity Ratio va Area Density
INCLUDE 17CHR
0.001 0.010 0.100 1.000
0.1
2 y = 0.0001x‘1'0737 e Flamelet Ratio
2 R2 = 0.9944 I Flame Front ihtio
E ‘ 0'01 -—Power (Ftarnelet Ratio)
; g —Power (Flan'e Front Ratio)
1 y = 0.0002x'O 6992
l R2 = 08752 I
g
‘ Area Density (glcm2)
l
I
l

Figure 137. Velocity ratio comparison for all fuel samples for 4 millimeter gap

spacing — log scale.

 

5mm Gap Spacings Velocity Ratio vs. Area Density

 

 

 

 

 

 

tmLunencm
0.001 0.010 0.100 1.000
0.1
y = 0.0002x'1-0795
R2 = 0.9901
g e Flannlet Ratio {
I? I Flam Front Ratio E
z- 0.01 _
g -——-Fbwer (Fhrraiet Ratio)
§ —Fbwer (Flame Front Ratio) p
y = 0.0003x‘0 7194 "“"‘"“““"""”""”"’ M” ”
R2 = 0.9994
0.001
Area Density (glcrnZ)

 

 

Figure 138. Velocity ratio comparison for all fuel samples for 5 millimeter gap

spacing - log scale.

145

 

iv 7' 7"" # _ —_ ___ ' _ ]

 

 

 

 

 

 

 

 

 

4mm Gap Spacings Velocity Ratio vs. Area Density
EXCLUDE 17CHR
0.001 0.010 0.100 1.000
0.1
o y = 0.0002x4’m
3 R2 = 09974 e Flamelet Ratio ' W
E 0.01 I Flame Front Ratio
8 i—Power (Flamelet Ratio)
3 g—Power (Flame Front Ratio)
y = 0.001x‘m
R2 = 0.9816
!
g 0.001
i Area Density (glam)
{

Figure 139. Velocity ratio comparison for all fuel samples for 4 millimeter gap

spacing - log scale (exclude 17CHR).

 

5mm Gap Spacings Velocity Ratio vs. Area Density

EXCLUDE 17CHR
0.001 0.010 0.100 1.000
~ 0.1

 

 

y = 0.0004x‘°‘°‘53
R2 = 0.9975
1 e Flamelet Ratio

 

 

 

 

 

 

o
g _ 0 01 I Flame Front Ratio
3 ’ —Power (Flamelet Ratio)
g y = 0.M3X-o'n‘3 L—PM (Flame Fm Ratio )
R2 = 0.9985
0.001
Area Density (glanZ)

 

 

Figure 140. Velocity ratio comparison for all fuel samples for 5 millimeter gap

spacing — log scale (exclude 17CHR).

146

Figures 135 to 140 show that the thicker fuel sample will produce smaller velocity
ratios. As with Whatman 44, the velocity ratios for flamelets are higher than the velocity
ratios for flame fronts.

Notice that in Figures 135 and 136, both velocity ratios for flame front and
flamelets seem to converge to the same asymptotic point as the fuel sample thickness
becomes large. This suggests (in accordance with experience) that as the fUCl sample
becomes thicker, the distinction between flame fi'ont and flamelets becomes less obvious

(i.e., they become more alike), as shown visually by examining Figures 133 and 134.

147

Chapter 10

Miscellaneous Analysis

This section will present additional flame spread behavior analysis, and is divided
into two sub-sections. The first part analyses the position of the flame relative to its
starting point with respect to time. The second part directly compares the oxidizer flow to

the flame spread velocity.
10.1. Flame Position vs. Time

Figures 100 to 104 of the ‘Ramp Up Test’ section show the graphs of flame
spread velocity versus time. Later, these data were subjected to a process called
Mollification to smooth the data and making them better visualization tools to relate the
flow velocity to the flame spread velocity, as shown by Figures 105 to 109.

Another way to present these data is by plotting the flame position versus time
from the flame spread velocity versus time data. In other words, the flame spread velocity

plots with respect to time is integrated, according to the formula:

I

x(t)flame = [v(t),,mdt (43)

10

where x(t) flame and v(t) flame are the position of flame and the flame spread velocity,
respectively, where t is time. Figures 98 and 99 show the process of gathering raw data,

which results in discrete data points instead of a continuous curve. In this case, to obtain

148

the position of flame with respect to time, each position datum is obtained according to

Equation (44).

 

v.
x = xt—lflame + (t _'t ) (44)

l i—l

i flame

Equation (44) is also commonly known as the backward differencing method, which is

first-order accurate. Figures 141 to 145 show the result of this process. The oxidizer flow

is plotted on the same time axis, labeled ‘Flow Velocity’.

 

 

Position vs. Time

4mm Ramp Up 2.16cmls2

 

 

 

‘ .
1
i
1

 

 

 

 

 

 

 

Position (cm)

 

 

 

 

 

 

 

 

 

 

 

-._ ___. _._..______-.___.. _ -... . ._ _. ___.____._._J

 

Figure 141. Flame position versus time graph — 4 millimeter gap spacing,

2.16 cmls2 acceleration rate.

149

 

8 8

N
O

(7+Fosition i;
1 i .
— FW “3'9“! H

Position (cm)
3

Flow Velocity (cm/s)

_a
O

01

 

 

 

 

Tim. (sec)

 

Figure 142. Flame position versus time graph - 4 millimeter gap spacing,

5.05 cm/s2 acceleration rate.

 

 

Position vs. Time
4mm Ramp Up 7.58cmlsz

 

 

j-o-' ”'Eo's'im‘n’ "T"
‘—-. _fiWYs'aciv

Position (cm)
Flow Velocity (cmis)

 

 

 

 

 

l Time (see)
Figure 143. Flame position versus time graph - 4 millimeter gap spacing,

7.58 chs2 acceleration rate.

150

‘ Position vs. Time ‘
“ 4mm Ramp Up 10.1cmls2 ‘

T;
. . E
1 g a .7 W”
‘l E E t-O-Position
2 —§ ;—F|owVeIocity{
s > ' ,, ,
‘* 3
it

 

Figure 144. Flame position versus time graph — 4 millimeter gap spacing,

10.1 cm/s2 acceleration rate.

