}V153I_J RETURNING MATERIALS:
Place in book drop to
LIBRARJES remove this checkout from
—c—. your record. FINES will

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

 

 

 

EFFECT OF WATER ON IGNITION OF
CELLULOSIC MATERIALS

BY

Mahmood Abu-Zaid

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Mechanical Engineering
Michigan State University
East Lansing, Michigan 48824

1988

Lfl
m
KW
\\ k

L1

3‘
‘.

KW

ABSTRACT
EFFECT OF WATER ON IGNITION OF CELLULOSIC MATERIALS

BY
Mahmood Abu-Zaid

This experimental study is an attempt to understand and quantify
the effect of water on fire extinguishment, thermal decomposition, and
piloted ignition of wood.

For the extinguishment part, the effect of water as externally
applied on hot porous and non-porous ceramic solids was studied. Both
solids were instrumented with several surface and in-depth
thermocouples. Similarities and differences in their thermal behavior,
and the heat transfer from the non-porous solid during the evaporation
of a single droplet of pure water were investigated. This study will
help determine the minimum water application, and the application
strategy to cool as-yet-unburned objects.

Thermal decomposition of wood as a function of sample moisture
content and externally applied radiation was investigated. Experiments
with three different moisture contents were performed at different heat
fluxes in a controlled atmosphere combustion wind tunnel. Simultaneous
measurements of weight loss rate; surface, bottom, and in-depth

temperatures; 0 depletion; production of CO', CO, total hydrocarbons,
2 2

and water were made. It was found that the presence of adsorbed water

delayed the decomposition process and diluted the decomposition
products.

For piloted ignition, experiments on Douglas fir with four
different moisture contents were performed at different levels of
externally applied radiation. In these experiments, in addition to the
quantities measured in the decomposition experiments, the time for

piloted ignition was also recorded. An absolute minimum mass flux of

approximately 0.22 mg/cmzs was found to be necessary for piloted
igntion to occur. It was also found that the presence of moisture
increases the ignition time, surface temperature and evolved mass flux
at ignition, and the critical heat flux. A single equation was derived
to correlate all of the ignition data. This correlation accounts for

the moisture-dependent thermal properties and the heat loss from the

sample.

To my father; Zaal, mother; Nadwa, my wife;
Ensherah, and children; Sawsan, Anas, Ala, Esra,

whose patient made the work possible.

ii

ACKNOWLEDGMENTS

I express my sincerest gratitude to Professor Arvind Atreya for
suggesting to me the fire extinguishment area as my dissertation topic
and for his encouragement and guidance. Professor Arvind Atreya has
been a source of profound inspiration, and working with him has itself
been a rewarding experience.

I am also very grateful for comments, suggestions, and
encouragement provided by Professors James V. Beck, John R. Lloyd,
Indrek S. Wichman and David Yen. I am thankful to Professor John J.
McGrath for providing the video system used in the experiments. I am
especially thankful to Dr. David D. Evans of NIST for his helpful
suggestions and encouragement.

I would like to thank my friends and colleagues, especially Mr.
Said Nurbakhsh and Mr. Kamel El Mekki, for their valuable help and
useful discussions. I am also thankful to the technical staff of the
Department of Mechanical Engineering, especially Mr. Robert Rose and
Mr. Leonard Eisele, for building the experimental facilities and
machining the wood samples. Finally, I would like to thank Ms. Julia
Pond for her diligent typing of the manuscript.

This work was supported by the National Institute of Standard and
Technology under the Grant No. 60NANBSOOS78, my graduate studies were

supported by Mu'tah University, Jordan.

iii

TABLE OF CONTENTS

LIST OF TABLES ................................................. vii
LIST OF FIGURES ................................................ viii
NOMENCLATURE ................................................... xiv
1. INTRODUCTION ................ . ............................... 1
1.1 Description of the Problem ................................. 2
1.2 Related Literature ......................................... 5
1.3 Present Work ............................................... 7
2. EXPERIMENTAL APPARATUS AND PROCEDURE ....................... 9
2.1 Droplet Evaporation Experiments ............................ 9
2.1.1 Apparatus ............................................ 9
2.1.1.1 Heated Ceramic Block ......................... 10

2.1.1.2 Droplet Generating System .................... 15

2.1.1.3 Data Acquisition Equipment ................... 16

2.1.2 Experimental Procedure ............................... 16

2.2 Wood Decomposition and Piloted Ignition Experiments ........ 17
2.2.1 Apparatus ............................................ 17
2.2.1.1 Combustion Wind Tunnel ....................... 17

2.2.1.2 Gas Analysis Equipment ....................... 21

2.2.1.3 Data Acquisition Equipment ................... 21

2.2.2 Procedure ............................................ 23
2.2.2.1 Sample Preparation ........................... 23

2.2.2.2 Calibration .................................. 24

2.2.2.3 Experiments .................................. 25

iv

2.2.3 Data Reduction ....................................... 26

2.2.4 Experimental Error ....... ‘ ............................ 27
3. DROPLET EVAPORATION ON HOT POROUS AND NON-POROUS SOLIDS ..... 28
3.1 Introduction .................. ’ ............................. 28
3.2 Background ................................................. 29
3.3 Literature Review .......................................... 30
3.4 Results and Discussion ..................................... 33
3.4.1 Transient Cooling of Hot Porous and Non-Porous Solids 34
3-4.1.1 Non-Porous Solid ...................... _ ....... 39
3.4.1.2 Porous Solid ................................. 40

3.4.2 Heat Transfer During Droplet Evaporation on the
Non-Porous Solid ..................................... 58
4. DECOMPOSITION OF WOOD ....................................... 74
4.1 Background ................................................. 75
4.2 Literature Review .......................................... 76
4.3 Products Analysis .......................................... 78
4.4 Results and Discussion ..................................... 79
4.4.1 Sample Mass Flux ..................................... 80
4.4.2 Sample Temperature ................................... 85
4.4.3 Decomposition Products ............................... 90
5. PILOTED IGNITION ............................................ 102
5.1 Background ................................................. 102
5.2 Previous Literature ........................................ 104
5.3 Apparatus and Procedure .................................... 109
5.4 Results and Discussion ..................................... 112
5.4.1 Ignition Delay Time .................................. 112
5.4.2 Correlation of Results ............................... 114
5.4.3 Surface Temperature .................................. 122

5.4.4 Sample Mass Flux ..................................... 129

5.4.5 Products Evolved ..................................... 132
6. CONCLUSIONS ................................................. 140
6.1 Droplet Evaporation Experiments ............................ 140
6.2 Thermal Decomposition and Piloted Ignition Experiments ..... 142'
6.3 Recommendations for Future Work ............................ 143

APPENDIX A. METHOD OF ESTIMATING LOCATIONS OF IN-DEPTH
THERMOCOUPLES AND THERMOPHYSICAL PROPERTIES OF

THE CERAMIC ........................................ 145
A.1 In-depth Thermocouples Locations ........................... 145
A.2 Porosity ................................................... 146
A.3 Thermophysical Properties .................................. 146
APPENDIX B. CRITICAL NOZZLES ........................ ........... 152
3.1 Mass Flowrate .............................................. 152
B.2 Calibration ................................................ 153
APPENDIX C. THE GAS ANALYSIS SYSTEM ........................... 157
C.1 H20 - CO2 Analyzer ......................................... 158
0.2 CO Analyzer ................................................ 158
0.3 02 Analyzer ................................................ 159
0.4 Total Hydrocarbon Analyzer ................................. 159
LIST OF REFERENCES ............................................. 160

vi

Table

Table

Table

Table

5.1

5.2

1A

2A

LIST OF TABLES

Comparison of Calculated and Experimental Values

of 0* and L ................ ‘. ......................... 122
Comparison of Mass Flux at Ignition for Different

Moisture Contents at Various Heat Fluxes ............. 132
In-depth Thermocouples Locations ..................... 148
Magnesium Oxide Ceramic-Refractory Grade ............. 149

vii

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

LIST OF FIGURES

Schematic of Experimental apparatus for-Droplet
Evaporation ...........................................

Ceramic Block Configuration ...........................

Apparatus for Thermal Decomposition and Piloted
Ignition Experiments ..................................

Gas Analysis Equipment ................................

Surface Temperatures vs Time for a Non-Porous
Solid During the Evaporation of a 30 pL Droplet .......

Indepth Temperatures vs Time for a Non~Porous
Solid During the Evaporation of a 30 pL Droplet .......

Surface Temperatures vs Time for a Porous Solid
During the Evaporation of a 30 pL Droplet .............

Indepth Temperatures vs Time for a Porous Solid
During the Evaporation of a 30 pL Droplet ...... , ......

Non-Dimensional Surface Temperatures vs Non-
Dimensional Time for a Non-Porous Solid at an

Initial Surface Temperature of 100°C During

the Evaporation of a 51 pL Droplet ....................

Non-Dimensional Indepth Temperatures vs Non-
Dimensional Time for a Non-Porous Solid at an

Initial Surface Temperature of 100°C During the
Evaporation of a 51 uL Droplet ........................

Non-Dimensional Indepth Temperatures vs Non-
Dimensional Time for a Non—Porous Solid at an

Initial Surface Temperature of 100°C During the
Evaporation of a 10 pL Droplet .........................

Non-Dimensional Surface Temperatures vs Non-
Dimensional Time for a Porous Solid at an Initial
Surface Temperature of 100°C During the Evaporation

of a 51 pL Droplet .....................................

viii

11

12

20

22

35

36

37

38

41

42

43

44

Figure

Figure

Figure

Figure

Figure

Figure
Figure

Figure

Figure
Figure
Figure
Figure
Figure

Figure

.10

.11

.12

.13

.14

.15

.16

.17

.18

.19

.20

.21

.22

Non-Dimensional Indepth Temperatures vs Non-

Dimensional Time for a Porous Solid at an Initial

Surface Temperature of 100°C During the Evaporation

of a 51 pL Droplet ...................................... 45

Non-Dimensional Surface Temperatures vs Non-

Dimensional Time for a Non-Porous Solid at an

Initial Surface Temperature of 200‘C During the
Evaporation of a 51 pL Droplet ......................... 46

Non-Dimensional Indepth Temperatures vs Non-

Dimensional Time for a Non-Porous Solid at an

Initial Surface Temperature of 200’C During the
Evaporation of a 51 uL Droplet ......................... 47

Non-Dimensional Surface Temperatures vs Non-

Dimensional Time for A Porous Solid At An Initial

Surface Temperature of 200’C During the Evaporation

of a 51 uL Droplet ..................................... 48

Non-Dimensional Indepth Temperatures vs Non-
Dimensional Time for A Porous Solid at an
Initial Surface Temperature of 200'C During the

Evaporation of a 51 uL Droplet ......................... 49
Evaporation Time vs Initial Solid Surface
Temperature for Various Droplet Volumes ................ 52

Recovery Time vs Initial Solid Surface Temperature
For Various Droplet Volumes ............................ 53

Non-Dimensional Maximum Radial Influence vs
Initial Solid Surface Temperature for Various

Droplet Volumes ........................................ 54
Non-Dimensional Maximum Axial Influence vs Initial

Solid Surface Temperature For Various Droplet Volumes.. 55
Non-Dimensional maximum Volume of Influence vs Initial
Solid Surface Temperature For Various Droplet Volumes.. 56
Parameter B vs Initial Solid Surface Temperature For
Various Droplet Volumes ................................ 57
Surface Temperatures For A Non-Porous Solid vs Time

During the Evaporation of a 51 pL Droplet .............. 60
Indepth Temperatures for a Non-Porous Solid vs Time

During the Evaporation of a 51 pL Droplet.........,.... 61

Contact Temperature vs Initial Solid Surface Temperature 64

ix

Figure 3.23 Heat Transfer Coefficients at the Surface of the Ceramic
block vs Initial Solid Surface Temperature ............. 65

Figure 3.24 Average Evaporative Heat Flux vs Initial Solid Surface
Temperature For Various Droplet Volumes ................ 66

Figure 3.25 Average Evaporative Mass Flux vs Initial Solid Surface
Temperature for Various Droplet Volumes ................ 67

Figure 3.26 Evaporation Time vs Initial Solid Surface Temperature
for Various Droplet Volumes ............................ 68

Figure 3.27 Recovery Time vs Initial Solid Surface Temperature
For Various Droplet Volumes ............................ 69

Figure 3.28 Non-Dimensional Maximum Volume of Influence vs Initial
, Solid Surface Temperature for Various Droplet Volumes.. 72

Figure 3.29 Parameter B vs Initial Solid Surface Temperature for
Various Droplet Volumes ................................ 73

Figure 4.1 Effect of Incident Heat Flux on Time-Dependent
Mass Flux For Dry Condition ........................... 81

Figure 4.2 Effect of Incident Heat Flux on Time-Dependent
Mass Flux for Dry Condition (Log-Log basis) ........... 81

Figure 4.3 Effect of Incident Heat Flux on Time-Dependent
Mass Flux for 11% Moisture Content Condition ........... 82

Figure 4.4 Effect of Incident Heat Flux on Time-Dependent
Mass Flux for 11% Moisture Content Condition
(Log-Log basis) ....................................... 82

Figure 4.5 Effect of Incident heat Flux on Time-Dependent
Mass Flux for 17% Moisture Content Condition .......... 83

Figure 4.6 Effect of Incident Heat Flux on Time-Dependent Mass
Flux for 17% Moisture Content Condition
(Log-Log basis) ....................................... 83

Figure 4.7 Effect of Moisture Content on Surface Temperature
At Various Incident Radiant Flux ....................... 87

Figure 4.8 Temperature Versus Time at Various Locations in the
Sample at 2 W/cm2 for 9% Moisture Content Condition ... 88

Figure 4.9 Temperature versus Time at Various Locations in the
Sample at 3 W/cm2 for 17% Moisture Content Condition... 89

Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure

Figure

Figure

Figure

Figure

Figure
Figure
Figure

Figure

.10

.ll

.12

.13

.14

.15

.16

.17

.18

.19

.20

.21

Time-Dependent Mass Flux of Evolved Products at

2 W/cm2 for 9% Moisture Content Condition ............. 91
Time-Dependent Product as Percent of Mass Flux at

2 W/cm2 for 9% Moisture Content Condition ............. 91
Time-Dependent Mass Flux of Evolved Products at

3 W/cm2 for 9% Moisture Content Condition ............. 92
Time-Dependent Mass Flux of Evolved Products at

3 W/cm2 for 17% Moisture Content Condition ............ 92
Time-Dependent Mass Flux of Evolved Products at

4 W/cm2 for Dry Condition ............................. 93
Time-Dependent Mass Flux of Evolved Products at

4 W/cm2 for 19% Moisture Content Condition ............ 93
Time-Dependent Mass Flux of Evolved Products at

1 W/cm2 for Dry Condition ............................. 9a
Time-Dependent Mass Flux of Evolved Products at

1 W/cm2 for 17% Moisture Content Condition ............ 94
Time-Integrated Product Mass and Composition (shown

as percentages next to bar) As Function of

Incident Heat Flux for 9% Moisture Content Condition .. 97
Time-Integrated Product Mass and Composition

(shown as percentages next to each bar) As Function

of Moisture Content at a Heat Flux of 2 W/cm2 ......... 98

Time-Integrated Product Mass and Composition
(shown as percentages next to each bar) As Function
of Incident Heat Flux for 17% Moisture Content Condition 99

Time-Integrated Product Mass and Composition (shown
as percentages next to each bar) As Function of

Moisture Content at A Heat Flux of 3 W/cm2 ............. 100
Schematic of the Ignition Process ..................... 110
Effect of Moisture on Ignition Delay Time ............. 113
Terms of Equation 5.4 vs Incident Heat Flux ........... 117
Correlation of Ignition Delay Time .......... . ......... 120

xi

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

.10

.11

.12

.13

.14

.15

.16

.17

.18

.19

.20

Effect of Incident Heat Flux on Surface Temperature
For Dry Condition ..................................... 124

Effect of Incident Heat Flux on Surface Temperature

for 11% Moisture Content Condition .................... 125
Effect of Moisture Content on Surface Temperature at

an Incident Heat Flux of 2.65 W/cm2 .................... 126
Effect of Moisture Content on Surface Temperature

at Ignition ........................................... 127
Sample Temperatures Histories for 11% Moisture Content

at 1.8 W/cmz .......................................... 128
Effect of Incident heat Flux on Time-Dependent Mass

Flux for Dry Condition ................................ 130
Effect of Incident Heat Flux on Time-Dependent Mass

Flux for 17% Moisture Condition ........................ 131
Mass Flux of Evolved Products at Incident Heat Flux

3.5 W/cm2 for Dry Condition ........................... 133
Products as Percent of Mass Flux at Incident Heat

Flux 3.5 W/cm2 for Dry Condition ...................... 133
Mass Flux of Evolved Products at Incident Heat

Flux 2.6 W/cm2 for Dry Condition ...................... 134
Products as Percent of Mass Flux at Incident Heat

Flux 2.6 W/cm2 for Dry Condition ...................... 134
Mass Flux of Evolved Products at Incident Heat Flux

2.75 W/cm2 for 11% Moisture Content Condition ......... 135
Products as Percent of Mass Flux at Incident Heat

Flux 2.75 W/cm2 for 11% Moisture Content Condition .... 135
Mass Flux of Evolved Products at Incident Heat Flux

2.77 W/cm2 for 27% Moisture Content Condition ......... 136
Products as Percent of Mass Flux at Incident Heat

Flux 2.77 W/cm2 for 27% Moisture Content Condition .... 136
Mass Flux of Evolved Products at Incident Heat

Flux 1.85 W/cm2 for Dry Condition ..................... 137

xii

Figure 5.21 Mass Flux of Evolved Products at Incident Heat

Figure

Figure

Figure

Figure

1A

2A

13

2B

Flux 3.21 W/cm2 for 17% Moisture Content Condition .... 137
Lease Square Fit to the Steady State Temperatures ..... 150
Transient Indepth Temperatures During a Step

Increase and Cut-Off of Heat Flux ..................... 151
Configuration of the Critical Nozzles .................. 155

Experimental and Theoretical Flow Rates of Air Through
a Critical Nozzle of Diameter 0.0785" ................. 156

xiii

‘ l e \
IEIHMCLATURE

a constant
D/d
specific heat

specific heat at constant volume
specific heat at constant pressure

diameter of droplet before impact
maximum diameter of droplet on surface
incident heat flux

heat flux entering the solid

release height, i.e., the distance between the needle tip and
the solid surface
heat transfer coefficient

overall heat transfer coefficient

thermal conductivity
liter

heat flux

rate of heat

radius of the disk

maximum radius of droplet on the surface

solid
temperature

equilibrium temperature

xiv

sat

(X.y.2)
(€.n.¢)

W

a

contact temperature
initial solid surface temperature in droplet evaporation

analysis, solid surface temperature in piloted ignition
analysis

ambient temperature

saturation temperature

droplet temperature before impact
time averaged surface temperature

TWS - Tsat

time

dimensionless time, t/r
initial volume of droplet

volume of droplet influence in solid

cartesian coordinates
oblate spheroidal coordinates
water

thermal diffusivity

Jpck

penetration depth

critical droplet thickness, i.e., droplet thickness on the
non-porous solid at the time the evaporation rate starts to
increase sharply

emissivity

T-T

T -T in droplet evaporation analysis, T-Tco in piloted
s e

ignition analysis

Ts-Tan

density

Stefan-Boltzmann constant

evaporation time of droplet

CHAPTER 1
INTRODUCTION

Cellulosic materials such as wood, cardboard, paper and cotton
often constitute the bulk of the fuels in many fires. .A common fire
starts in a residential room by a source of ignition. Thermal
radiation from the flames and the hot gases may heat and cause ignition
of nearby combustible materials allowing the fire to grow. The process
of fire spread is closely related to piloted ignition, and may be
regarded as a rapid succession of piloted ignitions.

