CONSTRUCTION OF A HELE-SHAW COMBUSTION APPARATUS
AND EXAMINATION OF LOW-FLOW NEAR-EXTINCTION REGIME
USING HIGH SPEED VIDEOGRAPHY
By
Jeffrey Stricker

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Mechanical Engineering
2012

ABSTRACT
CONSTRUCTION OF A HELE-SHAW COMBUSTION APPARATUS
AND EXAMINATION OF LOW-FLOW NEAR-EXTINCTION REGIME
USING HIGH SPEED VIDEOGRAPHY
By
Jeffrey Stricker
To ensure the safety of astronauts aboard the Space Shuttle and, more recently, the
International Space Station, the propagation of fire in microgravity environments must be well
understood, as it would be significantly different from fires on Earth.

Thus, the National

Aeronautics and Space Administration (NASA) Glenn Research Center in Cleveland, Ohio
studies micro-gravity flame spread through its Analysis of Thermo-diffusive and Hydrodynamic
Instabilities in Near-extinction Atmospheres (ATHINA) initiative. The facility has a drop tower,
which can be used to demonstrate flame propagation in free-fall, but the drop tests are expensive
and limited to the two to five second duration of the free-fall. To simulate a microgravity
environment without using the drop tower, thus significantly extending the test time from a few
seconds to several minutes, a device called the Narrow Channel Apparatus (NCA) was
developed at MSU. In the NCA a thin solid fuel is burned in a channel with a vertical height
small enough to suppress buoyant effects. In essence, the NCA allows us to study microgravity
flame behavior that is anticipated in space using a test apparatus on Earth under 1-g conditions.
The Michigan State University department of Mechanical Engineering is currently running its
third iteration of the NCA.

ACKNOWLEDGEMENTS

First of all, I wish to thank my advisor, Dr. Indrek Wichman, for providing me access to
his extensive references on combustion and for his continued guidance on this project.
I wish to thank Dr. Sandra Olson and her colleagues at the NASA Glenn Research Center
for their assistance in familiarizing me with the experimental apparatus and data analysis portion
of this project. I also wish to thank Dr. Fletcher Miller and his students, Karen Hung and Jacob
Pepper, for their support in this experiment.
Finally, I must thank my parents for the financial and moral support that they have
provided me throughout my graduate school experience.

iii

TABLE OF CONTENTS

List of Tables ................................................................................................................................. vi
List of Figures ............................................................................................................................... vii
Chapter 1 – Introduction ..................................................................................................................1
Chapter 2 – Literature Review .........................................................................................................3
Chapter 3 – Experimental Setup ......................................................................................................7
Overview ..................................................................................................................................... 7
Test Section................................................................................................................................. 8
Plenum Chamber Pressure Relief ............................................................................................... 9
Flow Control ............................................................................................................................... 9
Sample Ignition ......................................................................................................................... 10
Data Collection ......................................................................................................................... 10
Apparatus Photos ...................................................................................................................... 12
Chapter 4 – Procedure....................................................................................................................18
Flow Controller Setup ............................................................................................................... 18
Channel Setup ........................................................................................................................... 18
Camera Setup ............................................................................................................................ 19
Running Test ............................................................................................................................. 19
Post-Test ................................................................................................................................... 20
Data Analysis ............................................................................................................................ 20
Chapter 5 – Verification and Calibration Tests .............................................................................22
Chapter 6 – Constant Flow Velocity burn Tests ............................................................................34
Chapter 7 – Fast Ramp Down Burn Tests .....................................................................................42
Chapter 8 – Slow Ramp Down Burn Tests ....................................................................................58
Chapter 9 – Lessons Learned .........................................................................................................81
Flow System ............................................................................................................................. 81
Sample Holder .......................................................................................................................... 81
Sample Ignition ......................................................................................................................... 82
Preliminary Apparatus Photos .................................................................................................. 83
Chapter 10 – Conclusions and Future Work ..................................................................................84
Appendices.....................................................................................................................................87

iv

Appendix A – Stoichiometry Analysis ..................................................................................... 88
Appendix B – Pressure Relief Calculation ............................................................................... 91
Appendix C – Test Data............................................................................................................ 92
References ....................................................................................................................................129

v

LIST OF TABLES

Table 1. Flow Rate Test: Time Elapsed to Displace 1 L of Water ............................................... 23
Table 2. Overview of Test Data .................................................................................................... 92
Table 3. Flame Behavior ............................................................................................................... 94

vi

LIST OF FIGURES

Figure 1. Full View of Apparatus ................................................................................................. 12
Figure 2. View of Plenum Chamber with Top Removed, Highlighting Distribution Pipe and
Flow Guide.................................................................................................................................... 13
Figure 3. View of Test Section Fully Assembled ......................................................................... 13
Figure 4. View of Test Section with Top Window Removed....................................................... 14
Figure 5. View of Test Section with Top Plate Removed ............................................................ 14
Figure 6. Close-up View of Test Section with Sample Holder Removed .................................... 15
Figure 7. Close-up View of Sample Holder Guide Pins ............................................................... 15
Figure 8. View of Pressure Relief System .................................................................................... 16
Figure 9. Close-up View of Igniter Wire and Tensioner .............................................................. 17
Figure 10. Full View of Test Setup, Highlighting Cameras and Mirror ....................................... 17
Figure 11. Water Displacement Test Setup .................................................................................. 23
Figure 12. Hot Wire Calibration, 2011-10-24 .............................................................................. 24
Figure 13. Hot Wire Calibration, 2011-10-25 .............................................................................. 24
Figure 14. Channel Exit Flow Velocities, 2011-10-25: z = 50 mm.............................................. 25
Figure 15. Channel Exit Flow Velocities, 2011-10-25: z = 0 mm................................................ 26
Figure 16. Channel Exit Flow Velocities, 2011-10-25: z = -50 mm ............................................ 26
Figure 17. Hot Wire Calibration, 2011-11-11 .............................................................................. 27
Figure 18. Channel Interior Flow Velocities, 2011-11-11: z = 50 mm ........................................ 28
Figure 19. Channel Interior Flow Velocities, 2011-11-11: z = 0 mm .......................................... 28
Figure 20. Channel Interior Flow Velocities, 2011-11-11: z = -50 mm ....................................... 29

vii

Figure 21. Channel Interior Flow Velocities, 2011-11-18: z = 80 mm ........................................ 29
Figure 22. Channel Interior Flow Velocities, 2011-11-18: z = 0 mm .......................................... 30
Figure 23. Channel Interior Flow Velocities, 2011-11-18: z = -80 mm ....................................... 30
Figure 24. Hot Wire Calibration, 2011-11-21 .............................................................................. 31
Figure 25. Channel Interior Flow Velocities, 2011-11-21: z = 80 mm ........................................ 31
Figure 26. Channel Interior Flow Velocities, 2011-11-21: z = 0 mm .......................................... 32
Figure 27. Channel Interior Flow Velocities, 2011-11-21: z = -80 mm ....................................... 32
Figure 28. Test 2012-04-18_2, 70s, Flame Front (25 cm/s) ......................................................... 35
Figure 29. Test 2012-04-18_2, 180s, Flame Front (25 cm/s) ....................................................... 36
Figure 30. Test 2012-04-18_2, Burned Sample ............................................................................ 36
Figure 31. Test 2012-04-18_3, 70s, Flame Front (25 cm/s) ......................................................... 37
Figure 32. Test 2012-04-18_3, 170s, Flame Front (25 cm/s) ....................................................... 37
Figure 33. Test 2012-04-18_3, Burned Sample ............................................................................ 38
Figure 34. Test 2012-04-18_4, 75s, Flame Front (15 cm/s) ......................................................... 38
Figure 35. Test 2012-04-18_4, 200s, Flame Front (15 cm/s) ....................................................... 39
Figure 36. Test 2012-04-18_4, Burned Sample ............................................................................ 39
Figure 37. Test 2012-04-18_5, 70s, Flame Front (15 cm/s) ......................................................... 40
Figure 38. Test 2012-04-18_5, 180s, Flame Front (15 cm/s) ....................................................... 40
Figure 39. Test 2012-04-18_5, Burned Sample ............................................................................ 41
Figure 40. Test 2012-04-19_1, 70s, Flame Front (25 cm/s) ......................................................... 44
Figure 41. Test 2012-04-19_1, 190s, Weak Flame Front (10 cm/s) ............................................. 45
Figure 42. Test 2012-04-19_1, Burned Sample ............................................................................ 45
Figure 43. Test 2012-04-19_3, 85s, Flame Front (25 cm/s) ......................................................... 46

viii

Figure 44. Test 2012-04-19_3, 265s, Weak Flame Front (10 cm/s) ............................................. 46
Figure 45. Test 2012-04-19_3, Burned Sample ............................................................................ 47
Figure 46. Test 2012-04-19_4, 80s, Flame Front (25 cm/s) ......................................................... 48
Figure 47. Test 2012-04-19_4, 240s, Weak Flame Front (10 cm/s) ............................................. 48
Figure 48. Test 2012-04-19_4, Burned Sample ............................................................................ 49
Figure 49. Test 2012-04-20_1, 85s, Flame Front (25 cm/s) ......................................................... 50
Figure 50. Test 2012-04-20_1, 225s, Weak Flame Front (9 cm/s) ............................................... 50
Figure 51. Test 2012-04-20_1, Burned Sample ............................................................................ 51
Figure 52. Test 2012-04-20_3, 70s, Flame Front (25 cm/s) ......................................................... 51
Figure 53. Test 2012-04-20_3, 240s, Flamelets (9 cm/s) ............................................................. 52
Figure 54. Test 2012-04-20_3, Burned Sample ............................................................................ 52
Figure 55. Test 2012-04-20_4, 80s, Flame Front (25 cm/s) ......................................................... 53
Figure 56. Test 2012-04-20_4, 230s, Flamelets (9 cm/s) ............................................................. 53
Figure 57. Test 2012-04-20_4, Burned Sample ............................................................................ 54
Figure 58. Test 2012-04-24_1, 80s, Flame Front (25 cm/s) ......................................................... 54
Figure 59. Test 2012-04-24_1, 210s, Flamelets (5 cm/s) ............................................................. 55
Figure 60. Test 2012-04-24_1, Burned Sample ............................................................................ 55
Figure 61. Test 2012-04-19_1 (High Speed Camera), 257s (286s in standard rate video),
Flamelets Extinguishing (10 cm/s) ............................................................................................... 56
Figure 62. Test 2012-04-20_1 (High Speed Camera), 230s (346s in standard rate video),
Flamelets Bifurcating (9 cm/s)...................................................................................................... 56
Figure 63. Test 2012-04-20_3 (High Speed Camera), 273s (300s in standard rate video),
Flamelets Extinguishing (9 cm/s) ................................................................................................. 56
Figure 64. Test 2012-04-24_1 (High Speed Camera), 57s (98s in standard rate video), Flamelets
Bifurcating (17 cm/s) .................................................................................................................... 57

