Master's Thesis

POLYMETHYLMETHACRYLATE COMBUSTION IN A NARROW
CHANNEL APPARATUS SIMULATING A MICROGRAVITY
ENVIRONMENT
_______________
A Thesis
Presented to the
Faculty of
San Diego State University
_______________
In Partial Fulfillment
of the Requirements for the Degree
Master of Science
in
Mechanical Engineering
_______________
by
Garrett Randall Bornand
Fall 2014

iii
Copyright © 2014
by
All Rights Reserved

iv
DEDICATION
To my family for their continuous love, support, and encouragement.

v
ABSTRACT OF THE THESIS
Polymethylmethacrylate Combustion in a Narrow Channel
Apparatus Simulating a Microgravity Environment
by
Master of Science in Mechanical Engineering
San Diego State University, 2014
Fire safety is an important part of engineering when human lives are at stake. From
everyday homes to spacecraft that can cost hundreds of millions of dollars. The research in
this thesis attempts to provide scientific evidence that the apparatus in question successfully
simulates microgravity and can possibly replace NASA’s current test method for spacecraft
fire safety.
Flame spread tests were conducted with thermally thick and thermally thin
polymethylmethacrylate (PMMA) samples to study flame spread behavior in response to
environmental changes. The tests were conducted using the San Diego State University
Narrow Channel Apparatus (SDSU NCA) as well as within the Microgravity Science
Glovebox (MSG) on the International Space Station (ISS). The SDSU NCA can suppress
buoyant flow in horizontally spreading flames, and is currently being investigated as a
possible replacement or complement to NASA’s current material flammability test standard
for non-metallic solids, NASA-STD-(I)-6001B Test 1. The buoyant suppression attained in
the NCA allows tests to be conducted in a simulated microgravity environment-a
characteristic that NASA’s Test 1 lacks since flames present in Test 1 are driven by buoyant
flows. The SDSU NCA allows for tests to be conducted at various opposed flow oxidizer
velocities, oxygen percent by volume, and total pressure to mimic various spacecraft and
habitat atmospheres.
Tests were conducted at 1 atm pressure, thin fuel thickness of 50 and 75 microns,
thick fuel thickness ranging from 3 mm to 5.6 mm, opposed oxidizer velocity ranging from
10 to 25 cm/s, and oxygen concentration by volume at 21, 30, and 50 percent. The simulated
microgravity flame spread results were then compared to true microgravity experiments
including; testing conducted on the International Space Station (ISS) under the Burning and
Suppression of Solids (BASS) research, NASA’s 5.2 second Drop Tower, and Micro-Gravity
Laboratory’s (MGLAB) 4.5 second Drop Tower. Data was also compared to results found by
Michigan State University’s NCA. Flame spread results from the SDSU NCA compare
closely to that of the other experimental techniques. Additionally, an infrared camera and
species concentration sensors were added to the SDSU NCA and initial results are provided.
Fire Dynamics Simulator (FDS) was used to model the combustion of PMMA within
the SDSU NCA. Both thin and thick fuel beds were simulated and the numerical results were
compared to experimental data. The simulation was then used to determine various results
that cannot easily be found with experimentation, including how effectively the NCA
simulates microgravity under certain environmental conditions, gas and fuel bed
temperatures, heat fluxes, species concentrations, pyrolysis rate, and other various data. The

vi
simulation was found to give reasonable results and overall flame spread trends, but could be
improved upon with further detailed kinetic parameter studies.

vii
TABLE OF CONTENTS
PAGE
ABSTRACT...............................................................................................................................v
LIST OF TABLES.....................................................................................................................x
LIST OF FIGURES ................................................................................................................. xi
NOMENCLATURE ................................................................................................................xv
ACKNOWLEDGEMENTS................................................................................................. xviii
CHAPTER
1 INTRODUCTION .........................................................................................................1
1.1 Microgravity Flame Propagation Testing ..........................................................2
1.2 An Overview of the Problem.............................................................................3
1.3 Thesis Contribution............................................................................................5
2 LITERATURE SURVEY..............................................................................................8
2.1 Progression of Diffusion Flame Spread Theory and Experimentation..............8
2.1.2 Diffusion Flame Spread in a 1-G Environment......................................11
2.1.3 Microgravity Flame Spread ....................................................................12
2.2 Simulating Microgravity with a Narrow Channel Apparatus..........................15
2.3 Numerical Modeling........................................................................................16
3 EXPERIMENTATION................................................................................................18
3.1 Narrow Channel Apparatus..............................................................................18
3.2 Description of Experiments .............................................................................24
3.2.1 Thermally Thin Polymethylmethacrylate ...............................................24
3.2.2 Thermally Intermediate Polymethylmethacrylate...................................26
3.2.3 Burning and Suppression of Solids II on ISS .........................................28
3.2.4 Burning and Suppression of Solids II within the NCA...........................31
3.3 Description of Comparative Experiments........................................................33
3.3.1 MGLAB Drop Tower .............................................................................34
3.3.2 NASA Zero Gravity Research Facility...................................................35
3.3.3 Michigan State University NCA.............................................................36

viii
3.3.4 DARTFire Sounding Rocket...................................................................37
4 THIN PMMA EXPERIMENTAL RESULTS.............................................................39
4.1 Effect of Opposed Oxidizer Velocity and Oxygen Concentration ..................39
4.2 Effect of Gap Height........................................................................................41
4.3 Visual Observations.........................................................................................41
5 THICK PMMA EXPERIMENTAL RESULTS ..........................................................44
6 BASS II EXPERIMENTAL RESULTS......................................................................49
6.2 Species Concentrations....................................................................................52
6.3 BASS II NCA Infrared Imaging ......................................................................52
7 NUMERICAL MODEL...............................................................................................59
7.1 Numerical Model Description..........................................................................59
7.1.2 Computational Domain...........................................................................60
7.1.3 Boundary Conditions ..............................................................................61
7.1.4 Hydrodynamic Model.............................................................................62
7.1.5 Energy Transport ....................................................................................62
7.1.6 Pyrolysis Model ......................................................................................63
7.1.7 Combustion Model..................................................................................63
7.2 Selection of Baseline Polymethylmethacrylate Properties ..............................64
7.3 Domain Sensitivity Analysis............................................................................64
7.4 Grid Sensitivity Analysis .................................................................................64
7.5 Parametric Study of Chemical Kinetics...........................................................66
7.6 Solid Temperature Mapping ............................................................................67
7.7 Effect of Environmental Conditions on Thin PMMA .....................................68
7.8 Effect of Environmental Conditions on Thick PMMA....................................80
8 CONCLUSIONS..........................................................................................................95
9 FUTURE RESEARCH................................................................................................98
9.1 Experimental Research and Preliminary NCA Redesign ................................98
9.2 Numerical Model ...........................................................................................101

ix
REFERENCES ......................................................................................................................103
APPENDIX
A THIN PMMA INPUT FILE.......................................................................................107
B THICK PMMA INPUT FILE....................................................................................117

x
LIST OF TABLES
PAGE
Table 3.1. List of Available Sensors........................................................................................34
Table 7.1. Properties of Polymethylmethacrylate....................................................................65
Table 7.2. Domain Sensitivity Analysis ..................................................................................65
Table 7.3. Grid Sensitivity Analysis........................................................................................66

xi
LIST OF FIGURES
PAGE
Figure 1.1. Schematic of NASA’s test 1 used to conduct flammability tests on non-
metallic solid materials. .................................................................................................4
Figure 1.2. Flame comparison between a 1g environment and a microgravity
environment. ..................................................................................................................4
Figure 1.3. Side view schematic of opposed flow flame spread in a Narrow Channel
Apparatus. ......................................................................................................................5
Figure 2.1. Physical description of a diffusion flame spreading over a stationary fuel
bed..................................................................................................................................9
Figure 2.2. Flammability map for 5 cm wide, 7.6 µm Kimwipes®
. ........................................13
Figure 2.3. Flammability map for PMMA at different half-thicknesses, oxygen mole
fractions and opposed flow velocity. ηg is the non-dimensional flow velocity
and Ro is the radiation number for a quiescent environment.......................................14
Figure 3.1. Section view of the SDSU Narrow Channel Apparatus........................................18
Figure 3.2. SDSU Narrow Channel Apparatus........................................................................19
Figure 3.3. Schematic of the SDSU Narrow Channel Apparatus and flow system.................19
Figure 3.4. NCA fully developed flow solution. Left: side view. Right: top view..................20
Figure 3.5. Normoxic curve.....................................................................................................21
Figure 3.6. Infrared camera calibration device. .......................................................................22
Figure 3.7. Camera mount for the CMOS and infrared camera...............................................23
Figure 3.8. Example position vs. time plot..............................................................................23
Figure 3.9. Thin fuel sample holder.........................................................................................24
Figure 3.10. Schematic of thermally thin PMMA. ..................................................................25
Figure 3.11. Schematic of thermally intermediate PMMA......................................................26
Figure 3.12. NCA false bottom cut-out. ..................................................................................27
Figure 3.13. NCA false bottom insert......................................................................................27
Figure 3.14. Cutout-sample holder. .........................................................................................27
Figure 3.15. Microgravity Science Glovebox (MSG). ............................................................29
Figure 3.16. BASS II inside the Microgravity Science Glovebox (MSG). .............................30
Figure 3.17. Schematic of the BASS II duct............................................................................30

xii
Figure 3.18. BASS II sample holder mount. Dimension H controls the gap height. W
places the sample in the width of the channel (centered). The dimension t
matches the fuel thickness as the mount slides between the sample cards..................31
Figure 3.19. BASS II sample mounted in the SDSU NCA. ....................................................32
Figure 3.20. Exhaust gas filtration system...............................................................................33
Figure 3.21. Schematic of the experimental apparatus used for drop tower
experiments at MGLAB...............................................................................................35
Figure 3.22. Schematic of NASA’s drop tower experimental apparatus.................................36
Figure 3.23. DARTFire schematic...........................................................................................38
Figure 4.1. Effect of opposed velocity and oxygen concentration on thin PMMA
flame spread rate. Error bars are applied using the student’s t-test with a 95
percent confidence interval..........................................................................................40
Figure 4.2. Thin fuel gap height comparison...........................................................................42
Figure 4.3. Side-view flame comparison. Top: 18 mm gap height. Bottom: 6 mm gap
height............................................................................................................................42
Figure 4.4. Top view flame comparison. Left: 30% oxygen, 30 cm/s opposed flow
velocity. Right: 21% oxygen, 7 cm/s opposed flow velocity. .....................................43
Figure 5.1. Effect of opposed velocity on Thick PMMA flame spread rate............................45
Figure 5.2. Effect of opposed velocity and oxygen concentration on Thick PMMA
flame spread rate..........................................................................................................46
Figure 5.3. Residence time as a function of opposed velocity and oxygen
concentration. Where residence time is defined as ...........................................47
Figure 5.4. Opposed flow velocity and oxygen concentration effects on flame length...........48
Figure 5.5. Top view flame comparison of thick PMMA. Left: 21% oxygen, 25 cm/s
opposed flow velocity. Right: 50% oxygen, 10 cm/s opposed flow velocity..............48
Figure 6.1. Example BASS II position vs time plots. (A) 20.6-20.0 O2%, 2 mm thick,
2 cm wide, 1 sided. (B) 17.9-16.9 O2%, 3 mm thick, 2 cm wide, 2 sided...................50
Figure 6.2. Single-sided BASS II flame spread rate................................................................50
Figure 6.3. Double-sided BASS II flame spread rate. .............................................................51
Figure 6.4. Double-sided BASS II NCA with bottom flame lifting due to buoyancy.............52
Figure 6.5. Carbon dioxide and oxygen variation....................................................................53
Figure 6.6. Infrared imaging of the single-sided 3 mm, 2 cm wide BASS II NCA test.
(A) 6 minutes after ignition. (B) 10 minutes after ignition. (C) 17 minutes
after ignition. (D) 22 minutes after ignition.................................................................54
Figure 6.7. Infrared imaging of the double-sided 3 mm, 1 cm wide BASS II NCA test.
(A) 3 minutes after ignition. (B) 10 minutes after ignition. (C) 13 minutes
after ignition. (D) 17 minutes after ignition.................................................................55