Today, in the most technologically advanced civilization, we still
face the threat of unwanted fires. A recent estimate shows an annual
loss of $10 billion of property and forest in the United States. More
important and not expressible in dollars, is the annual loss of
thousands of lives, and millions of people sustaining serious burn and
smoke inhalation injuries. A fundamental study to understand the
chemical and physical process that occur during combustion of
cellullosic materials will result in achieving effective control and
prevention of unwanted fires.

In the burning of Cellulosic materials, water plays an important
role from ignition to extinguishment due to their porous and
hygroscopic nature and the availability of a large area to absorb

moisture from surrounding air. The moisture content of these materials

is variable, and is influenced by the climate, season and location.
Also, water is the most common agent used to extinguish unwanted fires.
Thus, a study of ignition and extinguishment of combustible cellulosic
materials is of great importance from the point of view of fire safety
because these materials constitute the bulk of fuels in many building
fires.

This study is an attempt to understand and quantify the effect of
water on piloted ignition and extinguishment of cellulosic materials.
Piloted ignition has been chosen because it is closely related to fire
spread and occurs at lower critical temperatures relative to
spontaneous ignition. Consequently, it is more hazardous. This work
will be useful in determining the onset of piloted ignition, the fire
growth rate under different humidity conditions as well as the minimum
water application rate and the application strategy to restrain the

fire growth.
1.1 Description of the Problem

The problem of extinguishment by water may be divided into two
parts; (i) extinguishment of the already burning objects and (ii)
prevention from burning of as-yet-unburned objects [Atreya (1985)]. In
the cooling of unburned solids, the physical process of water
evaporation on hot porous and non-porous solids is different. Results
from some early experiments conducted by Atreya (1985) indicate that a

water droplet impinging on a hot surface of wood (surface temperatures

0 .
in piloted ignition range from 300 - 400 C ) behaves differently than

it does on a hot copper plate at the same temperature. On the copper
plate, a thin vapor film is formed between the droplet and the hot
surface. As a result of this layer, the heat transfer is considerably
lower than in the case of direct contact (liquid has much better
thermal transport properties than its associated vapor). On the other
hand, in wood the droplet is quickly absorbed and re-evaporated. Thus,
in-depth cooling of the solid occurs.

The investigation of ignition of cellulosic materials must be
preceeded by the study of thermal decomposition of such materials. The
physical aspects of decomposition are associated with the heat and mass
transfer, processes which occur as the solid is heated. The external
heat is primarily transferred to the surface of the solid by radiation.
This radial energy will be transferred into the solid by conduction
process. As the temperature at any point rises, hygroscopic water is
first released, followed by decomposition of the solid in a complex
fashion. About two hundred compounds have been identified as the
products of decomposition of wood [Coos (1952)]. The rate at which
these products are generated depends upon the response of the solid
phase to the applied heat flux. These products are driven to the
surface of the sample by high pressure generated inside the solid where
they mix with the surrounding air.

A good quantitative understanding of the factors controlling the
decomposition of the solid, and the composition of the evolved gases,
is important for ignition and fire spread and for fire prevention in
general. It is also important in the context of minimizing air

pollution from wood burning stoves, Ohlemiller (1987).

The ignition mechanism is influenced by factors both external and
internal to the sample. Factors external to the sample are the
environmental variables, such as temperature, composition and velocity
of the surrounding gases, etc. Factors internal to the sample are its
thermophysical and thermochemical properties and its moisture content.
Experimentally, all variables other than the thermochemical and
thermophysical can be controlled.

The actual process of ignition is quite complicated. A simplified
phenomenological picture of piloted ignition by radiation has been
clarified by Atreya (1986) and Kashiwagi (1981). The solid must first
chemically decompose to inject fuel gases into the boundary layer.
These fuel gases must then mix with the surrounding air, and the local
mixture ratio must be near or within the flammability limits. At this
instant, a premixed flame, originating from the pilot flame, flashes
across the surface of the solid through the fuel-air mixture formed in
the boundary layer. Further heating of the solid results in evential
ignition and finally the establishment of a diffusion flame in the
boundary layer.

The adsorbed water in the wood matrix could effectively delay the
wood decomposition process, and reduce the flammability of wood (i.e.,
increase the ignition time) due to the high heat of vaporization of
water. In addition, water vapor could complicate decomposition
reactions, such as the reaction between water vapor and wood char to
produce carbon monoxide and hydrogen at high temperature.

In view of the above observations, the research problem is

logically divided into two parts:

(i) Effect of water as externally applied in the form of droplets,
(ii) Effect of adsorbed water on wood decomposition process, and
piloted ignition as a function of sample moisture content and

externally applied radiation.

1.2 Related Literature

In this section, only the general literature related to the problem
is discussed. Literature related to the specific component of the
problem is discussed in the appropriate chapters.

For droplet vaporization a number of studies on hot non-porous
metallic solids have been reported in the literature. These studies
show that the vaporization mode of a liquid droplet on a hot surface
depends on many factors. These are: (1) initial surface temperature of
the solid, (ii) isothermal or non-isothermal condition of the solid,
(iii) thermal properties of the droplet and the solid, and (iv) the
droplet momentum at impact. Few studies [diMarzo and Trehan (1986)]
have focused on in-depth cooling of the solid (which is important for
predicting the rate at which fuel gases are produced), and none have
reported transient in-depth temperature measurements.

A substantial portion of the combustible building materials are
porous with low thermal conductivity and thermal diffusivity.
Vaporization of water droplets on such solids is expected to be
different from that for the non-porous metallic solids. However,
despite the need to address the cooling of hot porous chars and unburnt
wood during fires, the author has been unable to find relevant

literature for porous solids.

A considerable amount of work has been done on the thermal
decomposition of wood and other cellulosic materials. These studies
were performed using different materials and under various conditions.
Martin (1965) investigated the decomposition of pure cellulose in
helium, Atreya (1983) investigated the decomposition from different
kinds of wood in air. The decomposition of PMMA and particle board
subjected to variable heat flux under an inert atmosphere was
investigated by Volvelle et a1. (1984) Kashiwagi (1987) studied the
products generated from wood decomposition in 3 atmospheres of

different oxygen concentration (N2, 10% O2 - 90% N2, and air).

Lee and Diehl (1981) investigated the effect of absorbed water on
samples containing 50% of oven-dry weight as water. They indicated
that moisture not only changes the solid-phase thermal properties but
also substantially dilutes the decomposition products. Atreya (1983)
from his recent work concluded that the desorption of moisture, which
has been ignored by previous investigators, has a considerable effect
on the energetics of the decomposition process.

Ignition of cellulosic materials (both spontaneous and piloted) has
been an active area of research in the past. Several excellent reviews
have been published on this subject [Welker (1970), Kanury (1972), and
Steward (1974)]. Although numerous techniques have been developed to
investigate piloted ignition phenomena, the experiments essentially
consist of exposing a sample to a known external radiant flux and
recording the time required to ignite it in the presence of a pilot
flame. Of the proposed ignition criteria, critical fuel mass flux at

ignition [Bamford et a1. (1946)] seems to be the most correct

physically (since it can be related to flammability limits), whereas
the critical surface temperature at ignition has proved to be the most
useful [Atreya (1983)]. However, only a few investigators [Atreya
(1983), Kashiwagi (1981), and Garden (1953)] have actually measured the
surface temperature at ignition, and measurements of the critical fuel
mass flux at ignition for cellulosic materials are not available due to
the complications caused by the simultaneous evolution of absorbed
moisture.

The importance of moisture content in relation to fire tests has
been recognized. Researchers have carefully controlled the moisture
content of their samples, but little attention has been devoted to the
effect of sample moisture on decomposition and piloted ignition. This

study is a systematic investigation of its effect.
1.3 Present work

In this work, the effect of water on piloted ignition and
extinguishment is investigated experimentally. The apparatus and
procedure used, experimental errors incurred, and data reduction
procedures are described in Chapter 2.

For the droplet experiments, the heated solid was cast from ceramic
and instrumented by several surface and in-depth thermocouples. In
order to compare the results of porous and non-porous solids under the
same conditions two identical solids, porous and non-porous, were cast
from the same ceramic powder. Using different droplet sizes, a

systematic set of experiments (initial solid temperature ranged from

75-325 0C) were performed. In these experiments, transient surface and
in-depth temperatures, and the droplet evaporation process were
recorded. The results of these experiments and a discussion is
presented in chapter 3.

Having recognized the importance and need for understanding the
phenomenon of thermal decomposition for the piloted ignition problem, a
set of experiments with three different moisture contents were
performed at different levels of externally applied radiation. In these
experiments simultaneous measurements of several physical and chemical
quantities were recorded. The results of these experiments are
described in Chapter 4.

For piloted ignition, a set of experiments with four different
moisture contents were performed at different levels of externally
applied radiation. In these experiments, besides the quantities
measured in the decomposition experiments, the time for ignition was
also recorded. Critical surface temperature and critical fuel mass flux
ignition criteria were discussed, and a correlation for ignition data
was derived and discussed. Chapter 5 describes the results of these
experiments.

Finally, conclusions for the various components of this research
problem and some recommendations for future work are presented in

Chapter 6.

CHAPTER 2

EXPERIMENTAL APPARATUS AND PROCEDURE

The primary objective of the present work is to understand and
obtain useful data for droplet evaporation on hot porous and non-porous
ceramic solids, wood decomposition, and piloted ignition of wood.
Experiments for these components were performed using different
facilities. This chapter describes the apparatus and procedure used for
each component, the data processing methods, and errors encountered in

these experiments.

2.1 Droplet Evaporation Experilents.

2.1.1 Apparatus

Experimental study of cooling of a hot, low-thermal-diffusivity
porous solid by water droplets requires an apparatus that satisfies the
following conditions:
(1) The solid should be nearly isothermal, porous, and made of a low
thermal conductivity and diffusivity material, and the temperature of
the heated solid must be controllable.
(2) The heated solid must be instrumented with surface and in-depth
thermocouples. From these transient temperature measurements during
droplet evaporation the following information can then be determined:

(1) recovery time i.e. the time for the surface to recover to its

10

initial temperature, and (ii) the size of "surface and in-depth zones"

affected by the droplet.

(3) A system capable of generating different size droplets is also

required. This system must be capable of delivering droplets on the

solid surface at a specified rate and at a specified release height.
An experimental apparatus was built to satisfy the above

requirements, a schematic is shown in Figure 2.1.

2.1.1.1 Heated Ceramic Block

To obtain a nearly isothermal solid three different shapes of the
ceramic blocks were considered. These were (i) Hemi-sphere: Here, to
make the flat surface isothermal, a large variation of the heat flux
applied to the hemispherical surface is necessary. (ii) Cylinder: It is
possible to obtain a flat isothermal surface in this geometry if the
cylinder is heated from below with a constant heat flux, and the
cylindrical surface is kept well insulated. The main disadvantage is
that a very small cylinder height is needed for the solid to be nearly
isothermal. (iii) Oblate spheriodal: This geometry has been chosen
because it combines the advantages of both cylindrical and spherical
geometry. It also provides a natural co-ordinate system for this
problem.

A castable porous ceramic (MgO) was used to cast a semi-oblate
spheroidal solid. This solid was used to simulate an isothermal semi-
infinite solid. The configuration of this block along with the
calculated temperatures and heat flux is shown in figure 2.2. The block
has a major semi-axis of 8.48 cm and a minor semi-axis of 7.22 cm.

Theoretical calculations described below show that this ceramic block

11

> oHH

53.33;» 3339 you. monsoon? Hanson—Coax” no causaosom a.” can»:

Hmunm um<mamm I : mm<m o<hm=m A<Hzouumo=

\ 0<Hm<> “N30“

 

 

ommmm

ou

           

mAHz<x ozHH<m=

 

V\\\\\V\\\\\\\_\§

I~ >0:

/.

 

«mamnho<
Hmoumm

 

 

 

 

 

 

 

 

 

onH<AszH

moaoza<mo moszww
on > on wqueao onm

12

.musoauuonxo ooanouv

page: new swoon «Houses on» ma Aolmv wanna use .ooseaon «any o>ownue ou

woumasuauo as: Auodo.mofiv unease: ago no ousuuuonaou one .muoumo: one aoum

oomuusm man» as msw>wuum Roam use: one On assoc ea aoaumwvou use souuoo>aoo

an commune cocoons on» aouu umoa usau one: one .Aeuov muses en» su .nH.H
.35 .36 .o um 09 unannouuou mane—Loom.“ 05. 303333.30». 3001— 38930 ~.~ was»:

 

  

 

    

 

 

Aemzou use
nmzHa wane
u
see Aamzoo out
mammmaomH
ama<mm
nulnvnu
=om.~
x nu
«A -:: -:
Arunuu oHaemeHne

 

 

 

 

.
_

 

 

xann a<mm
.o.~ n<onmzmanzoz

L

 

 

13

is nearly isothermal, despite the poor thermal conductivity (0.023
W/cm.k) of the porous ceramic. For the solid shown in Figure 2.2, let 5,
n, and d describe the orthogonal oblate spheroidal co-ordinate system.
Transformation of the cartesian to the oblate-spheroidal co-ordinate
system is given by: x - a cosh 6 cos n cos d, y'- cosh 6 cos n sin d,

and z - a sinh 5 sin n, where a - constant, 5 z 0, - g 5 n s g, 0 5 d 5

2«. As shown in Figure 2.2, 5- constant, represents the isotherms and n
- constant, represents the heat flux lines. As required by their
definition, they are orthogonal at every point in the domain. Under
steady-state conditions, the heat conduction equation reduces to the

following Laplace equation

2
v T - [ 2 1

a. £1
] (cos 6 )
a (sinh2 f + sin2 101/2 cos 5 85

55

1 ] E ii

+[ 1 2 (cos n )
a2(sinh2 5 + sin2 n) / cos n a" 3"

 

2
+ l 2 21 2 1 fl'1 ' 0 (2'1)
a cosh e + cos a 8d

Assuming that temperature is not a function of q and ¢, solution of the

above equation with the boundary conditions: T - T1 at £- 0, T - T2 at

£- 1.15, is given by:

T - 2.09m:2 - T1)tan'1[exp(£)]+2.64l T -1.641 T for e s 1.15 (2.2)

1 2

l4

and the heat flux is given by:

________;E________

(2.3)
(sinh2 5 + sin2 n)1/

2 [M—l
l + exp(2£)

where E -

As shown in Figure 2.2, the calculated heat flux on the exposed
surface between a and b is nearly constant and it is matched to the

radiative and convective heat losses to the environment. This yields a

maximum temperature across the solid of 9.01 0C which is considered to
be nearly isothermal relative to the dimensions of the droplet.

The above described block is instrumented with several surface and
in-depth thermocouples. Chromel-Alumel thermocouples of 76 um diameter
were used. The thermocouples were flattened to a film thickness of about
12 pm. Seven surface and six in-depth thermocouples were placed inside
the mold. The location of the thermocouples prior to casting was
measured to within i 0.1 mm. Due to the uncertainties introduced during
the casting process and shrinkage, the thermocouple locations were
evaluated by a least square fit to the steady state temperatures.
Detailed description for evaluating the thermocouple locations is
included in Appendix A.