ix

Figure 65. Test 2012-04-26_1, 100s, Flame Front (25 cm/s) ....................................................... 60
Figure 66. Test 2012-04-26_1, 140s, Flamelets (10 cm/s) ........................................................... 61
Figure 67. Test 2012-04-26_1, Burned Sample ............................................................................ 61
Figure 68. Test 2012-04-26_2, 90s, Flame Front (25 cm/s) ......................................................... 62
Figure 69. Test 2012-04-26_2, 120s, Flamelets (10 cm/s) ........................................................... 62
Figure 70. Test 2012-04-26_2, Burned Sample ............................................................................ 63
Figure 71. Test 2012-04-26_3, 185s, Flame Front (25 cm/s) ....................................................... 63
Figure 72. Test 2012-04-26_3, 215s, Flamelets (10 cm/s) ........................................................... 64
Figure 73. Test 2012-04-26_3, Burned Sample ............................................................................ 64
Figure 74. Test 2012-04-26_4, 180s, Flame Front (25 cm/s) ....................................................... 65
Figure 75. Test 2012-04-26_4, 240s, Flamelets (10 cm/s) ........................................................... 65
Figure 76. Test 2012-04-26_4, Burned Sample ............................................................................ 66
Figure 77. Test 2012-04-26_5, 100s, Flame Front (25 cm/s) ....................................................... 66
Figure 78. Test 2012-04-26_5, 160s, Flamelets (10 cm/s) ........................................................... 67
Figure 79. Test 2012-04-26_5, Burned Sample ............................................................................ 67
Figure 80. Test 2012-04-26_6, 10s, Flame Front (25 cm/s) ......................................................... 68
Figure 81. Test 2012-04-26_6, 70s, Flamelets (10 cm/s) ............................................................. 68
Figure 82. Test 2012-04-26_6, Burned Sample ............................................................................ 69
Figure 83. Test 2012-04-26_7, 90s, Flame Front (25 cm/s) ......................................................... 69
Figure 84. Test 2012-04-26_7, 150s, Flamelets (10 cm/s) ........................................................... 70
Figure 85. Test 2012-04-26_7, Burned Sample ............................................................................ 70
Figure 86. Test 2012-04-27_1, 90s, Flame Front (25 cm/s) ......................................................... 71
Figure 87. Test 2012-04-27_1, 240s, Flamelets (10 cm/s) ........................................................... 71

x

Figure 88. Test 2012-04-27_1, Burned Sample ............................................................................ 72
Figure 89. Test 2012-04-27_2, 90s, Flame Front (25 cm/s) ......................................................... 72
Figure 90. Test 2012-04-27_2, 240s, Flame Front (10 cm/s) ....................................................... 73
Figure 91. Test 2012-04-27_2, Burned Sample ............................................................................ 73
Figure 92. Test 2012-04-27_3, 80s, Flame Front (20 cm/s) ......................................................... 74
Figure 93. Test 2012-04-27_3, 180s, Flamelets (10 cm/s) ........................................................... 75
Figure 94. Test 2012-04-27_3, Burned Sample ............................................................................ 75
Figure 95. Test 2012-04-27_4, 80s, Flame Front (20 cm/s) ......................................................... 76
Figure 96. Test 2012-04-27_4, 180s, Flamelets (10 cm/s) ........................................................... 76
Figure 97. Test 2012-04-27_4, Burned Sample ............................................................................ 77
Figure 98. Test 2012-04-27_5, 80s, Flame Front (20 cm/s) ......................................................... 77
Figure 99. Test 2012-04-27_5, 180s, Flamelets (10 cm/s) ........................................................... 78
Figure 100. Test 2012-04-27_5, Burned Sample .......................................................................... 78
Figure 101. Test 2012-04-26_1 (High Speed Camera), 90s (141s in standard rate video),
Flamelet Extinguishing (9 cm/s) ................................................................................................... 79
Figure 102. Test 2012-04-26_6 (High Speed Camera), 125s (83s in standard rate video),
Flamelets Bifurcating (6.75 cm/s)................................................................................................. 79
Figure 103. Test 2012-04-27_2 (High Speed Camera), 117s (164s in standard rate video),
Flamelets Bifurcating (17.6 cm/s)................................................................................................. 79
Figure 104. Test 2012-04-27_4 (High Speed Camera), 234s (257s in standard rate video),
Flamelets Extinguishing (2.3 cm/s) .............................................................................................. 80
Figure 105. View of Plenum Chamber with Top Removed, Highlighting Old Seal Design........ 83
Figure 106. View of Test Section with Top Plate Removed, Highlighting Old Igniter Design ... 83
Figure 107. Steady State Flame Spread Rate vs. Oxidizer Flow Velocity for All Tests .............. 95
Figure 108. Steady State Flame Spread Rate Relative to Oxidizer Flow Velocity vs. Oxidizer
Flow Velocity for All Tests .......................................................................................................... 96

xi

Figure 109. Test 2012-04-09_1 Track .......................................................................................... 97
Figure 110. Test 2012-04-09_2 Track .......................................................................................... 97
Figure 111. Test 2012-04-18_2 Track .......................................................................................... 98
Figure 112. Test 2012-04-18_3 Track .......................................................................................... 98
Figure 113. Test 2012-04-18_4 Track .......................................................................................... 99
Figure 114. Test 2012-04-18_5 Track .......................................................................................... 99
Figure 115. Test 2012-04-19_1 Overall Track ........................................................................... 100
Figure 116. Test 2012-04-19_1 Initial Track.............................................................................. 100
Figure 117. Test 2012-04-19_1 Final Track ............................................................................... 101
Figure 118. Test 2012-04-19_3 Overall Track ........................................................................... 101
Figure 119. Test 2012-04-19_3 Initial Track.............................................................................. 102
Figure 120. Test 2012-04-19_3 Final Track ............................................................................... 102
Figure 121. Test 2012-04-19_4 Overall Track ........................................................................... 103
Figure 122. Test 2012-04-19_4 Initial Track.............................................................................. 103
Figure 123. Test 2012-04-19_4 Final Track ............................................................................... 104
Figure 124. Test 2012-04-20_1 Overall Track ........................................................................... 104
Figure 125. Test 2012-04-20_1 Initial Track.............................................................................. 105
Figure 126. Test 2012-04-20_1 Final Track ............................................................................... 105
Figure 127. Test 2012-04-20_3 Overall Track ........................................................................... 106
Figure 128. Test 2012-04-20_3 Initial Track.............................................................................. 106
Figure 129. Test 2012-04-20_3 Final Track ............................................................................... 107
Figure 130. Test 2012-04-20_4 Overall Track ........................................................................... 107
Figure 131. Test 2012-04-20_4 Initial Track.............................................................................. 108

xii

Figure 132. Test 2012-04-20_4 Final Track ............................................................................... 108
Figure 133. Test 2012-04-24_1 Overall Track ........................................................................... 109
Figure 134. Test 2012-04-24_1 Initial Track.............................................................................. 109
Figure 135. Test 2012-04-24_1 Transient Track ........................................................................ 110
Figure 136. Test 2012-04-24_1 Final Track ............................................................................... 110
Figure 137. Test 2012-04-26_1 Overall Track ........................................................................... 111
Figure 138. Test 2012-04-26_1 Initial Track.............................................................................. 111
Figure 139. Test 2012-04-26_1 Transient Track ........................................................................ 112
Figure 140. Test 2012-04-26_2 Overall Track ........................................................................... 112
Figure 141. Test 2012-04-26_2 Initial Track.............................................................................. 113
Figure 142. Test 2012-04-26_2 Transient Track ........................................................................ 113
Figure 143. Test 2012-04-26_3 Overall Track ........................................................................... 114
Figure 144. Test 2012-04-26_3 Initial Track.............................................................................. 114
Figure 145. Test 2012-04-26_3 Transient Track ........................................................................ 115
Figure 146. Test 2012-04-26_4 Overall Track ........................................................................... 115
Figure 147. Test 2012-04-26_4 Initial Track.............................................................................. 116
Figure 148. Test 2012-04-26_4 Transient Track ........................................................................ 116
Figure 149. Test 2012-04-26_5 Overall Track ........................................................................... 117
Figure 150. Test 2012-04-26_5 Initial Track.............................................................................. 117
Figure 151. Test 2012-04-26_5 Transient Track ........................................................................ 118
Figure 152. Test 2012-04-26_6 Overall Track ........................................................................... 118
Figure 153. Test 2012-04-26_6 Initial Track.............................................................................. 119
Figure 154. Test 2012-04-26_6 Transient Track ........................................................................ 119

xiii

Figure 155. Test 2012-04-26_7 Overall Track ........................................................................... 120
Figure 156. Test 2012-04-26_7 Initial Track.............................................................................. 120
Figure 157. Test 2012-04-26_7 Transient Track ........................................................................ 121
Figure 158. Test 2012-04-27_1 Overall Track ........................................................................... 121
Figure 159. Test 2012-04-27_1 Initial Track.............................................................................. 122
Figure 160. Test 2012-04-27_1 Transient Track ........................................................................ 122
Figure 161. Test 2012-04-27_2 Overall Track ........................................................................... 123
Figure 162. Test 2012-04-27_2 Initial Track.............................................................................. 123
Figure 163. Test 2012-04-27_2 Transient Track ........................................................................ 124
Figure 164. Test 2012-04-27_3 Overall Track ........................................................................... 124
Figure 165. Test 2012-04-27_3 Initial Track.............................................................................. 125
Figure 166. Test 2012-04-27_3 Transient Track ........................................................................ 125
Figure 167. Test 2012-04-27_4 Overall Track ........................................................................... 126
Figure 168. Test 2012-04-27_4 Initial Track.............................................................................. 126
Figure 169. Test 2012-04-27_4 Transient Track ........................................................................ 127
Figure 170. Test 2012-04-27_5 Overall Track ........................................................................... 127
Figure 171. Test 2012-04-27_5 Initial Track.............................................................................. 128
Figure 172. Test 2012-04-27_5 Transient Track ........................................................................ 128

xiv

CHAPTER 1 – INTRODUCTION

Combustion is a vital part of our everyday lives. While much of the current research in
combustion involves improving the efficiency of our electric power generation, heating,
transportation, and other such processes, it is equally important to study the detrimental aspects
of combustion, such as pollutant formation and damage caused by uncontrolled fires, in order to
improve the safety of our structures and vehicles. Combustion is a complex process involving
several research disciplines, including chemistry, fluid mechanics, and thermodynamics, and all
of its details are not yet fully understood. To ensure the safety of astronauts aboard the Space
Shuttle and, more recently, the International Space Station, the Micro-gravity Science Division
(MSD) team at the National Aeronautical and Space Administration (NASA) Glenn Research
Center in Cleveland, Ohio studies micro-gravity flame spread through its Analysis of Thermodiffusive and Hydrodynamic Instabilities in Near-extinction Atmospheres (ATHINA) program.
This initiative is necessary because combustion in micro-gravity environments is different from
that on Earth due to the lack of buoyant effects, whereby flames tend to propagate upwards as a
result of hotter gases rising above cooler gases due to gravity.

Currently, NASA tests

flammability of non-metallic solids using NASA-STD-(I)-6001A Test 1, but since this test
consists of an upward flame spread in normal gravity, it is not representative of a fire occurring
in microgravity conditions.
The most ideal way to study micro-gravity combustion would be to actually conduct
combustion tests in space, but while this will be done on a limited basis, it is very expensive and
subject to scheduling constraints. To replicate a micro-gravity environment on Earth, the NASA
facility has 2.2 and 5 second duration drop towers, which can be used to demonstrate flame

1

propagation in free-fall, but these tests are also expensive and are limited to the 2.2 and 5 second
duration of the free-fall.