xiii
Figure 6.8. Opposed flow velocity effects. (A) 10 cm/s, top view. (B) 6 cm/s, top
view. (C) 10 cm/s, side view. (D) 6 cm/s, side view. ..................................................56
Figure 6.9. Fuel bed width and thickness effects. (A) 1 mm thick, 2 cm wide. (B) 1
mm thick, 1 cm wide. (C) 4 mm thick, 2 cm wide. .....................................................56
Figure 6.10. Opposed flow oxygen concentration effects. (A) 20% oxygen by volume.
(B) 17.5% oxygen by volume......................................................................................57
Figure 6.11. Attempted flame blowout at 10 cm/s. (A) Top view. (B) Side view...................57
Figure 7.1. Computational domain of the thin PMMA simulation..........................................60
Figure 7.2. Computational domain of the thick PMMA simulation........................................60
Figure 7.3. FDS inlet velocity profile comparison. .................................................................61
Figure 7.4. Flame spread rate improvement due to changes in PMMA pyrolysis
properties......................................................................................................................67
Figure 7.5. Parametric study on increasing chemical kinetic parameters individually
by 10% and the effect on flame spread rate.................................................................68
Figure 7.6. Example Matlab solid temperature surface plots. Shown with Smokeview
gas phase temperature output.......................................................................................69
Figure 7.7. Flame spread rate as a function of relative velocity..............................................70
Figure 7.8. HRRPUA showing visualization technique used for area and length
measurements...............................................................................................................70
Figure 7.9. Flame area as a function of relative velocity.........................................................71
Figure 7.10. Flame length as a function of relative velocity. ..................................................71
Figure 7.11. Thin PMMA flame HRRPUA comparison. ........................................................73
Figure 7.12. W-velocity comparison. (A) 1.75 cm/s, 0g. (B) 1.75 cm/s, 1g. (C) 35
cm/s, 1g. (D) 35 cm/s, 0g.............................................................................................75
Figure 7.13. Thin PMMA carbon dioxide comparison............................................................76
Figure 7.14. Thin PMMA temperature comparison.................................................................77
Figure 7.15. Thin PMMA temperature variation with opposed flow. .....................................78
Figure 7.16. Thin PMMA surface temperature........................................................................79
Figure 7.17. Thin PMMA burn rate.........................................................................................80
Figure 7.18. Thin PMMA fuel bed thickness. .........................................................................81
Figure 7.19. Thick PMMA flame spread rate as a function of relative velocity. ....................82
Figure 7.20. Thick PMMA flame area as a function of opposed flow velocity. .....................83
Figure 7.21. Thick PMMA flame length as a function of opposed flow velocity. ..................84
Figure 7.22. Thick PMMA HRRPUA comparison..................................................................85

xiv
Figure 7.23. Thick PMMA carbon dioxide comparison..........................................................86
Figure 7.24. Thick PMMA temperature comparison...............................................................88
Figure 7.25. Thick PMMA maximum temperature variation with opposed flow
velocity.........................................................................................................................89
Figure 7.26. Fuel bed heat flux. (A) Net heat flux. (B) Incident heat flux. .............................89
Figure 7.27. Fuel bed heat flux. (A) Convective heat flux. (B) Radiative heat flux. ..............90
Figure 7.28. Thick PMMA surface temperature profile. .........................................................91
Figure 7.29. Thick PMMA burn rate. ......................................................................................91
Figure 7.30. Thick PMMA fuel bed thickness.........................................................................92
Figure 7.31. Quartz window heat flux. (A) Net heat flux. (B) Incident heat flux. ..................93
Figure 7.32. Quartz window heat flux. (A) Convective heat flux. (B) Radiative heat
flux...............................................................................................................................94
Figure 7.33. Quartz window temperature profile.....................................................................94
Figure 9.1. Gap height adjustment redesign. Left: Improved design allowing removal
of false bottom without the need to readjust gap height. Right: Previous
design.........................................................................................................................100
Figure 9.2. False bottom sleeve and fuel adjustment redesign. .............................................101

xv
NOMENCLATURE
ACRONYMS
BASS II Burning and Suppression of Solids II
CCD Charge Couple Device
CFA Couette Flow Apparatus
CFD Computational Fluid Dynamics
CMOS Complementary Metal-Oxide Semiconductor
DNS Direct Numerical Simulation
DARTFire Diffusive and Radiative Transport in Fires
FDS Fire Dynamics Simulator
GUI Graphical User Interface
HP Hagen-Poiseuille
InSb Indium Antimonide
LES Large Eddy Simulation
MGLAB Micro-Gravity Laboratory
MPET Metalized Polyethylene Terephtalate
MSG Microgravity Science Glovebox
NASA National Aeronautics and Space Administration
NCA Narrow Channel Apparatus
NIST National Institute of Standards and Technology
NDIR Non-Dispersive Infrared
PMMA Polymethylmethacrylate
SDSU San Diego State University
SIBAL Solid Inflammability Boundary At Low-Speeds
SPICE Smoke Point In Co-flow Experiment
UV Ultraviolet
ZGRF Zero Gravity Research Facility

xvi
SYMBOLS
A Solid Phase Pre-Exponential Factor (
BOF Pre-Exponential Factor ( )
Constant Pressure Specific Heat Capacity ( )
E Activation Energy ( )
G Gravity (
g Gravity (9.81
HRRPUA Heat Release Rate per Unit Area ( )
Thermal Conductivity ( )
Characteristic Length of Forward Radiation ( )
N_S Arrhenius Exponents
NU Stoichiometric Coefficients
Pressure (atm.)
Conductive Heat Flux ( )
Convective Heat Flux ( )
Radiative Heat Flux ( )
Arrhenius Reaction Rate (
Upstream Radiative Heat Flux Constant ( )
Downstream Radiative Heat Flux Constant ( )
Temperature ( )
Burning Temperature ( )
Adiabatic Stoichiometric Flame Temperature ( )
Max Temperature ( )
Vaporization Temperature ( )
Oxidizer Free-Stream Temperature ( )
Fuel Bed Thickness ( )
Normalization Fuel Bed Thickness ( )
The x Component of Velocity ( )
The z Component of Velocity ( )
Opposed Flow Air Velocity ( )

xvii
Equivalent Flow Velocity ( )
Flame Spread Rate ( )
Normalized Flame Spread Rate ( )
Opposed Flow Oxidizer Velocity ( )
Horizontal Cartesian Coordinate ( )
Oxygen Mole Fraction ( )
Vertical Cartesian Coordinate ( )
GREEK SYMBOLS
Ignition Region ( )
Gas Phase Conductivity ( )
Hydrodynamic Coefficient
Fuel Bed Conductivity ( )
Density ( )
Fuel Bed Thickness ( )
SUBSCRIPTS
Fuel Bed
Flame
Gas
s Solid
SUPERSCRIPTS
Gas Phase Arrhenius Exponent
Gas Phase Arrhenius Exponent

xviii
ACKNOWLEDGEMENTS
First, I would like to thank my thesis advisor, Dr. Fletcher Miller, for providing
direction and support throughout the duration of this work. I would also like to thank Dr.
Sandra Olson of NASA’s Glenn Research Center and Dr. Indrek Wichman from Michigan
State University for their continuous guidance and encouragement throughout my research.
Their participation in weekly teleconferences has provided countless ideas and suggestions
that continuously helped in the progression of this research.
I also want to thank my committee members, Dr. Subrata Bhattacharjee and Dr. Nagy
Nosseir, for their valuable insight in improving this thesis. Further, Dr. Subrata Bhattacharjee
and the graduate students in the San Diego State University Computational Thermodynamics
Laboratory are due thanks for providing many helpful discussions, supplying the thin fuel
used in this research, and sharing knowledge of key equipment used throughout my work.
Recognition is due to graduate student Jacob Pepper, who wrote the IGOR Pro
interface and setup a large amount of the data acquisition equipment used. Also, graduate
student Derrick Martinez and machinist Michael Lester for constructing the Narrow Channel
Apparatus. I would also like to thank graduate student Gregory Sullivan for his help in
conducting thick fuel experiments, setting up the infrared camera, and data analysis of the
BASS II results. Greg’s help has been a key to my success and is very much appreciated.
Last, but certainly not least, my friends in the San Diego State University Combustion and
Solar Energy Laboratory for their day to day support of my work and for providing an
endless number of outstanding memories.
This research was funded by NASA grant NNX10AD96A under the technical
observance of Dr. Sandra Olson.