Position vs. Tlme
4mm Ramp Up 15.15cmls2

  

E 20 35
1 1a — 30
1 16
i 14 25 T;

.. E
L 5 12 20 3 .A_____w__,
& E 10 g +Position 1
I E 15 3 tiff'g‘lvelosny.‘
| 8 8 >
. °- g
l 5 10 E
j 4

5

! 2
: o o
a o 50 100 150
i 11me(sec)
L , z,“ u. -.-. ..-..7 .. .. ,_ .__.s..mm..

 

Figure 145. Flame position versus time graph — 4 millimeter gap spacing,

15.15 chs2 acceleration rate.

151

Figures 141 to 145 show that the positions of flame are progressing at a
reasonably constant rate. This means that during a certain value of oxidizer flow, there
are not many variations of the flame spread rate. There are two distinctive slopes on the
figures above: one is the slope during the lower oxidizer flow (i.e., flamelet region), and
the other is the slope during the higher oxidizer flow (i.e., flame front region). These
slopes of course translate to the average velocity values. Table 16 listed the slope

(average velocity) values for the ramp up tests.

flble 16. Average Fljame Spread Velocity for The Ramp Up Tests

 

 

 

 

 

 

 

 

 

Slope (Average Flame Spread Velocity) (cm/s)
Acceleration Rate Flame Front Flamelet

2.16 cm/sz 0.196 0.071

5.05 cm/s’ 0.206 0.073

7.56 vm/s2 0.205 0.073

10.1 cm/32 0.194 0.076

15.15 chs2 0.188 0.068
Maximum Difference (96) 9.526 11.958

 

 

 

 

The maximum differences are listed to show how much variation occurs between
tests. All the lower oxidizer flow velocities are the same (3 cm/s), and all the higher
oxidizer flow velocities are all the same (33.3 cm/s). The maximum difference in flame
spread velocities (i.e. the slope of position of flame graphs) on the flame front region is
less than 12 percent, and in the flamelet region, it is less than 10 percent. This means

there is reasonable consistency in the experimental results.

152

 

10.2. Flame Spread Velocity vs. Oxidizer Flow Velocity

The objective of this experiment is to study the direct relationship between the
oxidizer flow velocity and the flame spread velocity. Up to this point, comparisons
between the flame spread velocities were made with the higher (flame front region) and
lower (flamelet) oxidizer flow velocity, as shown in Table 15. For this section, the flame
spread velocity measurements were taken for each test at the lower oxidizer flow
velocity. Note that not all lower oxidizer flow velocities result in the formation of
flamelets. If the lower oxidizer flow velocity is not low enough to create the stability
required for the flamelet formation, a flame front or flame cusps are formed instead. For
the range of lower oxidizer flow velocities that will result in flamelet formation, refer to
Figure 57, the ‘Flammability Map’.

The 5 millimeter gap spacing data were chosen since at that gap spacing value
more test were performed than on any other gap spacing values. A ruler was placed next
to the fuel sample before the beginning of each test. The image of this ruler was recorded
for sealing, as described in the ‘Experimental Procedure’ section. A known distance
(from the ruler) was selected and the flame spread across that distance was measured for
each test, resulting in an average flame spread velocity. These measurements were taken
for both the normal and inverted positions of the 5 millimeter gap spacing. Figures 146
and 147 show the raw data from those measurements. These raw data are then averaged
for each oxidizer flow velocity value, and presented in Figures 148 and 149. These
Figures also include the linear trend lines (along with their equations and R values) for

these averaged data points.

153

 

Normal Position l

Flame Spread Velocity vs. Oxidizer Flow Velocity [
1

l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 04000 t i
I ° I
03500 ,_______ .__.--” Flamelet Regime “4”“ ___. __fl
I | .
30.30001 — —— —-: ___ — __ m -__ L ___...e-__- -..-.. _ -
s
3 0.2500 «F —-——— : — —— — ~—— —— 4— ._.: ___. .__. - _-
'0
C I I l
30.200044 — .—--—_ ___ ..__ e3- ,4 .. - _ -
I I
20.1500 4 —_ —.+_ w:- L-M- --.9_-__i -_. _ _ __
I I
5 0.1000 —_ 1—3 . — —.v:—— Flame Front —-—-~« 1
0.0500 Quench ”mm—{em Regime __
l
; 0.0000 . -, , . - , . ,
l 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
% Oxidizer Flow (cmis)
i,

 

 

 

Figure 146. Raw data for the normal position flame spread velocity vs. oxidizer flow

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

velocity.
’ Inverted Position
Flame Spread Velocity vs. Oxidizer F low Velocity
0.3000 . . .
: Flamelet Regime :
e
0.2500 : 4'
g I . I
3 0.2000 : :
g l l
2 01500 1 1 _ - i
G I I
'2 - .
0.1000 . I »
g l. e ' : Flame Front
0.0500 4—4 Quench L n. l . Regime ._.
I I
I I
0.0000 . : t . 1. T
0.0 2.0 4.0 6.0 6.0 10.0 12.0
Oxidizer Flow (cmis)

 

 

 

Figure 147. Raw data for the inverted position flame spread velocity vs. oxidizer

flow velocity.

154

l Normal Position
Averaged Flame Spread Velocity vs. Oxidizer Flow Velocity

 

 

l
l 0.400000
1 0.350000

A

0.300000 .

  
 
   
 

0.250000

 

y = 0.0239x + 0.036

0.200000 ~ . 1&2 509519

0.150000

 

Flame Spread (cmls

Flame Front —~~-r l
Regime

0.100000 1

 

 

 

 

 

l
1 0.050000
1 0.000000

 

10 12 14
Oxidizer F low (cm/s)

l
_v 7 .__7 ... ,,_’W .7, ___._, ,7 ..7., , _7 , 1

Figure 148. Averaged normal position flame spread velocity vs. oxidizer flow

 

 

 

 

 

 

 

 

 

 

 

velocity.
1 "—— TF _-_ ‘ .__ "WW
i Inverted Position
| Averaged Flame Spread Velocity vs. Oxidizer Flow Velocity .
l l
i
i 3 y = 0.029x - 0.0162
E B’ = 0.9mm-
9.
'0
I
2
D.
d)
O
E 70.7- fiiiiiiv
g
It Flame Front
_ Regime P
10 12
.. l

 

 

Figure 149. Averaged inverted position flame spread velocity vs. oxidizer flow

velocity.