Additionally, burn tests can and have been conducted aboard

conventional aircraft flying parabolic trajectories, which provides about half a minute of
microgravity testing conditions, but are also subject to substantial expense, scheduling
constraints, and relatively high “g-jitter” (fluctuations in acceleration, particularly on the aircraft
in parabolic flight but also to a lesser extent in the drop tower and aboard the space shuttle). To
simulate a microgravity environment inexpensively and conveniently, an apparatus was
developed in which a solid fuel could be burned in a channel with a vertical height small enough
to suppress buoyant effects but not so small as to suppress the flame through heat loss to the
chamber walls. Versions of this apparatus are being studied at NASA’s Glenn Research Center
at Lewis Field (which subsequently shall be called NASA-Lewis), San Diego State University
(SDSU), and at Michigan State University (MSU). Each team’s experimental apparatus is
slightly different and is intended for different purposes. MSU’s Narrow Channel Apparatus
(NCA) has evolved through three iterations. Details of the differences among these appear in the
Literature Review and Experimental Setup sections.

2

CHAPTER 2 – LITERATURE REVIEW

This project continues the previous experimental research done at Michigan State
University by Oravecz [1] and Tanaya [2]. Prior to the research at MSU, theoretical analyses of
thin solid fuel combustion were conducted by deRis [3] and Fernandez-Pello and Williams [4, 5].
Additionally, works by Wichman and Williams [6, 7] present a simplified model, based on the
results of deRis, that applies to solid fuel flame propagation with an opposed oxidizer flow.
Experimental research on diffusion flame spread over thermally thin solid fuels with
opposed oxidizer flow in microgravity has been performed by Olson et al. at NASA-Lewis [8, 9,
10].

The results show that, as expected, at high flow velocities extinction occurs due to

convective heat loss, but contrary to deRis’s model, at low velocities quenching occurs due to
radiation and other heat loss. In the latter situation, the assumption by deRis of infinitely fast
chemical kinetics becomes invalid. These results were not apparent until conducting simulated
microgravity tests because achieving sufficiently low flow velocities in normal gravity is
difficult due to buoyant effects, which tends to entrain oxidizer at speeds up to 30 cm/s near the
flame front. Of interest in this project is combustion in this near-extinction limit of low oxidizer
flow, where the flame front dissociates into narrow, finger-shaped flamelets. The studies by Zik
et al. [11] show that the spacing between these “fingers” (flamelet paths) is determined by the
Péclet number, which is the ratio of advection to diffusion, and that the width of the fingers is
determined by heat loss.
The research conducted by Oravecz consisted of the development of a narrow channel
combustion chamber, also known as a Hele-Shaw apparatus, based on the preceding work and
others [1]. Oravecz designed the MSU Flame Rig based on research that showed that buoyant

3

effects would be suppressed if the gap spacing, or the vertical height in the combustion chamber
between the plates, is below a certain critical value. This critical value occurs where the Raleigh
number, which is representative of the ratio of convective to conductive heat transfer, is no
greater than 1708.

Oravecz’s experiments showed that a Hele-Shaw apparatus can be an

acceptable supplement to existing methods of studying microgravity flame spread, as indicated
by the similarity of flame instabilities in the MSU Rig and the NASA Drop Tower tests. At
sufficiently low oxidizer flow velocities, flames would disintegrate into flamelets that oscillated
on the order of 1 Hz and spread slower than the flame front. They always propagated in the
direction opposing oxidizer flow and would generally converge or auto-extinguish.

Most

importantly, these flamelets survived in flow velocities found on the Space Shuttle and
International Space Station, indicating that such a fire could propagate in those environments.
The discussion section in [1] includes a reminder that the Hele-Shaw apparatus is not a true zerogravity environment but rather just simulates one by suppressing buoyant flow, and it has the
additional disadvantages of a large heat sink near the burning sample and a channel geometry
that produces a viscous-dominated flow. However, it offers versatility that the true microgravity
testing methods do not, specifically a long testing time, easy interchangeability of test samples,
optical access, and variation of test geometry.
The research conducted by Tanaya [2] continues the work of Oravecz. These tests
utilized a new apparatus, called the Simulated Microgravity Flamelet Tunnel (SMFT). It had the
same plenum and contraction chambers as the previous MSU Flame Rig but a different, although
functionally similar, test section. Rather than standing on legs, it had a rolling cart to which it
was mounted along its center axis (the axis in line with the flow) at each end, so that it could be
rotated about that axis. By rotating the apparatus to the inverted (upside-down) position and

4

observing that the flame behavior is the same, it could be inferred that buoyant effects are in fact
largely (but not totally) suppressed. Like the MSU Flame Rig, the SMFT test section contains a
copper heat sink on one surface and a glass window for viewing on the other. The Flame Rig
utilized shims to maintain gap spacing; to allow for rotating to the inverted position, the SMFT
used bolts that secured to threaded holes in the sample holder, wall plate, aluminum base plate.
The ignition system consisted of a straight wire across the sample mounted to tensioning clips at
each end, unlike the old rig which initially used a rigid wire that was formed into many
sinusoidal curves to ensure complete contact (later versions utilized the wire-tensioning concept).
The SMFT had a stand secured to its top plate for mounting the camera and blocking excess
ambient light.
Tests conducted in the SMFT involved observing the flame breakup and convergence
under linearly decreasing and increasing opposed flow velocities, respectively, which the paper
called “ramping” flow. Test results indicated that flame stability in the opposed flow setup is
strongly dependent on oxidizer flow rate and gap height; this is consistent with the previous tests.
The ramp down tests showed that the number of flamelets formed is initially dependent on the
flow deceleration rate but, after about 20 seconds, loses that dependency while also being
bounded within a consistent range. The ramp up tests showed that the flamelets propagate slower
than the flame front, but the ratio of flame spread velocity to oxidizer flow velocity is higher for
the flamelets than the flame front, since the oxidizer flow is slower in the flamelet regime. For a
steady flow test condition, the flame spread rate is nearly linearly proportional to the oxidizer
flow velocity in the near-extinction (low flow) regime, then remains approximately constant and
eventually decreases with increasing flow velocity.

5

Following Tanaya, Aditjandra [12] conducted numerous experiments using the SMFT.
These tests focused on analyzing the number of flamelets formed and proportion of the sample
burned in the low-flow near-extinction limit. Most were conducted in both the normal and
inverted positions, which resulted in slight but statistically significant differences. The flamelet
population tests showed that about half of the flamelets must be productive (that is, split into
additional flamelets) in order to sustain the flamelet propagation (that is, for the rate of flamelet
propagation to equal or exceed the rate of flamelet extinction). The rate of flamelet formation
was slightly greater in the normal position than inverted and increased with increasing oxidizer
flow. The study of burned area showed that about 40% to 70% of the fuel sample was consumed
on average in the normal position versus 20% to 50% for the inverted position. Many of
Aditjandra’s results [12] were summarized in a detailed article by Olson et al. [13].
Concurrently with the MSU NCA project are two projects at SDSU. A narrow channel
similar to but smaller than the one at MSU is being used to test the effect of air pressure on the
flame spread rate [14]. This setup utilizes bottled oxygen and nitrogen, along with a mass flow
controller for each, to test different normoxic conditions (that is, varying the total air pressure
and oxygen concentration while keeping the partial pressure of oxygen in the mixture constant),
which mimics the conditions found on spacecraft. Another channel is being used to test an
alternative fuel, trioxane (C3H6O3) [14]. Trioxane is physically thick but behaves as a thermally
thin fuel because only a thin layer burns off the top surface of the fuel in each test. This fuel has
the notable advantage that a single sample can be burned up to thirty times with no change in its
behavior.

6

CHAPTER 3 – EXPERIMENTAL SETUP

OVERVIEW
The new setup discussed in this thesis is known as the Michigan State University Narrow
Channel Apparatus (MSU NCA), to parallel the setup currently being studied at San Diego State
University (SDSU), called the SDSU NCA. (SDSU actually has two narrow channels: one that
resembles the NASA and MSU NCA’s, and another that utilizes a moving op wall to produce an
opposed flow with a linear velocity gradient.) The MSU NCA employs the rolling stand and
rotating shaft of its predecessor but has a new, wider channel mounted to it, as well as a greatly
streamlined inlet plenum and a shorter outlet plenum section. See Figure 1 for a full view of the
apparatus. The system consists of a ¼” inner diameter hose originating at the shop air supply
(140 psi static pressure) and passing through a brass pressure regulator (reducing the air supply
to 60 psi), a rotameter (to verify the flow rate), and Y-junction (to divide the flow evenly among
two tubes), after which the tubes are fed into opposite ends of a distribution pipe in the plenum.
The distribution pipe releases air toward the back of the plenum chamber, which reflects the
airflow forward, effectively pressurizing the entire plenum chamber and driving a uniform flow
out its exit. Just upstream of the plenum exit, the air passes through a plastic honeycomb, which
has a 5 to 1 aspect ratio, and a fine metal screen, the combination of which should eliminate most
turbulence and other flow nonuniformities associated with this setup. See Figure 2 for a view of
the hardware in the plenum chamber. The air then traverses the test section where combustion of
the sample is occurring and is finally released into the atmosphere through a converging exit
chamber.

7

TEST SECTION
An entirely new channel has been constructed for this experiment.

Unlike its

predecessor’s, the new test section now has quartz glass windows on both the top and bottom
plates instead of a copper heatsink on the bottom plate, which allows simultaneous viewing from
both directions (eliminating the need to rotate the channel to the inverted position to view the
combustion from below) and ensures symmetrical heat loss. As mentioned in the overview, the
test section is also wider than its predecessor’s in order to accommodate a larger sample, which
can yield more flamelets. The new test section also has side windows for verifying vertical
symmetry. The top window is now removable from the top plate for fast and easy sample
reloading, identical to the NASA NCA. See Figure 3 for a view of the new test section and
Figure 4 for the same view with the top window removed.
Gap spacing is maintained using 1 mm thick shims (as in the original MSU Flame Rig)
rather than threaded bolts (as in the SMFT) because the sample holder, which is made of 0.032inch (0.8 mm) thick stainless steel, is somewhat flexible, and so it takes many supports to keep it
rigid. Shims are stacked in ten different places from the floor of the test section to the sample
holder and then eight places from the sample holder to the ceiling. See Figure 5 for a view of the
shims above the sample holder and Figure 6 for a view of the shims below it. The extra pair of
shim locations below the sample holder accommodates pins that secure the sample holder in
place when the top plate is removed, a detail of which is shown in Figure 7. Additional shims
can be placed if the sample holder sags between supports. The channel plates are clamped to
each other in two places along each side and secured by two bolts at each end, ensuring that
neither plate deforms in the middle of the test section, which would cause a variation in the gap
spacing. The apparatus overview in Figure 1 shows these clamps.