1
CHAPTER 1
INTRODUCTION
Humanity’s interest in the heavens has been universal and un-diminishing over the
centuries. The desire to explore the unknown, discover new worlds, push the boundaries of
science and engineering, and then push further is ever-growing. This longing has caused
many to dream of space travel and the ability to explore beyond Earth’s atmosphere. With the
advancement of technology we now have the ability to send manned and unmanned vehicles
into space.
Human space exploration continues to provide answers to many of the questions
about our Universe. It provides knowledge not only of what lies beyond, but expands
technology in ways that where once completely unrelated to space travel. Curiosity and the
eagerness to explore have built peaceful connections between the nations, and continue to
create new industries worldwide. While, the good outweighs the bad, there are many
challenges involved with space exploration that continue to prove dangerous to the astronauts
involved.
Throughout the history of space exploration, there have been incidents of fires taking
place on board spacecraft. On Apollo 13, an oxygen tank ruptured and caused an explosion
that led to all oxygen stores, water, electrical power, and use of the propulsion system to be
lost within 3 hours [1]. The astronauts in danger miraculously made it back to Earth safely,
but the terror of what could have been is an outstanding reminder of the dangers of space
travel. On the Mir Russian Space Station, an oxygen generator failed and caught fire [2]. The
module the fire occurred in was closed and the fire eventually burned out, but not before
causing major damage to some of the hardware on board. The astronauts survived the horrific
incident, but the situation could have easily been much worse.
A crucial piece of the design process is correctly determining the materials that can
safely be used to construct the spacecraft. This is where a key environmental effect causes
problems with ground-based experimentation of such materials. The issue is that the gravity
on Earth leads to buoyancy effects not seen in flames studied in space. The flames are largely

2
dependent upon the buoyant flow produced, and therefore act much differently within the
two environments.
1.1 MICROGRAVITY FLAME PROPAGATION TESTING
When it comes to testing materials in actual microgravity, the National Aeronautics
and Space Administration (NASA), is limited to a small selection of experiments. One option
is to conduct experiments on the International Space Station (ISS). The ISS provides plenty
of time to run tests at true microgravity, but is obviously very costly; both in astronaut time,
and in the cost of transporting the apparatus, equipment, and samples to the ISS. While,
NASA is currently conducting flame spread tests on the ISS under the Burning and
Suppression of Solids II (BASS II) experiments [3], a major goal of these tests is to provide
data for comparisons with simulated microgravity results from ground based
experimentation.
A second option is to conduct flame spread experiments in the interior of a rocket in
freefall. This provides microgravity testing conditions for several minutes, but as with
conducting experiments in Earth orbit, the cost is quite high. An example of this is in the
NASA experiment, Diffusive and Radiative Transport in Fires (DARTFire), where the
effects of low velocity flow, oxidizer concentration, and weak external radiative heat flux
were studied on spreading flames over thick PMMA [4].
Another means of experimenting in true microgravity is to conduct tests in an aircraft
traveling on a parabolic path, which provides approximately 25 seconds of consistent
microgravity [5]. This option is much cheaper than conducting tests in orbit, but limits the
experiment to thin fuels with short experimentation times. Further, this option is still quite
expensive and cannot easily be implemented as an everyday testing method.
A key problem with the methods surveyed so far is the fact that they are all non-
ground-based methods of testing. This means that the tests are not well suited for everyday
experimentation, and a large team is needed to conduct a test. In order to have a ground-
based method of experimentation NASA designed and built the Zero Gravity Research
Facility (ZGRF) at the NASA Glenn Research Center in Cleveland, Ohio. The ZGRF
contains a drop tower that provides approximately 5.18 seconds of true microgravity [6]. The
drop tower, a 143 meter steel chamber that is evacuated to allow experiments to be conducted

3
in free-fall, has provided NASA scientist and engineers with a valuable testing method for
microgravity experimentation, but is limited in the short time span at which it provides a true
microgravity environment. This limits microgravity testing to quick burning solids, not
allowing researchers to experiment with thick solid materials due to the length of time
needed to reach steady flame spread.
1.2 AN OVERVIEW OF THE PROBLEM
As previously mentioned, material flammability testing is an important part of the
design process of any spacecraft. For the obvious safety reasons each and every material
needs to be analyzed at its conditions of use and approved before it can be used to construct
the spacecraft or allowed to be taken into space.
The Narrow Channel Apparatus (NCA) is being studied as a possible replacement for
or complement to NASA-STD-(I)-6001B Test 1, NASA’s current test method used to
conduct material flammability tests on non-metallic solid materials to be used onboard
spacecraft [7]. Since Test 1 is an upward flame propagation test conducted in normal gravity
(see Figure 1.1), buoyancy effects play a large role in the flame behavior. The effect that
gravity has on flames can be seen in Figure 1.2 [8]. Flames are sustained in a gravitational
environment by the natural convection produced in the region near the flame. This natural
convection feeds oxygen to the flame and causes the flame to elongate. In the absence of
gravity the natural convection produced by buoyancy is no longer present, causing the
combustion process to be controlled only by the diffusion of oxygen into the flame region.
These effects cause the flammability tests to typically be conservative in their
determination of whether a particular material, at a given thickness, can be used in spacecraft
or not. This is due to the increased flame spread rate, flame size, and temperatures produced
by the buoyant flow caring additional oxygen to the flame.
The SDSU NCA has the ability to more appropriately simulate actual spacecraft
ventilation conditions by effectively suppressing buoyant effects and allowing the opposed
flow oxidizer velocity to be controlled [9]. Buoyant effects are suppressed in the NCA by
spatially confining the flow with the top plate of the channel (see Figure 1.3), a quartz
window in this case, so that the flame can be observed.

4
Figure 1.1. Schematic of NASA’s test 1 used to conduct
flammability tests on non-metallic solid materials.
Figure 1.2. Flame comparison between a 1g environment and a microgravity
environment. Source: NASA Science Casts. “ScienceCasts: Strange Flames on the
International Space Station.” Last modified June 17, 2013. https://www.youtube.com/
watch?v=BxxqCLxxY3M.

5
Figure 1.3. Side view schematic of opposed flow flame spread in a Narrow Channel
Apparatus.
Research on flame spread in a NCA has been shown to be of importance to not only
microgravity environments, but also fire safety in normal gravity under similar geometric
conditions. In 1998 Swissair Flight 111 crashed into the Atlantic Ocean killing all 229
passengers onboard [10]. The Transportation Safety Board of Canada investigated the crash
of Flight 111, determining that a fire behind the cockpit bulkhead due to a wire arc ignition
of a metalized polyethylene terephtalate (MPET) covered insulating blanket propagated
through a narrow gap starved of oxygen until it reached a vent cap which allowed the fire to
grow rapidly. The creeping flame produced in the narrow gap matched that seen in the NCA,
producing the fingering flamelets formed in the near extinction limit regime.
1.3 THESIS CONTRIBUTION
In order to replace or supplement the current NASA test method a NCA is being
studied to determine if the apparatus can acceptably simulate a microgravity environment
allowing flame propagation testing and an improved material selection process. The
experimental and numerical work presented here attempts to answer many of the questions
about the ability of the NCA to mimic microgravity.
Chapter 2 is a review of the literature on laminar diffusion flames on solids and flame
propagation over nonmetallic solid materials. This review provides the necessary background
on flame spread, microgravity testing, and numerical modeling to contextualize the research
presented in this work.
Chapter 3 provides a detailed description of the experimental apparatus, software, and
equipment used in this research, including an infrared camera and gas analysis sensors. Thin,

6
thick, and BASS II type testing methods are described including the modifications made to
the apparatus that allow testing of each. In order to better understand the experiments the
NCA is being compared to a brief description of each is provided.
Chapter 4 presents experimental results of the burning of thin PMMA sheets in the
NCA at standard ambient conditions over a range of opposed oxidizer velocities. Flame
spread rate comparisons are made to true microgravity experimentation and found to be in
good agreement. The effect gap height has on flame spread rate and the ability for the NCA
to successfully simulate microgravity is studied through visual evaluations and comparisons
to true microgravity flame spread data.
Chapter 5 presents experimental results of the burning of thick PMMA in the NCA at
standard ambient conditions over a range of opposed oxidizer velocities. SDSU NCA results
are compared to results found in a NCA at Michigan State University. A single data point
from a sounding rocket experiment known as DARTFire, and results from the first round of
BASS II experimentation provide true microgravity comparisons.
Chapter 6 presents experimental results of BASS II type testing in the NCA and true
microgravity BASS II results obtained on the International Space Station. Comparisons and
the problems that arise in comparing the two testing methods are examined.
Chapter 7 presents a two-dimensional numerical model of the flame spread over both
thin and thick PMMA. Fire Dynamics Simulator (FDS) developed by the National Institute
of Standards and Technology (NIST) was used to model the process. FDS is a computational
fluid dynamics (CFD) software package that solves transient, low Mach number, buoyant,
reactive Navier Stokes equations. The solid phase decomposition of PMMA is modeled using
a single-step Arrhenius reaction and the gas phase chemical kinetics are modeled using a
finite-rate single-step combustion process. Results are then compared to experimental data
from the SDSU NCA. One of the key motives in simulating combustion in the NCA is the
ability to turn off gravity and study how effective the narrow gap is in simulating a
microgravity flame.
Chapter 8 summarizes the key conclusions of the experimental and numerical
investigation completed in this work.
Chapter 9 provides suggestions for future experimental and numerical modeling
research. A preliminary redesign of the NCA offers initial design ideas on improving from

7
the current NCA including electronic control of the gap height and fuel thickness
adjustments.

8
CHAPTER 2
LITERATURE SURVEY
In general, flames fall into one of two classifications: premixed flames, where the fuel
and oxidizer are mixed prior to ignition, and diffusion flames, where the fuel and oxidizer
meet in a so called flame sheet. Much of the research done on flames has been on premixed
flames due to the increased complexity of diffuse flames. One of the first studies on diffusion
flames was conducted by Burke and Schumann in 1928, suitably titled, “Diffusion Flames”
[11]. The paper presents experimental results for both cylindrical and flat diffusion flames.
Further, theory is introduced that supports the behavior and geometry seen in diffusion
flames. Following the research completed by Burke and Schumann numerous studies have
built upon the theory and experimentation they introduced. The following literature survey
will attempt to give an overview of the research progression from non-solid fuel diffusion
flames to the research presented in this thesis: the flame propagation of a diffusion flame
over thermally thick and thin solid fuel in a simulated microgravity environment.
2.1 PROGRESSION OF DIFFUSION FLAME SPREAD
THEORY AND EXPERIMENTATION
The first article on spreading diffusion flames was presented by de Ris in 1969 [12].
In this article de Ris conducts an analysis of a steady diffusion flame spreading over a solid
fuel bed (thin or thick fuel; each yields different formulae). An illustration from de Ris’
article (Figure 2.1 [12]) provides a physical description of a spreading diffusion flame. As
seen in Figure 2.1 [12], and previously stated, a diffusion flame is defined as a flame where
the fuel vapor and oxidizer meet in a flame sheet. In this flame sheet, the fuel vapor and
oxidizer react in stoichiometric proportions, where the reaction rate is assumed to be
dominated by the diffusion rate of the reactants, rather than by the chemical kinetics.
De Ris made several assumptions in his analysis of a spreading diffusion flame. The
model employed by de Ris used an opposed oxidizer flow. Therefore, there is no forward
(upstream) convection heat transfer from the flame to the fuel bed. All forward heat transfer
to the unburned fuel surface must then come from gas phase conduction and radiation. The