In order to compare the results of porous and nonporous solids under
the same conditions, two identical solids of the same material were
casted. The non-porous solid was obtained by spreading a small amount of

solution of very.fine ceramic powder of the same material on the surface

15

of the porous solid. This procedure closes the surface pores, rendering
the surface nonporous.

As shown in figure 2.1, the ceramic block is heated by a heating
mantle. In order to achieve a perfect contact, the heating mantle has
the exact dimensions of the ceramic block. Furthermore, the heating
mantle has been configured to deliver the heat flux to the ceramic solid
according to equation (2.3). The temperature of the ceramic block is

controlled by a variac.

2.1.1.2 Droplet Generating Syste-

The droplet generating system consists of the following items (i)
stainless steel needles (ii) high quality syringe (iii) 90 V DC gear-
motor and (iv) speed controller.

The water droplets were generated by a high quality motor driven
syringe. To obtain different sizes 10, 12, 15, 19, and 25 gauge
stainless steel needles were used. These needles were filled to a smooth

flat tip. Care was taken to keep the main body of the syringe which

contains water at ambient temperature (25 0C).

The syringe plunger was driven by a gear-motor whose speed was
controlled by a speed controller. As the syringe plunger slowly moved, a
droplet was formed on the needle tip and increased in size, until its
weight became sufficient to detach it from the tip.

The droplet generating system has been calibrated using electronic
digital scale (Mettler AE 100) with an accuracy of i 0.1 mg. All needles
of different sizes were calibrated for the droplet size at ambient

temperature by individually weighing 15 droplets. The mean deviation was

16

found to be i 2% from the arithmatic mean. For example, a 30 pL droplet
was found to have a volume in the range 29.4 - 30.6 ML.

The motor-driven syringe system was placed on an adjustable height
platform to facilitate the adjustment of the needle tip height from the

hot solid surface. The adjustable range is from (0-10cm).

2.1.1.3 Data Acquisition Equip-eat

An analog-to-digital converter which provides 22-bit resolution is
linked to the HP data acquisition unit by IEEE interface bus. Signals
from the thermocouples were fed into a HP3497A controller with 39
differential input channels. The data was collected in real time by
using Digital Equipment corporation PDPll-73 microcomputer. Further, the
data was transferred to VMS-VAX 750 for permanent storage on magnetic

tapes.

2.1.2 Experimental Procedures

The experimental procedure consists of setting the variac to a fixed
value to heat the ceramic block to the desired temperature. It took
between 10 to 12 hours for the solid block to reach steady state
conditions.

Pure water was used for these experiments. This water was double
distilled, deionized and degassed. Degassing process consists of boiling
the water and cooling it under vacuum of 23 inch mercury. Before

starting the experiment, some water in the needle was ejected to ensure

0
that the droplets were at ambient temperature (~ 25 C).

17

Tests were performed with initial surface temperature ranging from

75-325 oC. The droplets were deposited gently on the heated surface from
a release height of 6-9 mm (from the tip of the needle to the hot
surface). This released height was varied to achieve a constant distance
between the bottom of the droplet and the surface of the solid. During
these tests transient measurements of temperature were collected by the
data acquisition system.

A video camera was used to record the life time of the droplet on
the hot surface. This provides a precise measurement of the evaporation
time and droplet size. In addition, it shows the droplet behavior on the

hot-surface.
2.24 Wood Deco-position.and Piloted Ignition.Experilents

2.2.1 Apparatus

The apparatus used in these experiments is schematically shown in
figures 2.3. It consists of three main units:
(1) Combustion wind tunnel, to perform experiments in a well-controlled
atmosphere and radiation environment.
(ii) Gas analysis equipment, to measure concentration of chemical
species in the exhaust.

(iii) Data acquisition equipment, to collect and store data.

2.2.1.1 Combustion'Wind.Tunnel
This facility is the heart of the apparatus. It is capable of

providing a flat-plate boundary layer flow on the sample surface at a

l8

desired composition and free stream velocity. This tunnel consists of
three main sections: the inlet section, the test section, and the
exhaust section.

(1) Inlgt_§gggign: Cases are supplied to the inlet section by two
sources; (a) pressure regulated laboratory air supply (up to 95 psig),

and (b) high pressure gas cylinder (02 or N2 ). These gases pass

separately through high pressure tanks. In these tanks any oscillations
in the flow would be damped out. To control the gas composition, these
gases are metered by critical nozzles. Detailed description and
calibration of these nozzles is presented in Appendix B. The resultant
mixture enters the tunnel through eight symmetrically distributed
turbulent jets. It then passes through a set of screens and a permeable
cloth into a large settling chamber. In this settling chamber the gas
velocities are small and the pressure is only slightly above
atmospheric. It then passes through turbulance manipulators to suppress
the turbulance in the flow.
(ii) Ig§§_§eggign : The test section consists of three modules: radiant
heaters, tunnel top, and tunnel mainframe.

The external radiation was simulated by the use of nine radiant
heaters. Three high temperature quartz heaters and six U-shaped
chromalox coil heaters. The heaters are controlled by a 3-phase 440 volt

variable transformer. These heaters have a maximum black body radiation

0
temperature of 1200 K and are therefore well suited to simulate

external radiation in building fires. Reflective chromium plated steel

sheets were configured to obtain a maximum radiant flux of 4 W/cm2 at

the sample surface.

19

The water-cooled tunnel top houses five pieces of the infrared
optical glass (Kodak IRTRAN-II). The glasses were securely held in a
copper frame lined with flexible silicon rubber to allow for expansion.
The combination of these glasses make a 6" * 30” infrared glass window.
This glass window was cooled convectively by compressed air.

The tunnel mainframe is made of thick water-cooled brass walls. The
mainframe is connected to the heater system to maintain the position of
the heaters fixed relative to the sample surface. It consists of four
observation windows for optical access and three slots for probe access.
The latter slots were lined with high temperature silicon rubber to
prevent leakage. Two non-combustible blocks (Kaowool insulating boards)

mounted flush with the sample surface, are placed symmetrically on both

sides of the sample. On top of the insulating boards, a %% thick

aluminum plate was placed in order to create a smooth boundary layer.
Except for the sample space (6" * 3“) all tunnel base was covered with
the combination of the ceramic boards and aluminum plate.
(iii) Exhau§§_§ggtign : The primary purpose of the exhaust section is
to mix the products of decomposition (or combustion) to obtain a
representative gas sample for chemical analysis. This gas samplewas
used to determine the production or depletion rates of the species
measured at the gas analyser. The gas sample is extracted through a
cross-shaped probe, and carried by heated lines to the continuous gas
analysers.

The mixing process takes place in a steel mixing chamber lined with
the insulating boards. The mixing is accomplished by using heated

electrical resistance tapes as louvers upstream in the mixing chamber.

20

nucosuuonxm couuacwu vouoaqm
can couuunonaouon Hassock new nauuumnn< n.~ ouswuu

 

 

 

 

Amz<m
xMHhHauzbzuo domkzoo oz< an2<=0
8n ‘.
mH<
222355“ L l m
Aun
E E

mxmguawo IV I. .1

N2 xo~o =9;

«Omwmxmroo
xH< team

4

 

 

 

 

Ill

xH< mmmoomm

     
               

  

 

 

 

mo<¢owm
<H<o onHmHDO
nuauzoo <H<a m

 

 

 

 

 

o
:

 

 

mmo<o xDAh H<mm

mmnz<=o
ozHAaHmm

onHomw
ngzH

wmamsoooxx
wmoa 9:0

moH>mn

 

uzh
Hm:

mxmh<mx
Hz<Ha<d

 

 

 

onHomm
Hmmh

 

mxoa<qamuz<z muzwgbmxbh :9

<
3002a: onH<>¢mmmo 3o

i—“v

(snows svo)
31m aaivan

f

 

  

me<Am ozHAooo

onHomm
Hm:<:Xm

mmmoo< mmomm

HmazH guzzzh Ha
Hun Hzmqamxak we
mmm<h oz~a<mz H:
mmmuz ozHH<mz 3:

mmmz<xo oszH: oz
nwonH<H>me<

z<m
zo~Hoam

 

Mz<H

Hma<xxw .w
Hwa<=xm

 

 

 

21

Further downstream a net of electrical resistance tapes and wires were
used for heating. This combination results in a hot mixing chamber.
This heating is necessary to prevent condensation of heavy molecular
weight hydrocarbons.

The suction introduced by the exhaust fan is adjusted to obtain the
desired pressure above the sample surface. To reduce the turbulent
fluctuations inside the tunnel, the fan is isolated from the tunnel by

using 24" * 24" * 36" metal chamber.

2.2.1.2 Gas Analysis Equipment:
The gas sample was sucked from the exhaust section by a pump into

the gas analysis equipment to continuously measure the depletion of 02

and production of CO CO, total hydrocarbons [THC] and H O. A 1/4 inch

2’ 2
outside diameter sampling line that was about 70 inches long was used to
supply the gas sample to the analysis equipment. This resulted in
transport delay times. The transport time for the gases from the sample
location to the instrument was measured for each analyser. These times

were 3.5 sec. for [THC], 6.5 sec. for O 4.5 sec. for CO, and 5.0 sec.

29

for H20 and 002. A schematic of the system is shown in Figure 2.4, and a

detailed description is given in Appendix C.

2.2.1.3 Data Acquisition Equipment:

In these experiments the same data acquisition equipment described
in section 2.1.1.3 was used. In addition to the analog signal from the
thermocouples, signals from the instruments, heat flux gauge, and the

electronic balance were also collected and stored.

22

 

02 TO EXHAUST

 

 

 

 

 

 

 

 

 

 

 

 

 

ANALYZER I
I co
I ANALYZER
com)
BY-PASS TRAP
mo
ANALYZER

 

FILTER

 

 

-’-—""""‘l

 

    

ZERO GAS
OR
CALIBRATION GAS

Figure 2.4 Gas Analysis Equipment.

23

2.2.2 Procedure:

2.2.2.1 Sample Preparation:

Wood samples were cut to the desired size (6” * 3" * 1.5”) from 12”
* 12" * 1.5" boards of Douglas fir. This wood was obtained from the
Forest Products Laboratory at Madison Wisconsin. The samples were
carefully machined to smooth surfaces to eliminate any effect caused by
surface roughness. The samples were instrumented by surface, bottom and
in-depth thermocouples. (The thermocouples were made from fine chromel
and alumel wires (0.003" dia). The method employed to install the
thermocouples was developed and successfully used by Atreya (1983). For
surface and bottom thermocouples, the thermocouple wires flattened to a
film thickness, and electrically welded. The bead was polished by a
fine emery cloth. A very fine layer of wood was skinned off from the
sample surface using a sharp razor blade, and a portion of the
thermocouple was secured underneath this layer using Tight-bond wood
glue. The extra glue was wiped off and the assembly was allowed to dry
under a heavy weight for 2-3 hours. The in-depth thermocouples were
installed at different locations. Holes were drilled (1 mm dia, 1 inch
length) at desired locations on the side of the wood sample. Ceramic
insulating tubing with two holes were used to house the thermocouples
inside the holes.

The samples were then dried in a temperature-controlled chamber at

o
105 C until a constant weight was achieved. This weight was recorded. A

temperature-humidity-controlled chamber at a known relative humidity was

24

used to fix the moisture content in the samples to the desired value.

This chamber (Model TH3, BMA Inc.) has controllable humidity range of 10

to 98% relative humidity for temperatures between 10 0C and 85 0C.
Samples were conditioned for 5-6 weeks until a constant weight was
obtained. The difference between this weight and the dry weight was used
to determine the sample moisture content. Samples edges and bottom were

taped with aluminum foil to avoid any edge effects.

2.2.2.2 Calibration:
The gas analysis instruments were first turned on and allowed to

warm up until a steady value was reached. Zero gas (21.3% 02 and rest
N2) was allowed to flow through the gas analysers, and all the

instruments were electrically adjusted to their reference value. These

values were recorded. Calibration gas (0.201% methane, 1.97% CO 0.201%

2.
CO and rest N2 ) was allowed toflow through the gas analysers. The
output of all instruments was again recorded and a calibration factor
was calculated and stored. The water meter was calibrated using room
air. The mole fraction of this air was determined from the wet and dry
bulb temperatures. Periodically, the meter was checked against a third
point, by introducing an air stream at known condition from the humidity
chamber. The calibration factors for some instruments, namely
thermocouples and the radiometer, were constant, thus they did not

require calibration each time.

25

2.2.2.3 Experiments:

The heaters were turned on at the desired radiant heat flux, and
were allowed to warm up for about 15~20 minutes. The air valve was
opened, and the pressure upstream the critical nozzle was set to the
desired value by adjusting the air regulator. The air was allowed to
flow through the tunnel for about 10 minutes to attain a steady state
condition for all instruments. The pressure above the sample was
adjusted by the exhaust fan to obtain atmospheric pressure. The leakages
from the gap around the sample and from the top of the tunnel were then
determined. A known volume of methane gas tracer was introduced in the
main flow downstream of the critical nozzle, and leakages were
determined from the difference in the methane concentration measurements
upstream and downstream of the tunnel.

The cooling plates were opened, and the tunnel was allowed to warm
up for about 20 minutes to drive out the absorbed moisture from the
insulating boards. The tunnel was then allowed to cool to room
temperature. The sample was placed on the weighing table with its top
surface flat along the tunnel base. The sample was then exposed to a
known external radiation by opening the cooling plates, and data
collection was simultaneously started.

The heat flux sensor located upstream in the tunnel continuously
monitored radiation from the heater during the experiment. A calibration
between the heat flux at the sensor location and the heat flux at the
sample location was used to determine the incident radiation on the

sample surface.

26

For a given moisture content and external radiation, the following
measurements were recorded as a function of time: (i) surface, bottom
and in-depth temperatures, (ii) mass evolution rate of water vapor,
(iii) production rate of combustible gases and (iv) weight loss rate of
wood sample.

For piloted ignition experiments, in addition to the above
measurements, the time required for ignition was also recorded. A video
camera was focused on the sample to record the experiment. This provided

a precise measurement of the time of ignition.

2.2.3 Data Reduction:

The output of the instruments were stored by microcomputer in real
time. These data represent the mole fraction of the species measured.
It must be processed because not all instruments register an event at
the same time due to time lag and response time of the instruments. The
procedure used for data reduction is the same as developed and used by
Atreya (1983). The final result of this procedure is that all measured
quantities are corrected and brought to the same time base. This data
consists of instantaneous mole fraction of all the species at 5 sec.
intervals. These values were then converted to percentages on mass

basis. Finally, knowing mass flow rate of air, surface area and weight

loss rate of the wood sample, the mass flux (mg/cm2.s) for each specie

was determined.

27

2.2.4 Experimental Error:

The difference between the true value of a quantity and the measured
value is known as the measurement error. These errors may be systematic
or random. Systematic errors must be eliminated or corrected, whereas,
it is difficult to predict and correct random errors.

In these experiments, a systematic error introduced by the leakages
in the tunnel was corrected by the tracer method discussed in Section
2.2.2.3. Random errors from electrical noise, limitation in accuracy
and response time of measuring instruments, condensation of heavy
hydrocarbons in the tunnel and sample lines, and oscillations in weight
loss measurements were difficult to estimate since they vary with time
due to the transient nature of the experiments. Therefore no corrections

were made for those errors.

CHAPTER 3
DROPLET EVAPORATION ON HOT POROUS AND NON-POROUS CERAMIC SOLIDS
3.1 Introduction

The study of the behavior of impinging individual droplets on a hot
surface is essential in the analysis of problems involving liquid
sprays. Knowledge of how the droplet impacts and evaporates will help
in understanding and prediction of the effect of sprays on cooling of
hot surfaces. Such problems are encountered in a number of engineering
areas; fire extinguishment, nuclear reactor safety, cooling of turbine
blades, and cooling of continuously casted steel.

The evaporation of a liquid droplet on a hot solid surface, occurs
by extracting heat from the region directly below it. In quasi-steady
evaporation, the evaporation rate is the heat transfer rate divided by
the heat of evaporation [Renksizbulut and Yuen (1983)]. For such a
case, the heat of evaporation is equal to the heat transfered through
the droplet and also equal to the heat transfered from the solid to the
bottom of the droplet.

Water sprays are the most common method used in fire extinguishment.
Different types of solids with different surface characteristics are
involved in a fire. Natural solids such as wood are porous and produce
low thermal conductivity porous chars, whereas, plastics such as PMMA

are non-porous and remain non-porous during burning. Hence, the study

28

29

of thermal behavior of water droplet varporization on a hot and low-
thermal-conductivity solids is important for fire extinguishment.

Atreya (1985), divided fire extinguishment problem into two parts (i)
extinguishment of the already burning objects (ii) prevention from
burning of the as-yet-unburned objects. The present work primarily
addresses the latter part of the problem. However, it may be applicable
to the former part as well.