8

PLENUM CHAMBER PRESSURE RELIEF
Unlike previous iterations of the narrow channel apparatus, the plenum chamber employs
a U-tube manometer pressure relief system designed to prevent the static pressure in the test
section from breaking the quartz windows. A detail of this setup can be seen in Figure 8. Dr.
Fletcher Miller at San Diego State University, the Principal Investigator on this project, provided
the
P=

formula

for

the

maximum

allowable

pressure

differential

on

the

window

M ⋅ T 2 (x2 − y2 )
, where M is the modulus of rupture (M = 7000 psi for quartz glass), T is
x2 y2F

the plate thickness (T = ¼ inch), x is the unsupported length (x = 17 inches), y is the unsupported
width (y = 10 inches), and F is the desired factor of safety (F = 5). The result per these
parameters was P = 1.178 psi, which corresponds to a 32.6-inch water column. There is
additional safety in the fact that the pressure relief is in the plenum chamber, which is upstream
and therefore at a higher static pressure than the test section. The appropriate height of the water
column is marked on a ruler affixed to the manometer; the container should be checked and
refilled periodically due to water evaporation.
FLOW CONTROL
The flow is metered by an Alicat model MC-100SLPM-D/5M flow controller with a 100
L/min capacity, which is controlled by a PC running an Igor Pro experiment written by Mr.
Jacob Pepper at San Diego State University. The program calculates the necessary flow rate
based on the desired flow velocity and channel geometry. It can also control separate oxygen
and nitrogen flows to maintain a desired oxygen concentration. We cannot use this functionality
since we are using shop air and only one mass flow controller. A portable hygrometer indicated
that, with lab ambient conditions of 75.6oF and 70% relative humidity, the shop air was 74.9oF
and 13.8% RH, which is dry enough that we do not require a dehumidifier or bottled air.

9

SAMPLE IGNITION
Like its predecessor, the NCA igniter is a straight wire strung between two spring-loaded
clips that maintain the wire tension as it stretches due to the temperature increase, and the sample
is placed on top of the wire such that it is heated from below. See Figure 9 for a close-up view
of the igniter wire and tensioner. The wire is 0.0071-inch-diameter (33-gauge) Kanthal A-1,
whose resistance is 16.88 ohms/foot, so the 17-inch (1.42-foot) length of wire has a total
resistance of 23.97 ohms. Current is supplied to the igniter wire by a Variac type W50M analog
voltage controller with a 115 volt, 50-60 Hz AC input and 0 to 135 volt (RMS), 50-60 Hz AC
output, up to 40 amperes. The sample ignites evenly when the voltage controller is set to 30 V,
which corresponds to a current of about 1.25 A (which has been verified by a multimeter in
series with the igniter circuit) and a power dissipation of about 37.6 W through the igniter wire.
See Figure 9 for a close-up of the igniter wire and tensioner.
DATA COLLECTION
The camera mount has been retained from the SMFT to accommodate a lightweight,
consumer-grade digital camcorder (recording approximately 30 frames per second) viewing the
combustion from above. Additionally, a high-speed (200 frame per second) Basler camera views
the combustion from below, in order to capture details of the flamelet dissociation process that
are not detectable by standard rate video. Because of the high-speed camera lens’s long focal
length, it must be placed far from the apparatus in order to view the entire sample, so a mirror
has been placed at a 45-degree angle to the sample under the apparatus to reflect the image of the
sample horizontally without any distortion (provided that the camera is perfectly level). The
camera is positioned a suitable distance from the mirror to capture the image of the entire sample
without much additional space around it (the tip of the camera lens positioned 23 inches from the
edge of the sample cart has been found to be ideal), and at a suitable height to capture the entire

10

sample with the camera level (the tripod, with legs and neck fully extended, positioned 2 inches
above the floor has been found to be ideal). See Figure 10 for camera position and mirror setup.
Video recordings of the burn tests are objectively analyzed using NASA’s open-source
Spotlight software, which can determine the flame propagation speed by tracking a specific point
along the flame front. Previously the flame speed was estimated manually, but we are now
employing Spotlight as suggested in Tanaya’s “Conclusions and Suggestions for Future Work”
section [2] in order to increase the accuracy of our data.

11

APPARATUS PHOTOS

Figure 1. Full View of Apparatus
For interpretation of the references to color in this and all other figures, the reader is
referred to the electronic version of this thesis.

12

Figure 2. View of Plenum Chamber with Top Removed, Highlighting Distribution Pipe and
Flow Guide

Figure 3. View of Test Section Fully Assembled

13

Figure 4. View of Test Section with Top Window Removed

Figure 5. View of Test Section with Top Plate Removed

14

Figure 6. Close-up View of Test Section with Sample Holder Removed

Figure 7. Close-up View of Sample Holder Guide Pins

15

Figure 8. View of Pressure Relief System

16

Figure 9. Close-up View of Igniter Wire and Tensioner

Figure 10. Full View of Test Setup, Highlighting Cameras and Mirror

17

CHAPTER 4 – PROCEDURE

FLOW CONTROLLER SETUP
1. Make sure air supply lines are connected (from shop air supply to pressure regulator to
Alicat Mass Flow Controller). Turn on shop air supply valve (in fume hood).
2. Plug in the Alicat Mass Flow Controller and start up the PC to which it is connected.
3. Load the appropriate program into IGOR (located in the WaveMetrics folder in “My
Documents”).
CHANNEL SETUP
1. Verify that gap spacing is as desired (each shim is 1 mm thick, and sample holder can be
approximated as zero thickness) and that test section bolts are tightened (such that there is
no sagging of the top or bottom plates).
2. Remove top window and place rubber support shims. Each of these shims is also 1 mm
thick, so there should be the same number of rubber shims as metal shims below the
sample. Position the rubber shims using a paper clip or other tool in order to prevent
fingerprints on bottom window.
3. Load sample (half sheet of 46 cm x 57 cm Whatman-44 filter paper) by placing it on top
of the sample holder and igniter wire, sliding its edges under the paper clips that are
attached to the sample holder. View through side windows to ensure sample is level and
making even contact with the igniter.
4. Dry sample by running a high air flow rate (at least 80 SLPM) through loaded channel for
at least 5 minutes.

18

CAMERA SETUP
1. Assemble high-speed camera, connect it to Ethernet port of PC, and start up PC.
2. Insert USB key and load StreamPix Single Camera program, then load “Basler Pylon
GigE cameras” module (under “Camera” tab).
3. Position high-speed camera on floor facing mirror such that camera body is perfectly
level (to prevent distortion of the reflected image) and image of sample is fully visible
and occupies nearly the entire camera view (to minimize wasted space). This generally
occurs when tip of lens is about 23 inches from edge of apparatus cart, and camera tripod
(with legs and neck fully extended) is mounted approximately 2 inches above the floor.
The ½-inch thick PMMA samples can be used as shims to place the tripod. Ensure
aperture is set to fully open position (f-1.4), and adjust focus until rubber shims under the
sample are clear. (Focusing the camera is easiest after the sample is loaded, as the
sample blocks the view of background clutter.)
4. Place camera mount on top of test section with open side facing the channel exit.
5. Position consumer digital camcorder on top of camera mount such that lens is facing
downward through viewing hole and igniter wire is at right edge of view (to match
orientation of high-speed camera). Ensure sample is fully visible and level relative to
edges of camera view. Focus camera manually if necessary.
RUNNING TEST
1. Plug in LED lights in camera mount.
2. Set desired parameters in IGOR flow control program, start the flow, and wait for mass
flow controller to stabilize at set point.
3. Begin video recording on both cameras. (For high-speed camera, create new sequence
file in StreamPix before recording.)

19

4. Turn on power strip that feeds the igniter wire power supply, check its setting (with
Kanthal A-1 33-gauge igniter wire, 30 volts RMS is appropriate), and turn on igniter wire
switch (on desk next to multimeter). Multimeter can be used to verify current through
igniter wire (1.25 amperes AC is appropriate). Check that igniter wire glows and sample
ignites, then turn off igniter.
5. When sample autoextinguishes, stop the air flow in IGOR and turn on lights, then stop
the video recordings. Ensure that both cameras have captured clear footage of the burned
sample before stopping. (The spotlight connected to the igniter’s power strip can be used
to illuminate sample if overhead lights are insufficient.)
POST-TEST
1. Remove top window and discard sample fragments. The camera mount has Velcroattached removable curtains so that the clean-up and sample replacement process can be
completed with the camera mount in place.
2. Remove rubber support shims, and then clean interior faces of both windows using glass
cleaner and paper towels.
DATA ANALYSIS
1. Collect the video footage of the combustion tests and convert to a usable format if
necessary. (If analyzing high-speed video sequence captured by StreamPix, exporting to
.avi using Cinepak Codec by Radius encodes fairly quickly, is a reasonable size and
quality, and is readable by Spotlight 8 and Windows Media Player.)
2. Open video in Spotlight 8 and track using “Local Maximum Tracking”, with Area of
Interest (AOI) centered on brightest area of light. Tracking in 0.1 to 0.5-second intervals
is reasonable, depending on whether the combustion is in the flamelet or flame front

20

regime. Watch the tracking process to ensure AOI does not veer away from desired
combustion area. (If this occurs, manual tracking may be used.)
3. Find tracking results file and import its data into Excel. Using an image editor (such as
Microsoft Paint), determine the width (in pixels) of the viewable sample area, which is
known to equal 43 cm. Based on that relationship, calculate each x-position from the
tracking table in centimeters and determine the flame velocity between each reading, then
calculate the overall flame spread rate for the test using the average of the individual
velocities.
4. Save screenshots from the burn test footage as desired.

21

CHAPTER 5 – VERIFICATION AND CALIBRATION TESTS

The system was checked for leaks by applying a soap bubble solution to the joints and
seams of the apparatus between the mass flow controller and test section, which verifies by
conservation of mass that the set flow rate is the same as the flow rate at the sample. When air
was flowed through the channel, any leaks should have become visible by the escaping air
creating bubbles in the applied solution. None were found.
The height of the test section was measured. For a nominal gap height of 8 mm (4 shims
above sample holder and 4 below it), the actual gap height was found to be 7 mm at the channel
entrance, 8 mm at the sample, and 6 mm at the channel exit. The IGOR flow program should
therefore be set to the nominal gap height since that is the height that corresponds to the sample
location.
Next, the flow rate of the mass flow controller was verified by running a water
displacement test, as depicted in Figure 11. Flow originated from the clear plastic tube that
ended at the throat of the soda bottle. As air floated up to the base of the soda bottle shown in
the diagram, it displaced water from the bottle at a rate equal to the air flow rate. For this test, a
beaker was used because the body of the soda bottle was too flexible. A stopwatch was used to
time the displacement of 1 L of water. The flow rates were verified at 5 L/min and 10 L/min
within the uncertainty associated with starting and stopping the timer. While higher flow rates
will be used in the burn tests, they would cause the beaker to empty too quickly to be accurately
timed, but since the mass flow controller is new and was calibrated at the factory, we can assume
that it is accurate throughout its entire range. The complete data from this test are presented in
Table 1.