9
Figure 2.1. Physical description of a diffusion flame spreading over a stationary fuel
bed. Source: de Ris, J. N. "Spread of a Laminar Diffusion Flame." Symposium
(International) on Combustion 12, no. 1 (1969): 241-252.
gas phase is assumed to have constant properties and a uniform flow (no boundary layer).
Downstream convection heat transfer is included in both the thin fuel and semi-infinite
(thick) models, but radiation is only included in the semi-infinite model. The temperature
distribution across the thickness of the thin fuel model is assumed uniform.
A few assumptions where made about the fuel bed. First, the unburned fuel is initially
in the condensed solid phase, and stays in this phase until the fuel reaches the vaporization
temperature ( ). At this point, the fuel surface continues to vaporize at with a
constant heat of vaporization. Fuel mass transfer in the vertical direction is assumed to be
purely from diffusion, while boundary conditions at the surface of the fuel are linearized to
provide a good approximation of the convection perpendicular to the surface.
An important assumption made by de Ris was the idea that combustion occurs only in
the gas phase. This assumption leads to the conclusion that the combustion process is
dominated by the mass transfer of reactants rather than the chemical kinetics since the fuel
vapor and oxidizer are considered to mix instantaneously.
De Ris then used the energy and species conservation equations, along with boundary
conditions at the unburned fuel surface, vaporizing fuel surface, and at infinity to find a
solution to the thin fuel bed problem. Non-dimensionalizing, performing a coordinate

10
transformation, taking the Fourier Transform of the governing partial differential equations,
and finally integrating the resulting ordinary differential equations provides de Ris’
approximation for opposed flow flame spread over a thin fuel. This solution is given in
Equation 2.1 as,
(2.1)
where , , and are the fuel-bed density, constant-pressure specific heat capacity, and
thickness, respectively. is the flame spread rate, is the oxidizer free-stream
temperature, is the gas-phase conductivity, and is the adiabatic stoichiometric flame
temperature. Inspection of Equation 2.1 shows a balance between the heat transfer rate
needed to raise the unburned fuel to its vaporization temperature and the gas phase forward
conductive heat transfer rate from the flame to the unburned fuel bed. Notably, in this regime
(referred to by many as the thermal regime), where the mass transfer rates of reactants govern
the combustion process, the flame spread rate is independent of the opposed flow oxidizer
velocity seen by the flame.
De Ris then extended the analysis to include the effects of fuel bed conduction and
radiation. Using the partial differential equations and boundary conditions from the thin fuel
solution de Ris then formulates the fuel bed equations and fuel surface boundary conditions.
Letting the net radiation heat transfer flux received by the fuel bed be,
(2.2)
where , , and are upstream and downstream Radiative heat fluxes, and characteristic
length of forward radiation constants respectively. The equations can then be converted into
three simultaneous Wiener-Hopf integral equations and solved exactly, leading to,
(2.3)
With the definition,

11
(2.4)
Where is the density, is the fuel bed conductivity, is the constant-pressure specific
heat capacity, and is the air velocity. Inspection of Equation 2.3 shows the left hand side
contains the flame spread velocity, the first term on the right describes the effect of forward
gas phase conduction, and the middle and last terms describe the effects of upstream and
downstream Radiative heat transfer, respectively. Notice that the flame spread rate is
inversely proportional to the fuel bed conductivity in the vertical direction.
2.1.2 Diffusion Flame Spread in a 1-G Environment
While de Ris was laying the foundation for the description and theory of steady
diffusion flame spread, McAlevy and Magee performed experiments to measure flame spread
rate and fuel surface temperature for flame spread over two types of thermoplastics in [13].
The experiments run studied the effects of varying the pressure, , and the oxygen mole
fraction, , for flame spread over the two thermoplastics, polystyrene and
polymethylmethacrylate. With the completion of testing a correlation between the flame
spread rate and the combined pressure-oxygen mole fraction was made. This correlation is
given in Equation 2.5 as,
(2.5)
where and are experimentally determined values dependent upon the fuel type and
oxygen diluents used. For PMMA burning in an oxygen-nitrogen mixture, and were
found to be 3.0 and 0.82, respectively. Additionally, they found that the fuel surface
temperature abruptly increases from its initial temperature to the “burning temperature”,
in a small distance, . For PMMA this was determined to be approximately 399˚C (750˚F)
and was independent of gas phase environmental conditions. While, varied from 1.524 mm
to 3.810 mm (0.06 to 0.15 in.) depending on the gas phase environmental conditions.
Following the steep temperature rise, the fuel surface temperature was found to remain
constant at . McAlevy and Magee postulated that the flame spread velocity is controlled by
the “ignition region”, at the leading edge of the flame. A simplified, continuous, diffusive
gas phase ignition model was analyzed and using experimental surface temperature profiles
in the region, a power-law relationship was determined that matched that of the experiment.

12
With this knowledge it was concluded that the gas phase process in the ignition region
strongly influenced the flame spread.
With the completion of their initial analysis, McAlevy and Magee joined Lastrina in a
similar study to find the critical fuel thickness where the thermally thin fuel approximation
can no longer be used [14]. As in McAlevy and Magee’s previous work, it was shown that
the major processes controlling flame spread lie in the ignition region of the leading edge of
the flame, primarily in the gas phase. Experimental correlations between flame spread and
the ratio of specific heats of the oxygen-inert gas mixture and the mole fraction of oxygen
were determined for both cellulose and thermoplastic fuels.
Following the analysis’ and experimentation of de Ris, McAlevy, Magee, and
Lastrina many researchers have further investigated diffusion flame spread and the major
contributing mechanisms involved. Wichman’s investigation on the effects of finite-rate
chemistry on flame spread in [15], highlighted the importance of Damkӧhler number (ratio of
the characteristic particle residence time to the characteristic chemical reaction time) when
the convective mass transfer rate is high enough to compare to the finite-rate reactions in the
gas phase of spreading flames. Rybanin concluded in [16], that when the Damkӧhler number
decreases, the flame spread rate and flame size also decrease. This in turn, can lead to flame
extinction due to heat loss to the surroundings when the flame becomes too small. This limit,
known as the blow-off limit, occurs when the forced flow becomes too high for the flame to
sustain itself.
2.1.3 Microgravity Flame Spread
In the normal gravitational environment of the Earth the extinction limit for a low
forced flow does not exist since the buoyant flows created by the spreading flame are present.
Under microgravity conditions, with the lack of these buoyant flows the flames may act
differently and the influence of small forced convective flows becomes of further interest.
These small forced flows can be seen in microgravity conditions within spacecraft, were the
cabin atmosphere is conditioned and enters the spacecraft at low flows through a ventilation
system.
In 2001, Olson analyzed the influence of oxygen and opposed flow on flame spread
in a true microgravity environment [17]. One of the major contributions of the research was a

13
flammability map showing the three flow regimes that exist in microgravity flame spread. In
Figure 2.2 [17], region I illustrates the thermal regime where the gas phase diffusion of
reactants is the controlling mechanism of flame spread, as seen in the de Ris model. In region
II the Damkӧhler number becomes small due to the high convective flow and the flame
spread is limited by the residence time of the gas phase reactants. In the final region, region
III the flame spread is limited by the oxygen transport to the reacting area, providing a lower
extinction limit, known as the quenching zone.
Figure 2.2. Flammability map for 5 cm wide, 7.6 µm Kimwipes®
. Source: Olson, S.
L. "Mechanisms of Microgravity Flame Spread over a Thin Solid Fuel: Oxygen
and Opposed Flow Effects." Combustion Science and Technology 76 (1991): 233-
249.
In a later study Bhattacharjee et al. presented a flammability map for PMMA which
includes the effects of opposed flow, oxygen concentration, and fuel half-thickness
(Figure 2.3 [18]). The PMMA flammability map [18] shows the transition from the
quenching limit and the thermal region where the necessary oxygen levels are provided for

14
Figure 2.3. Flammability map for PMMA at different half-thicknesses, oxygen mole
fractions and opposed flow velocity. ηg is the non-dimensional flow velocity and Ro is
the radiation number for a quiescent environment. Source: Bhattacharjee, S., R.
Ayala, K. Wakai, and S. Takahashi. "Opposed-Flow Flame Spread in Microgravity-
Theoretical Prediction of Spread Rate and Fammability Map." Proceedings of the
Combustion Institute 30 (2005): 2279-2286.
the fuel half-thickness. As in Olson’s results the oxygen transport to the reacting zone limits
the flame spread. Therefore, if the oxygen concentration is not high enough for a given flow,
the flame will not sustain itself.
In order to study thick PMMA flame spread a sounding rocket, known as the
Diffusive and Radiative Transport in Fires (DARTFire) experiment, was completed. The
research studied the effects of low flow velocities, oxidizer concentrations, and weak external
radiant heat flux on diffusion flames over 20 mm thick black PMMA. Olson determined in
[4], that an opposed flow on the order of diffusive velocities is sufficient to sustain
combustion where a completely quiescent environment would not. Further, flame weakening

15
was noted as the regression of the sample allowed the flame to sink below the floor of the
duct, where the effective flow velocity decreased and heat losses to the sample walls
increased. A key finding Olson mentions is that flame spread rate is more sensitive to oxygen
concentration than flow velocity or external radiant flux. This leads to the conclusion that
operational oxygen concentrations should be set as low as possible because any increase
dramatically increases the fire hazard of the material.
2.2 SIMULATING MICROGRAVITY WITH A NARROW
CHANNELAPPARATUS
In [19], an apparatus called a “Hele-Shaw Cell” was found to suppress the induced
buoyant flow seen in a spreading diffusion flame. NASA scientists revisited the Hele-Shaw
Cell design to develop an apparatus with the ability to simulate microgravity flame spread in
[20].The apparatus was then referred to as a Narrow Channel Apparatus, and research into
developing the NCA into a NASA material flammability test method began. The apparatus,
along with other NCAs, were designed to study flame spread in the near-limit regime, as well
as, a phenomenon known as flame fingering where individual flamelets are formed.
In [21], Olson et al. ran experiments in a NASA NCA and compared the results with
true microgravity flame spread results. Olson determined the NCA effectively suppresses
buoyancy for a thin fuel in opposed flow, allowing research scientist to study microgravity
flame spread characteristics in a simulated microgravity environment.
With the NCA now recognized as a method for simulating microgravity flame spread
researchers needed to study the effect that gap height and width have on flame behavior.
Sidebotham et al. found the gap height that provides the best compromise between buoyancy
suppression and heat loss to the top of the NCA for Whatman 44 filter paper, a cellulose fuel,
at 1 atmosphere is 10 mm [22]. In [23], Zhang and Yu determined the channel width had no
effect on flame spread rate for a “sufficiently wide” sample. Although, variations were seen
over many sample widths and flow rates where differences in heat loss and side oxygen
diffusion caused fluctuations in flame spread rate. In some cases finger-like flames were
produced in samples that, if narrower would not allow self-sustained flames.
A NCA aimed at achieving a linear velocity profile (Couette Flow Apparatus) in
order to provide a more accurate simulation of the flow conditions seen in a spacecraft was
designed, built, and tested at SDSU following the theory Wichman presented in [24]. Hung