155

Figures 146 to 149 show that there is no significant discontinuity (or ‘jump’) of
the flame spread velocity as the oxidizer flow velocity changes fi'om the flamelet region
to the flame front region. instead, the flame spread velocity data lie along a smooth curve
(or line) as the oxidizer flow is increased. This is explained by the fact that in between the
flamelet and the flame front regions, exists a transitional phase such as flame cusps (refer
also to the ‘Flammability Map’ section where this was discussed). When the flame is at
the transition phase, the flame speed bridges the difference between that in the flame
from region and in the flamelet region.

Next, a comparison between the normal and inverted positions flame spread
velocities is made. Figure 150 shows both trend lines plotted in Figures 148 and 149 on
the same graph. It is interesting to see that the two trend lines have almost the same slope,
with an offset of 0.005. There are no data available for the inverted position at the higher
final oxidizer flow velocities. Figure 150 also shows that at the normal position, the flame
spread is faster at the same oxidizer flow for the normal case (especially below 10 cm/s),
suggesting that the normal position flamelets are slightly stronger. This would accord
with the heat transfer discussions of differences between the normal and inverted

positions. Refer to the ‘Data Collection’ section.

156

 

Averaged Flame Spread Velocity vs. Oxidizer Flow Velocity

§

 

 

E

 

 

0.250000 ~

 

e Normal 1
A Inverted I
Linear (Normal)

[f - - Linear (inverted) ,

 

 

 

0.150000

 

 

Flame Spread (cmis)
0

0.100000« --

 

 

0.050000

 

 

 

 

0.000000

 

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
Oxidizer Velocity (cmls)

 

 

 

 

Figure 150. Averaged normal and inverted positions flame spread velocity vs.
oxidizer flow velocity. The trend lines for the averaged data (both positions) are

plotted too.

Up to this point, the oxidizer flow range included for the analysis are the ones on
the lower oxidizer flow values. The lower flow values ranges from 0 to 12.5 cm/s. It is
shown by Figures 146 to 149 that the relationship between the oxidizer flow velocity is
reasonable linear to the flame spread velocity. The next step is to investigate whether for
the higher flow velocity, this linear relationship is still true.

There are some normal positions experimental results that were analyzed for the
higher oxidizer flow velocity values. These averaged data, included in the flame spread

velocity vs. oxidizer flow velocity values, are presented in Figure 151. A trend line with

157

second order polynomial is used for data interpolation, and the curve equation is included

inFigure151.

 

Normal Position
Averaged Flame Spread Velocity vs. Oxidizer Flow Velocity

0.400000

 

0.350000

 

0.300000 7—
0.250000 *4
0.200000 4W.

 

 

0.150000

 

 

y“=' -0.0009x‘ + o.03§27¥0.00€9"
0.100000 —— ~ , R2 = 0-9464 *

 

Flame Spread (cmls)

0.050000%, -- , , _. ~44

 

 

 

0.000000

 

0 5 10 15 20 25 30
Oxidizer Flow (cmls)

 

 

 

._. ___ “___—V.i A. 7.“;

Figure 151. Averaged normal position flame spread velocity vs. oxidizer flow

velocity where the higher values of oxidizer flow are included.

Figure 151 demonstrates that linear relationship between the oxidizer flow and the
flame spread velocity is no longer true when the higher flow velocities are considered. In
other words, for the velocities not near the instability range, the relationship between the
oxidizer flow velocity and the flame spread velocity is not linear.

Since the oxidizer flows in the opposite direction to the flame spread (opposed

flow), when the flame is at the most stable form (i.e. flame front), its spread velocity is no

158

longer affected by stability regimes, as in the lower oxidizer flow velocity cases. When
the flames are in the form of flamelets, the spread velocity is automatically lower. When
the flame becomes more stable and forms flame front, the spread velocity will increase
automatically. When the flames are all in the form of flame front, instead of being
affected by its stability, the oxidizer opposed flow velocity solely affects the flame spread
rate.

According to F emandez-Pello (1981), for a thin fuel samples, the “flame spread
rate always decreases as the opposed velocity increases”. The thin fuel samples used by
F emandez—Pello were paper sheets. The experimental results fi'om MSU shown by Figure
151 are in accord to some extent with the experimental results of Fernandez-Fella. Figure
2 on Femandez-Pello (1981) shows the correlation between the flame spread rate and the
opposed flow velocity for a thin paper fuel sample. The oxidizer flow velocity range in
F emandez—Pello’s experiment were well above 10 cm/s. The flame spread rate is almost a
constant for the oxidizer velocity range from 10 cm/s to ~ 100 cm/s. Above 100 cm/s, the
flame spread velocity starts to decrease as the oxidizer flow velocity is increased. The
SMFT does not have the capability to produce an oxidizer flow above 100 cm/s, therefore
a direct comparison to the Femandez-Pello’s results are difficult to be made. If more data
points at the higher velocity is obtained for the MSU test, a relatively constant value of
flame spread velocities as observed in Femandez-Pello (1981) can be obtained. The
Figures 146 to 150 above shows that the proportional linearity between the oxidizer flow
velocity and the flame spread velocity is only valid for the lower oxidizer flow (near
instability range). The Femandez-Pello experiments will be affected more by buoyancy

compared to the MSU experiment due to the Fernandez-Pello’s experiments that were

159

carried out in a wind tunnel with a vertical spacing of 4 centimeter. The MSU data
presented above are tested in the wind tunnel with a vertical spacing (gap spacing) of 5
millimeter, which is smaller by a factor of 8 compared to Femandez-Pello’s wind tunnel.
This explains the discrepancy between the velocities scale.

In addition to the above, Olson (1991) states that for the velocity ranges near the
flame instability range, the spread velocity is controlled by an “oxidizer transport-limited
chemical reaction”. This is what was discussed above as if the flame is near its unstable
range, the stability of the flame governs its spread rate. Figure 6 on Olson (1991) shows
that for air as oxidizer (21% 02), the flame spread velocity increases with the oxidizer
flow velocity up to ~ 15 cm/s. Above that value, the flame spread rate decreases as the
oxidizer flow rate increases, which is somewhat in agreement with the MSU results. The
discrepancies between the MSU and Olson results, especially the velocity scaling, can be
explained by the difference in the experimental setup. Olson used a 20 centimeter
diameter combustion tlmnel while the MSU experiments were conducted in 3 to 7

millimeter gap spacings rectangular tunnel.