22

Figure 11. Water Displacement Test Setup
Set Flow
Actual Flow
Test #
Rate (L/min) Time (s)
Rate (L/min)
1
5
11.5
5.2
2
5
10.2
5.9
3
5
10.8
5.6
4
10
4.9
12.2
5
10
5.8
10.3
6
10
5.3
11.3
Table 1. Flow Rate Test: Time Elapsed to Displace 1 L of Water
The preceding tests demonstrate that the total flow rate through the test section is as
expected but not necessarily that the velocity profile is a fully developed parabola, as desired. A
constant temperature hot wire anemometer was used to check the flow velocity profile at various
average velocities, with probes placed at various points along the width (z axis) and height (y
axis) of the channel. The hot wires were first calibrated using a known flow rate through a
calibrator with a nozzle exit designed such that the velocity profile is uniform. The highest
average flow velocity measured was 50 cm/s, for which the peak flow velocity would be 75 cm/s
(1.5 times the average velocity for a parabolic velocity profile). Thus the peak Reynolds number
is Re =

V *D

ν

=

(0.75m / s ) * (0.0172m)
= 822 , which is well within the limit for laminar flow.
15.68 * 10 − 6 m 2 / s

See Figure 12 and Figure 13 for the calibration curves from October 24 and 25, 2011,
respectively. The raw data are in separate Excel files (not reproduced in this paper).
23

Figure 12. Hot Wire Calibration, 2011-10-24

Figure 13. Hot Wire Calibration, 2011-10-25
Note that while technically the voltage is dependent on the flow velocity, the axes are
reversed on the calibration because we need a formula to calculate velocity as a function of
voltage. For these two calibrations, the zero point and curve shapes are somewhat different, but

24

both fit strongly to second-order polynomials (R2 > 0.99). When the flow was tested, the second
calibration curve was used since the tests were conducted on the same day as the calibration.
The flow was tested at the channel exit in three locations across since the probe could be
placed there easily, but the resulting data seemed unreliable, possibly due to interference from
room air currents. Still, the calculated velocities from the channel exit flow tests are presented
below. See Figure 14, Figure 15, and Figure 16 for data at the three z-axis positions. The raw
data are in separate Excel files.

Figure 14. Channel Exit Flow Velocities, 2011-10-25: z = 50 mm

25

Figure 15. Channel Exit Flow Velocities, 2011-10-25: z = 0 mm

Figure 16. Channel Exit Flow Velocities, 2011-10-25: z = -50 mm
These results are not as expected. The curves should be parabolic with maxima in the
center (at y = 3 mm) and minima at the edges (y = 1 mm and 5 mm), which they are not. During
tests at zero flow, the voltages were fluctuating significantly, and in some flow tests, the voltage

26

fell below the calibrated zero point, which suggests the data are unreliable. (In situations where
the voltage was below the zero point, the calculated velocity was taken to be zero.)
The hot wire failed unexpectedly, so another one was calibrated on November 11, 2011.
Its results are below.

Figure 17. Hot Wire Calibration, 2011-11-11
This calibration is not strongly parabolic because the voltages were different for a given
flow velocity on the ramp-down test than for the ramp-up test. Still, this calibration curve was
used because it was fairly similar to the others.
This time the flow was tested above the sample holder (in the top half of the test section)
just upstream of the cutout for the sample. The tests from November 18 onward were conducted
at 80 mm left and right of center instead of 50 in order to measure closer to the edge of the
sample. Results from flow tests on November 11 and 18, 2011, are presented below.

27

Figure 18. Channel Interior Flow Velocities, 2011-11-11: z = 50 mm

Figure 19. Channel Interior Flow Velocities, 2011-11-11: z = 0 mm

28

Figure 20. Channel Interior Flow Velocities, 2011-11-11: z = -50 mm

Figure 21. Channel Interior Flow Velocities, 2011-11-18: z = 80 mm

29

Figure 22. Channel Interior Flow Velocities, 2011-11-18: z = 0 mm

Figure 23. Channel Interior Flow Velocities, 2011-11-18: z = -80 mm
These results are also not as expected. The results on November 18 are considerably
different from those of November 11 even though the configuration was the same. The hot wire
was recalibrated on November 21, 2011. Its results are below.

30

Figure 24. Hot Wire Calibration, 2011-11-21
This calibration is much closer to expectations, so flow testing was then performed. One
more data point in the vertical (y) direction was incorporated in this test. The results are below.

Figure 25. Channel Interior Flow Velocities, 2011-11-21: z = 80 mm

31

Figure 26. Channel Interior Flow Velocities, 2011-11-21: z = 0 mm

Figure 27. Channel Interior Flow Velocities, 2011-11-21: z = -80 mm
These results are again not as expected. During this test (after the reading at z = 80 mm,
y = 2.5 mm, and V = 50 cm/s), the probe failed. The next (and last remaining) probe was broken
during calibration.

32

Unfortunately, the hot wire tests have not yielded useful data. All of these tests were
conducted with the air distribution pipe facing the back of the plenum chamber, and the data
indicate that the flow is lower than desired. The data were not accurate enough to derive a
correction factor that would be used to generate the desired flow. Their measurements were
inconsistent even for known conditions. The bulk of the probe bodies limited where they could
be placed, and positioning them precisely was difficult with the equipment available in the lab.
When they were placed, they exhibited random uncertainty that in some cases exceeded the
expected range being measured, possibly due to other modes of heat transfer to nearby surfaces.
As a result of these issues, combustion tests were performed despite the absence of channel flow
data.

33

CHAPTER 6 – CONSTANT FLOW VELOCITY BURN TESTS

The first set of burn tests was conducted with constant flow velocities in order to verify
that the results were consistent with those from previous rigs. Data from all burn tests appear in
Table 2. Four tests were conducted at 25 cm/s and two at 15 cm/s. The first two 25 cm/s tests
were redone because the backlighting made it difficult to determine the proportion of sample
area burned, though their results have still been included in the average flame speed calculation.
The flame speed is determined as the slope of the linear regression curve for the flame track data,
all of which have correlations (R2 values) greater than 0.997.
In the 25 cm/s tests, the average flame spread rate is 0.216 cm/s, or 0.864% of the
oxidizer flow velocity, and the flame front appears bright white, as shown in Figure 28, Figure
29, Figure 31, and Figure 32. In the latter two 25 cm/s tests, virtually the entire sample was
consumed, as shown in Figure 30 and Figure 33.
In the 15 cm/s tests, the average flame spread rate is 0.192 cm/s, or 1.28% of the oxidizer
flow velocity. The flame appears dim and blue but is still a complete flame front, as shown in
Figure 34, Figure 35, Figure 37, and Figure 38. During these tests only 80 to 90 percent of the
sample was consumed, primarily because the sample ignited unevenly and required about one
fourth of the sample length to stabilize, as shown in Figure 36 and Figure 39.
In all of the constant velocity tests, the global mass air to fuel ratio was strongly fuel lean
(between 21 and 27, which correspond to equivalence ratios of approximately 4 to 5.5), which is
consistent with previous results.
It should be noted that, while the flame front was not always uniform due to incomplete
ignition and flow disruptions caused by the support shims, the flame spread rate was relatively

34

constant across the width of the channel; that is, even if the flame front was not uniform, it
tended to propagate uniformly.

Figure 28. Test 2012-04-18_2, 70s, Flame Front (25 cm/s)

35

Figure 29. Test 2012-04-18_2, 180s, Flame Front (25 cm/s)

Figure 30. Test 2012-04-18_2, Burned Sample

36

Figure 31. Test 2012-04-18_3, 70s, Flame Front (25 cm/s)

Figure 32. Test 2012-04-18_3, 170s, Flame Front (25 cm/s)

37

Figure 33. Test 2012-04-18_3, Burned Sample

Figure 34. Test 2012-04-18_4, 75s, Flame Front (15 cm/s)

38

Figure 35. Test 2012-04-18_4, 200s, Flame Front (15 cm/s)

Figure 36. Test 2012-04-18_4, Burned Sample

39

Figure 37. Test 2012-04-18_5, 70s, Flame Front (15 cm/s)

Figure 38. Test 2012-04-18_5, 180s, Flame Front (15 cm/s)
40

Figure 39. Test 2012-04-18_5, Burned Sample

41

CHAPTER 7 – FAST RAMP DOWN BURN TESTS

Ramp down tests were conducted in which the sample was ignited in a high, constant
oxidizer flow velocity condition, and then, once the flame front stabilized, the flow velocity was
decreased at a constant rate until reaching a low velocity steady state. The purpose of these tests
was to examine the steady state flame behavior at high versus low flow velocities. The flame
speed is determined as the slope of the regression curve for the flame track data (linear
regression for steady state and second order for transient conditions). These quantities have
correlations (R2 values) greater than 0.984.
The first three tests were initiated at 25 cm/s and ramped down at the deceleration rate 5
cm/s2 for 3 seconds to a final velocity of 10 cm/s. The average initial flame spread rate was
0.214 cm/s, or 0.856% of the oxidizer flow velocity. Here, the flame front appears bright white,
as shown in Figure 40, Figure 43, and Figure 46. The average final flame spread rate was 0.114
cm/s, or 1.14% of the oxidizer flow velocity. The flames appear dim blue and periodically
flicker or dissociate into flamelets, indicating that they are at the lower limit of the flame front
regime, as shown in Figure 41, Figure 44, and Figure 47. In these tests, two thirds to three
fourths of the sample was consumed, as shown in Figure 42, Figure 45, and Figure 48.
The next three tests were initiated at 25 cm/s and ramped down at 2 cm/s2 for 8 seconds
to a final velocity of 9 cm/s. The average initial flame spread rate was 0.217 cm/s, or 0.868% of
the oxidizer flow velocity, and the flame front appears bright white, as shown in Figure 49,
Figure 52, and Figure 55. The average final flame spread rate was 0.0953 cm/s, or 1.06% of the
oxidizer flow velocity, and as in the previous set of tests, the flames appear dim blue, as shown

42

in Figure 50, Figure 53, and Figure 56. In these tests, between two thirds and three fourths of the
sample was consumed, as shown in Figure 51, Figure 54, and Figure 57.
A final test was initiated at 25 cm/s and ramped down at 1 cm/s2 for 20 seconds to a final
velocity of 5 cm/s, as an interim between the fast and slow ramp down tests. Its initial flame
spread rate was 0.198 cm/s, or 0.990% of the oxidizer flow velocity. The flame front appears
bright white, as shown in Figure 58. Its final flame spread rate was 0.0739 cm/s, or 1.48% of the
oxidizer flow velocity, and the flames have dissociated into faint blue flamelets, as shown in
Figure 59. The test was stopped when the first flamelet neared the end of the sample. About
half of the sample was consumed, as shown in Figure 60. In this test, the ramping rate was slow
enough that the transient behavior could be examined; the flame spread rate during the flow
deceleration was found to be 0.720% of the oxidizer flow velocity.
In all but the last of these tests, the ramping rate was fast enough that only the steadystate behavior could be examined; that is, the transient period was very short. As expected, the
flame spread rate tended to be on the order of one hundredth of the oxidizer flow velocity
(slightly lower for flame fronts and higher for flamelets), and the flame front showed signs of
weakening and dissociation at flow velocities around 10 cm/s.
In all of these ramping tests, the global mass air to fuel ratio was strongly fuel lean
(between 28 and 43, which correspond to equivalence ratios of approximately 5.5 to 8.5), which
is consistent with previous results. Note that the air to fuel ratios are much higher in the ramping
tests than in the constant high velocity tests, indicating that the flamelets burn much more fuel
lean than the flame front.
Of interest in particular in these experiments was the behavior of the flame as it reaches
the lower limit of the flame front regime and dissociates into flamelets. High speed footage was

43

recorded to capture this behavior, but due to the low lighting conditions and the small lens on the
high-speed camera, little could be seen in these videos, though some screenshots are presented
below.