16
determined that the CFA’s flame spread results were lower than results found from a NCA
[25]. It was concluded that the additional heat loss to the moving belt used to create the linear
velocity gradient was the cause for the reduced spread rates.
2.3 NUMERICAL MODELING
In [26], Bhattacharjee numerically simulated downward flame spread over solid fuels
in a gravitational field and compared the results with available experimental measurements.
The two-dimensional study focused on correctly modeling the temperature and velocity
fields. The numerical model solves the mass, energy, species-mass, and momentum
equations in the gas phase and the energy equation in the solid phase and includes gas-phase
and pyrolysis kinetics, gas and surface radiation with radiation feedback. Flame spread over
thin cellulose fuels, and both thick and thin PMMA were considered and shown to reproduce
the correct flame structure for a diverse range of fuel and ambient conditions (1atm., 21-50%
O2, 0-75cm/s opposed flow).
Fereres-Rapoport studied the effect of environmental variables on the ignition of solid
fuel through experimental, analytical, and numerical analyses in [27]. The study’s focus was
on the influence of low pressure on ignition. Fereres-Rapoport used Fire Dynamics Simulator
(FDS) to correctly simulate the thermo-physical mechanisms leading to ignition of PMMA
and compare results to experimental findings. It was concluded that reduced pressure
environments result in smaller convective heat losses from the heated fuel to the
surroundings due to a thickening of the thermal boundary layer next to the solid fuel surface,
leading to faster fuel pyrolysis. Further, Fereres-Rapoport concluded the reduced pressure
results in a lower mass flux of volatiles required to reach the lean flammability limit of the
gases at the pilot, leading to a reduction in ignition time mainly due to an enlarged boundary
layer and a thicker fuel species profile. These findings indicate that the flammability of
combustible materials is enhanced at low pressures and elevated oxygen concentrations.
In this research a Narrow Channel Apparatus is being studied as a means to simulate
microgravity flame spread. In order to better simulate the flow conditions a surface flame
may experience. Hamdan, in [28], numerically studied the Couette flow in a finite length
channel, similar to a NCA that can be used to simulate a boundary layer due to a linear
velocity profile near the surface. Hamdan determined the Couette Flow Apparatus (CFA)

17
provided more of a pseudo-Hagen-Poiseuille-Couette flow because of the pressure
differential created along the channel attributed to the pull force along the entrance of the
channel created by the moving top plate as well as the pressure differential created by the
flow exiting the channel. The model was then used to study the combustion of a thin
cellulose sample. The gap height above the sample and the velocity of the top plate were
varied and the effect on the flame spread rate was investigated. 0g flames were found to
spread faster than 1g flames over varying top plate velocities at a set gap height. Varying the
gap height, while holding the top plate velocity constant, presented a crossover phenomenon
on the flame spread rate.

18
CHAPTER 3
EXPERIMENTATION
Chapter 3 presents the San Diego State University Narrow Channel Apparatus and the
key additions made to the NCA, including an infrared camera, species concentration sensors,
and the ability to study thick PMMA. Thermally thin, thick, as well as Burning and
Suppression of Solids II NCA experimentation is explained. The ISS BASS II experiment is
examined, and an overview of the comparative experiments is provided.
3.1 NARROW CHANNELAPPARATUS
The SDSU Narrow Channel Apparatus is an 8.3 cm wide by 100 cm long (in the flow
direction) black anodized aluminum duct with an adjustable gap height from 1 to 25 mm
(Figure 3.1 and Figure 3.2). The gap height is adjusted with the use of a false bottom. An
insert that runs the length of the channel with adjustment screws allowing adjustment up and
down within the channel, therefore increasing or decreasing the gap between the top and
bottom of the channel. A schematic of the NCA is provided in Figure 3.3, where the entire
flow system including, mass flow controllers, filtration system, vacuum system, and species
concentration sensors are shown. The cameras used for image processing are also provided.
Figure 3.1. Section view of the SDSU Narrow Channel Apparatus.

19
Figure 3.2. SDSU Narrow Channel Apparatus.
Figure 3.3. Schematic of the SDSU Narrow Channel Apparatus and flow system.
Alicat MC-50SLPM-D and MC-1SLPM-D mass flow controllers are sent commands
to control the oxidizer flow velocity and oxygen concentration by a remote computer with the
use of a graphical user interface (GUI) created in IGOR Pro by Pepper as explained in [29].
Each gas has a dedicated 50 SLPM and 1 SLPM mass flow controller, allowing calculations
done in the GUI to control the gas flow velocity and composition over the desired range.
Honeycomb flow straighteners are placed at the inlet and outlet to ensure uniform flow
conditions and minimize flow disturbances. The NCA is long enough to provide a fully

20
developed Hagen-Poiseuille flow to the sample region. (Figure 3.4) A top view of the flow in
Figure 3.4 shows the velocity is linear across the width of the sample.
Figure 3.4. NCA fully developed flow solution. Left: side view. Right: top view.
The SDSU NCA is the first of its kind to allow for testing at reduced pressures. The
reduced pressures are achieved with a Gast DOA-P708-AA vacuum pump that can
successfully reduce the pressure to about 28 kPa. The pressure reduction is set with a L.J.
Engineering 329S regulator. Pressure is measured with an Omega DPG1100B-100G pressure
gauge just downstream of the mass flow controllers. The ability to conduct tests at reduced
pressure in simulated microgravity is of interest to fire safety researchers because future
spacecraft cabin atmospheres are proposed to operate at reduced pressures along the
normoxic curve. The normoxic curve is made by holding the partial pressure of oxygen
constant at levels present on Earth as the total pressure is varied as seen in Figure 3.5 [30].
The most current planned spacecraft atmosphere (the red box) in Figure 3.5 [30] is around
34% oxygen with a pressure of 56.5 kPa. The normoxic curve is of great importance because
conditions that fall too far from the curve lead to decompression sickness, hypoxia, and/or
greater material flammability.
A quartz window inserted flush in the lid of the channel allows for 5 megapixel video
capture of the flame spread with a Silicon Video 5c10 CMOS video camera. Video is
recorded using the provided Epix Xcap version 3.7 for Windows software. Additionally,
there is another quartz window on the side of the NCA for images and viewing. The side
view allows for visual observations and comparisons of the flame shape, size, height, and
length.
0 10 20 30 40 50
0
0.1
0.2
0.3
0.4
0.5
Velocity (cm/s)
ChannelHeight(cm)
0 10 20 30 40 50
-4
-2
0
2
4
Velocity (cm/s)
ChannelWidth(cm)

21
Figure 3.5. Normoxic curve. Source: Campbell, P. Recommendations for
Exploration Spacecraft Internal Atmospheres: The Final Report of the NASA
Exploration Atmospheres Working Group. Houston: National Aeronautics
and Space Administration, 2006.
A Raytheon Radiance HSX high speed infrared camera with an indium antimonide
(InSb) sensor allows for infrared imaging between 3 and 5 microns. An Amber 25 mm lens
(F 2.3) and a Janos 50 mm lens (F 2.3) with a flame filter are used on the camera providing
the correct focus and wavelength range for the intended object. The flame filter blocks
wavelengths produced by the flame, therefore removing the flame from the image and
allowing the camera to see only the sample. The camera software, ImageDesk II lets the user
adjust a variety of setting to produce the best image possible. Calibration of the camera was
achieved with the use of a calibration device made in-house. Figure 3.6 shows the device that
consists of a sheet of ceramic fiber board machined to allow for a heating element and an
aluminum plate to sit inside. The temperature of the heating element is controlled with a
variable transformer that allows the user to select an input voltage to the heating element.
The thermocouples on the surface of the aluminum can then be used to determine the steady
state temperature of the plate. A calibration curve was made by holding the plate at a
constant temperature and recording the infrared intensity, then repeated over the entire range
(25-365˚C). Using the calibration curve a Matlab program was developed providing a means

22
Figure 3.6. Infrared camera calibration device.
to plot the temperature fields. The Matlab code was developed because ImageDesk II was
extremely non-user friendly and user manuals were not available.
A stronger camera mount was designed and built to allow the much heavier infrared
camera to be mounted above the NCA with the option to adjust the distance from the NCA
quartz window for proper focusing with the lens being used (Figure 3.7). The new design
allows for both, the infrared and standard CMOS camera to be simultaneously used during
testing. Further, the cameras are mounted on slides for easy positioning adjustment.
The Epix Xcap version 3.7 for Windows software outputs a video file of the flame
from above in AVI format. The AVI file is then compressed using VirtualDub 1.9.11
software. Spotlight-16 software is used to track the leading edge of the flame [31]. A scale is
set in the software allowing it to convert from pixels to distance. In most cases Spotlight-16
will automatically track the flame by following a set threshold value, but in some cases
(primarily with thick fuels) the user must manually click on the flame front for each frame.
At the completion Spotlight outputs data in the form of position vs. time. This data is then
copied into Microsoft Excel and plotted. Figure 3.8 shows an example of this plotted data for
a sample of thick PMMA. Results in Figure 3.8 are from a test at 15 cm/s with an initial
oxygen concentration of 21 percent followed by a change to 30 percent oxygen concentration

23
Figure 3.7. Camera mount for the CMOS and infrared camera.
Figure 3.8. Example position vs. time plot.
0 100 200 300 400 500 600 700
0
10
20
30
40
50
60
70
80
90
100
Position[mm]
Time [s]
x = 0.046t + 18.919
R2
= 0.997
x = 0.308t - 125.515
R2
= 0.994
AOI
21% O2
30% O2

24
at 490 seconds. It is noticeable that the flame spread follows a characteristic linear position
vs. time relationship, allowing for a linear curve fit to the data. A clear change in flame
spread velocity can be seen when oxygen concentration and/or opposed flow velocity is
changed. Multiple test conditions run on a single sample were found to agree with tests run at
a single uniform condition. Therefore, multiple oxygen concentrations and/or velocities were
performed per individual thick PMMA sample. This was not done with the quick spreading
thin fuels as the transition times are of greater overall time percentage.
3.2 DESCRIPTION OF EXPERIMENTS
Thin, thick, BASS II within the NCA and aboard the ISS techniques are all explained
in the following sections. Because thin fuel was the only fuel type previously tested in the
SDSU NCA the modifications made to the channel are also described. Further, a NCA
redesign follows in Chapter 9, where larger changes are suggested.
3.2.1 Thermally Thin Polymethylmethacrylate
The sample is overlapped with and taped to a 0.74 mm thick stainless steel sample
holder with a 5.1 cm x 30 cm cutout to match the sample size used in NASA Test 1
(Figure 3.9). Care is taken to keep the sample taut. The sample holder then holds the sample
in the center of the channel as shown in Figure 3.10. A 27 gauge Kanthal igniter wire with a
small piece of paper around it is used to ignite the PMMA. Without the paper, the wire can
slice through the PMMA without igniting it.
Figure 3.9. Thin fuel sample holder.