160

Chapter 11

Smoke Wire

The purpose of the smoke wire experiment is to observe the uniformity of the
flow inside the SMFT test section. Any non-uniformity such as flow recirculation and
separation can affect the flamelet behavior and add variability to the experimental results.
All of the data presented here are qualitative.

The smoke is created by using a similar system as the igniter wire, where a DC
current is passed through a high-resistance conductor (wire), resulting in a large quantity
of heat dissipation from the wire. The wire was coated with regular automotive bearing
grease that will vaporize once heat is applied to it, producing smoke. A new plate was
fabricated to replace the sample holder, since on the sample holder the terminal for the
igniter wire is in the downstream direction. In order to visualize the flow correctly, the

wire needs to be on the upstream direction. This setup is shown in Figure 152.

 

Figure 152. The smoke wire setup. The wire heats up and warms the grease

to vaporize it into visible smoke.

161

Figures 153 to 155 show the experimental snap shots. The smoke lines are not
perfectly straight, but this does not completely represent the streamlines of the oxidizer
flow inside the test section. There are several factors that caused the smoke lines not to be
perfectly straight. The first is the fact that the wire width is 10 centimeter less than the
plate width (5 centimeter offset on each side). This will allow some of the smoke to fill
the ‘empty’ space downstream of the area with no wire. The second reason is the wire
itself and the terminal creates an obstruction on the flow field and therefore affects the
smoke lines. Note that during the actual experiment, the igniter wire will be placed on the
downstream side, therefore any obstruction or perturbation to the flow will occur past the
sample and will not alter the flow over the sample (unlike the case here in the smoke

visualization).

 

Figure 153. The smoke lines when the flow is initiated. The smoke wire is at

the right and the flow is right-to-left.

162

 

Figure 154. The smoke lines when the flow is in a nearly steady state.

 

Figure 155. The smoke lines when the flow reaches a steady state.

163

There are no observable disruptions of the flow such as recirculation or separation
that will significantly affect the flamelet behavior, therefore the experiment can be

conducted according to the established theory of flame spread.

164

Chapter 12

Alternative F ucl Test

All of the fuel samples tested so far are pure cellulosic. This research also
includes series of tests where a fuel sample developed in the NASA Glenn by a research
group named SIBAL (Solid Inflammability fioundary At Low-Speed). The above-
mentioned file] sample is a 50-50 composite between fabric and fiberglass, with an area
density of 0.01805 g/cmz. The cotton is spun around each fiberglass thread core and the
resulting composite thread is woven into a fabric. This fuel has a thickness in the order of
1 millimeter. This fuel will be referred as the SIBAL fuel sample. The experimental
procedure will be identical to the procedure of testing cellulose samples.

Figure 156 to 15 8 show the experimental results. Interestingly, for this type of
fuel sample, flamelets are never formed. If the oxidizer flow is turned down too far, the
flame will extinguish. Turning the oxidizer flow at the minimum flow without
extinguishing the flame will result in a flame front traveling in the longitudinal direction
of the sample, consuming part of the fabric. As soon as the flame front reaches the end of
the sample, it turns into smolder. With cellulosic fuel sample such as Whatman 44, the
smolder usually consumes what is left of the sample and extinguish shortly after that. For
the SIBAL fuel, the smolder actually forms a smolder front and travel backward
(upstream) consuming the rest of the fabric that was not consumed by the flame front.

The above-discussed phenomenon is consistent with the experiments using
thicker fuel sample such as 17CHR (‘Ramp Up Test’ section). With the thicker cellulosic

fuel sample, the flamelet and flame front are becoming more alike and for the 17CHR

165

fuel, the flamelet physically resembles a small flame front. The flame front on the SIBAL
fuel initially consumes the cotton part since it has less density and is more combustible
than the fiberglass. When the smolder is formed and traveled upstream, it consumes the

leftover fabric. Note that Figure 158 shows the fiberglass not consumed by the flame.

 

Figure 156. Flame front after ignition for the SIBAL fuel sample.
Flame travel is from the left to the right, and oxidizer flow from the right to
the left. Note that the image above seems bit low resolution. That is not due to the
low resolution of the camera, instead it is showing the woven pattern of the fabric on

the SIBAL fuel sample.

166

Flame Front Consuming Fabric

 

Figure 157. Flame front consumes the cotton. The black area that is traveled

by the flame front is the unburned fiberglass.

Smoldering front consumes the
rest of the fabric that was not
previously consumed by the

flame front.

 

Figure 158. Smolder front consumes the fiberglass after all cotton is

consumed.

167

Chapter 13

Conclusions and Suggestions for Future Work

The experiments involving flame spread over a solid thin fuel were initially
performed at NASA’s “Zero-g” drop facility. The research at MSU required the
fabrication of the Simulated Micro-gravity Flame Tunnel (SMFT), which was capable of
simulating micro-gravity flame spread over a thin fuel, similar to the results in the drop
facility. It is important to note that the SMFT simulated the micro-gravity conditions by
suppressing buoyancy inside the test section. The SMFT has the capability to perform
tests in the normal position (right side up) and in the inverted position (upside down).
Using fluid mechanical principles, the oxidizer flow inside the SMFT’s test section is
regulated to produce different ranges of velocities what will create instabilities to the
flame. A programmable mass flow controller is part of the flow system, where the set
velocities, acceleration rates, and deceleration rates can be programmed to give an
accurate flow output. Hot wire probe and smoke wire tests were performed to determine
the uniformity and the condition of the flow inside the test section.

All of the data (experimental results) were collected according to the determined
experimental procedure. These results have different final velocity ranges and different
flame stability outcomes. They are grouped into different regimes. Flammability maps
were created for both the normal and inverted positions. It is evident fi'om the
flammability map that the flame stability in an opposed-flow setup is strongly dependent

on the oxidizer flow rate and the gap spacing of the tlmnel test section. The flammability

168

map is useful for future researchers since the stability of the flame is easily visualized
from such maps.

The ramp down tests prove that for the deceleration rate ranges used in the SMFT,
the initial flamelet formation count is dependent of the deceleration rate. This experiment
also shows that the steady state flamelet formation count is bounded in a consistent range.