During the flow deceleration, the flame front initially decreases in intensity

proportionally to the flow velocity and in the process changes hue from bright yellow to dim
blue, as shown in the burn test figures below. Nearing the flamelet regime, the flame front
begins flickering at a frequency that increases as the flow rate decreases. Finally, the flame
again glows bright yellow immediately prior to bifurcating or autoextinguishing, as shown in
Figure 61, Figure 62, Figure 63, and Figure 64. Note from the burned sample images that, at 9
and 10 cm/s flow velocities, small flamelets tended to propagate near the edges of the sample,
whereas the flamelets in the center of the sample were larger, possibly due to locally lower air
flow velocities around the edges.

Figure 40. Test 2012-04-19_1, 70s, Flame Front (25 cm/s)

44

Figure 41. Test 2012-04-19_1, 190s, Weak Flame Front (10 cm/s)

Figure 42. Test 2012-04-19_1, Burned Sample

45

Figure 43. Test 2012-04-19_3, 85s, Flame Front (25 cm/s)

Figure 44. Test 2012-04-19_3, 265s, Weak Flame Front (10 cm/s)

46

Figure 45. Test 2012-04-19_3, Burned Sample

47

Figure 46. Test 2012-04-19_4, 80s, Flame Front (25 cm/s)

Figure 47. Test 2012-04-19_4, 240s, Weak Flame Front (10 cm/s)

48

Figure 48. Test 2012-04-19_4, Burned Sample

49

Figure 49. Test 2012-04-20_1, 85s, Flame Front (25 cm/s)

Figure 50. Test 2012-04-20_1, 225s, Weak Flame Front (9 cm/s)

50

Figure 51. Test 2012-04-20_1, Burned Sample

Figure 52. Test 2012-04-20_3, 70s, Flame Front (25 cm/s)

51

Figure 53. Test 2012-04-20_3, 240s, Flamelets (9 cm/s)

Figure 54. Test 2012-04-20_3, Burned Sample

52

Figure 55. Test 2012-04-20_4, 80s, Flame Front (25 cm/s)

Figure 56. Test 2012-04-20_4, 230s, Flamelets (9 cm/s)

53

Figure 57. Test 2012-04-20_4, Burned Sample

Figure 58. Test 2012-04-24_1, 80s, Flame Front (25 cm/s)

54

Figure 59. Test 2012-04-24_1, 210s, Flamelets (5 cm/s)

Figure 60. Test 2012-04-24_1, Burned Sample

55

Figure 61. Test 2012-04-19_1 (High Speed Camera), 257s (286s in standard rate video),
Flamelets Extinguishing (10 cm/s)

Figure 62. Test 2012-04-20_1 (High Speed Camera), 230s (346s in standard rate video),
Flamelets Bifurcating (9 cm/s)

Figure 63. Test 2012-04-20_3 (High Speed Camera), 273s (300s in standard rate video),
Flamelets Extinguishing (9 cm/s)

56

Figure 64. Test 2012-04-24_1 (High Speed Camera), 57s (98s in standard rate video),
Flamelets Bifurcating (17 cm/s)

57

CHAPTER 8 – SLOW RAMP DOWN BURN TESTS

Ramp down tests were conducted in which the sample was ignited in a high, constant
oxidizer flow velocity condition, and then, once the flame front stabilized, the flow velocity was
decreased at a slow constant rate until reaching zero flow. The purpose of these tests was to
examine the transient flame behavior during slow ramping down flow. The flame speed is
determined as the slope of the regression curve for the flame track data (linear regression for
steady state and second order for transient conditions), all of which have correlations (R2 values)
greater than 0.966.
The first three tests were initiated at 25 cm/s and ramped down at 0.5 cm/s2 for 50
seconds. The average initial flame spread rate was 0.173 cm/s, or 0.692% of the oxidizer flow
velocity, and the flame front appears bright white, as shown in Figure 65, Figure 68, and Figure
71. The average transient flame spread rate was 1.15% of the oxidizer flow velocity, and the
flames appear dim blue and periodically flicker or dissociate into flamelets at around 10 cm/s
flow, as shown in Figure 66, Figure 69, and Figure 72. In these tests, about half of the sample
was consumed, as shown in Figure 67, Figure 70, and Figure 73.
The next four tests were initiated at 25 cm/s and ramped down at 0.25 cm/s2 for 100
seconds. The average initial flame spread rate was 0.213 cm/s, or 0.850% of the oxidizer flow
velocity, and the flame front appears bright white, as shown in Figure 74, Figure 77, Figure 80,
and Figure 83. The average transient flame spread rate was 0.860% of the oxidizer flow
velocity. The flames again appear dim blue and periodically flicker or dissociate into flamelets
at around 10 cm/s flow, as shown in Figure 75, Figure 78, Figure 81, and Figure 84. In these

58

tests, about half of the sample was consumed, as shown in Figure 76, Figure 79, Figure 82, and
Figure 85.
The next two tests were initiated at 25 cm/s and ramped down at 0.1 cm/s2 for 250
seconds. Since flamelets in these tests reached the end of the sample before the flow velocity
had reached zero, the next three tests were initiated at 20 cm/s and ramped down at 0.1 cm/s2 for
200 seconds. The average initial flame spread rate was 0.201 cm/s, or 0.806% of the oxidizer
flow velocity, for the first two tests and 0.204 cm/s, or 1.02% of the oxidizer flow velocity, for
the next three tests, and the flame front appears bright white, as shown in Figure 86, Figure 89,
Figure 92, Figure 95, and Figure 98. The average transient flame spread rate was 1.04% of the
oxidizer flow velocity, and the flames again appear dim blue and periodically flicker or
dissociate into flamelets at around 10 cm/s flow, as shown in Figure 87, Figure 90, Figure 93,
Figure 96, and Figure 99. In these tests, about half of the sample was consumed, as shown in
Figure 88, Figure 91, Figure 94, Figure 97, and Figure 100.
In all of these ramping tests, the global mass air to fuel ratio was strongly fuel lean
(between 22 and 42, which correspond to equivalence ratios of approximately 4.5 to 8.5), which
is consistent with previous results.
The transient behavior of the flame in these tests was similar to that described in the
previous chapter. During the flow deceleration, the flame front initially decreases in intensity
proportionally to the flow velocity and in the process changes hue from bright yellow to dim
blue, as shown in the burn test figures below. Nearing the flamelet regime, the flame front
begins flickering at a frequency that increases as the flow decreases. And again, the flame glows
bright yellow immediately prior to bifurcating or autoextinguishment, as shown in Figure 101,
Figure 102, Figure 103, and Figure 104.

Apart from the time scale, this behavior seems

59

relatively unaffected by the rate at which the flow is decelerated. Note from the burned sample
images that the flamelets did not survive for very long, and very few of them bifurcated, which
indicates that the range of flow velocities in which flamelets survive is very narrow.

Figure 65. Test 2012-04-26_1, 100s, Flame Front (25 cm/s)

60

Figure 66. Test 2012-04-26_1, 140s, Flamelets (10 cm/s)

Figure 67. Test 2012-04-26_1, Burned Sample

61

Figure 68. Test 2012-04-26_2, 90s, Flame Front (25 cm/s)

Figure 69. Test 2012-04-26_2, 120s, Flamelets (10 cm/s)

62

Figure 70. Test 2012-04-26_2, Burned Sample

Figure 71. Test 2012-04-26_3, 185s, Flame Front (25 cm/s)

63

Figure 72. Test 2012-04-26_3, 215s, Flamelets (10 cm/s)

Figure 73. Test 2012-04-26_3, Burned Sample

64

Figure 74. Test 2012-04-26_4, 180s, Flame Front (25 cm/s)

Figure 75. Test 2012-04-26_4, 240s, Flamelets (10 cm/s)

65

Figure 76. Test 2012-04-26_4, Burned Sample

Figure 77. Test 2012-04-26_5, 100s, Flame Front (25 cm/s)

66

Figure 78. Test 2012-04-26_5, 160s, Flamelets (10 cm/s)

Figure 79. Test 2012-04-26_5, Burned Sample

67

Figure 80. Test 2012-04-26_6, 10s, Flame Front (25 cm/s)

Figure 81. Test 2012-04-26_6, 70s, Flamelets (10 cm/s)

68

Figure 82. Test 2012-04-26_6, Burned Sample

Figure 83. Test 2012-04-26_7, 90s, Flame Front (25 cm/s)

69

Figure 84. Test 2012-04-26_7, 150s, Flamelets (10 cm/s)

Figure 85. Test 2012-04-26_7, Burned Sample

70

Figure 86. Test 2012-04-27_1, 90s, Flame Front (25 cm/s)

Figure 87. Test 2012-04-27_1, 240s, Flamelets (10 cm/s)

71

Figure 88. Test 2012-04-27_1, Burned Sample

Figure 89. Test 2012-04-27_2, 90s, Flame Front (25 cm/s)

72

Figure 90. Test 2012-04-27_2, 240s, Flame Front (10 cm/s)

Figure 91. Test 2012-04-27_2, Burned Sample

73

Figure 92. Test 2012-04-27_3, 80s, Flame Front (20 cm/s)

74

Figure 93. Test 2012-04-27_3, 180s, Flamelets (10 cm/s)

Figure 94. Test 2012-04-27_3, Burned Sample

75

Figure 95. Test 2012-04-27_4, 80s, Flame Front (20 cm/s)

Figure 96. Test 2012-04-27_4, 180s, Flamelets (10 cm/s)

76

Figure 97. Test 2012-04-27_4, Burned Sample

Figure 98. Test 2012-04-27_5, 80s, Flame Front (20 cm/s)

77

Figure 99. Test 2012-04-27_5, 180s, Flamelets (10 cm/s)

Figure 100. Test 2012-04-27_5, Burned Sample

78

Figure 101. Test 2012-04-26_1 (High Speed Camera), 90s (141s in standard rate video),
Flamelet Extinguishing (9 cm/s)

Figure 102. Test 2012-04-26_6 (High Speed Camera), 125s (83s in standard rate video),
Flamelets Bifurcating (6.75 cm/s)

Figure 103. Test 2012-04-27_2 (High Speed Camera), 117s (164s in standard rate video),
Flamelets Bifurcating (17.6 cm/s)

79

Figure 104. Test 2012-04-27_4 (High Speed Camera), 234s (257s in standard rate video),
Flamelets Extinguishing (2.3 cm/s)

80

CHAPTER 9 – LESSONS LEARNED

FLOW SYSTEM
The most significant issue with the flow system was air leaking out of the plenum
chamber. The original design, shown in Figure 105, had a rubber gasket between the walls and
top plate of the plenum chamber, but this gasket was never seated properly and allowed air to
leak between it and the top plate. Additionally, the sudden contraction at the plenum chamber
exit (in absence of the cardboard flow guide) likely created a significant constriction that
worsened the leak. When a pressure test revealed the location of the leak, a layer of petroleum
jelly (Vaseline) was used to seal the gap between the gasket and top plate, but this caused the
rubber gasket to warp and ultimately exacerbated the problem. Finally the rubber gasket was
discarded in favor of the bead of silicon adhesive sealant along the top of the perimeter walls of
the plenum chamber; this was effective.
Before the leakage issue was properly diagnosed, it was noted that the flow through the
test section was stronger with the distribution pipe rotated 180 degrees (that is, facing the plenum
exit instead of the back wall). Burn tests in this configuration produced flickering, unstable
flame fronts, an indication of nonuniform and unsteady flow through the test section. The pipe
was therefore returned to its original configuration when the flow issue was resolved.
SAMPLE HOLDER
The sample holder is currently not very rigid due to its thickness. The placement of the
shims keeps it fairly level, but it sags in the middle of the channel since shims cannot be placed
here due to the disruption they would cause to the flow. Shims were placed strategically to
minimize sagging, but a new sample holder constructed of a different material would be ideal, as
this issue has still not been fully resolved.