25
Figure 3.10. Schematic of thermally thin PMMA.
Note: Not to Scale
g
GAP HEIGHT
GAP HEIGHT

26
Thin fuel tests presented in this thesis, other than in the gap height comparison, have
a total gap height of 10 mm. The sample (50/75 micron) is placed directly in the middle of
the gap, providing an equal gap above and below the sample. This leads to a gap of 5 mm
from the top quartz window in the lid, and 5 mm from the aluminum false bottom of the duct.
3.2.2 Thermally Intermediate
Polymethylmethacrylate
All tests were performed at a gap height of 5 mm and pressure of 1 atm
(Figure 3.11). The PMMA was milled from a sheet of 0.220 inch thick clear Lucite Lux cast
acrylic.
Figure 3.11. Schematic of thermally intermediate PMMA.
A 5.08 cm (2 in.) by 10 cm cut-out in the false bottom is located 7 cm from the back
of the channel and is centered in the width direction to house the sample during testing. The
cutout (Figure 3.12) was made from the previous insert, that when removed, was designed to
allow liquid fuels in the NCA. This was done by machining the original insert shown in
Figure 3.13. The insert had the 10 cm section cut out, allowing the sample to sit between the
two original pieces as seen in Figure 3.14. The location allows the flow to fully develop and
the flame to be viewed from the side quartz window. A steel spacer and the PMMA sample
sit inside the cut-out so the top of the sample is flush with the top of the false bottom. As the
sample burns it tends to lift and bow. Strips of aluminum tape are used along the sides of the
sample to prevent the lifting and bowing which disturbs air flow. Care is taken to keep the
sample flat as well as to remove any minor scuffs or defects which could affect the results by
perturbing the air flow. Two 27 gauge Kanthal wires are twisted together to make the igniter.

27
Figure 3.12. NCA false bottom cut-out.
Figure 3.13. NCA false bottom insert.
Figure 3.14. Cutout-sample holder.

28
The igniter wire is pressed down flat and evenly across the top of the PMMA sample near the
back to create a flat flame front upon ignition. Ignition takes 10 to 30 seconds depending on
the flow (longer for slower flows and lower oxygen concentrations) at 9 volts and 11 amps.
During testing the CMOS camera gain was set to 17.8 decibels, exposure to 136
milliseconds and the frame rate to 2 frames per second, providing the best image quality and
slowest frame rate available in the Epix Xcap software.
3.2.3 Burning and Suppression of Solids II on ISS
The Burning and Suppression of Solids II (BASS II) experiment is a follow-on
experiment to the original BASS experiment that demonstrated the ability to investigate
flammability, flame spread, extinguishing, etc. in the Microgravity Science Glovebox (MSG)
working volume (Figure 3.15 [32]). The experiment utilizes slightly modified Smoke Point
In Co-flow Experiment (SPICE) hardware (shown inside the MSG in Figure 3.16). Key
personnel consist of Project Investigator Dr. Sandra Olson, four Co-Investigators including;
Drs. Fletcher Miller, Subrata Bhattacharjee, James T’ien, Carlos Fernandez-Pello, and
Project Scientist, Dr. Paul Ferkul. The experiment consisted of 100 fuel samples covering
thin and thick flat samples, rods, and solid spheres. The flat samples of different thicknesses
were made of PMMA and Solid Inflammability Boundary At Low-Speeds (SIBAL)
materials. The rod sample material was black and clear PMMA, and the solid spheres were
also PMMA.
The modified SPICE hardware in Figure 3.17 [33] consist of a 7.62 cm wide, 7.62 cm
tall, and 17.5 cm long duct with a forced flow from an inlet fan that pulls gas from the
MSG’s 255 liter working volume. An anemometer is placed to give the average flow velocity
of the incoming gases. A radiometer in the back corner provides flame radiation readings. In
the exhaust of the duct a heat sink and a filter cool and clean the exhaust gases before
allowing them to exit into the MSG working volume.
The BASS II work presented in this thesis is under Co-Investigator Dr. Fletcher
Miller. The work is on thick PMMA sheets ranging from 1 to 5 mm in thickness, 1 and 2 cm
in width, and consists of single and dual sided burns in opposed flow. The oxygen
concentration during testing is constantly decreasing due to small volume of the MSG and
the fact that the incoming gas is pulled from the same chamber.

29
Figure 3.15. Microgravity Science Glovebox (MSG).
Source: European Space Agency. "Space in Images." Last
modified November 29, 2005.
http://www.esa.int/spaceinimages/Images/2005/11/Microgr
avity_Science_Glovebox.

30
Figure 3.16. BASS II inside the Microgravity Science Glovebox (MSG).
Figure 3.17. Schematic of the BASS II duct. Source: Bhattacharjee, S. "Research on
Flame Spread at SDSU: The Bhattacharjee Group." Accessed November 5, 2014.
http://flame.sdsu.edu/.

31
During testing live video of the top view camera and the side view video camera
allowed viewing of the experiment. Through telecommunication with NASA instructions
could be sent to the astronaut allowing for interactive changes in flow, radiometer settings,
and other experimental settings. The astronaut would place the sample into the duct with the
igniter downstream, providing opposed flow of the incoming oxidizer. The oxygen
concentration would be changed to roughly the desired concentration by opening the MSG
door for a set time, allowing oxygen to flow into the chamber and therefore raising the levels
or by the addition of nitrogen into the chamber while releasing gases in the chamber to hold
the pressure constant. The fan could then be set to the desired flow rate and when ready
ignition was provided by a ribbon Kanthal wire at the back side of the sample directly
between the sample holders.
3.2.4 Burning and Suppression of Solids II within the
NCA
To better replicate the conditions from BASS II, sample holder mounts were made to
allow the standard BASS II samples to be tested within the SDSU NCA. Multiple aluminum
mounts were machined to provide the correct 5 mm gap height (H) for the different sample
thicknesses (Figure 3.18). The mounts place the samples directly in the middle of the width
of the channel with minimal flow disturbances in a way similar to the mount used in the
BASS II experiments.
Figure 3.18. BASS II sample holder mount. Dimension H controls the gap
height. W places the sample in the width of the channel (centered). The
dimension t matches the fuel thickness as the mount slides between the
sample cards.

32
The same twisted Kanthal igniter wire used to ignite the thick PMMA was used for
ignition of the BASS II NCA samples. The igniter was placed in the gap between the sample
holders (cards) just as in the BASS II experiments providing similar ignition and flow
conditions. The sample and igniter sandwiched between the sample cards and mounted to the
machined sample mount in the NCA can be seen in Figure 3.19.
Figure 3.19. BASS II sample mounted in the SDSU NCA.
The species concentrations of carbon monoxide, carbon dioxide, and oxygen are of
interest because of the unique microgravity simulation characteristics that the NCA allows.
Therefore, sensors were added to measure each of the species concentrations, allowing for
future comparisons to true microgravity experiments.
A filtration system was designed to remove particulate from the exhaust of the NCA,
allowing the sensors to be run inline without damage (Figure 3.20). The system successfully
removes particulate down to roughly 3 microns. This was achieved by installing a K&N air
filter inside clear acrylic piping. The exhaust gases of the NCA flow into one side of the
acrylic pipe, where the filter is installed as seen in Figure 3.20. The exhaust then must pass
through the filter where the soot is caught. Then the gases can travel up the pipe and are then
reduced back to the original sized tubing before reaching the species concentration sensors.
Carbon dioxide concentrations are measured with a K-33 ICB CO2 sensor from
CO2Meter. The non-dispersive infrared (NDIR) sensor can measure carbon dioxide
concentrations between 0 and 30 percent with a resolution of 0.001 %. During testing carbon
dioxide data is collected using the provided Gaslab software.

33
Figure 3.20. Exhaust gas filtration system.
Carbon monoxide and oxygen concentrations are measured with the use of an Enerac
M500. The standard carbon monoxide sensor did not provide a high enough range and was
replaced with a Membrapor CO/SF-40000-S electrochemical sensor with a range of 0-40,000
PPM. The Membrapor sensor has a resolution of 10 PPM providing high quality results.
Table 3.1 provides an overview of the sensors within the Enerac and the individual carbon
dioxide sensor used. The Enerac software used is Enercom. Enercom allows the user
complete control over the Enerac, as well as providing tables and graphs of all the data. After
testing the data can be exported to a CSV file for plotting and data analysis.
3.3 DESCRIPTION OF COMPARATIVE EXPERIMENTS
Results found in the SDSU NCA and aboard the ISS are compared to many other
experimental results found elsewhere. This section attempts to provide a brief overview of
the experiments and techniques used.

34
Table 3.1. List of Available Sensors
Measured Parameters Range Resolution Accuracy
Carbon Dioxide (CO2)
Non-Dispersive Infrared
0-30 % 0.001 % +/- 3 % M
Oxygen (O2)
Electrochemical Cell
0-25 % 0.1 % +/- 0.2 % M
Carbon Monoxide (CO)
0-40,000 PPM 10 PPM +/- 12 % M
Nitric Oxide (NO)
0-2,000 PPM 1 PPM +/- 2 % M*
Nitrogen Dioxide (NO2)
0-1000 PPM 1 PPM +/- 2 % M*
Sulfur Dioxide (SO2)
0-2,000 PPM 1 PPM +/- 2 % M*
Combustibles
Catalytic Sensor
0-5 % 0.1 % +/- 2 % (CH4) M
M = Measured, * +/- 1 to 2 PPM for less than 100 PPM range.
3.3.1 MGLAB Drop Tower
The experimental apparatus used is shown in Figure 3.21 [34]. The apparatus is a
closed-type wind tunnel with a 340 mm long x 100 mm wide x 190 mm high chamber. Flow
is created by a small fan driven by a brushless DC motor producing flow velocities between 0
and 150 mm/s. The test section where the sample holder is located is 80 mm x 80 mm. The
thin PMMA samples (60 mm long x 10 mm wide) are attached to the sample holder and
ignition is accomplished via Ni-Cr resistance wire. Three thicknesses were studied; 15, 50,
and 125 microns, and oxygen concentrations of 21%, 30%, and 50% by volume were used.
Flame spread was recorded with a CCD camera under an ambient temperature and pressure
of 300 K and 1 atm, respectively.
Microgravity is obtained with the 4.5 second drop tower of Micro-Gravity Laboratory
(MGLAB) in Gifu, Japan. Ignition starts 1.6 s before the drop and the apparatus is dropped
0.3 s later. When fuel thicknesses of 15 μm and oxygen concentrations above 30% are used
ignition is started after the drop because the flame spread rate is relatively fast.