The ramp up tests demonstrate that for the same test conditions, flamelets travel at
a slower rate than the flame front. However, when the velocity ratio between the flame
spread velocity and the oxidizer flow velocity is computed, the flamelets have a larger
ratio than the flame front. In other words, relative to the oxidizer flow velocity, the
flamelet travels faster than the flame fi'ont. This experiment also shows that as the
acceleration rate of the oxidizer flow is increased, the time constant needed to form a blue
flame front and a full (yellow) flame front is not conclusively inversely proportional to
the acceleration rate. Several cellulosic fuel samples were included in this experiment,
and the experimental results suggest that as the area density of the fuel sample increases,
both the flame spread velocity and the velocity ratio between flame spread and oxidizer
flow monotonically decrease. Also, as the area density of the fuel increases, the flamelets
are harder to form and if they do form, they are very similar physically in appearance to a
flame front.

When the data fiom the ramp up test was analyzed, it was shown that for a steady
state test condition, the flame position and time is reasonably linearly proportional,
regardless of flame fi'om (i.e., flamelet or flame front). Also, for an oxidizer velocity
range near the instability range of the flame, the flame spread velocity is reasonably

linearly proportional to the oxidizer flow velocity. This is not the case when the oxidizer

169

flow velocity is increased to a range that is ‘far’ from the instability range. For this
condition, and also based on previous research by Olson (2001) and Femandez-Pello
(1981), as the oxidizer flow velocity is increased, the flame spread velocity becomes
nearly constant, then decreases as the oxidizer flow is increased fiirther.

There are several shortcomings of the SMF T. The first is the sample holder that is
not stifl' enough. As a result, a great amount of effort is required to create a uniform gap
spacing in the test section during test preparation. A new sample holder from a thicker
material can be made so that the sample holder is always guaranteed to be flat at all
locations. The second shortcoming of the SMFT is the weight factor when inverting the
device upside down or right side up from the other positions. A structure or a linkage
system can be fabricated to ease this process, therefore improving laboratory safety.

There are no complete data for the flamelet size analysis at the moment. Some
analyses of the flamelet size can be performed in the future. The flamelet size can be
plotted as a function of the oxidizer flow velocity, gap spacing, and oxygen concentration
(another variable that can be introduced in future tests, since all of the test performed so
far use air for oxidizer).

A more complete analysis of the heat transfer mechanism involved can be made
in the future, and accurate surface temperature data can be gathered using an infra-red
(IR) camera. This has been done in the past at the NASA Glenn Research Center, so a
similar system can be adapted for the MSU apparatus.

So far all of the flamelet velocity analyses have been done manually as described

in the “Ramp Up Test” section. There is a software called TRACKER and SPOTLIGHT

170

that will do this analysis mechanically. The use of this software will greatly increase the
accuracy and quality of the experimental results.

The SMFT is a valuable addition to the ATHINA project since it will simulate a
combustion process and flame spread under a reduced gravity environment. The data
from the SMFT experiments can aid NASA in future studies on the micro-gravity

combustion and flame spread over solid thin (and possibly thick) related fuels.

171

 

Appendices

172

Appendix A

Ramp Down Test - Raw Data

 

 

 

 

Initial Flamemet Count vs. Deceleration Rate 1

16 ’

)1 g 14 —»-~~ -~‘w—-~A-~— » 7 7 ~ » - ~ 1
5 3 12 A A—--A—— - — — — L
'1 0 A A I
‘ *510 ~ ~ A , — ~ 7 A 7 77 1|
'5 ‘ 1A3I‘Tlfl‘l 1|

[ E 8» A—n— ~ ~ — 7 . 1:
.1, 1A3mmnew41

l u—_ 6 . ‘ 7‘.-. _ - ~ 77
i a 4 L- 7 7‘ 7 g 7 7 7 7 7 7 7 7 7 7 l
1 g 1
1 _. 2 .- ~ — — - ~— -» - —— _ —~ 7 ~ 1
1 o I 1
1 0 5 10 15 20 I
1 Deceleration (cmls2) I
1 1

Figure Al. 3mm Gap Spacing Initial Flamelet Count vs. Deceleration Rate

‘ ‘ ‘ ’ 1
Initial STEADY Flamemet Count vs. Deceleration Rate I
l

|
|

 

 

 

 

16 I

1 ‘g’ 14 ~1~—-- MFA 7 A» 7 ~- 99 - l
l 8 12 fin—r ‘ I— 97..“ ‘*#—' 7;“ Vii? Vii—A # A— 1
I § 10 9%: 7- A ‘9 E‘em"m““fl1
i E 2 7,55: ,7 7_"'_..;.,‘.77 Modem
1 E A A l
1 a 4 " ‘ — * 1
s 2 - 7 ~ 7 1
| 0 ' ‘ I
0 5 10 15 20 1

Deceleration (cmlsZ)

Figure A2. 3mm Gap Spacing Initial Steady Flamelet Count vs. Deceleration Rate

173

 

 

Steady-State Flame let Count vs. Decele ration i

 

 

 

 

|
16 '
14 - eeeeeeee A ~ A 7 7 7 A; 7 ,7 1
15 A A A 1
=12 A A 7 7 ~ 7 7 w
1 o A A 1
1 o 10 -A A -7 7 ,, 7 7 7 5 777-7
‘6 A 1A3mm 1,
1 '3 8 - 7 A .7 ~ 7 ~~~~ 7. 7 7 7 *7 .7 ~ 1 '1
| E A 1A3mm new11
6 . 7 7 7 7 7, .7 77 7 .7 777 7-7.. .
1' 3 1
IL
1 g 4 '9 T " ’ ’ T ” 1
2 -. — 7 — 7 7 —- 7 —— 7 f 1
I 0 1 1
1 0 5 10 15 20 1
1 Deceleration (cmls2)

Figure A3. 3mm Gap Spacing Steady State Flamelet Count vs. Deceleration Rate

 

 

 

.__ 77...... .777. 777—..—

Initial Flamemet Count vs. Deceleration Rate 1
1
1

 

 

 

 