81

While the sample holder currently has paper clips attached to it for securing the sample in
place, initially the sample was affixed using electrical tape. This method allowed for more
tension across the sample, which could reduce the sagging due to gravity, but had the
considerable disadvantage of not being adjustable after the sample was placed. Because of this
issue, the sample was difficult to place uniformly, and so the paper clip method was found to be
easier to set up.
SAMPLE IGNITION
The igniter wire was originally positioned forward of its current position such that it
spanned the open region of the sample holder. Since the position of the tensioning clips was
constrained by the chamber design, the wire was routed to its desired location by a pair of
guides, as shown in Figure 106. This configuration was thought to be advantageous since it
positioned the igniter within the viewing area of the window and in the path of the air flow
(instead of adjacent to the sample holder), allowing for easier and more uniform ignition.
However, it turned out that, without the backing of the sample holder, the igniter wire made very
little contact with the sample.

Furthermore, the igniter wire in this configuration would

frequently slip out of place, and repositioning it required removing the top plate of the test
section. Finally, the guides, which were made of rubber due to its insulating properties, were
found to be scorched by the igniter wire after repeated tests. For these reasons, the igniter wire is
now placed straight across the test section.

82

PRELIMINARY APPARATUS PHOTOS

Figure 105. View of Plenum Chamber with Top Removed, Highlighting Old Seal Design

Figure 106. View of Test Section with Top Plate Removed, Highlighting Old Igniter Design

83

CHAPTER 10 – CONCLUSIONS AND FUTURE WORK

In constant velocity flows, the flame propagation rate was approximately 1% of the
oxidizer flow velocity (slightly lower for the flame front and slightly higher for flamelets), which
is consistent with previous experiments. The track is strongly linear, indicating that the flame
speed is constant. In the ramping flows, the flame propagation rate was also approximately 1%
of the oxidizer flow velocity but was variable, particularly in the faster ramp down tests. This
behavior is possibly due to the lag time between the change in flow velocity and the response of
the flame. The track is strongly second order, indicating that the flame speed decreases linearly
with respect to the flow deceleration.
Flamelets in the near-extinction regime tend to be a dim blue hue (which was often
difficult to see in the video) and often flickered with increasing frequency as the oxidizer flow
velocity decreased. However, immediately before bifurcating or extinguishing, the flamelet
would glow bright yellow, possibly indicating that the flame is cooling. The flame front began
dissociating into flamelets at oxidizer flow velocities of approximately 10 cm/s. The flamelets
tended to extinguish at approximately 4 cm/s flow rate. Thus, the range of flow velocities that
supports the propagation of flamelets is fairly narrow, as indicated by the short life span of the
flamelets in the tests that ramped the flow down to zero.
While the new Narrow Channel Apparatus is an improvement over its predecessor, the
wider channel presents challenges that need to be addressed. The highest priority is a more
effective sample holder and ignition system. Currently, the sample holder sags due to gravity in
the center of the channel, and supporting it with shims is not an option as that would disrupt the
flow. Thus, a new sample holder constructed of a more rigid material is needed. Furthermore,

84

the rubber shims that support the center of the sample disrupt the flame propagation and the
uniformity of the sample, but leaving the sample unsupported in the center is not an option as the
sample sags enough to touch the bottom glass, which severely disrupts the air flow and flame
propagation. An ideal solution would be the use of thin, non-conductive and heat-resistant wire
strung between the clips to uniformly and non-disruptively support the sample across its entire
width.
The next priority is improvements to the video recording technology. In particular, the
high-speed camera is not very useful in its present state. At the very least, it needs a lens with a
larger aperture for the flamelets to be visible in the video, as we are attempting to view very
small points of light from a considerable distance. Ideally, the camera would also be higher
resolution, as the flamelets currently appear pixelated due to their small size relative to the area
being viewed. However, this would require a new camera, which would constitute a substantial
expense. An alternative would be to install additional lighting in the camera mount, but too
much light could overpower the light of the flamelets and render them invisible.
Additional long-term upgrades include a second mass flow controller, equipment to
measure the temperature at various points in the channel, and equipment to analyze the chemical
content of the exhaust gases. With two mass flow controllers and bottled gases, the flow rate of
oxygen and nitrogen can be controlled independently, allowing for testing the effects of varying
oxygen concentrations. The IGOR program used to control the flow is already written to
accommodate two MFCs. Accurate temperature measurements can verify the FLUENT models
and the heat release formula derived by Aditjandra [12]. Accurate analysis of the exhaust gases
can verify the oxygen consumption model derived by Aditjandra [12] and also provide real-time
data regarding the combustion conditions. NASA-Lewis already has equipment to measure the

85

oxygen consumption rate. In the long-term future, a second high-speed camera and a PC with
dual Ethernet ports and more processing power to run them would be ideal, as this setup would
provide comparable video from both above and below the sample. Currently, the two videos are
different speeds, different resolutions (with different aspect ratios), and offset from each other
(due to the method of starting and stopping the video recording), which makes direct
comparisons difficult.

86

APPENDICES

87

APPENDIX A – STOICHIOMETRY ANALYSIS

The samples are Whatman-44 filter paper (cellulose) with chemical formula C6H10O5. Its
combustion process in air is
C6H10O5 + 6O2 + xN2  6CO2 + 5H2O + xN2
where coefficient x depends on the oxygen concentration. (In this case, x is the number of moles
of nitrogen per six moles of oxygen.)
Let n be the oxygen mole fraction. The nitrogen mole fraction is then 1 – n (assuming
negligible concentrations of other gases in air). That makes the molar ratio of nitrogen to oxygen
(1 – n)/n, which from the above formula is also equal to x/6. Therefore
x = 6/n – 6
Then the general form of the combustion equation is
C6H10O5 + 6O2 + (6/n – 6)N2  6CO2 + 5H2O + (6/n – 6)N2
For combustion in pure oxygen, n = 1, and so the formula simplifies to
C6H10O5 + 6O2  6CO2 + 5H2O
For the typical atmospheric molar oxygen concentration of 21% which was used in this
experiment, n = 0.21, and so the formula simplifies to
C6H10O5 + 6O2 + 22.57N2  6CO2 + 5H2O + 22.57N2
While the nitrogen is not part of the combustion process, its concentration is necessary to
determine the oxygen flow rate based on the known total air flow rate. The air flow rate is
reported as volumetric, and the volumetric concentration of oxygen in the air is equal to its molar
concentration.

88

The stoichiometric mass ratio of air to fuel is then
mair
mF

6⋅ MW O2 +

 6 − 6 MW
n

N2



6⋅ 31.9988

MW F

kg
kmol

+

 6 − 6 28.0134 kg
n

kmol



162.1402⋅

kg

0.1475 +

1.0366
n

kmol

For atmospheric air (n = 0.21), the stoichiometric mass air to fuel ratio is
mair/mF = 5.08
The actual mass ratio of air to fuel can be approximated by the known air flow rate and
elapsed time and the estimated mass of fuel burned.
Let v be the average oxidizer flow velocity and h be the total gap height. The gap width
is a constant 46 cm. The oxidizer flow rate is
F = (46 cm)*v*h.
The density of atmospheric air is approximately 1.225 kg/m3 (or 0.00125 g/cm3), and the
gap height for all of the tests in this paper is 8 mm (0.8 cm), so the oxidizer mass flow rate for
these tests is (0.04508 g/cm)*v. Then the total mass of air is
⌠ tf

  v ( t) d
mair 0.04508
t

cm  ⌡t
 0

This integral simplifies to v*t for constant velocity flows or (vf + v0)/2*t for ramping
g

(constant acceleration) flows, where t is the elapsed time.
The approximate mass of fuel burned is simply the estimated percentage of the sample
area burned times the mass of the unburned sample, which is approximately 10.4 grams, as
measured by the load cell in the lab’s cone calorimeter. (Each sample theoretically could have
been weighed before and after the test to determine the mass burned, but as the sample tended to
disintegrate into small pieces during the combustion process, determining the remaining mass
post-burn would have been difficult.)

89

The approximate mass air to fuel ratio for atmospheric air and an 8 mm gap height is then
mair
mF

tf

0.0043346 ⌠

Pb ⋅ cm ⌡t

v ( t) d
t
0

where Pb is the percentage of the sample area burned. The solved integral will have units of
length, which cancels the length units in the denominator of the coefficient, resulting in a
dimensionless quantity, as expected.

90

APPENDIX B – PRESSURE RELIEF CALCULATION

Physical properties of quartz window -

Length: x := 17in

Plate thickness: T :=

M := 7000psi

Modulus of rupture:

Width: y := 10in

Selected Factor of Safety -

4

in

Unsupported surface area:

A := x⋅ y

F := 5

Properties of water - ρ := 1000

kg
3

P

ρ⋅ g ⋅ h

m

P :=

(

M⋅ T x + y
2 2

x y F
h :=

P
ρ⋅ g

P

or h
2 2

Calculation using Dr. Miller's formula which corresponds to:

1

h = 32.601⋅ in

91

ρ⋅ g
2

)

P = 1.178⋅ psi

A = 170⋅ in

2

APPENDIX C – TEST DATA

Test Name
2012-04-09_1
2012-04-09_2
2012-04-17_1
2012-04-18_1
2012-04-18_2
2012-04-18_3
2012-04-18_4
2012-04-18_5
2012-04-19_1
2012-04-19_2
2012-04-19_3
2012-04-19_4
2012-04-20_1
2012-04-20_2
2012-04-20_3
2012-04-20_4
2012-04-24_1
2012-04-26_1
2012-04-26_2
2012-04-26_3

Initial
v
(cm/s)
25
25
25
25
25
25
15
15
25
25
25
25
25
25
25
25
25
25
25
25

Initial
Initial
Ramp Ramp
Flame
Flame
Rate
Time
Final v Speed
Speed /
Transient Flame
2
Initial v
Speed (cm/s)
(cm/s ) (s)
(cm/s) (cm/s)
0
N/A
25
0.2180
0.872%
N/A
0
N/A
25
0.2369
0.948%
N/A
0
N/A
25
N/A
N/A
N/A
0
N/A
25
N/A
N/A
N/A
0
N/A
25
0.2056
0.822%
N/A
0
N/A
25
0.2053
0.821%
N/A
0
N/A
15
0.1876
1.251%
N/A
0
N/A
15
0.1963
1.309%
N/A
-5
3
10
0.2039
0.816%
N/A
-5
3
10
N/A
N/A
N/A
-5
3
10
0.2231
0.892%
N/A
-5
3
10
0.2137
0.855%
N/A
-2
8
9
0.2205
0.882%
N/A
-2
8
9
N/A
N/A
N/A
-2
8
9
0.2157
0.863%
N/A
-2
8
9
0.2161
0.864%
N/A
-1
20
5
0.1984
0.794%
0.0072*v + 0.0599
-0.5
50
0
0.1576
0.630%
0.0096*v + 0.0186
-0.5
50
0
0.1938
0.775%
0.0124*v - 0.0139
-0.5
50
0
0.1686
0.674%
0.0124*v - 0.0340
Table 2. Overview of Test Data
(Red background indicates discarded tests.)