The outline of this chapter is briefly given below. Section 3.2
describes the factors effecting droplet evaporation, and the different
regions of droplet vaporization process. Section 3.3 reviews the
literature on the subject. Section 3.4 describes similarities and
differences in the thermal behavior of the porous and the non-porous
solids, and the heat transfer from the non-porous solid during droplet

evaporation.
3.2 Background

Vaporization mode of a liquid droplet on a hot solid surface depends
on many factors. These factors would have a significant effect on the
evaporation time, the heat transfer rate from the solid surface, and the
droplet behavior during evaporation. Such factors are (1) initial
surface temperature of the solid (ii) isothermal or non-isothermal
condition of the solid (iii) porosity of the solid (iv) thermal
properties of the droplet and the solid, and (v) the droplet momentum at
impact.

The droplet vaporization process on a hot solid surface could be

divided into 4 regions; evaporation, nucleate boiling, transition

30

boiling, and film boiling (Leidenfrost phenomena). In the evaporation
region, a direct contact between the droplet and the hot surface exists.
Heat is transfered to the droplet by free convection, and evaporation
takes place from the top and sides of the droplet. In the nucleate
boiling region, vapor bubbles begin to form on the hot surface. As the

initial surface temperature (Ts) increases, the rate of heat transfer

from the hot solid increases. This is because the bubbles formation
become more dynamically which cause an increase in fluid mixing near the
hot surface. In the transition boiling, a vapor film start to form
between the droplet and the hot surface due to the extensive formation

of bubbles. As Ts increases, the rate of heat transfer decreases. This

is because the direct contact between the liquid and the surface is
decreasing, and the vapor has a lower thermal conductivity than of the

liquid. In film boiling, the droplet set on a vapor layer. As Ts

increases the rate of heat transfer increases, because thermal radiation

is the major heat transfer mode from the hot surface to the liquid.'
3.3 Literature Review

Numerous studies of droplet vaporization process on hot metalic
solids have been reported in the literature. These studies have
focussed mainly on evaporation time and heat transfer from the hot solid
during droplet evaporation. Few studies [di Marzo and Trehan (1986)]
have focused on indepth cooling of the solid (which is important for
predicting the rate at which fuel gases are produced), and none-have

reported transient in-depth temperature measurements.

31

Seki et al. (1978), reported surface temperature measurement of a
hot wall cooled by a single droplet. They gave no description of the
cooling effect produced by evaporating droplets. They demonstrated good
agreement between the theoretical computation of solid-liquid contact
temperature and experimental measurements. According to the solution
for a semi-infinite body with a sudden change in surface temperature

[Carslaw and Jaeger (1959)], the contact temperature is given by

T 7w + T 1
.£L______§__§
T - (3'1)!

c 1w + 18

where 1 - Jpck; p,c,k represents density, specific heat, and thermal

conductivity respectively, and T8, Tw are the initial solid and droplet

temperatures.

Makino and Michiyoshi (1979, 1978), studied the effect of initial
size of the water droplet on it's evaporation time. They also
determined the time-averaged heat flux from the time-dependent
measurements of the projected droplet base area, the solid surface
temperature just below the droplet and the liquid temperature within the
droplet. Since their work includes nucleate boiling, transition
boiling, and film boiling regions, they present a wide range of the so

called the boiling curve (q vs A T ). Thus this work establishes the

sat
thermal behavior of droplets during the vaporization process.

Bonancina et a1. (1975), studied the heat transfer characteristics
of atomized liquids on hot surfaces in a high pressure environment.

Rizza's (1981) theoretical work on a non-isothermal wall was very useful

32

for a spray evaporators. However, the work of [Bonancina (1975), Rizza
(1981)] is of limited significance in fire extinguishment studies.

Gottfried and Bell (1961), determined the evaporation rate of the
droplet from the measured evaporation time. The disadvantage of this
method is that it assumes the evaporation rate to be constant which
actually increases as the droplet becomes smaller. This non-constancy
of droplet evaporation was noticed by Abu-Zaid and Atreya (1988a).
Wachters et al. (1966), determined the evaporation rate by photographing
the droplet from the side at small intervals. They obtain the volume
and the rate of change of volume as a function of time. The
disadvantages of this method is that very accurate volume measurements
are necessary because their rate of change needs to be determined and
they are multiplied by a large heat of evaporation. .Also, the errors
become worse as the volume decreases.

Gottfried et a1. (1966), investigated the Leidenfrost phenomenon of
some common fluids. They used droplets of volume < 0.1 mL on stainless
steel plate at surface temperature ranging from ISO-500°C. They
determined the Leidenfrost point for water, carbon tetrachloride,
ethanol, benzene, and n-octane. Furthermore they found that the
Leidenfrost point was independent of droplet size for the range of
droplet volumes studied.

Avedisian and Koplik (1987), studied Leidenfrost boiling of methanol
droplets on hot porous ceramic surfaces. Their work was motivated by
the recent interest in ceramic combustion engines. They measured
evaporation times in the wetting and film boiling regions on a polished
stainless steel surface and three ceramic/alumina surfaces of 10%, 25%,

and 40% porosity. They found that Leidenfrost temperatures increased as

33

surface porosity increased. They also indicated that in the Leidenfrost
regime, droplets evaporated faster on the porous surfaces than on the
stainless surface, and the evaporation time decreased with increasing
surface porosity at the same surface temperature. They explained this
reduction in evaporation time by the decrease of the vapor film
separating the droplet from the ceramic surface due to vapor absorption
and flow within the ceramic material.

The work most directly related to the present work is that of di
Marzo et a1. (1987). They investigated experimentally the thermal
behavior of droplets on hot aluminum and macor solids. In continuation
of the above work, di Marzo and Trehan (1986), theoretically
investigated the cooling effect of a hot semi-infinite aluminum block
and obtained isotherms and envelope of the droplet influence. Evans and
di Marzo (1986), presented a mode for the evaporation of a draplet in
contact with high thermal conductivity and diffusivity solids. They
determined that the evaporation time is in reasonable agreement with the
experimental data. The model also describes the spatial and temporal

behavior of the evaporation heat flux.
3.4 Results and Discussion

The data discussed here were collected during the large number of
systematic experiments performed on porous and non-porous ceramic solids
at various initial surface temperatures. Droplet sizes of 10-51 pL were
used, because they are representative of the droplets delivered by the
fire suppression equipment [Evans and di Marzo (1986)]. Section 3.4.1

describes the transient cooling of hot porous and non-porous solids.

34

The similarities and differences in the thermal behavior of both solids
during droplet evaporation is discussed. This work was presented in a
paper [Abu-Zaid and Atreya (l988a)] submitted to the Journal of Heat
Transfer. Section 3.4.2 describes the heat transfer during droplet
evaporation on the non-porous solid. An extended abstract of this work
[Abu-Zaid and Atreya (l988b)] was accepted for presentation at the

Eastern Session of the Combustion Institute (Florida, December 1988).

3.4.1 Transient Cooling of Hot Porous and.Non-Porous Solids

Results of a typical test composed of 3 droplets are shown in
Figures 3.1-3.4. Transient surface and indepth temperatures
measurements for a 30 uL droplet are shown in Figures 3.1 and 3.2 for a
non-porous solid, and in Figures 3.3 and 3.4 for a porous solid. These
figures show the repeatability of the experimental measurements during
the droplet evaporation process. Figures 3.1 and 3.3 also show the
uniformity of surface temperature prior to droplet impact.

Individual traces of surface and indepth temperatures for different
droplet sizes and for different initial surface temperatures are shown

in Figures 3.5-3.13. In these figures a non-dimensional temperature 0 -

T-T
[T -T was plotted against a non-dimensional time t* (t/r). The non-
s e

dimensional temperature 0 is chosen because 0 versus t* curves for each
case of the porous and the non-porous solids, for all initial solid

surface temperatures (Ts) and droplet sizes that were investigated, fall

approximately on top of each other.

dean—own A: an e no coauouooeaé one
9555 ozom 383.602 c you on; Er nonsense-.3. snowman H6 93»:

 

 

35

 

 

 

 

 

 

 

 

Aommv mEfi
com 00¢ com com 00— o
— D - m - D - b - n
_ ION.
i=8 328.8: a 3.. 2382 39825.5. .
. as we peso;
58 p o . seizes—u. , 1 DB
.55. n . £885.95. ..
SE n v. 0.9805
as v n . £88235. 1 cm
H E:— V I
2.50
I: ..nn
Tom

 

 

 

 

 

 

 

 

 

 

 

 

 

(3°) SGJnlDJadUJSl eooyng

dean—80 A: on e me .3330er one
wot—:0 0.38 333.52 a you on: g monsoon—oases noose—AH «.n 3503

 

 

36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aommv 0:5
000 00¢ 000 00m 00 F 0
— p — p — p — . — .
. . on
.38 «8.898.. a 3.. 2382 3.589520. .
as»: : 2388.5: Inn
88 we 2 . 33805.85. 0.
es 2 a. 39885:.
SE 2 a . 830835 I Om
osEn 2 295853: .
100
~ 100
w ..
11 a low
T\ < : < A
.IIIIIII <1] < I‘ll-II...

 

 

 

(3°) SGJnlDJedUJGl qidepul

37

on» 06:30 0.300 esouom e u

Joan—cum A: on a mo soaueueneaé

em 06:. e> nousueuoaaeu. edema—am n6 swam:

 

 

Aommv oEfi
0mm . 03. 0am 0am 0m. F

.38 308.. a .8 8382 29885.0. . ea 3 a. e. caucuses
E... a. o . oESoofiafi
as a. n . £38535.
as n . v . 3:82.55
as c n . snoooosafi
es 4 a . £88285.

£50 a . €885.25. _

 

 

 

 

 

 

 

 

 

 

 

 

 

(3°) SSJnloJedUJSi 3:30an

.uoaaoua a: On a Mo coaunuonfifi
05 95.3.5 v.30» 39.3." I you 2.3. g nousunuogoa 5955 in can»;

Aoomv oEt.

 

oem 00¢ com com 00. o
. _ . F b — p — p _ p on
.38 328 a 8.. 2882 390863;...
.5: Q: 2 . 033853 ..
E... in .2 u 03:08:55. IDN

 

as; 3%;

883 «.0333 on

.33 2%8853 f
.l

 

 

 

 

 

 

 

 

 

 

 

 

(3°) seJnmJeduJal HldGPUI

39

For both the porous and the non-porous solids at an initial surface

temperature (Ts) 5 164°C, the hot surface immediately cooled to the

theoretical contact temperature upon contact with the droplet. For TS >

164°C, the surface temperature, upon contact with the droplet falls
considerably below the theoretical contact temperature. Seki et a1.
(1978) found that on stainless steel, the measured contact temperature
increases with increase in the surface temperature up to 180°C, and was

in agreement with the theoretical contact temperature. For 200 5 TS 5

300°C, they found that the measured contact temperature was less than
the theoretical contact temperature and.was approximately constant. The
present results for the ceramic are in qualitative agreement with Seki's
data except that the present data shows a deviation from the theoretical
value at 164°C rather than at 180°C.

After achieving the contact temperature the solid surface approaches
the "equilibrium temperature" (interface temperature during
evaporation). During the droplet evaporation period the thermal
behavior of the porous and non-porous solids are different. These

differences are described below:

3.4.1.1 Ronrporous solid

Measurements show that for TS - 75°C, the surface and indepth

temperatures remains nearly constant during the droplet evaporation
process. This implies that the heat flux during the evaporation time is

also nearly constant. For 100 5 T8 5 175°C, the interface temperature

*
remains nearly constant until a critical drOplet thickness (6 ) is

40

reached on the hot surface. Then, the temperature starts to decrease

, sharply until the droplet has completely evaporated as shown in Figure
3.5. After the droplet has completely evaporated, the solid recovers
Quickly to its initial surface temperature. This sharp decrease in the
surface temperature implies an increase in the evaporation rate. This
will be discussed in more details in section 3.4.2.

For T - 200°C, the interface temperature drops to a value few

8
degrees above the saturation temperature of water (see Figure 3.10).
Then the surface recover to its initial value. The evaporation times at
this temperature were very short, since the droplet diameter after
impact on the surface was very large (see Figure 3.19). This large

droplet diameter is due to lower surface tension.

3.4.1.2 Porous solid

For porous materials the interface temperature never clearly attains
an equilibrium value. Both the surface and the indepth temperatures
continue to decrease until the droplet vanishes from the surface (as can
be seen from Figures 3.8 and 3.9). The time taken for the droplet to
vanish from the surface is defined as the evaporation time. By this
time a part of the droplet has already been evaporated, and the rest has
penetrated in both the axial and the radial directions. The effect of
the droplet penetration is clear from following two observations: 1) A
thermocouple in the porous matrix at the same location as the non-porous
matrix cools faster under otherwise indentical conditions. This implies
that, for the porous matrix, there must be an energy sink in the

neighborhood of the thermocouple. Thus confirming the indepth

61

no 3332—38.
onus Hosea-soaao

.uoaaoua A: an e no soaueuoaq>m as» mcauso 0.00“

coauuam ueuuusn so us 00000 nsouom.soz a you
.soz n> copay-unmask oonuusm Hosea-soaao.Coz n.n ouawum

...”. mEz. _oco_mcmEElcoz

 

 

 

 

 

 

 

             

 

 

e; N; 0; 0.0 0.0 v.0 N0 0.0 N0!
— . p . _ . _ . _ . _ p _ . _ O.Nll
6:8 388.8,. .. 3. .8532 338252... .
EE «.0 h . 20:08:55. 0 8 no“. ..
.88 h 0 . 30:08:55.
ES» .n.o_a=o8=5.c. r043:
E...“ v. 0.038.505.
as v n. 30800535. 1
.5... 2033853.: no.7.
2.50 ._ 238852....
1.0.0...»
0
. 10.0
a .
v I00
2......1.--\ -1 .Flt- this .
- a. x 10;

 

 

0 seJmoJeduJal aoopn'g louogsuang—UON

42

.uoanoua A: an a mo souuuuomefi on» 05.30 0.00."
no assuage—.3. 030.30 .2335 as us 3.3m asouoméoz a won

053. assoausoaaaéoz 2' masseuse—Sn. sudopaa decode—550502 06 9.5»:
*0 mEC. _oco_mcmE_olcoz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.3 a; .o._ 90 0.0 ....o do 0.0 «.0:
_ . — . — . P . p. p — p — — O.NII
i=8 388.8: a 3.. 209.8. 39.8853... .
EE 0.: = 0 03:83:20. 0 we I on. ..
2&3 223885.... 10.7.
65nd massage... ..
SEQ.— «co—aggfi
» 88...” :2038593 i.e.—...:
+ [0.0:
rod
10.0
re;
a

 

 

uoN

a seJmoJeduJel Lndepm louogsuowgg—

43

 

.uoaaoua A: 0H e no soauauoau>m onu usausa 0.00.
no cuss-nomeoa ooeuusm Heauusu an as vaaom esouom.soz a you
06.9 Assad-606.0.soz u> nouauquoaaoa nuaovsu Hence-606.0.soz s.n ousuwm

.... mEc. _oco_mcmE_olcoz

0.. .... a; 0.. who who .30 «“0 £0 «.0:
_ . . 0.0..
. i=8 328.80 a .6. 22.8. 20.68505 5. a... .. . 20.6852... 0 8 0.. ..

SE 0.0 o. 0 296852.... 1. m. P I
5.. n... 0 . 20.68505 .
EE 0.. 0 0.296858... .

85:0 . 0 208852.... ... O. F ....

10.0..

p p n -

 

 

 

 

 

 

 

 

 

 

 

H . 2 Jr” 10..

 

 

 

l
0.
O
eseanoJedwai Lndepul louogsuew‘gQ—UON

(+4

.00a0000 A: an 0 00 souuauo0o>m on» 00.»:0 0.00.
«0 00500000809 0000000 Huuuucu 00 00 0HHom asouom 0 you
05.8 Husoaosoa.0.:oz 0> 00u00000050a.0000030 Hucoaaaoeua.soz 0.n ouswam

.... 0E: _o:o_mcmE_olcoz

 

 

 

 

 

 

 

0.\. 0.0 0.0 0.... 0.0 0.N 0.. 0.0 0. T...
. . . _ . _ . _ . _ . 0 . . 0. _..1
.0000 26.00 a .6. 30080. 20080520. 5: 0.0 .. 0 2000852.... 0 a0 10... .
.EE 0 0 0 0.0000052... ....w.0.l
.5... 000.00.08.52... .
:50 v0 0.0080505. Wum 0'
.500 22008052.... r.v.O.l
SE 0 a 0 0.0008500... 00.8.0... ..
850 ..208852... . TN.O.1
.00
N .
1N.0
10.0
10.0
1.. .. ....w 0
10._.
a
1N.—
.. 0..

 

 

BSGJnlDJedLUGi aooyng louogsuewgg—UON

115

.uoa0000 A: an 0 00 00.000000>m on» mCHuso 0.00H
mo 00000000309 0000050 Hquuacm ca 00 0.Hom nsouom 0 now
05.9 Hucoamc0e.0.coz m> 00u=unu0050a £00000H H000.m¢0e.0.coz 0.n ouswam

... 0E... _0co.mcmE_01coz

 

     
      

 

 

 

 

 

 

 

 

 

 

 

0.0 0mm 0mm .0? own 0N 0“. oflo 021

0:8 26.00 :0. 2.0.68. 20008505. 0 . O —.l
.50... 30208853.. 000... ...lw.O.l

65.0.0.0. 02000805.... . .
5.; 0.208850... ”.0 0|
EEQ— ..w§o_na§§fi o l¢.0|1

2030 .uoaSSEip 028- a .

.1N.O.1

10.0

1N0

-....0

10.0

10.0

w

111111 2| 0..