35
Figure 3.21. Schematic of the experimental apparatus used for drop
tower experiments at MGLAB. Source: Takahashi, S., M. Kondou,
K. Wakai, and S. Bhattacharjee. "Effects of Radiation Loss on
Flame Spread over a Thin PMMA Sheet in Microgravity."
Proceedings of the Combustion Institute 29, no. 2 (2002): 2579-2586.
3.3.2 NASA Zero Gravity Research Facility
The NASA Zero Gravity Research Facility at Glenn Research Center is a 142 m long
steel vacuum chamber with a 132 m free-fall distance that provides 5.18 seconds of
microgravity. A 5-stage vacuum pumping process reduces the pressure in the tower to 0.05
torr in approximately one hour, reducing the aerodynamic drag on the freefall vehicle to less
than 0.00001 g. A crane is used to position the vehicle and release mechanism at the top of
the vacuum chamber. At the completion of the fall the experiment vehicle is stopped in the
decelerator cart, located at the bottom of the chamber. The decelerator cart is 6.1 m deep and
filled with 3 mm diameter expanded polystyrene beads that dissipate the kinetic energy of the
2500 lb. experiment vehicle bringing it to a stop in about 4.6 m with a peak deceleration rate
of about 65 g.
The experimental apparatus (Figure 3.22 [35]) sits inside the drop vehicle and consist
of a low-speed flow tunnel that provides up to ~30 cm/s forced flow of gas through a 20 cm
ID duct at 0-16 psia pressure and 0-100% in diluent. Two separate bottles contain the

36
Figure 3.22. Schematic of NASA’s drop tower experimental
apparatus. Source: Olson, S. L., and G. Ruff. "Microgravity
Flame Spread over Non-Charring Materials in Exploration
Atmospheres: Pressure, Oxygen, and Velocity Effects on
Concurrent Flame Spread." Society of Automotive Engineers
Technical Paper Series 1 (2009): 1-7.
oxygen and diluents that allow for the desired oxygen concentration and pressure. A back
pressure valve controls the total pressure within the system. Prior to the drop the flow is
started to establish a steady flow and pressure in the tunnel. Ignition of the fuel starts just
before or at the start of the drop allowing the fuel to ignite and reach a steady flame spread
rate before reaching the decelerator cart. At the completion of the drop the test section is
vented to vacuum to extinguish the flame.
3.3.3 Michigan State University NCA
The Narrow Channel Apparatus at Michigan State University (MSU) is similar to the
SDSU NCA with only a few key differences. The most influential difference of the MSU
NCA is the width of the channel. To better study flame “fingering” a 30.5 (12 in) wide
channel was made that allows for much wider samples. Although the MSU NCA has the

37
ability to study much wider samples the comparisons in this thesis were tested with 3.81 cm
(1.5 in) wide samples. Another difference is the MSU NCA bottom (sample holder in the
case of thick fuels) is made of steel, reducing the heat loss from the hot fuel to the channel.
3.3.4 DARTFire Sounding Rocket
A sounding rocket was used to study the effect of low velocity flow (on the order of
diffusive velocities), oxidizer concentration, and weak external radiant heat flux on the flame
spread over thermally thick PMMA under the Diffusive and Radiative Transport in Fires
(DARTFire) experiment [4]. The experiment consist of twin flow tunnels each providing 1-
10 cm/s flow through the 10x10 cm cross-section and 15 cm long tunnel, as shown in
Figure 3.23 [4]. 20 mm long x 20 mm thick x 6.35 mm wide, black PMMA samples are
placed in the floor of the duct 4 cm from the inlet flow straighteners. The PMMA was
insulated from the aluminum floor with ~ 0.7 mm thick Fiberfrax®
insulation. The flow was
controlled by setting the pressure upstream of a flow orifice. The velocity profile was
checked for smooth uniform flow across the duct using a hot-wire anemometer at 5, 10, 15,
and 20 cm/s. The pressure within the duct was held at a constant 1 atmosphere.
A near-infrared laser diode (812 nm) with custom lenses were mounted in the ceiling
allowing for experiments under a uniform external radiant flux of up to 2 W/cm2
on the
sample surface. The irradiation provided to the sample was uniform to within 5%.
Gas phase thermocouples (0.025 mm diameter type-R) were mounted at 1, 2, and 3
mm above the sample. Three bare bead thermocouples (0.076 mm diameter type-K) provide
surface temperatures and one cylindrical bump thermocouple records subsurface
temperatures. Side posts block glowing of the thermocouples from camera views. An
intensified array UV video camera with a resolution of 0.1 mm records at 5 frames/s from an
edge view. The camera takes images of the chemiluminescence of OH* radical species in the
flame using appropriate filters (20-nm bandwidth, centered at 310 nm).
Two experiments are conducted simultaneously (one per tunnel) during each of the
four ~ 6 minute microgravity flights. When the sounding rocket enters the microgravity
portion of its flight , the flow begins; once flow has been established
throughout the system, the igniters ignite the samples simultaneously, and the laser diode
irradiates one of the two samples. Five seconds later the igniters turn off. At a prescribed

38
Figure 3.23. DARTFire schematic. Source: Olson, S., U. Hegde, S. Bhattacharjee, J.
Deering, L. Tang, and R. Altenkirch. "Sounding Rocket Microgravity Experiments
Elucidating Diffusive and Radiative Transport Effects on Flame Spread Over
Thermally Thick Solids." Combustion Science and Technology 176 (2004): 557-584.
time the laser turns off and then back on again at a later time, allowing evaluation of the
effect of a heat flux change on flame spread rate. The test ends when the rocket begins to
reenter the atmosphere by forced extinction via vacuum exposure.

39
CHAPTER 4
THIN PMMA EXPERIMENTAL RESULTS
Thermally thin PMMA flame spread experiments were conducted in the SDSU NCA.
Opposed oxidizer velocity and oxygen concentration effects were studied and the results
compared to true microgravity experiments from NASA’s 5.18 s drop tower and the ISS. Gap
height effects are analyzed and a visual comparison of flames is provided.
4.1 EFFECT OF OPPOSED OXIDIZER VELOCITY AND
OXYGEN CONCENTRATION
Figure 4.1 shows the flame spread rate as a function of the relative velocity between
the flame and the opposed flow. For an opposed oxidizer flow, the relative velocity is the
opposed flow velocity plus the flame spread rate. Tests conducted with fuel thicknesses other
than 75 μm were normalized to better compare the data. Normalization was achieved using a
simple thickness ratio as follows:
Where , is the normalized flame spread rate, is the actual flame spread rate,
is the fuel thickness, and is 75 μm (the thickness to which all other fuels are normalized).
This is based on the idea that for a thermally thin fuel the flame spread rate is inversely
proportional to fuel thickness.
Flame spread results for a total gap height of 10 mm are compared to those of [35]
and [34]. In [34] an equivalent flow velocity ( ) was defined in order to correct for
boundary layer development. Equations were taken from [36] because of inaccurateness in
[34]. The equivalent flow velocity is therefore defined as:
where is the hydrodynamic coefficient and is the opposed oxidizer velocity. While a
range of values (average 0.28) for the hydrodynamic coefficient were used in [34], a single
value of 1/3 is used for comparisons within this work, as in [35].

40
.
Figure 4.1. Effect of opposed velocity and oxygen concentration on thin
PMMA flame spread rate. Error bars are applied using the student’s t-
test with a 95 percent confidence interval.
In Figure 4.1 we can see good overall agreement between the data sets. Normalization
for fuel thickness seems to collapse the data fairly well except for a few of the thickest 125
μm data points at 30 % oxygen and a single 50 μm data point at 21 %. Where there is overlap
between true microgravity and NCA data there is generally very good agreement. The NCA
data also agree well with the one NASA microgravity test at 30% oxygen and 30 cm/s
relative velocity. Poor agreement is seen at the very low end of the relative velocity scale
where the flames extinguish. The NCA flames suffer more heat loss due to the proximity of
the quartz window and the bottom plate, so that those flames tend to go out sooner than the
true microgravity flames which do not have this loss mechanism.
The effects of oxygen concentration can be seen throughout Figure 4.1. At 21 percent
oxygen concentration it is noticed that the flame spread rate raises, plateaus, and then begins
to drop once again, so that there is an optimal relative velocity that maximizes the flame
spread rate. As oxygen concentration is raised to 30 percent it is noticeable that the drop
immediately following the plateau no longer exists within the velocities tested here. At 50
0 5 10 15 20 25 30 35 40
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Relative Velocity [cm/s]
NormalizedFlameSpreadRate[cm/s]
SDSU-75m,21%
SDSU-75m,30%
SDSU-75m,50%
SDSU-50m,21%
SDSU-50m,30%
SDSU-50m,50%
MGLAB-15m,21%
MGLAB-50m,21%
MGLAB-125m,21%
MGLAB-15m,30%
MGLAB-50m,30%
MGLAB-125m,30%
MGLAB-15m,50%
MGLAB-50m,50%
MGLAB-125m,50%
NASA-25m,21%
NASA-25m,30%
NASA-25m,30%,70.3kPa

41
percent oxygen there is a short plateau, followed by a second rise in the flame spread rate.
Further, the oxygen concentration plays a large role in the flame spread rate.
4.2 EFFECT OF GAP HEIGHT
From Figure 4.2 it is clear that gap height plays a large role in the flame spread rate.
As the gap height is lowered buoyancy effects become less dominant and the flame begins to
experience simulated microgravity conditions. While this is sought after in a NCA, there is a
point when the heat loss to the top and bottom plates will cause unrealistic spread rates. The
effect gap height plays on a flame can be seen in Figure 4.3. Both tests were for 75 μm
PMMA at 1 atm pressure, opposed oxidizer velocity of 15 cm/s, and 21% O2 concentration
by volume. The top test was set to a total gap height of 18 mm (9 mm above and 9 mm below
the sample). The bottom test was set to a total gap height of 6 mm. It is obvious that the two
flames act very differently from one another, while all other conditions were the same. In the
18 mm test it is visible that the buoyancy effects are still largely acting on the flame. The
flame visibly slopes upward and is bright yellow. In the 6 mm test the buoyancy effects are
obviously suppressed and the flame turns much bluer. It is unclear from Figure 4.3, but the
flame length in the flow direction also shortens greatly as seen in Figure 4.4.
In Figure 4.2 there is a noticeable difference between each gap height. At a total gap
height of 6 mm the flame spread rate is greatly reduced and the maximum spread rate shifts
toward a slower relative flow. As the gap height is raised the difference in flame spread rate
is reduced. This is expected, since the heat loss to the top and bottom plates will be reduced
as the plates are distanced from the flame. At some point the buoyancy effects will no longer
grow stronger and the flame spread rate will be that of an open flame.
4.3 VISUAL OBSERVATIONS
During testing it was noticed that the non-charring PMMA would melt and bubble as
it was burned. Afterword, some of the melted PMMA would be left, unburned on the bottom
plate of the narrow channel apparatus. This could possibly cause a change in the flame spread
rate. Other noticeable effects during the burning process consist of changes in brightness,
color, and length of the flame in the flow direction as shown in Figure 4.4. In the lower
extinction limit and blow-off regions the flame became much less yellow and turned bluer in
color. It would also shrink in length and brightness. As with opposed flow velocity, the