‘ 16
1 g 14 — 7 — 4— — — 1
1 8 12 7- .7 7- 77-.. 7 7.. 7 1
1 s- 10 ..7 -7 -77 .77-. ,7. -7 7 .7- 7.. 7 7 .7, _7. _7 -71
1 % a l 1I5mm H
. E 8 . 7 a 7 7 7 7 g . 1'
1 a la a 1.5mm new 1
1 E 6 . I ~ I - 7 ~: , ~—»-- 7., 77 ~—- 777..
1 E 4 —~ l~-I~ ""1.” 7 a 1
+ E 2 — ----- ~ .1 - — ~— I» ~ ‘
5 0 5 10 15 ‘
1 Deceleration (chZ) 1

7 . 77 7 7 7 77.1

Figure A4. 5mm Gap Spacing Initial Flamelet Count vs. Deceleration Rate

174

Initial STEADY Flamemet Count vs. Deceleration Rate 1

 

 

 

 

16 1
1§14 777 77777777 777 1
01277777 ~777 7- 777 1
0" 7 7 7 .77 7 7 7 . 1 7 7 1 7 7 7 ,7 77 .7 .7
1 % 1g f at 1.5mm 11
«En ' 1‘“ 7 ‘ a 5mm new 1
E 6 .. id 77 ..7 :: 7. 7 7 7. . 5 7 7. 7 7 7 7 7 . 1
E 4 " 7 II I 7 7 n 7 s 7 7 1
in! $4 ‘
'5 2 7 5 7 . —
0
1 0 5 10 15

Deceleration (cmls2)

Figure A5. 5mm Gap Spacing Initial Steady Flamelet Count vs. Deceleration Rate

Steady-State Flamelet Count vs. Deceleration

 

 

 

 

1
1

1 16 1
1 14 7 7 7 77 7 7 7 1
1 g 12 7 7 7 7 7 7 1
o 1
o 10 .7 77 7 7 a 7 77 7 .7 7 . .77 .777 .1
1'3 I a 1.5mm 11
g 8 7': .1, :11", 7 IW 7 1l5mm new11

1 I_' 6 7 m 7 , 7 fl 7 I 7 7' ' 77, W 1

u.

2 7 77 7 1

0 1

0 5 10 15

1
Deceleration (chZ) 1

Figure A6. 5mm Gap Spacing Steady State Flamelet Count vs. Deceleration Rate

175

Appendix B

Ramp Down Test — Weighted Average

 

.54.;

N#O’
J
J
J
J
J
J

».: — J “A A 3mm J
__ , -__. ,. J A 3mm new "3
'7‘" AVG. -.. J:

J i

..s
O

O)

 

lnltlal Flamelet Count
A on

 

 

 

Deceleration (ch32) J

Figure B1. Weighted Average 3mm Gap Spacing Initial Flamelet Count vs.

Deceleration Rate

 

 

 

 

 

J Initial STEADY Flamemet: Count vs. Deceleration Rate J

 

 

 

 

 

16 W ~m ”— — J
E 14 J ~ ~~ ‘ -- _ _g I

J 3
J o 12 J WW __J
J 16' 1o , 'J A 3mm JJ
. ‘5’ 3 J A 3mmnewJJ
‘ a 6 JTAVQ _J1
J i" 4 J
J 5 2‘ *w 7 J
J 0 i J
0 5 1O 15 20 J

J Deceleration (cmlsZ) |

 

Figure BZ. Weighted Average 3mm Gap Spacing Initial Steady Flamelet Count vs.

Deceleration Rate

176

SS Flamelet Count

Figure B3. Weighted Average 3mm Gap Spacing Steady State Flamelet Count vs.

Initial Flamelet Count

Steady-State Flamelets Count vs. Deceleration

 

 

l

 

 

 

Deceleration (cm/$2)

Deceleration Rate

Initial Flamemeu Count vs. Deceleration Rate

 

 

l

 

 

:- I
7 7 I 7 7
7 77 7 ‘j 7 E 77
5 10 15

Deceleration (cm/52)

*7 AVG.

I 5mm ‘5

A 3mm g,

A 3mm new \

a 5mm new H
AVG ,,,,, ll

»
l
l
l
l

Figure B4. Weighted Average 5mm Gap Spacing Initial Flamelet Count vs.

Deceleration Rate

177

Initial STEADY Hamemets Count vs. Deceleration Rate \

l

 

 

 

 

‘ § 15 ~ 7 ~ 7 ~ ~ 7 fl l
‘ l
. § 10 , 7 ; ' 5mm M
“E, " " I 5mmnewM
‘ E 5 , \ ' 7 , , l AVG, ,, in
_ I :3 ‘
.2 a.
= l
E o 1
o 5 1o 15 ;

Deceleration (cm/32)

Figure BS. Weighted Average 5mm Gap Spacing Initial Steady Flamelet Count vs.

Deceleration Rate

Steady-State Flamelets Count vs. Deceleration

 

 

 

 

‘ 12 l‘

l '5 1o 7 ,7 7,, .- 7, fl l

l 3 “ l l 5mm ‘;

l o a 7 .- .7 l l

.1 S _4, I l l 5mmnewl

‘ E 6 7, 7, 7 7 I 7, 1 i'AVGW U
I!

1 E 4 ’ ’ ’ ‘

3 2 7 7 _7 7 1

l

0 l

0 5 10 15 '

Deceleration (cm/32)

Figure B6. Weighted Average 5mm Gap Spacing Steady State Flamelet Count vs.

Deceleration Rate

178

Appendix C

Ramp Down Test — Normalized Average

 

 

 

 

Initial Flamemet: Count vs. Deceleration Rate
l
2 l
‘E 13 7, WW W WW WW WW W WW W W—
l 3 1.6 WW W ~ 7 7. WW-W-W W— W WW ., --
o A 1
U 1.4 J- 77 .7. .-7. -7 7 77 _V 7
J g 4 ... A 1-47 ...... , ». A-.._- _ A .1
a 0.84 7 -WJ»‘ ~—W- -- W A WWWWW 7 WW ‘ " “ i
l 70 0.6 L t - i 7 W 7 7 % l
l g 0.4 — » WWW — W 7
S 0.2 4r — - 7 ,7 7 77 _7 7 7 7 >
0
O 5 10 15 20 t
DeceleratioMchsZ) l

Figure Cl. Normalized Average 3mm Gap Spacing Initial Flamelet Count vs.