....

e SQJnloJedel mdepu; {Duogsuewlg—UON

46

.00H0ou0
A: an a mo coaueuo00>m 0&0 00.000 0.00N mo ousuuu00aoa
oouuusm Huauusu :0 00 ouaom 0:0u00.coz 0 now 03.9

..50..a:050.:oz 0> 00.309.000.00. 0000.30 1:05:030002 0.6 9.50...

.... 0E... 600.058.0152

0.0. 0._N— 0M0 0“0

- b

0.0....

 

......s 58.8.. . a. 23.8. 200852.... 5.2 .. . 20.38520. 0 .0. .. ....
.E... .. 0 0 0.000850...
5.. 0 0 0 2000852....
:5 0 v 0 20000052....
02:0. n..§:622sa#h
E... v N 0 2000805....
.. ... ..rww . m ”1m 2.000 ..296852F

 

 

 

 

 

 

 

 

 

0.0

 

 

9 SGJnlDJadUJSl eoopng jDUO1SU8UJ10—UON

£17

.0000000
A: an 0 00 00.000000>m 000 00.000 0.000 00 00000000309
0000000 00.0.0. :0 00 00.00 000000.002 0 000 05.9

00000000600.coz 0> 000300000509 £00000. H0000000300.coz HH.n 0030.0

.. 08: 6020000501002

 

     

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o.

 

0.0. 0.0. 0.0 0.0 0.0...
n — _ , — . — p 0.0
... 8 30.00.02. 0 .0 2.0 08 0 9.0850 0
= .. .0 . . .0. e... 0.: .. . 20885.0. 0 .0. u 0. .
. .5. 0.0 0. . 20385.0.
...... 0.... 0 . 20885.... 1 N0
...... 0.. a 0 0.968500.
08.5.0 . 0 0.0008520. .
1 ...0
... 0 .0
1 0.0
0 .
1M / ...:

 

 

e SGJnlDJadLUGi Lndepul [ouogsuewgg—UON

48

00.30.00 A: an 0 m0 03000000>u 05 900030 0.00“ «0
00300000600. 000.0030 H3300" 00 u< 0.300 030000 0 00m 0.0.3.
100000030002 0> 000900000500. 000003 0000000050002 36 00030

.L 08:. 6020000051002

 

 

 

 

 

 

 

 

 

 

 

 

o.m 0.0 0.0 o.m 0.0 o.n 0N o; 0.0 0.7.
. _ . . . . . L 0 — 0 _ 0 _ 0 L . N.OII
.2000 3800 0 00.. 0002000— 3000000.0005. 0
5300 .20—088590. 002.. 0. .
.0000. 0.0—00000055 p .
E0. 0. n 0 3000000005.. IO 0
3&0 1300840500. .
0.0; n00.00800cofi .
.00; «0030085800. N TNO
880 22088690. ..
4.0
T06
..md
..... -00
m.—

 

_e SQJmpJedUJSl eooyng [ouogsuewgg—UON

49

.0000000 .00 0m. 0 .00 00000000020 000 000000 0.000 .00
00000000000. 0000000 0000000 00 00 00000 000000 0 00.0 0000.
00000000300002 0> 000000000500. £00005 00000000200002 00.0.. 00000.0

*0 00.5 6020000051002

 

   

 

 

 

      
 

 

 

 

 

 

 

0.0 0K. one ohm 0.0 0?... ohm 02 ouo oql
. . . . . _ . _ . 0.01
.33 0000000000 200000.0—00800005. EE 0.: : 00.088055. 08— no.0. ..
SE in o— 0000008§ _ 1.0.0
00...; 0030080055.
.552 0000088590. .
000050 00.800.890.05. ...N.O
SP“; .. 0 0 0
L Tao
10.0
r
10.0
“IO.—
j‘ll .
: r0 0

 

 

e SSJrnDJSdLUQl Lndepul louOgsuewig—UON

SO

penetration of moisture. (ii) The evaporation time for the porous case
is lower than the non-porous for the same droplet size and under the
same condition. After the droplet vanishes from the surface, the porous
solid recovers slowly, as opposed to the quick, pure conduction,
recovery for the non-porous solid. The porous indepth temperature
profiles during recovery are not as smooth as the non-porous solid, due
to the migration of moisture inside the matrix (as can be seen from
Figures 3.4 and 3.9). The presense of the moisture changes the thermal
properties of the solid. This change makes the recovery behavior not
only different but also longer than the non-porous solid.

The evaporation time for various droplet sizes versus intial solid
surface temperature is shown in Figure 3.14. This time is as defined
previously, the time taken for the droplet to vanish from the surface.
Figure 3.14 shown that this time is a function of the droplet size and
the intial surface temperature. As expected, lower the initial surface
temperature and larger the droplet volume, more time is required to

evaporate the droplet. Figure 3.14 shows also that for TS 5 164°C, the

evaporation time of the non-porous is a larger than the porous. For

Ts > 164°C the evaporation time for the porous is larger. For this

range the parameter B for the non-porous solid was larger than the
porous solid (see Figure 3.19). Generally, the evaporation time on the
non-porous solid is larger than the porous solid for the same droplet
under the same conditions. This indicates that in the process of
cooling hot porous and non-porous solids, the porous solid needs

droplets at higher frequency.

51

The recovery time is shown in Figure 3.15. This time is defined as
the time for the surface to recover to its initial temperature. This
time was determined by evaluating the time it took the surface
temperature to recover to 20 percent of the maximum temperature drop.
The maximum temperature drop is the difference between the initial
surface temperature and the equilibrium temperature of the thermocouple
underneath the evaporating droplet. The non-porous solid took slightly
less time to recover than the porous solid despite the fact that for the
non-porous solid, the evaporation process is steady and the evaporation
time is larger.

Figures 3.16 and 3.17 show the non-dimensional maximum radial (x/rd)
and axial (z/rd) influence distances of a single droplet plotted against

the initial surface temperature. The influence distance encloses all
locations where the temperature drop due to the droplet evaporation is
at lest 20 percent of the maximum temperature drop i.e., the non-
dimensional temperature 0 - 0.8. Figures 3.16 and 3.17 show that the
influence distance decreases at higher solid temperatures as more energy
is available to evaporate the droplet. Also these influence distances
are larger for the porous solid.

The non-dimensional maximum volume of influence (Vi/V) is shown in

Figure 3.18. This quantifies the cooling effect of a single droplet.
The volume of influence is defined as the volume of the semi-oblate
spheroid formed by the axial and radial influence distances. Due to the
low thermal conductivity/diffusivity of the ceramic, the droplet will
produce intensive local cooling, because heat recovery from other

portions of the solid matrix is limited. Thus, the influence zone is

52

. 3532, unseen «so?!»

you 339.3309 ooumusm 33m .2335 a> on; 60388926 36 0563

AOL 9.383th mootam 2.0m BEE

 

now 02 m: cm. 0%. cm? m—... 2.: mm

... druid”. a I | I.
t i ' '

-H A I. I

 

 

  

-

    

3L SEEKEB. oootam 2.0m .03....
u 9" - mm. . ofL at . 3. . om. . 3...

 

 

 

r at
1' II lo'""
I 1.0 .
IIII 1” III,
II .. I IQ IO.
D. llllllll II
I», m
I u
I I
I . I.
.8u2.8.. I I; 11 m
...... I. x ,d .2 h
is; o I; s
.12; . z a
.184 a I I.
12.5 .
13.5 .
on

 

I
8
(ass) emu uogoJodD/g

 

53

. .3332, 3398 28.3.3
you 8333309 3556 3.3m H335" u> as: hugged 36. enough

8.. otabotanoH mootam 30m 625
now cm. wt of. .22. .23. .01. .03. mm on

 

 

 

 

 

 

 

 

 

 

 

 

 

1|).l1» . — . b 0
row
// Too. m.
8.. $322.53. 8325 2.8 3...... .0 . o
2...... . . . . / sea
4 . M
/ / ICON “.4
a x . w
m m. .63 a
. s.)
.... a
I m , ..
ogunssmueun /// III N) /// soon {3-\
L .I one / .6 I
...“.mnmn / a m .. -3...
“Hum”; on. // . V ..
> a s 8..

5h

.urasao>.uoanou0 33.25 you 333.038. 33.30
03o» 7.3.25 2r 9233.35 H363. 55x3. announces—3&3. 35 can»:

Aoov 9.30..an0... montam 2.0m BEE
new. .03. MAL. .0m... .92. .23. .mt. .0n.:. mm 0N.

 

. _ . . in.“

 

 

 

 

 

 

 

......nn. H . .. M/ -3.
.12.... 0.... /.//u .
12.5 To MEN/MU.

4100.; nun. . N I lwmn H 10.0
.12->01. , IIdI/r/M ..
.... 5...... 7» I 0.0

. 9n

(pJ/X)
aouenuug [ogpoa 'ngow louogsueng—uo

55

.uossHo> uoaaouo nsouue> you ousuauodaoa ooeuusm
vqaom asuuunn o> cocosducn anqx< essaxaz Hacodacoe«0.coz s~.n ousmmm

8.9 238380... montam 2.0m .025
new 0? 0m... 0m. 0...; 0m: 0: 0a.: 00

h 0 — p b

0n.

 

 

 

.aouoa.coz
.souoa

sao.n>
saa.n>
sLonn>
41n¢n>
...nn>

 

 

 

aouenuul [0ng ways»; youogsuewgg—UON

56

 

..oa:~o> uoanoua nsoauu> uom ousuuuoaaoa ooauuam page»

Houuacn u> cocoaaucn «o oasao> asauxdl uncoaucoa«0.:oz 0H.n ousuum
8L. 830..an3 mootam 23w .025
com .02. .0? 3.. .92. .9»... .m... .3... .mm. .o

 

naouoaucox I

059.30 0.16
....o.n> o
.32.; a
....on..> a
.32.; o
.16.; p

 

 

 

 

 

 

 

( A/EA)

aouanuu; ,Lo awnloA 'xow [ouogsuewgg—UON

.35."; nod—85 3325
you 3313939 .0336 3.?» .7313 2. n nova-.33 36 33»:

8L 9.30..an2. 32:5 2.8 .025
2.x . om. . om. . o... . o? . 2.... o.» . 8 2 o

 

_ . - .

S7

 

\»II I... II
\h/
\x d\ \ \\
\ %\ \ .4.‘ I... &
mu \4\ zlox ...
\- \| \
... \0\
0..
osouoaucoz I
:6»: 9.6
.32.... o
.32.; 4
1.8.5 o
4:23.. o
.3 .9... p

 

 

Jagawwod

8

58

expected to be relatively high. The influence zone of the porous solid
is larger than the non-porous solid. This is due to the enetration of

the droplet into the solid matrix, thus causing more cooling in both the
radial and the axial directions. Figure 3.18 shows that the non—porous

solid at T < 100°C cools more with a 10 pL droplet than with larger

S
droplets. This is because the evaporation process for larger droplets
attains a steady state and the extra fluid in the larger droplets just
increases the evaporation time. This phenomena was not observed in
porous solids because the evaporation process never quite attains a
steady state.

Figure 3.19 shows the parameter 3 plotted against the initial solid
surface temperature. This parameter is defined as the ratio of maximum
droplet diameter during evaporation to the droplet diameter before
impact. The diameter of the droplet on the surface was determined from
the measured maximum wetted area. The droplets after impact had a
circular shape during the evaporation process except for the non-porous
solid at temperatures greater than 164’6. At these temperatures the
droplets explode due to the intensive vapor formation between the
droplet and the solid surface. As the surface temperature increases, B
tends to increase due to the decrease in surface tension. B for the
non-porous solid was larger than that for the porous solid due to

migration of a part of the droplet into the porous matrix.

3.4.2. Heat Transfer During Droplet Evaporation on the Ron-Porous Solid
During the experiments, the evaporation phenomena was recorded by a

video camera. The droplet evaporation behavior changed with the initial

59

solid surface temperatures (Ts). For 75 5 T8 5 125°C, upon impact, the

droplet have a disk configuration, and direct contact between the
droplet and the solid exists. The evaporation takes place from the top
and the sides of the droplet. The droplet diameter on the surface
remains constant for up to approximately 80 percent of the evaporation

time, and then starts to shrink. At Ts - 125°C, after the initial

contact, nucleate boiling was visually observed (see Figure 3.20; note
that the equilibrium temperature is above the boiling point of water).

As Ts increases (i.e. 150 5 T8 5 175°C), the boiling became more

dynamic, and the extensive formation of vapor sometimes scatters the
droplet. Seki et a1. (1978) reported similar droplet behavior at these

temperatures. For Ts > 200°C, the droplets spread very fast on the hot

surface, forming a thin water film. The evaporation time at this
temperatures was very short.
Individual traces of the surface and indepth temperatures for a 51

pL droplet impacting on the solid at T8 - 125°C are shown in Figures

3.20 and 3.21. As discussed in section 3.4.1, the hot surface
immediately cooled to the contact temperature upon contact with the
droplet. The contact temperature was evaluated at the temperatures
studies using equation 3.1. This temperature was compared to the

experimental values. For T8 < 164°C, both temperatures were in
excellent agreement. For Ts > 164°C, the measured contact temperature

is approximately constant (slightly above the boiling point) and is
lower than the theoretical contact temperature. The theoretical and

experimental contact temperatures for the range studied are shown in

60

9535 03.2.

Joan—0.8 A: an a no souuouoegu on»

I? ”MAO” gOHOm

:32 < ooh sensuouonaoa ooomusm owé 35mg

Aoomv oEc.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o —... F . 0.0 °_B O—W . .0.” . o. P. . o _. Ilom
6:... 328.8: a 8.. 82.82 3§o8§o§ as n... . p . 258.52...
. .. S
n
..om u.
D
3
u 9
II.—
roo F 8
I. w
r d
m
..o. . m.
a m
u 9
D p c - s
It .1 4 WI! IONF \d
I) l O
N . (

 

 

 

on...

61

 

60398 A: a... a mo scavenger» on“.
95.39 3.3. Er 63o» osouoméoz a you nouguuooaoa snoop—nu .36 83»:

A03. oEc.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

om. o... o.» 0.. o...
f - . . on
as a... z . €385.25 ..
55...... o. . £885....

.5... ....» a . 933853.... low
E... a; a . 2&08585.
cue...” . . 2:88:55. ..
..oo .
fl 4)., ‘1!
u
no . .
, o ..
O~ jONF
11 .. UH

 

 

on...

(30) semimadwel mdepu'

62

Figure 3.22. Seki et a1. (1978) found that on stainless steel, the
measured contact temperature was in agreement with the theoretical
contact temperature up to 180°C. For higher temperatures the measured
contact temperature was lower than the theoretical temperature.

Figure 3.20 shows that after achieving the contact temperature the
solid surface approaches the ”equilibrium temperature," (i.e. interface

temperature during evaporation). This equilibrium temperature remains

nearly constant until a critical droplet thickness (6*) is reached on
the hot surface. Then, the surface temperature starts to decrease
sharply until the droplet has completely evaporated. This sharp
decrease in the surface temperature implies an increase in the
evaporation rate. As the droplet thickness reduces, the heat transfer
through the droplet increases allowing more evaporation to take place
from the top surface of the droplet. The time for evaporation after the
droplet has achieved the critical thickness is longer for larger droplet
at a specified initial surface temperature. This is because larger
droplets spread more on the surface, and this large volume require
longer time to evaporate. As the initial surface temperature increases
this time decreases due to higher heat flux. After the droplet has'
completely evaporated, the solid recovers smoothly to its initial

surface temperature. For T8 2 200°C, the interface temperature (which

is equal to the contact temperature) was a few degrees above the boiling
temperature of water. Since, the droplet diameter after impact on the
surface for this temperature range was very large (see Figure 3.29) the
evaporation times were therefore very short. After evaporation, the

surface recovers to its initial temperature.

63

Prior to droplet impact, the measured temperatures were compared
with steady state calculations. The agreement was within 2°C. This
temperature distribution was used to calculate the heat flux inside the
solid which was then evaluated at the solid surface. From the surface
heat flux, the overall heat transfer coefficient was determined
according to equation (3.2), the convective heat transfer was determined

according to equation (3.3)

q” - ho (Ts - To) (3.2)

q" - h (Ts - T”) + co (T: - T2) (3.3)

Figure 3.23 shows the heat transfer coefficients at the surface of the
ceramic block versus surface temperature.

Surface and indepth temperature measurements during droplet
evaporation show that the isotherms and the heat flux lines inside the
solid matrix may be approximated by oblate spheriodal coordinate system.
In such a coordinate system a temperature distribution [adapted from

[Keltner (1973)] is given by:

2(T -T )

2
T<g,t*) - -——3——§— (1-2 5— + L-)cot‘1 g + T

x 6 62 s ’ (3'4)

Where g represents the isothermal lines, t* - §§ is the dimensionless

time, a is the thermal diffusivity, R is the radius of the disk, Ts and

64

.ousueuomaoa oceans 3.3m .13”.qu Er 333350... 33:8 36 was»...

8... 9.323th oootam 23m .03....

 

 

on» own 0mm com 02 00 F on
n p b b L b P h b
. 0...
wow 0
O
...om w.
. w
109. 1.
. . l
..ou. m
.. w
:0:
. m
100. m1
. a
10m. \w
. (O
.oacoEtoaxm o ICON
.ooztoofi. II .