42
Figure 4.2. Thin fuel gap height comparison.
Figure 4.3. Side-view flame comparison. Top: 18 mm gap height.
Bottom: 6 mm gap height.
oxygen percentage caused the same effects. The higher oxygen percentages caused brighter,
yellow, and longer flames.
0 5 10 15 20 25 30 35
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Relative Velocity [cm/s]
NormalizedFlameSpreadRate[cm/s]
SDSU - 6mm,75m
SDSU - 10mm,75m
SDSU - 14mm,75m
SDSU - 18mm,75m
MGLAB - 15m
MGLAB - 50m
MGLAB - 125m
NASA - 25m
18 mm
6 mm

43
Figure 4.4. Top view flame comparison. Left: 30% oxygen, 30 cm/s
opposed flow velocity. Right: 21% oxygen, 7 cm/s opposed flow velocity.
1.5 cm

44
CHAPTER 5
THICK PMMA EXPERIMENTAL RESULTS
Flame spread experimentation on thick PMMA in the thermally intermediate range
was conducted. Opposed oxidizer velocity and oxygen concentration effects were analyzed
and the results compared to related experiments from Michigan State University and NASA.
Flame length, residence time, and visual observations offer further analysis and
understanding of the flame characteristics.
San Diego State University data was plotted against data from Michigan State
University where similar tests were conducted within a wider Narrow Channel Apparatus
(Figure 5.1). Error bars are applied using the Student’s T-Test with a 95 percent confidence
interval at all 21 percent oxygen concentrations. Error Bars applied to the 30 and 50 percent
tests were found at a few points and the largest applied throughout the range of values. Flame
spread below an opposed flow of 12 cm/s is difficult to achieve. In the SDSU NCA
extinction was determined to be close to 11 cm/s, although MSU was able to record results at
a slightly reduced flow of 10 cm/s. Opposed flows above 25 cm/s were not researched as the
main interest lies in spacecraft ventilation flows that do not reach levels high enough to
achieve blow-off. Further, as the opposed flow increases the flame spread reduces and
becomes less of a fire hazard.
Initial results follow closely to what MSU has determined for their 0.5 inch thick
PMMA (Black and Clear). While MSU’s sample thickness is relatively large compared to the
samples used here, the agreement is good. It is noticeable that the majority of the SDSU data
fall just below the results found by MSU. This is expected, because the MSU sample holder
is made from steel and therefore, heat losses are reduced compared to the aluminum false
bottom of the SDSU NCA. In the SDSU NCA at low opposed flows the flame spread rate is
higher than the MSU results and a large decrease in flame spread rate is not shown.

45
Figure 5.1. Effect of opposed velocity on Thick PMMA flame spread rate.
Further flame spread rate testing was done at 30 and 50 percent oxygen concentration
by volume. A single data point from the NASA DARTFire experiments is plotted along with
the SDSU data [6]. The DARTFire result is for a 50 percent oxygen concentration at an
opposed flow of 10 cm/s.
The 30 percent and 50 percent oxygen concentrations along with the DARTFire
results are plotted with 21 percent oxygen concentration in Figure 5.2. The large effect
oxygen concentration has on the flame spread rate can be easily seen. While, the flame
spread rates of the higher oxygen concentration levels are much higher the extinction limit
seems to be approximately the same at roughly 9cm/s. The effects of oxygen concentration
can be seen throughout Figure 5.2. In Figure 5.1 with 21 percent oxygen concentration it is
noticed that the flame spread rate rises, reaches a peak, and then begins to drop once again,
so that there is an optimal opposed velocity that maximizes the flame spread rate. As oxygen
concentration is raised to 30 percent it is noticeable that the drop immediately following the
peak no longer exists within the velocities tested. At 50 percent oxygen there is no peak, but
rather a constant rise in the flame spread rate with opposed flow. It is notable that the results
8 10 12 14 16 18 20 22 24 26 28
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Opposed Flow Velocity [cm/s]
FlameSpreadRate[mm/s]
SDSU-5.6mm,21%
MSU-12.6mm,21%
MSU-12.6mm,21%,Black

46
Figure 5.2. Effect of opposed velocity and oxygen concentration on
Thick PMMA flame spread rate.
found are similar to that of the thin PMMA testing, and do not follow a simple linear trend as
one would expect from thick fuels.
During testing it was observed that bubbles are formed in the pyrolysis layer of the
thick PMMA as the flame passes over the sample. At completion and throughout the test it is
noticeable that the amount of bubbling or foaming depends on the depth of the pyrolysis
layer. Therefore, foaming appears to be a characteristic that is dependent upon the length of
the flame and residence time of the test condition. The mean residence time is found by
dividing the visible flame length by the rate of flame spread. From Figure 5.3 it is observed
that the residence time greatly increases as the opposed flow is increased and as the oxygen
concentration is reduced. This correlation seems to hold true to the amount of foaming seen
throughout testing. Figure 5.4 shows that the visible flame length is not dependent upon the
oxygen concentration, but is simply dependent only on the opposed flow velocity seen by the
flame. As the incoming flow is increased the flame elongates. At flows of approximately
22.5 cm/s and above the flame descends within the burrowed out section of the fuel sample.
8 10 12 14 16 18 20 22 24 26 28
0
0.2
0.4
0.6
0.8
1
1.2
SDSU-5.6mm,21%
SDSU-5.6mm,30%
SDSU-5.6mm,50%
MSU-12.6mm,21%
MSU-12.6mm,21%,Black
DARTFire-20mm,50%,Black

47
Figure 5.3. Residence time as a function of opposed velocity and
oxygen concentration. Where residence time is defined as
Perceptibly, because the depth of the pyrolysis layer decreases as oxygen concentration
increases, the flame is less able to hide behind the melt front at higher oxygen concentration,
but because the amount of oxygen available is much higher the flame is sustained.
5.2 VISUAL OBSERVATIONS
Other noticeable effects during the burning process consist of changes in brightness,
and color (Figure 5.5). Once again, in the lower extinction limit and blow-off regions the
flame became much bluer in color with a reduction in the amount of yellow and orange
colors. The higher oxygen percentages caused brighter, yellow, and longer flames. As the
flame reduced in size and became blue the flame front became slightly more curved in some
cases. The start of an individual flame “finger” as studied by Olson, Miller, and Wichman in
[37] was seen at the lowest of opposed flow velocities, where the flames struggled to stay
formed. It is believed that a wider sample would create fingers because the flame could
breakup and separate into regions far enough away from each other to receive the necessary
10 15 20 25
0
50
100
150
200
250
300
350
400
450
FlameResidenceTime[s]
SDSU-5.6mm,21%
SDSU-5.6mm,30%
SDSU-5.6mm,50%

48
Figure 5.4. Opposed flow velocity and oxygen concentration effects on
flame length.
Figure 5.5. Top view flame comparison of thick PMMA. Left: 21% oxygen, 25 cm/s
opposed flow velocity. Right: 50% oxygen, 10 cm/s opposed flow velocity.
oxidizer to sustain itself. When the small flame finger was present it would frequently travel
side to side burning fuel before moving forward against the opposed flow.
10 15 20 25
4
6
8
10
12
14
16
18
20
22
FlameLength[mm]
SDSU-5.6mm,21%
SDSU-5.6mm,30%
SDSU-5.6mm,50%

49
CHAPTER 6
BASS II EXPERIMENTAL RESULTS
BASS II testing was completed aboard the International Space Station under multiple
sample and flow conditions. Aiming to replicate the tests conducted, BASS II style test were
conducted in the NCA using the same sample holders and the previously mentioned sample
mounts. Opposed oxidizer velocity and oxygen concentration are studied first, followed by
species concentration change during experimentation aboard the International Space Station.
A visual comparison of infrared images and flames under varied conditions is also provided.
To better understand the key results the data were broken into multiple figures. Both
single-sided and dual-sided tests were conducted where the flame was either allowed to burn
on both sides of the sample (similar to thin fuels) or only on the top side (similar to the work
presented in Chapter 5 on thermally intermediate fuels). To achieve single-sided flame
spread the bottom fuel surface was blocked with a thin sheet of mica.
Due to boundary layer growth and oxygen concentration depletion during BASS II
testing aboard the ISS the flame spread rate is much less constant and the position vs time
plots are not as linear as previously shown in Figure 3.8. Examples of such cases are
provided in Figure 6.1.
In Figure 6.2 single-sided flame spread data was plotted against opposed flow
velocity. Clear PMMA samples ranging from 1 to 5 mm in thickness were tested. As
expected the flame spread rate reduces as the fuel thickness increases. With increased
opposed flow velocity the flame spread linearly increases in the region studied except for the
5 mm thick fuel where a slight decrease was seen. It is noteworthy that this is the only test
where the opposed flow was increased instead of decreased during testing.
The oxygen concentration during testing changed due to the limited MSG chamber
size (255 liters). Experimental results show an average change of -1.185 mole % with a
maximum change of -3.2 mole % and a minimum of 0.3 mole %.

50
Figure 6.1. Example BASS II position vs time plots. (A) 20.6-20.0 O2%, 2 mm
thick, 2 cm wide, 1 sided. (B) 17.9-16.9 O2%, 3 mm thick, 2 cm wide, 2 sided.
Figure 6.2. Single-sided BASS II flame spread rate.
An individual BASS NCA test on 3 mm thick PMMA at 20 cm/s opposed flow with
the same oxygen concentration (average of 20.4 mole %) is shown where the spread rate is
slightly lower than the experimental results from ISS. It is believed that the flow differences
experienced by the flame and the increased heat loss to the channel walls are to blame. The
0 200 400 600 800 1000 1200
0
20
40
60
80
100
120
Position[mm]
Time [s]
x = 0.062t + 50.489
R
2
= 0.997
x = 0.055t + 52.795
R
2
= 0.996
x = 0.035t + 62.519
R
2
= 0.995
AOI
15cm/s
14cm/s
10cm/s
0 500 1000 1500
0
20
40
60
80
Position[mm]
Time [s]
x = 0.0373t + 9.394
R
2
= 0.998
x = 0.0299t + 14.187
R2
= 0.999
AOI
15cm/s
10cm/s
0 5 10 15 20 25
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
1mm
2mm
3mm
4mm
5mm
3mm, NCA

Master's Thesis

Recommended

Recommended

More Related Content

Similar to Master's Thesis

Similar to Master's Thesis (20)

Master's Thesis