92

Final
Flame
Speed
(cm/s)
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.1257
N/A
0.1066
0.1101
0.0995
N/A
0.0978
0.0887
0.0739
N/A
N/A
N/A

Final
Flame
Speed /
Final v
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
1.257%
N/A
1.066%
1.101%
1.106%
N/A
1.087%
0.986%
1.478%
N/A
N/A
N/A

2012-04-26_4
2012-04-26_5
2012-04-26_6
2012-04-26_7
2012-04-27_1
2012-04-27_2
2012-04-27_3
2012-04-27_4
2012-04-27_5

25
25
25
25
25
25
20
20
20

-0.25
-0.25
-0.25
-0.25
-0.1
-0.1
-0.1
-0.1
-0.1

100
100
100
100
250
250
200
200
200

0
0
0
0
0
0
0
0
0

Table 2 (cont’d)
0.2053
0.821%
0.1908
0.763%
0.2320
0.928%
0.2223
0.889%
0.1885
0.754%
0.2148
0.859%
0.2082
1.041%
0.2049
1.025%
0.1996
0.998%

93

0.0072*v - 0.0005
0.0096*v + 0.0247
0.0064*v + 0.0424
0.0112*v + 0.0034
0.0080*v + 0.0324
0.0060*v + 0.0499
0.0120*v - 0.0192
0.0140*v - 0.0372
0.0120*v - 0.0072

N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A

N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A

Test Name
2012-04-18_2
2012-04-18_3
2012-04-18_4
2012-04-18_5
2012-04-19_1
2012-04-19_3
2012-04-19_4
2012-04-20_1
2012-04-20_3
2012-04-20_4
2012-04-24_1
2012-04-26_1
2012-04-26_2
2012-04-26_3
2012-04-26_4
2012-04-26_5
2012-04-26_6
2012-04-26_7
2012-04-27_1
2012-04-27_2
2012-04-27_3
2012-04-27_4
2012-04-27_5

Initial
v
(cm/s)
25
25
15
15
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
20
20
20

Ramp
Rate
(cm/s2)
0
0
0
0
-5
-5
-5
-2
-2
-2
-1
-0.5
-0.5
-0.5
-0.25
-0.25
-0.25
-0.25
-0.1
-0.1
-0.1
-0.1
-0.1

Ramp
Time
(s)
N/A
N/A
N/A
N/A
3
3
3
8
8
8
20
50
50
50
100
100
100
100
250
250
200
200
200

Final v Number of Number of
(cm/s) Bifurcations Extinctions
25
N/A
N/A
25
N/A
N/A
15
N/A
N/A
15
N/A
N/A
10
0
0
10
4
7
10
1
3
9
1
0
9
4
8
9
7
9
5
4
10
0
0
14
0
0
11
0
0
1
0
0
14
0
1
13
0
1
14
0
0
11
0
2
7
0
1
9
0
0
8
0
2
11
0
0
4
Table 3. Flame Behavior

94

Approx.
Sample Area
Burned (%)*
100%
100%
90%
90%
70%
60%
60%
70%
70%
70%
60%
40%
40%
50%
60%
40%
60%
60%
70%
80%
70%
70%
70%

Approx.
Mass Air/
Fuel Ratio
26.87
26.33
21.75
23.55
28.47
42.93
36.86
31.65
32.50
33.67
28.61
36.57
31.15
45.51
41.54
40.64
N/A
25.29
33.28
29.12
22.29
22.29
22.29

*In Table 3, the sample area burned was approximated visually. Tests 2012-04-09_1 and 2012-04-09_2 are not included
because their burned sample area could not accurately be determined.

Figure 107. Steady State Flame Spread Rate vs. Oxidizer Flow Velocity for All Tests

95

Figure 108. Steady State Flame Spread Rate Relative to Oxidizer Flow Velocity vs. Oxidizer Flow Velocity for All Tests

96

Figure 109. Test 2012-04-09_1 Track

Figure 110. Test 2012-04-09_2 Track

97

Figure 111. Test 2012-04-18_2 Track

Figure 112. Test 2012-04-18_3 Track

98

Figure 113. Test 2012-04-18_4 Track

Figure 114. Test 2012-04-18_5 Track

99

Figure 115. Test 2012-04-19_1 Overall Track

Figure 116. Test 2012-04-19_1 Initial Track

100

Figure 117. Test 2012-04-19_1 Final Track

Figure 118. Test 2012-04-19_3 Overall Track

101

Figure 119. Test 2012-04-19_3 Initial Track

Figure 120. Test 2012-04-19_3 Final Track

102

Figure 121. Test 2012-04-19_4 Overall Track

Figure 122. Test 2012-04-19_4 Initial Track

103

Figure 123. Test 2012-04-19_4 Final Track

Figure 124. Test 2012-04-20_1 Overall Track
(Outliers have been deleted.)

104

Figure 125. Test 2012-04-20_1 Initial Track

Figure 126. Test 2012-04-20_1 Final Track
(Outliers have been deleted.)

105

Figure 127. Test 2012-04-20_3 Overall Track
(Outliers have been deleted.)

Figure 128. Test 2012-04-20_3 Initial Track

106

Figure 129. Test 2012-04-20_3 Final Track
(Outliers have been deleted.)

Figure 130. Test 2012-04-20_4 Overall Track

107

Figure 131. Test 2012-04-20_4 Initial Track

Figure 132. Test 2012-04-20_4 Final Track

108

Figure 133. Test 2012-04-24_1 Overall Track
(Outliers have been deleted.)

Figure 134. Test 2012-04-24_1 Initial Track

109

Figure 135. Test 2012-04-24_1 Transient Track

Figure 136. Test 2012-04-24_1 Final Track
(Outliers have been deleted.)

110

Figure 137. Test 2012-04-26_1 Overall Track

Figure 138. Test 2012-04-26_1 Initial Track

111

Figure 139. Test 2012-04-26_1 Transient Track

Figure 140. Test 2012-04-26_2 Overall Track

112

Figure 141. Test 2012-04-26_2 Initial Track

Figure 142. Test 2012-04-26_2 Transient Track

113

Figure 143. Test 2012-04-26_3 Overall Track

Figure 144. Test 2012-04-26_3 Initial Track

114

Figure 145. Test 2012-04-26_3 Transient Track

Figure 146. Test 2012-04-26_4 Overall Track

115

Figure 147. Test 2012-04-26_4 Initial Track

Figure 148. Test 2012-04-26_4 Transient Track

116

Figure 149. Test 2012-04-26_5 Overall Track

Figure 150. Test 2012-04-26_5 Initial Track

117

Figure 151. Test 2012-04-26_5 Transient Track

Figure 152. Test 2012-04-26_6 Overall Track

118

Figure 153. Test 2012-04-26_6 Initial Track
(Video recording was not initiated until after burn test had started.)

Figure 154. Test 2012-04-26_6 Transient Track

119

Figure 155. Test 2012-04-26_7 Overall Track

Figure 156. Test 2012-04-26_7 Initial Track

120

Figure 157. Test 2012-04-26_7 Transient Track

Figure 158. Test 2012-04-27_1 Overall Track

121

Figure 159. Test 2012-04-27_1 Initial Track

Figure 160. Test 2012-04-27_1 Transient Track

122

Figure 161. Test 2012-04-27_2 Overall Track

Figure 162. Test 2012-04-27_2 Initial Track

123

Figure 163. Test 2012-04-27_2 Transient Track

Figure 164. Test 2012-04-27_3 Overall Track

124

Figure 165. Test 2012-04-27_3 Initial Track

Figure 166. Test 2012-04-27_3 Transient Track

125

Figure 167. Test 2012-04-27_4 Overall Track

Figure 168. Test 2012-04-27_4 Initial Track

126

Figure 169. Test 2012-04-27_4 Transient Track

Figure 170. Test 2012-04-27_5 Overall Track

127

Figure 171. Test 2012-04-27_5 Initial Track

Figure 172. Test 2012-04-27_5 Transient Track

128

REFERENCES

129

REFERENCES

1.

Oravecz, L.M. “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, 2001.

2.

Tanaya, Stefanus A. “Examination of a Simulated Micro-Gravity Device for Evaluating
Flame Instability Transitions and Flame Spread Over Thin Cellulosic Fuels.”
M.S. thesis. Michigan State University, 2004.

3.

DeRis, J. N. “Spread of a Laminar Diffusion Flame.” Twelfth Symposium (International)
on Combustion, The Combustion Institute, 1969, pp. 241-252.

4.

Fernandez-Pello, A. C., and F. A. Williams. “Laminar Flame Spread over PMMA
Surfaces.” Fifteenth Symposium (International) on Combustion, The Combustion
Institute, 1975, pp. 217-231.

5.

Fernandez-Pello, A. C., and F. A. Williams. “A Theory of Laminar Flame Spread over
Flat Surfaces of Solid Combustibles.” Combustion and Flame, 1977, Vol. 28, pp.
251-277.

6.

Wichman, I.S. and F.A. Williams. "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.

7.

Wichman, I.S. and F.A. Williams. "Comments on Rates of Creeping Spread of Flames
Over Thermally Thin Fuels." Combustion Science and Technology, 1983, Vol. 33,
pp. 207-214.

8.

Olson, S.L. "Near Limit Flame Spread over a Thin Solid Fuel in Microgravity." TwentySecond Symposium (International) on Combustion, The Combustion Institute,
1988, pp. 1213-1222.

9.

Olson, S.L. "Mechanisms of Microgravity Flame Spread over a Thin Solid Fuel: Oxygen
and Opposed Flow Effects." Combustion Science and Technology, 1991, Vol. 76,
pp. 233-249.

10.

Olson, S.L., and H.R. Baum, and Kashiwagi, T. "Finger-Like Smoldering over Thin
Cellulosic Sheets in Microgravity." Twenty-Seventh Symposium (International) on
Combustion, The Combustion Institute, 1998, pp. 2525-2533.

11.

Zik, Ory, Zeev Olami, and Elisha Moses. "Fingering Instability in Combustion." Physical
Review Letters, 2 November 1998, Vol. 81, #18, pp. 3868-3871.

130

12.

Aditjandra, Karin Laksmi. “Population and Sample Burned Area Analyses for Near Limit
Flames in a Simulated Low Gravity Environment Over Thin Cellulosic Fuels.”
M.S. thesis. Michigan State University, 2005.

13.

Olson, Miller, Wichman Olson, S. L., F. J. Miller, and I. S. Wichman. “A New Species
of Fire: Characterizing Fingering Flamelets Using Biological Population
Measures.” Combustion Theory and Modeling, 2006, Vol. 10, No. 2, pp. 323-347.

14.

Pepper, Jacob M., and Fletcher J. Miller. “Characterizing the Narrow Channel Apparatus
as a NASA Standard Material Flammability Test.” 2011 Fall Technical Meeting
of the Western States Section of the Combustion Institute, The Combustion
Institute, 2011.

15.

Hung, Karen W., Sandra L. Olson, and Fletcher J. Miller. “Flame Spread Across
Trioxane in a Narrow Channel Apparatus Simulating Microgravity Flow
Conditions.” 2011 Fall Technical Meeting of the Western States Section of the
Combustion Institute, The Combustion Institute, 2011.

131