 

 

 

65

can

655309.39 33.5» 3.25 H335" 2r 300.3
oaseuoo on» no ooeuuam on» us cocoaoqumooo nomuceua use: n«.n ousmum

Gov 239.382. mootnm 2.0m BEE

 

 

oon omN DON omw 00F om
. — . _ . —. . _ . — . n
O O [D
o o o o o o o o o o ..
In
0 1.0
o
o .
n.o o 1..
o r
o
o In.
0 I
o .caoEooo 3.39.... .31 02.02600 0
o “5.0—taco L832... .00: .335 o in.

 

 

(M'zw/ M) Siuagoggeog .1949qu _|_ 309};

66

non—32’ uodooun 2.6.3.5 you 3332.509
ocean—am 3.?» H335 e> no.8 use: 3332—26 gauge 3..” 9.53...

EL 838an2. montam 26m 625

 

on... can emu own. om. 2... on
. p . . . . lo
0 a a
O 1
m .
w 4 O rOON
D O
O D .
D d
r.00...
. . o
D d u
o
p o r.80
o > 4 o r
u a. ..u o. o
o ._ a. a ..
4 q o .1 on a com
o .3 Q. o .
o .3. .n p

 

 

 

coop

(aw/MM) an meH eAgoJodo/xg efioJeAv

67

£25.25 noun—9.8 26.32, you 333330....
333» v.30» 1.3.2: -> “3.8 .3: 2,3309% omwuo>< mm. n 0.3»;

Gov 230..an8. mootam 26m 62:.

 

 

 

 

_ V
on» can emu 8m . 09 8. on m
r p n p — . — . — filo J
o a a r ,m
a r 9
o ”.00.. 3
... m
o NOON d
o a . m
D noon m...
.. A.
d i. 9

o -oo¢
n o n .W
4 noon m.
> p n H
o p a o .Sopouooon
a o 41 3 4 .. X
o m. .3 on a woos m1
..u 2 o. ... w
P .. B p fl.oom W
. .z
S
(

68

Jog—30> uoaaoua 28.33, .
you 3330960.“. commas» 33m #335 -> 3:. 603.9235 36 as»;

Aoov BBEmQEm... mootaw 2.0m BEE

 

 

can com com com om. cm: on
[illicit-4m . . o
a a wow
r
89 239.3th 8826 2.0m .025 N
can . own L can . Ra . om“ . omu P a? . aged 0 o How
. ION—
o o o 0 Had «A... N “nom—
o a a a 1.6 .m < f
. a a a . 1 w 52
D N O [0.0 W. t
u w a r n ..
n a a .3 m. fi3N
‘ a
..a o. o o #0.. \sl IO¢N
i 2 « 1 x o u
an ... . r2 ( 1.2m
4: 3. . Y r
t.

 

 

 

 

 

l>
o
o
F)

(098) awu uogoJodo/g

69

...-5.25 uni—93 33...;
won 03532—18. 35.3» 3.?» H335 u> as: .9263... mm... 353

Aoov 9.30..an0... mootam 2.0m .03....

 

 

own 000 com DON OMF 00—. on
III}; . . _ . o
. a o o
m m 0 .
d 0 F00
D d r %
m -8. 3
8b SEEKS: 823m 2.0m .02.... o m
o...” a.» . om - a? , emu . omu . 3a . om. . mso D .. m
0 G o 1 0 Q TOWP
M a m m ... . .. p . u...
o u u p q we w
p o n o v o .l 9
n 4 1.- m 0 com )
a . s
1 o 1 u. 3
”.3 M" u ém 10mm mw
.3 on a . m o .
41 fl? 0 o o. ...-COW
.3 B o . .
2 b

 

 

 

 

70

Te respectively are the solid surface temperatures before and after the

step change, and 8 is the penetration depth. This is the solution for a
step change in temperature over a disk-shaped area on the surface of a
half-space. Using this equation, temperatures were evaluated at various
locations and at different times. The results were in fair agreement
with the measured values. From equation (3.4) it can be shown that the
rate of heat arriving at the surface (Q), is given by the following

equation

I.
Q - 4 R K (Te - T8) (1 + 6) . (3.5)

The instantaneous average evaporative heat flux is then determined by
dividing by the instantaneous wetted area. The instantaneous
evaporation rate may also be determined from Q by dividing it by the
heat of evaporation.

The average evaporative heat flux and the average evaporative mass
flux were correlated with the initial solid surface temperature as shown
in Figures 3.24 and 3.25. Both quantities follow the same pattern; the
higher the initial surface temperature and the smaller the draplet, the
higher is the average evaporative heat flux and the average evaporative
mass flux. This is because at higher temperatures, more energy is
available for droplet evaporation, and small droplets have larger
conductance due to their smaller thickness on the hot surface. Figures

3.24 and 3.25 also show a slight decrease in both quantities for T8 >

300°C. The evaporation time and the recovery time curves (Figures 3.26

and 3.27) also show a slight increase at the same temperature. Since,

71

this temperature corresponds to the maximum average evaporative heat
flux and the minimum evaporation time, it is likely the beginning of the
transition regime. Unfortunately, the experiments could not be
continued for higher temperatures, (i.e. in the transition regime), due
to the burn out of the heating mantle;

The non-dimensional maximum volume of influence (Vi/V) is shown in
Figure 3.28. This quantifies the cooling effect of a single droplet.

Figure 3.28 shows that the solid at T8 5 lOO‘C cools more with a 10 pL

droplet than with larger droplets. This is because the evaporation
process for larger droplets attains a steady state and the extra fluid
in the larger droplets just increases the evaporation time. Thus, the
smaller droplets are more efficient for cooling materials at this range
of temperatures.

Figure 3.29 shows the parameter B plotted against the initial solid
surface temperature. This parameter is defined as the ratio of maximum
droplet diameter during evaporation to the droplet diameter before

impact. For T8 2 200 °C, B is quite large, since the droplets spread

fast and form a thin water film on the hot surface. For this range B

decreases as T8 increases. This is because the droplet is evaporating

very fast, and does not have enough time to spread on the hot surface.

72

own

.3552, unack— naoauu> pom ouauuonaou. accuser. v.38

iguana .9 00:05.35 mo 25.3., .555: announces—3&3“ 36 936......

A0... 8320an... 83:6 2.0m .03....

 

 

8... SN 0.8 on. 8. on
T . . _ . _ r _ . b . .
om
m m m s
a -06.
m 4
m e (
w m m 10.0?
a o .
» Q 10....
D t
..1 o— o > m 06—
D .II
..1 3 4 m o
..m on a 4 a ..
.. n... o I .
..1 .n p o 0 CNN

 

 

(AW-A)
eouenum ;o ewmoA °xow louogsuewgg—UON

73

.3533 0030.8 30.725
you 33300309 003.50 030m ~03qu 0> a 302.30.. 36 0.33..

A0... 0.30.0083. mootam 2.0m .05....

 

 

 

own . own . OMNL owu . om? p Wm. . 0.0 . so...
0 u .
m a m p 10..
m ..o.m we
. w
m m -3 .w.
1?
o w . m
w a
m M Tod 8
m ._a o. 0 .
.... a. 4
c o .... on 0 106
a m .3. n... 0 ..
P D ...: PD 9

 

CHAPTER 4
DECOMPOSITION OF WOOD

In Chapter 2 it was noted that, prior to ignition investigation, it
is important to understand the response of the solid phase to heat, and
the resulting production of decomposition products. In this work,
products generated when a wood sample heated under controlled
conditions are examined in a systematic set of experiments. The set

includes 12 experiments, performed at 4 different heat fluxes

(approximately 1,2,3, and 4 w/cmz), and at 3 different moisture
contents (dry, 9%, and 17%). Conditions for these experiments were
suggested by the specific application of interest. In room fires, the
wood burning process may be treated as transient heating and
decomposition of a semi-infinite wood sample. The moisture content of
wood varies and is influenced by the climate, season and location. The
heat flux incident on the wood samples also varies with time and their
location relative to an already existing flame in a room fire.

The results of the decomposition experiments are presented in this
chapter. Section 4.1 describes the physical aspects of the problem and
discusses the importance of the moisture content and its relation to

the overall decomposition problem. Section 4.2 reviews the previous

74

75

literature. Section 4.3 outlines the procedure used to analyze all the

measured species. Section 4.4 includes results and discussion.
4.1 Background

Decomposition of wood refers to the process that produces various
chemical species and a residual charred solid upon heating. The
physical properties of wood are highly dependent upon its microscopic
structure. Wood is predominantly made up of a lignocellulosic
structure with various infiltrated substances. It is composed of three
major constituents: (i) cellulose (50% by wt), (ii) hemicellulose
(25%), and lignin (25%) [Stamm (1964)].

The physical aspects of wood decomposition are associated with heat
and mass transfer. Consider a semi-infinite wood sample initially at
ambient temperature, subjected to a prescribed heat flux. The sample
heats up by pure transient conduction, in which moisture starts to
evaporate and an evaporation zone begins to travel into the solid.

When the sample surface layer becomes sufficiently hot, it starts to
decompose. This decomposition zone moves into the interior of the
solid. Consequently, the products of decomposition are driven to the
surface of the sample, and flow upward due to buoyancy and mix with the
surrounding air. As the interior of the sample becomes hotter the
decomposition zone penetrates deeper into the virgin solid, leaving
behind a thermally insulating layer of char.

The rate at which the decomposition products and water vapor are
generated depends on the response of the solid-phase to the applied

heat flux. As pointed out by Simms and Law (1967) and Lee et al.

76

(1974), the presence of moisture in the solid phase changes its
thermophysical properties. This will influence the heat and mass
transfer process because of its large heat of vaporization. This heat
of vaporization depends upon how the water is held inside the solid
matrix. There are four ways by which water can be held in wood: (i)
Free or absorbed water: this water is mechanically held in the
capillary structure as a result of surface tension forces. Its
quantity is limited by the porosity of the wood. The energy required
to evaporate this water is only slightly greater than its latent heat.
(ii) Bound or adsorbed water: this water is held by hydrogen bonds to
the cell walls. Its quantity is limited to 30 percent of the oven-dry
weight of the wood. The energy required to evaporate this water
increases as moisture content decreases. At the fiber saturation point
(~30 percent moisture content) the energy required is the same as for
absorbed water. (iii) Water vapor: this is present in the air filling
the cell cavities. The quantity of this vapor is normally a very small
fraction of the total moisture content. (iv) Water of constitution:
this is a part of the molecular structure of cell walls and it is
firmly held by chemical bonds. In reality it is not water at all, and

is only released upon thermal degradation of wood.
A.2 Literature Review

Many investigators have studied decomposition of wood and
determined the various products. Schwenker and Beck (1963) detected 37

- volatile compounds, and they identified only 18 of them. Coos (1952)

77

listed 213 compounds which have been identified as products of wood
decomposition.

Martin (1965), investigated the decomposition of pure cellulose in
helium. He reported a time-dependent analysis of permanent gases and
tar. He indicated that there are at least two fundamentally different
ways in which cellulose thermally decomposes; one supplies the bulk of
the volatile fuel which supports the flaming combustion of the
material, the other produces mainly water and the oxides of carbon.
Vovelle et al. (1986) studied the decomposition from PMMA and particle
board subjected to a variable radiant heat flux under an inert
atmosphere. They indicated that the mass loss rate is directly
proportional to the instantaneous value of the heat flux absorbed by
the surface of the sample.

Lee et al. (1976), investigated wood decomposition in air at two

heat fluxes (3.2 and 8.4 w/cmz). They indicated that the decomposition
process is dependent upon the external heating rate, the total time of
heating and the anisotropic properties of wood. Lee and Diehl (1981)
investigated the effect of absorbed water (samples with 50% of oven-dry
weight) on the decomposition of wood. They indicated that moisture not
only changes the solid-phase thermal properties and delays the
decomposition process, but also substantially dilutes the decomposition
products.

Atreya (1983) , investigated wood decomposition experimentally and
theoretically. He performed numerous experiments for 10 different
kinds of wood in air, and reported transient measurements of surface

temperature and evolved mass flux. In his theoretical investigation he

78

included energetics and kinetics of moisture desorption in his model.
He indicated that the desorption of moisture, which has been ignored by
previous investigators, has a considerable effect on the energetics of
the decomposition process.

Kashiwagi et a1. (1987), studied the products generated from wood

decomposition in 3 atmospheres of different oxygen concentrations (N2,

10% 0 - 90% N 'and air). They indicated that ambient oxygen

2 2’
significantly increases the desorption mass flux and surface
temperature. They showed that decomposition rates in air were nearly

double those in nitrogen, and surface temperature in air increased as

much as 200°C over that in nitrogen at 4 w/cmz, due to oxidation of

char.
4.3 Products Analysis

It is important to know the production and the depletion of major
species during decomposition. The equipment used for continuous

measurements of C02, C0, 02, H20 and total hydrocarbons (THC), as well

as the necessary time lag and response time corrections, are described
in Chapter 2. As a result, the mole fractions of all measured species
were obtained as a function of time. Since the instrument used to
measure CO require a dry sample stream, the mole fraction of CO was on
a dry basis. This value was converted to a wet basis using the

measured mole fraction of H20. The following equation was used:

mole fraction on wet basis - measured mole fraction on dry basis (4.1)
x (l-mole fraction of water removed by the drier)

79

The evolved mass fluxes of all the measured species were then determined
using the total mass flow rate of air. The total mass flow rate of air
includes the air flow through the critical nozzles (equation B.l in
Appendix B), and leakages from the gap around the sample and top of the
tunnel. As explained in Section 2.2.2.3, these leakages were determined
by the methane trace method, and they had been calculated from the
difference in the methane concentration measurements upstream and down-

stream of the tunnel. The instantaneous 02 depletion was determined by
subtracting the initial steady state value of 02 from the measured value

at any time.

The sample mass flux was calculated by dividing the time derivative
of the measured transient weight-loss by the initial front surface area
of the sample. Although it was not possible to measure the organic
condensibles (tar), its instantaneous value was determined by difference

[(sample mass flux + 02 depletion) - (sum of the mass evolution rates of
C02, CO, H20 and THC)]. Finally, the mass fluxes of all the measured
evolved species were normalized by the sum of sample mass flux and O2

depletion to obtain the percent mass flux.
4.4 Results and.Discussion

This section describes the results obtained from the set of
systematic experiments done on Douglas fir. In these experiments as
described in Chapter 2, simultaneous measurements of weight loss,
surface, bottom, and indepth temperatures, oxygen depletion, production

of C02, C0, total hydrocarbons (THC), and water were made. The

80

experiments were parameterized with the sample moisture content and the
external radiation flux. All experiments were performed in air under
non-flaming conditions. A duration of 30 minutes was assumed to be long
enough to reveal enough information about the decomposition process.

For all three moisture contents, spontaneous ignition occurred at 4

W/cmz, and data was continuously recorded even after ignition.

4.4.1 Sample.lass Flux

Mass flux histories for dry, 9%, and 17% moisture contents at 1,2,3,

and 4 W/cm2 are shown in Figures 4.1, 4.3, and 4.5 respectively. These
figures show that higher evolved mass flux corresponds to higher heat

flux. The general trend for mass flux at high heat fluxes (4, 3, and 2

W/cmz) is rise to a peak, fall-off and then attaining a nearly quasi-
steady state value.

Wichman and Atreya (1987) have developed a simplified model for the
decomposition of charring materials in an inert atmosphere. They
predict a fall-off in mass flux proportional to the negative one-half
power of time. Figures 4.2, 4.4, and 4.6 show the mass flux histories
plotted on a log-log basis. The fall-off rule agrees only for some
cases and for only a short period of time. For example, the curve for

the dry case at 4 W/cm2 follows such a time-dependence from 100-240

seconds. This disagreement may occur because of the effect of ambient
oxygen concentration on the decomposition process. The behavior of
evolved mass flux in the presence of oxygen is expected to be different
than in an inert atmosphere, because oxygen not only causes a strongly

exothermic reaction at the sample surface, but it could also diffuse

MASS FLUX (mg/cm2 3)

MASS FLUX (mg/cm2 8)

81

 

 

 

 

 

 

 

 

 

 

1.4 .
1.2-1
1.0-1 1.. ‘ .
Z -, a W/cm2 3, 3',
III .0 :0 q. 0 0 . ’9
0.8-‘5. \ '3 ", if. .‘5 315- ":\ °
2 ' .‘g‘g' 7‘fl€?§ 1‘... o,’ '3' ‘1‘, 5"
O.6-: ' °' ' - 'r' a", .
: . . 3 w/Cllz . : 3’0 0". .
04-3 .."-- ’3. 1"}. ,.~’;.' '1-513
2. : ":2: WW“ ' '
002-:‘05 Mm ‘
'3’ t 1 w/cm2 . A ...
0.0 It"!!! If! llrviil IT! 733 III
0 200 400 600 800 1000 1200 1400 1600 1800
TIME (s) ‘
Figure 4.1 Effect of Incident Heat Flux on Time-Dependent Haas Flux
for.Dry Condition.
1.00 0..“ .....'0~.. (O w/sz 5"-
o.80-E \/\t"m‘{\’?§‘{1"‘a
0.60— ,5
. n....r*""”‘~\“"~\.s}}{5%
0.40- ° “ W
020- 2 ulcmz
0.101
0.081
0.06:
0.04- It
.. ’ fw\w‘W-;‘3M§
0.02— .
m2
0001 U U r '2'; I I A ltwlc I I I TU'
10 50 '1' O 500 1000

TIME (s)

Figure 4.2 Effect of Incident Heat Flux on Time-Dependent Hess

Flux for Dry Condition

 

MASS FLUX (mg/cm2 3)

MASS FLUX- (mg/cm’ 8)

82

 

 

1.0
0.8-
;; 4 U/cnz
05': - NW w‘w‘. "WWW-.4” «JAM
:0. \x: vxflwflhlxflw’ 3"“
I
00‘: o... 3 W/sz 'W
.. y 1'
b# 1 WICIIIZ
.— .;_-'A—-:“ A L4— __ —— .. A TAvA_M_A_.__’
000'000l101'00000' '0' '0' '1‘ '0' 0"

 

 

 

200 400 600 800 1000 1200 1400 1st 1800
TIME (s)

Figure 4.3 Effect of Incident Bent Flux on Tine-Dependent

Hus Flux for 112 Moisture Content Condition

 

 

 

 

1.00
008° ... ‘0. 4 "/m2
0060'“ .. . LWAWVW'
0-40- .° ..-°---:" 3 w/ «X3 va‘W‘J”

. 3 ' ° 3 0 ::°.. WW
0.20-a

I 1
0.101‘
0.084 --.-.... 1 111/an

.1 ,. '
0008- ”‘11.. ~11» . .‘

.. «I- . :‘Afgfiwfi
0004 l T l I I I I i g ‘ I U I I 'l

10 50 100 500 1000

TIME (s)

Figure 4. 4 Effect of Incident Heat Flux on Time-Dependent

Mace Flux for 112 Moisture Content Condition

‘ MASS FLUX (mg/cm2 3)

MASS FLUX (mg/cm”1 s)

83

 

 

 

 

0"]

000 1200 1400 16'00 1800

0.0 IIIIIIFIII'III 0"l000l000l000
0 200 600 300

TIME (s)

Figure 4.5 Effect of Incident Bent Flux on Tine-Dependent
Mess Flux for 172 .Ioieture Content Condition

 

1 000 '-
0.80%

0.60-
0.40-

0.20-

0.10:
0.08-

0.061
0.04--

 

 

 

 

0.02 I .' U-IYIIUI I I I III!
10 50 100 500 1000

TIME (s)

Figure 5.6 Effect of Incident Heat Flux on Time-Dependent
Mass Flux for 172 Moisture Content Condition

 

84

through the porous char to the decomposition front of virgin wood.