Deceleration Rate

 

Initial STEADY Flamemet: Count vs. Deceleration Rate

 

 

 

 

l
l
l 2 ‘
E

! 8 15 ‘ 77777 t 777‘ _ 777777777 7 7 ‘
l 1‘ ‘ A ‘ — l
l g 1 71... A7--- . - .3...
r g A A A A l
i '2? °5
5 l

l 0 .
‘ 0 5 10 15 20 l
Deceleration (cmlsZ) f
777777777 _ 7- .77 _ ._l

Figure C2. Normalized Average 3mm Gap Spacing Initial Steady Flamelet Count

vs. Deceleration Rate

179

 

 

Steady-State Flamelets Count vs. Deceleration

 

 

 

 

2 l
'g', l
1 8 1.5 W W W _7 W 7
z z; . . . A . W- *
‘ T1 1 7_ ‘ 4 ~—~—- -- 7 —~ ~ A3mml1
r. s A t 7 g ,2
1 g A A
: ,,, 0.5 W W W 7 I
a)
0 1 l
0 5 10 15 20 ‘
Deceleration (chsZ) l

Figure C3. Normalized Average 3mm Gap Spacing Steady State Flamelet Count vs.

Deceleration Rate

lnItlsI Flamemets Count vs. Deceleration Rate l

 

 

 

 

 

 

2
1 'g' 1.8 , W WIW WWW ~. . I 7
; o1.64»~ WWIWWW W WW 7 1
u 9.. 1.4 7 ~ I — _ I~-——— W W ,
.2 1.2 I — WIW— I W WW W 7
g 1* -------—- - — —— ~~—-—-~ — il5mmJ
.2 0.8 « WWI WWI -- _7 I l
I g 0.. W 7
1 a 0.4 -77 - — _ I W I W 1
1 § 0.2 W WW WWW W W W WWW I
: " 0 .
l 0 5 10 15 .
5 Deceleration (chZ) l

 

Figure C4. Normalized Average 5mm Gap Spacing Initial Flamelet Count vs.

Deceleration Rate

180

 

Initial Flamelet Count

Figure C5. Normalized Average 5mm Gap Spacing Initial Steady Flamelet Count

88 Flamelet Count

Figure C6. Normalized Average 5mm Gap Spacing Steady State Flamelet Count vs.

Initial STEADY Flamemet: Count vs. Deceleration Rate

 

 

 

 

2
1.8 — W W W — _ _ _ _
1.6 WW W W W r - WWW W W W .7 7
1.4 WWW-WW W- W IW W W W WW _ .
1.2 I 7IW 7 V, r
1 _ M I WI 7 I 7, _ I,,, f
0.8 IL I -__. : “I .
0.6 —-- ~ W WWW _ I 7 z ,
0.4 1 7.7 . 7 , , . A. 7 77. on- . n 7 ,
0.2 7-- 7 - 7 7 -_ , _ _
0 .
0 5 10
Deceleration (cm/32)

 

vs. Deceleration Rate

 

Steady-State Flame lets Count vs. Decele ration

 

 

 

 

2
1.5» W WW WW W W W
I
I I
1 WI IW , WW ~ [WW—
I I I I
I I
0.5 W WW W W
o
o 5 1o

Deceleration (cm/32)

Deceleration Rate

181

15

15

 

Appendix D

Ramp Up Test — Flame Spread Velocity

f i 7 _ 7 7 , , i

3mm Ramp Up 2.1chl32 <

 

Flame Spread (cmls)

 

 

 

010203040506070809010111213 I
0 0 0 0

Time (a)

Figure D1. Flame Spread Velocity Graph - 3mm Gap Spacings,
2.16 cm/s2 Acceleration Rate

3mm Ramp Lb 5.05cmlaz

 

0.4000 40
0.3500 ’ 35

g 0.3000 , 30 ‘—+;V1

3 0.2500 25 ‘ +V2
{ 3 0.2000 20 ‘_._v3
‘ t72.0.1500 15 —AVG
1 a 0.1000 10 :—'—F|0W11
p 0.0500 5

 

 

 

 

‘ 0.0000 ‘ _ o
‘ 0 20 40 60 80 100 120140 160

.‘ Time (a) }
\7, 7, , 7 ,, ,WV *7
Figure D2. Flame Spread Velocity Graph - 3mm Gap Spacings,

5.05 cm/s2 Acceleration Rate

182

3mm Ramp Up 7.58cmls2

 

0.4000 40
0.3500
0.3000
0.2500
0.2000 »

0.1500

Flame Spread (cmls)

0.1000

 

0.0500

 

 

 

0.0000 . O
0 20 40 60 80 100 120 140 '
Tlmo (a)

Figure D3. Flame Spread Velocity Graph - 3mm Gap Spacings,
7.58 cm/s2 Acceleration Rate

3mm Ramp Up 10.1ml32

 

\

0.4000 40 ‘

0.3500 35 ‘

0.3000 30 ,. l
, ‘ —e—V1

~ 25 +V2 ‘ w

20 .+V3 i

Flame Spread (cmls)
O O
N N
0 0|
0 O
o o

 

 

0.1500 15 ._AVG;‘
0.1000 10 —Flpiwr‘
0.0500 < 5

0.0000 0 ‘

 

 

Time (s) i

Figure D4. Flame Spread Velocity Graph - 3mm Gap Spacings,
10.1 cm/s2 Acceleration Rate

183

 

‘ 3mm Ramp Up 15.15cmlsz

 

‘ 0.4000 40
1 0.3500
‘ 0.3000
0.2500
0.2000

0.1500

Flame Spread (cmls)

0.1000

0.0500

 

 

0.0000

 

01020304050607080901011121314
00

Tlme (e)

Figure D5. Flame Spread Velocity Graph - 3mm Gap Spacings,
15.15 cm/s2 Acceleration Rate

5mm Ramp Up 2.20mle2

 

 

 

 

 

 

0.4500 30
0.4000
0.3500 ,.
'3' 0.3000 20 +V1
0
3 0.2500 V2
2 1+V3
Q 1
a, 0.2000 ‘
0.1500 10 .
5 1—F|OW“
0.1000
0.0500
0.0000 . , . o
0 10 20 30 40 50 60 7o 00 90 100110120

Tlme (s)

Figure D6. Flame Spread Velocity Graph - 5mm Gap Spacings,
2.2 cm/s2 Acceleration Rate

184

5mm Ramp Up 4.84chsZ

 

 

 

 

 