Thus, to theoretically investigate wood decomposition in air, it is
important to include the heat released by char oxidation in such a

model.

From Figures 4.1-4.6, we note that the mass flux curves do not start.
at the origin. This is because the mass flux is an average quantity,
determined from the time derivative of the measured mass loss. Since
the initial mass evolution from the samples is essentially pure water,
the starting point is higher for higher moisture content. The
desorption of water from the solid matrix is quickly followed by the
evolution of a large amount of tar. This increases the mass flux, and
it continues to increase until a thin char layer forms. In this heating
stage, the mass flux attains its maximum value. .As the sample heating
rate increases at higher heat fluxes, the heat-up time is decreased.
Thus, for higher heat fluxes the mass flux peak occurs earlier. Vovelle
et al. (1984), observed that this peak increases linearly with heat
flux.

The production of volatile gases, which escape from the solid,
leaves behind a layer of low thermal conductivity char. Formation of
the char increases both the absorptivity and emissivity of the surface.
But since the thermal conductivity of the char is much lower than that
of wood, the net heat transfer through the char is reduced but the net
heat loss by radiation is increased. The oxygen in the ambient air also
reacts with the char in a highly exothermic reaction and supplies
additional heat to the solid. This heat compensates for the reduction
of the net heat transfer through the solid due to the formation of char.

Furthermore, the char cracks permitting a larger quantity of volatiles

85

to escape. A combination of these factors is believed to produce the
quasi-steady value attained by the decomposition rate after attaining

the peak.

4.4.2 Sample Temperatures

Measurements of surface, bottom, and indepth temperature of the
sample were also made during the experiments. These measurements
contain the history of heating of the solid matrix, and they reflect the
thermal response of the solid to the irradative heat flux. For example,
surface temperature measurement is very important with regard to the
heat generated by char oxidation. These reactions take place primarily
in a thin surface layer, because their rates strongly depend on
temperature and the presence of oxygen. Thus, surface temperature is
very useful in determining when the heat of char oxidation is being
released as well as at the rate that this heat is transferred to the
surroundings.

Surface temperature profiles for dry and 17% moisture content at

1,2,3 W/cm2 are shown in Figure 4.7. In this figure higher surface
temperature corresponds to higher heat flux. The rate of rise for these
temperatures is determined by the moisture content and the irradiative
heat flux. This rate increases with higher heat flux and lower moisture

content. As an example for the dry case, it took 240 seconds exposure

2 for the surface temperature to reach 300°C and Only 16

at l W/cm
seconds exposure at 3 W/cmz, whereas for 17% moisture content it took

1600 sec to reach the same temperature at 1 W/cm2 (see Figure 4.7).

86

Char oxidation occurs when the surface temperature is greater than
400°C [Kashiwagi (1987)]. This reaction is strongly exothermic, and is
expected to rapidly accelerate the local temperature rise. Figures 4.7
and 4.13 show that, when surface temperature suddenly increases, a
significant increase in oxygen depletion is noticed at the same time, as
expected. Figure 4.7 shows a sharp rise in-surface temperature occurs
for the 17% moisture content case at 245 seconds. Simultaneously oxygen
depletion increases significantly at the same time (see Figure 4.13).
Thus an inflection point in the surface temperature at a value greater
than 400°C indicates the start of char oxidation.

Figure 4.7 shows that the start of the oxidation time increases as
incident heat flux decreases, and decreases as moisture content
decreases at the same incident heat flux. For example, char oxidation

2

occurs for the dry case at 550 seconds when exposed to 2 W/cm , and at

215 seconds for 3 W/cmz. For 17% moisture content the oxidation occurs

at 1050 seconds for 2 W/cm2 and at 245 seconds at 3 chm2. This is
because the char develops more quickly at higher heat fluxes and lower
moisture contents. The effect of moisture is expected to be less
significant near the heated surface after char oxidation starts, because
for this region the temperature profile is primarily determined by the
exothermic heat generated from oxidation of carbon.

Temperature profiles at various locations in the sample

2

corresponding to 9% moisture content at 2 W/cm and 17% moisture content

at 3 W/cm2 are shown in Figures 4.8 and 4.9, respectively. The indepth
temperature rose slightly above the boiling point of water, and remained

constant for some time before continuing its rise. This is due to the

.xsdm panacea acovaoca naoaun>
no unsunuunaoe ooeuusm so usoucoo unsumaoz mo uouuum u.e «woman

Amy .2:

com. com. 00¢. com. 000. com com 00... DON

bub—DIPP_-P—Pbb|—

o

 

>mo --..
2 R up I

87

I L

00000000000000.0000 0.00 C 8 00“.. 00

M

 

¢ .

 

nub PEI. DP. —-P-bPPL—o

nae-fir

NEO\3 .—

I}

1111'

‘3'“)

.

111

.............................:.....

NEO\3 H

a 4......
wao\3 ‘swts

5

«an; ~\‘\~30\3 n \

00.10”...

. .0113 s. ...

a
N

«J

.
‘fi

..

EO\3 m

oo_.

 

 

(0.) ElaanaZ-IdWZ-Il BOVJEIDS

88

000m 0

.souuuvsoo acoueoo ousueqoz.no you Nao\3 N as «Haanm
on» c. enoaueooa esouue> an «8.9 nouo> oueunuonsoa m.< ouswum

00.. 00.0

3 Us:
.00.: 003 000. 000 000 00¢ 00m 0

up. bun—b-n—nnn—ppb—ppn—-p-PP—o

 

 

1.111111111111111

     

3an .838 - 1

aacu

ooouuam

 

 

 

(0.) aamvaadwai BTdWVS

89

.eoauaveoo usousoo «asunder as. you «aux: n on «Haaam
on» e. «coaueooa esoaue> on «sub usouo> ouaunuoaaoa a. e shaman

 

 

 

 

 

 

3 m2:
000.. 000—. 00¢. 00m. 000. 000 000 00¢ 00m 0
DD. :DDDBLInbb-D-L-Ir. b-n b-P —-DL_0PLI—o
.3... 333. , I 1. \

1 En... 00—
Hoom

m
noon

m
H00¢
moon
8.2%. .....000

aam.~ u

 

 

005

(0.) aamvaadwai 31dwvs

 

90

endothermic vaporization of the adsorbed water. This plateau in
temperature-time curves became longer at greater depths in the sample
and at higher moisture contents. The upward inflection of the indepth
temperatures at 600 seconds in Figure 4.8, is caused by the heat
generated from char oxidation (notice the sudden increase in the
surface temperature at this time). For different moisture contents it
is difficult to compare local indepth temperatures, since thermocouples

are not at the same locations for all cases.

4.4.3 Decomposition Products

In this study, oxygen depletion, production of 002, CO, total

hydrocarbons (THC), and water are closely examined in order to quantify
the effect of the adsorbed water on the decomposition process. The data
are presented as mass flux of the permanent gases, and as percent of
mass flux. Plots for different moisture contents at different heat
fluxes are shown in Figures 4.10-4.17.

The water has a similar trend at 2 and 3_W/cm2. An initial mass

flux peak arises from the desorption water in the solid matrix, followed
by a rise to another peak, and eventually a gradual fall-off in the

water production. This latter water may include the water of

constitution. The water curve at 1 W/cm2 shows only one peak (see
Figures 4.16 and 4.17). The quantity of water produced is generally
larger for larger heat fluxes and higher moisture contents.

The water production is quickly followed by a large amount of tar
whose quantity was determined by difference, as mentioned previously.

This organic condensate has an extremely complex in composition.

MASS FLUX (mg/cm” 3)

PRODUCT AS PERCENT OF MASS FLUX

0.24

0.16

0.08

0.00

91

 

 

 

 

320
o 10“
0. “H . .. 0‘ 02 DE?
. .°'0'o‘o. m
c ‘ .1 0. 0.1 002 “-"JV‘L:
. ...-«"3.» .1...- __
«9'7: 0 0, CO
" ’5”: ' ° " °
. .. m
n'. fig?‘ I
0 200 400 600 800 1000 1200 1400 1600 1800
TIME (s)

Figure 4. 10 Time-Dependent Mass Flux of Evolved Products
at z ale-2 for 9: Moisture Content Condition

 

100
80-
q
60 "
320
O‘."
40 I [d 0‘
“I "1
a 1'-
a 1‘. TAR
\
0 .‘0 - CO .e
2 8 ‘ ‘ .2”-’.o..0'-r,w
.' ~04“..-..._‘ Lou—....-
quY~ 3119' CO #
“...,”W’. 7 T110
0 I 0" '0' '0' '0' ll. UIUIUIUIIIU

 

 

 

 

0 200 400 600 800 1000 1200 1400 1600 1800

TIME (s)

Figure 4.11 Time-Dependent Product as Percent of Mass Flux at

2 1J/c1n2 for 92 moisture Content Condition.

MASS FLUX (mg/cm2 3)

MASS FLUX (mg/cm’ 3)

 

 

 

 

 

 

-OoM'III'III'TIr'III II III rII rII'Ifi

0

 

200 400 600 800 1000 1200 1400 16001800
TIME (s)

Figure 4. 12 Time-Dependent Hess Flux of Evolved Products at

3 11]an for 9: Hoieture Content Condition.

 

0.32
320
0.28
0.24 -‘ 1
0 . g
a 1 ,
0.20 . 1 o ‘0 '.\
0 1 o " ~°‘TAR
0.16 .0 \o 0 ‘\. ‘ .... ...‘ 02 DEPLETION
. ..°‘O..I‘.‘ "s. . ‘9 I
Oe12 . 3 co . ‘0 \ 0| .
0 as ' ' x ' ‘-~"‘""": .331.
~ I A . ......--........~.' - -.
0.04 ., I ' VAT 4 fr”
0.00 :“
-OoO‘IIVI'III'III'III IIiIIfiI III'ITI'II

 

 

 

 

 

0

200 400 600 800 10001200 140016001800
TIME (s)

2
Figure 4.13 Time-Dependent Mass Flux of Evolved Products at 3 H/cm

for 172 Moisture Content Condition.

93

 

 

 

 

 

 

 

 

1.2-3 .1
A I I
.0) . 1
1'4 '1 E
a - 1.
01 0.8: 00,1. co2 -.
.l p 0“ 1'. “A 0' ‘m 's .0--
3 2 . ‘\ 1"" ...11. : ' “ “x.
x 016': '1 0; nmmon
D m .‘ .
..J . '1 '
h- 0.4- ; ..
‘ ' I . 1
g : : . : \.. . ... maz . 0" :0
s I, 1
2 0.2:: : ‘3 so 1“”...0.“ “. '0': as“: .. 1‘: . .“.‘
- ' ' co and mo ’. Crc' .,' ,. '.' ‘ '
0.0 ’ “" ‘
0 200 400 600 800 1000 1200 1400
TIME (s)
Figure 4.14 Time-Dependent Mass Flux of Evolved Products at 4 chmz
for Dry Condition.
1.0
.I
r-\ i
m 1
n 0.8-1
E .1
O '1 g
b. 0,1
E ' 4
t. y l
X 1 . . “2°
3 0.4-1 \ - . '3. '- '0'
'4 fl 'o."‘V Vvv‘vv 12:”
la. ' J 1
1 0‘ I. c02:.: \. ,‘M.’ *Vu .. '0.."\.:.'~
m - . 0‘ ~ .. 1" '. '. ... ..W". 0..., 0‘130':
‘0 0.2-1 Ig‘.o"'o' ' 'Q . 1' T“ o 1' .
< .I 0 e.
2 .4 , m
0.0 ,g 1‘ .14-m _r00 and THC ,
0 200 400 600 800 1000 1200 1400 1600 1800

TIME (s)

Figure 4.15 Time-Dependent Mass Flux of Evolved Products at 4 W/cm2
for 91 Moisture Content Condition.

94

 

 

MASS FLUX (mg/cm” 3)

 

 

0.040 A
0.032
n 0
0.024 I
I I
a I ,
’0: m ' '5“. O 0‘
0.016 . e 1"...1.‘ 1.. .l' :0"; ".0 I ' '1' .
. 1 -",.1‘."1'1. 1- ,{L'°°',: 5.11"." "' 7.1111111." '.'
-: ':. ' . °I E
0008‘ . .1 1 THC -
. , '1'. .. ,- 02 flananou
d o ' rrmm' 1““ Hui '1 "‘7
0.000 .. ’. Jill ‘M‘.. I s
0 200 400 600 800 1000 1200 1400 1600 1800

TIME (s)

Figure 4.16 Time-Dependent Mhss Flux of Evolved Products at 1 Wen2
for Dry Condition

 

 

 

 

0.08
A 2
m
N ..
E 0.06-
0 .. 320
E 1
E " .
V 0.04- :, v, .
>< ,1 ,."11An . ‘ s
D " 1 ' 1' ’1 ‘1 ‘0 '1
d .I c I .I. . fl .0 3... ".I II
' I. ' ' . g. 1"." H" '4’;'1 III
(I) 0.02 ° I. ' ‘I'? a ‘1'”.1 I" ‘ III I I
Q I. ”I...” :.'I ' I ... b I .: :
2 A '0‘ '. 1 . . ° II
DEPLET 0N
21g. .fll HALL“. -1- 2 $691,311.
0.00 - -1«—

' 200 400 600 800 10001200140016001800
TIME (s)

Figure 4.17 Time-Dependent Mass Flux of Evolved Products at 1 II/cm2
for 171 Moisture Content Condition.

95

Kashiwagi et a1. (1987) investigated its composition by means of
capillary gas chromatography coupled with mass spectroscopy. They were
able to pass only 20% by weight through the gas chromatograph. They

identify more then 40 different species.

The production of combustible gases (C0 and THC) at 2 and 3 W/cm2
shows similar trends; gradual increase until attaining a steady value.
For both gases this steady state value is slightly higher for higher
heat fluxes and lower moisture contents. But a major difference is the
time it took to attain this value. For example, for the 9% moisture

content case the steady state value was attained after 950 seconds at

2W/cm2 (see Figure 4.10), 200 seconds at 3 W/cm2 (see Figure 4.12),

whereas for the case with 17% moisture content it took 400 seconds at 3

W/cm2 (see Figure 4.13). At 1 U/cmz, there were only a small amount of
combustible gases produced for the dry case (see Fig. 4.16), and a not
noticabe amount for the 17% moisture content (see Figure 4.17). Thus
the presence of moisture significantly affects the production of
combustible gases by both delaying and decreasing their production.

The oxygen depletion is the amount of oxygen used to react with
carbon to form carbon oxides. As mentioned earlier in Section 4.4.2,
the time when a significant increase in oxygen depletion occurs
corresponds to the time of an inflection point in surface temperature.
Generally, the higher the heat flux and lower the moisture content, the

larger the oxygen depletion and C02 production.
The mass flux of all the measured species was time-integrated for

all three moisture contents and for l, 2, and 3 W/cmz. Since

96

spontaneous ignition occurs at 4 W/cm2 for all cases, this set of
experiments will be discussed separately later in this section.