0.4000 40

0.3500 35
g0.3000 30 1+V1
3 0.2500 25 +V2
202000 , 20 +V3
20.1500 15 _AVG:
5 0.1000 10 —F|ow

0.0500 . -. 5

0.0000 , 0

0 20 4o 60 80 100 120
Time(s)

Figure D7. Flame Spread Velocity Graph - 5mm Gap Spacings,
4.84 cm/s2 Acceleration Rate

5mm Ramp Up 8.07cmlsZ ‘

0.5000 30 1
0.5500
0.5000
0.4500
0.4000
0.3500
0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000

 

Flame Spread (cm/e)

 

 

 

 

 

0 10 20 30 40 50 60 7O 80 90100110120
Time(s)

Figure D8. Flame Spread Velocity Graph - 5mm Gap Spacings,
1.07 cm/s2 Acceleration Rate

185

1 5mm Ramp Up 12.1cm/32

 

 

0.4000 40
0.3500 35
: 0.3000 . 30
j ‘3‘
1 E 0.2500 25
; z
1 C
3 g 0.2000 < 20
i m
E 0.1500 15
E
1 0.1000 10
1
0.0500 , 5
1 0.0000 .

 

 

 

‘ 0 10 20 30 4O 50 60 70 80 90 100110120
‘ Time(s)

+V1

+V3
L—AVG ‘1
—Flow

Figure D9. Flame Spread Velocity Graph - 5mm Gap Spacings,

12.1 cm/s2 Acceleration Rate

186

Appendix E

Ramp Up Test — Velocity Ratio

Flame Spread I Flow Velocity Ratio ‘
3mm Ramp Up 2.16ch32 1

 

1 0.025
0.02

0.015

Ratio

0.01

0.005

 

 

 

 

O 20 40 60 80 1 00 120
1 Time (s)

Figure E1. Velocity Ratio Graph - 3mm Gap Spacings,
2.16 cm/s2 Acceleration Rate

Flame Spread I Flow Velocity Ratio
3mm Ramp Up 5.05ch32

 

Ratio

 

 

 

 

Figure E2. Velocity Ratio Graph - 3mm Gap Spacings,

5.05 cm/s2 Acceleration Rate

187

Flame Spread I Flow Velocity Ratio
3mm Ramp Up 158ch32

 

 

 

 

 

Time (s)
Figure E3. Velocity Ratio Graph - 3mm Gap Spacings,
7.58 cm/s2 Acceleration Rate
1 Flame Spread I Flow Velocity Ratio
3mm Ramp Up 10.1cmlsz

 

 

 

 

 

Time (s)

Figure E4. Velocity Ratio Graph - 3mm Gap Spacings,
10.1 cm/s2 Acceleration Rate

188

7* "1

Ratio

Ratio

Flame Spread I Flow Velocity Ratio
3mm Ramp Up 15.15cmlsZ

 

 

 

 

 

Time (s)

Figure E5. Velocity Ratio Graph - 3mm Gap Spacings,
15.15 cm/sz Acceleration Rate
Flame Spread I Flow Velocity Ratio
5mm Ramp Up 2.2cmlsz

 

0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01

 

 

 

 

 

Time (s)

Figure E6. Velocity Ratio Graph - 5mm Gap Spacings,
2.2cm/s2 Acceleration Rate

189

0 20 40 so 80 100 120 140 3

0 20 4o 60 80 100 120 140 1

Flame Spread I Flow Velocity Ratio
5mm Ramp Up 434ch32

 

 

 

 

 

0 20 40 60 80 100 120
Time (s) 1

Af_ .i .__ ,7. ._. .t ;,,1

Figure E7. Velocity Ratio Graph - 5mm Gap Spacings,
4.84cm/s2 Acceleration Rate

Flame Spread I Flow Velocity Ratio 1

5mm Ramp Up 8.07cmlsz

 

Ratio

 

 

 

 

0 20 4o 60 80 100 120 1
Time (s)

Figure E8. Velocity Ratio Graph - 5mm Gap Spacings,
8.07cm/s2 Acceleration Rate

190

 

Flame Spread I Flow Velocity Ratio
5mm Ramp Up 12.1ch32

 

 

 

 

 

 

0 20 40 60 80 100 1 20

Figure E9. Velocity Ratio Graph - 5mm Gap Spacings,
12.1cm/s2 Acceleration Rate

191

REFERENCES

Cengel, Y.A., 2003, “_Iiegt Transfer : A practical Approa_c_h”, McGraw-Hill, New York.

Femandez-Pello, A.C., Ray, SR, and Glassman, 1., 1982, “Flame Spread in an Opposed
Forced Flow: The Effect of Ambient Oxygen Concentration”, Eighteen Symposium
(International) on Combustion, 1981, pp. 579-587.

Incropera, RP. and DeWitt, DP, 1996, “Introduction to Heat Transfer”, Wiley, New
York.

Olson, S.L., 1990, “Mechanism of Microgravity Flame Spread over a Thin Solid Fuel:
Oxygen and Opposed Flow Effects”, Combustion Science and Technology, 1991, Vol.76,
pp. 233-249.

Oravecz, L.M., 2001, “Instabilities of Spreading Diffusion Flames in Microgravity and
the Design and Construction of a Hele-Shaw Apparatus That Produces Flames in the
Near Extinction Limit Regime Under Simulated Low Gravity Conditions”, M.S. Thesis,
Michigan State University, Department of Mechanical Engineering.

Potter, MC. and Wiggert, DC, 1997, “Mechanics of Fluid”, Prentice-Hall, New Jersey.

Schlichting, H., 1960, “Boundary Layer Theory”, Boundary Layer Theory, McGraw-Hill,
New York.

Wichman, LS. and Williams, F .A., 1982, “A Simplified Model of Flame Spread in an
Opposed Flow along a Flat Surface of a Semi-infinite Solid”, Combustion Science and
Technology, 1983, Vol. 32, pp. 91-123.

Wichman, 1.8. and Williams, F.A., 1983, “Comments on Rates of Creeping Spread of
Flames Over Thermally Thin Fuels”, Combustion Science and Technology, 1983, Vol.
33, pp. 207-214.

Zik, O., Olami, Z, and Moses, E., 1997, “Fingering Instability in Combustion”, The

Weizmann Institute of Science, Department of Physics of Complex System and
Department of Chemical Physics.

192

 

   

11311111111111111111111