The effect of the heat flux and the moisture content on the
integrated mass products are shown in Figures 4.18 - 4.21. These
Figures show that a larger amount of water production corresponds to
higher heat flux and higher moisture content. Also larger oxygen

depletion and 602 production correspond to higher heat flux and lower

moisture content. Finally, the production of combustible gases is
larger for high heat fluxes and lower moisture contents, as expected.
From these results, it is clear that the presence of adsorbed water in
the solid matrix reduces the average fuel rate evolved during the

decomposition process, by delaying their production.

At 4 W/cmz, and for all moisture content cases (dry, 9% and 17%
moisture content), spontaneous ignition occurred. In reality, these
experiments belong to the discussion on ignition, but since they were
performed in the absence of a pilot flame like the rest of the
decomposition experiments, discussion of these result is presented here.

The general trend of the measured species in the three cases is very
similar (see Figures 4.14 and 4.15). At time of ignition there is a
sudden increase to a sharp peak, then fall-off, and finally attainment
of a steady state value. The sharp peak is caused by the large amount
of fuel available in the gas phase at time of ignition. From visual
observation, the flame then reduces in size, because it is now fed by
the fuel evolved from the solid phase. As noticed previously in the
decomposition experiments, the fuel produced attained a steady state

value even with the char formation. This fuel with a constant value of

97

 

 

 

 

 

 

120
110’ ‘
100— ‘ .
/ mc
90— {I -1
A. // CO
5 80- If”, -.
8 ”,1. c.
D -
8 7°” l// / 2
n: gl/
1:. / /
m 50.— ’18 02 .-
c: I ‘/ 1’ A
a: / /
ca 50.. .—
1 [1’ 97 m
/
5 4°” /4 J ‘
c: / / a’
“ /f/ z’
30— // / 2 ’ 52 a 0 '-
/// 2
20 ' IA?/,I “
,W
.. 29 ’ -
10 5*
77
1 2 3

Figure 4.18 Time-Integrated Product Mass and Composition (shown
as percentages next to each bar) as Function of
Incident Heat Flux for 9! Moisture Content Condition.

98

 

 

 

 

 

 

 

120
110 '1 "’
100
’3’ 90 ' T
3
{-1 80 " 0
O
D
c: ‘_,
2 70 - 9 P-
m 4‘ \ \\
m —
o 60 _ 24 \\9~I_-~ ..
3 P‘ ~ ~11--.—
50 - 18 -
j H'- — _l.£ ...-D"
31 -
S 40 - 21 _.
{-1 y’u'
r,
30 " 47 21 ~~ ‘
/
20 — / -1
/
I
10 b 29 61 -’
Dry 9: 17:

Figure 4.19 Time-Inetegrated Product Mass and Composition
(shown as percentages next to each bar) as Function
of Moisture Content at 2 "[032.

99

 

 

 

 

 

120
1101. d
1110
100 - ’1 «
Ill 00
a 90 " C’ / "
. I: ,1 CO2
g; 80 ' , I] (>/ '
§70_ {ll/1602 ..
/
$3 / fl ’ 27 TAR
'3} l’
E1 /418 ,rl
g 40 - / , 1'
[ll/I
30 (I, 35/
4’ ’
.. I / ..
20 I” ,
/
27 / 5
10 " "
82
1 *2 3

HEAT FLUX (W/cmz)

Figure 4.20 Time-Integrated Product Mass and Composition (shown
as percentages next to each bar) as Function of
Incident Heat Flux for 17% Moisture Content Condition.

100

 

 

 

 

 

 

 

120
110- x‘mc -
100- III 7} .
... ’T I I
.v 6” life
90" 7 ’a“ /
I" 13 ,’ 12 co2 .7
sur- ,
13 18 // 15 02
70 ...--- / -‘
~ /
25 18 //
60 '1 / ‘ 27 TAR .1
*“‘“f
501’ 59 27 I 1
/
40" ,r’ -
, 54 1120
30 - // .
/ 52
20 b 1’ -
10 ' 33 '-
Dry 9: 172

MOISTURE CONTENT

Figure 4.21 Time-Integrated Product Mass and Composition
(shown as percentages next to each bar) as Function
of Moisture Content at A Heat Flux of 3 W/cm’i

101

the oxygen concentration in ambient air produces nearly constant values

of CO2 and H20.

The rate of water release in all three cases is very close. This
because the adsorbed water has been evolved prior to the time of
ignition and the combustion water is much greater than adsorbed water.

The C02 production was larger for the dry case, because more fuel

evolved in this case is expected to be more. Oxygen depletion was
larger for the dry case, which was needed to produce a large amount of

002 and H20.

The time of ignition for the dry, 9%, and 17% cases was 88, 130, and
158 seconds respectively. Thus the presence of moisture increase the
time required for spontaneous ignition by diluting and delaying the
production of fuel. The average surface temperature at spontaneous
ignition was approximately 530°C. This temperature is expected to be

[lower for piloted ignition, which is the subject of the next chapter.

Chapter 5
Piloted Ignition

In the previous chapter, it was shown that the presence of moisture
in the wood matrix delayed the decomposition process and diluted the
.decomposition products. These factors are expected to have a
significant effect on the ignition process. The work presented in this
chapter attempts to experimentally quantify the effect of adsorbed
water on the piloted ignition of wood. The experiments were performed
for four moisture contents (dry, 11%, 17% and 27%) and at various

incident heat fluxes ranging from 1.8 to 3.7 W/cmz.

This chapter is divided into four major sections. Section 5.1
describes the physical aspects of the problem, factors influencing the
ignition process and the relation of moisture to ignition. Section 5.2
reviews the previous literature. Section 5.3 describes the
experimental apparatus and procedure. Section 5.4 describes the
results of the experiments; a model to correlate the ignition data is

also derived and discussed.
5.1 Background

Ignition refers to the appearance of flame in the volatile gas

stream evolved from a solid exposed to external heating (usually

102

103

radiative). Depending upon whether the ignition occurs with or without
the aid of an external ignition source, the result is accordingly
classified as spontaneous (auto-) or piloted (forced) [Welker (1970)].
For piloted ignition the actual process, as noted in Chapter 1, is
quite complicated. The solid must first decompose to produce
combustible gases. The combustible gases then mix with the surrounding
air to produce a mixture within the flammability limits of composition.
The mixture needs an external source of heat, such as a pilot flame, to
initiate combustion. Spontaneous ignition will occur only when two
conditions are satisfied at some point in the proximity of the exposed
surface. First the air-fuel mixture must attain flammability limits,
and second, this mixture must achieve a thermal condition that enables
it to automatically react in an exothermic manner to yield a flame
without the aid of any external source of heat. It is clear that the
attainment of spontaneous ignition is more difficult than piloted
ignition.

The factors influencing ignition of a sample exposed to a radiant
heat flux may be separated into two general categories namely, those
external and internal to the sample. Factors external to the sample
are environmental variables such as temperature, composition and
velocity of the surrounding gases, and the magnitude and spectral
quantity of the external radiant flux. Factors internal to the sample
are its thermophysical and thermochemical properties, its moisture
content, and whether the sample is thermally thick or with negligble
temperature gradient (thin materials). Experimentally, except for the

sample properties, all other variables can be controlled.

104

A common feature of cellulosic materials is that they are
hygroscopic and porous. Thus, a large area is available for them to
adsorb moisture from the surrounding air. Moisture content of these
materials is variable, and is influenced by the climate, season and
location. Normally, wood at room temperature and humidity contains
about 10 to 15% water by weight, and may contain up to 30% (fiber
saturation point) if it is in equilibrium with air saturated by water.
As discussed in Chapter 4, this water is held by hydrogen bonds to the
cell walls, and is called agggxhgg_fiatg1. Its presence in the wood
matrix is expected to have a significant effect on the ignition
process. The obvious effect is to increase the time required for
ignition by changing the heat transfer and hence the rate of _
temperature rise. This is because the presence of moisture increases
the values of the thermal properties such as the thermal conductivity
and the volumetric specific heat. Furthermore, since the specific heat
of water vapor is twice that of nitrogen, additional fuel is required

to attain the same limit flame temperature.
5.2 Previous Literature

The ignition of cellulosic material (both auto- and piloted) has
been an active area of research, and several excellent reviews have
been published on the subject [Welker (1970), Kanury (1972), and
Steward (1974)] .

Numerous techniques have been developed to investigate the ignition
phenomenon. Various heat sources has been used, namely gas-fired

radiant panels [Simms and Co-workers (1960,1963,l967)], high tungsten

105

filament lamps [Smith et al. (1970)], carbon arc [Martin (1969)], 002

laser [Kashiwagi (1981)], electrical heating elements [Atreya (1983)],
etc. The experiment essentially constituts of exposing a sample to a
known external radiant flux and recording the timee required to ignite
the sample in the presence of a pilot flame. The ignition times
measured by various investigators have not always been in good
agreement. For example, the time for ignition of wood irradiated by a
tungsten lamp is about four times larger than by diffusion flames at
the same incident irradiance [Wesson et a1. (1971)]. The primary
reason for this disagreement was found to be due to the difference in
the spectral quality of the incident radiation relative to the spectral
absorptance of wood. Thus the spectral nature of the heat source to be
used in the experiments must closely match the intended application.
For building fires this implies that the heat source should have an
effective blackbody radiation temperature of about 1200’K.

Several ignition criteria have been proposed such as: critical
surface temperature at ignition [Simms (1963)], critical mass flux
[Bamford et a1. (1946)], critical char depth [Sauer and Interim
(1956)], critical mean solid temperature [Martin (1965)], etc. Of
these, critical fuel mass flux at ignition seems to be physically the
most correct since it can be related to flammability limits. Surface
temperature has proved to be the most useful ignition criterion since
it can be conveniently related to fire spread [Atreya (1983)].
However, only a few investigators [Atreya (1983)], Kashiwagi (1981),
Garden (1953)] have directly measured the surface temperature at

ignition; others have estimated it by extrapolating measured indepth

106-

temperatures [Martin (1965)], or by using a linear heat conduction
theory [Simms (l960,l963,l967)]. The reported surface temperature at
piloted ignition for wood ranges from 300-540'0. Atreya's recent work
(1983) has shown that with the correct interpretation of measured
surface temperatures a somewhat narrower range [375 i 30°C] is
obtained. In this regard, it is important to note that all of the
above criteria are based on an indirect quantity which is assumed to be
closely related to ignition.

The importance of moisture content in relation to fire tests has
long been recognized. Researchers have carefully controlled the
moisture content of their samples. However, few investigators have
addressed its effect on ignition. Martin et a1. (1958), studied auto-
ignition of a-cellulose conditioned at different relative humidities
using carbon arc as a radiant source. Simms and Law (1967) performed
both auto- and piloted ignition tests on samples with different
moisture contents using gas-fired radiant panels. In both these
studies moisture was found to significantly affect the ignition
process.

Atreya (1983) investigated (experimentally) piloted ignition for
different types of wood. He showed that critical surface temperature
is a reasonable and convenient criterion for ignition. He indicates
that this temperature increases for low heat flux experiments due to
the build up of char. Later Atreya et a1. (1987) investigated the
effect of sample orientation on piloted ignition and opposed-wind flame_
spread. They used two types of wood (red oak and mahogany) to
investigate two orientations (horizontal and vertical). They reduced

the experimental data according to the thermal flame spread theory

107

using the measured surface temperature. They indicated that as long as
the temperatures are defined consistently with the thermal theory, the
results are orientation independent within the measurement error.

The mathematical description of the heat transfer for the ignition
problem has been widely investigated. The problem is usually
considered to be one dimensional with all heat transfer occurring
normal to the sample surface. The cases investigated produce
correlations of different forms, because different boundary conditions
and various assumptions were used in each case.

Butler et a1. (1956), assumed that the sample is inert, undergoes
no decomposition, front surface is opaque to the incident radiation, no
heat lossess, and the rear surface is insulated. They used the
solution of this problem as the basis for correlating their ignition
data. They omitted the surface temperature rise at ignition because of
the difficulties they encountered in attaching physical significance to
it. They also indicated that the temperature rise at ignition was
relatively constant. Thus, they claimed no significance error was
introduced by omitting it.

Simms and Law (1967), included convective cooling losses at the
front surface in the boundary condition. However, in order to
correlate their ignition data for wet and dry woods, they had to fix
two unmeasured parameters, the convective heat transfer coefficient (h)

and surface temperature at ignition (Ts). For piloted ignition they
used Ts - 360°C and h - 8.6x10'4 cal/cm2.sec°C; for spontaneous

ignition they used Ts - 525°C and h - 14x10.4 cal/cmz. sec°C. The

108

convective coefficients used in their work are larger than those used
by other investigators. For example, Alvares et a1. (1969) used h -

4 cal/cm2.sec°C based on free convection heat transfer theory.

2.8 x 10'
In the opinion of Welker (1970), the reason for a large value of h
(required to obtain a correlation) in Simms work may be that the heat
losses by reflection of incident radiation and emission of radiation
from the surface were ignored.

Atreya and Wichman (1987) developed an approximate analytical model
for the piloted ignition process. They solved for time, surface
temperature and mass flux at ignition, and showed that the predictions
were in good agreement with the experimental measurements. Tzeng et
a1. (1988) developed a theoretical model for the piloted ignition
process. They solved the equations numerically, and examined the
location of the ignition source, fuel mass evolution rate from the ‘
surface, and the surface temperature of the solid. They showed that
the (1) model adequately explains the pre-ignition flashes that-are
often observed experimentally, (ii) provides a rational criterion for
positioning the pilot flame, (iii) indicates that the heat losses to
the surface play an important role, and shows that the fuel flow rate
by itself is insufficient for predicting the onset of piloted ignition.

Recently, Mikkola and Wichman (1988) suggested a correlation that
relates the incident heat flux (F) to the negative one-half power of
ignition time (t). They deduced this relation from the solution of the
one-dimensional heat conduction equation for a semi-infinite solid with
constant external heat flux, linearized surface heat loss, and a

constant initial sample temperature. They used wood ignition data

109

0.5

available in the literature and plotted t' vs F, and showed that the

data collapsed onto separate straight lines whose common abscissal

intercept is approximately constant (12.5 KW/mz). They explained
this constancy for the different types of wood by the similarity in the

chemical structure of wood. They indicated that this value (12.5

W/mz) is the minimum heat flux for ignition, and provided additional
support for their correlation, by showing that the data for
polypropylene and polyurethane, have different slopes and intercepts.
They explained that this is because their chemical structure are

different from wood.
5.3 Apparatus and Procedure

A schematic of the experimental arrangement used for these tests is
shown in Figure 5.1. Since the experimental apparatus used is
described in detail in Chapter 2, only the relevant aspects are
considered here. The external radiation was provided by electrical
radiant heaters. These heaters have a maximum backbody temperature of
1200°K and are therefore well suited to simulate external radiation in
building fires. A small natural-gas flame was located in the mixing
layer, pointing downward. The pilot flame was off the sample edge at a
height of 0.5" from the sample surface (see Figure 5.1). The
configuration and the position of the pilot were chosen in order to
avoid heating the sample.

The samples were prepared and instrumented with thermocouples. The

samples were then dried by a standard procedure in a temperature-

110

onoooum nowuasmu can no owuqaonom _~.m enough

 

 

Mooqm mummoo afigooolmu9<§

zH momzmm Rank H<mm

a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m<o
A<MDH<Z

«348 \ \
65,5335 30:32 \ \
. \\ \ é
\ \
\
\\
..sx . ..Hé \
. \1 x a gamma. J]\
\111 .kV\
‘
as 932 \\ 1 \\
\.\ 111
.\ . 1111
x x \ :3” 3.8an
\.\.\ \\\
\\\\
\
\

ZOHH<Hn<m A<2MHHNM

 

 

111

controlled chamber at 105°C until a constant weight was achieved. Then
they were conditioned in a temperature-humidity-controlled chamber for
5-6 weeks until a constant weight was obtained. This procedure of
fixing the moisture content in all samples used for a particular set of
experiments minimized the changes in the results due to variations in
moisture content. The sample edges and the bottom were wrapped with
aluminum foil to minimize any edge effects.

All the experiments were performed under identical conditions of
air flow rate and atmospheric pressure above the sample. This is
because for piloted ignition the mixture of the products of
decomposition generated by the solid, and the surrounding air, should
be within flammability limits. The flow rate of air above the sample
may considerably affect the results.

The procedure consisted of placing the instrumented sample flush
with the Kaowool insulating boards on a load cell to continuously
monitor its weight loss. The sample was then exposed to a known
external radiation. This external radiation was also continuously
monitored by a heat flux sensor located upstream the sample (see Figure
5.1). A calibration between the heat flux at the sensor location and
the heat flux at the sample location was used to determine the incident
radiation on the sample surface. A video camera was focused on the
sample, and recorded the entire experiment; this provided information
about ignition behavior and a precise measurement of ignition time.

For a given moisture content and external radiation, in addition to
the time required for ignition, the following measurements were also
recorded as a function of time: (1) front and back surface

temperatures, (ii) mass evolution rate of water vapor, (iii) production

112

rate of combustible gases, and (iv) weight loss rate of the wood

sample.

5.4 Results and Discussion

This section describes the results obtained from the piloted
ignition experiments. These experiments were parameterized with the
sample moisture content, and the external radiant flux. Thus, except
for these two parameters, all the experiments were performed under

identical conditions.

5.4.1 Ignition delay time

Ignition delay times were measured for the samples with different
moisture contents under various external radiant fluxes. This value
was provided from recorded tapes for all the experiments. A summary of
all the ignition data is shown in Figure 5.2. This figure shows that
the ignition time is a function of moisture content and the incident
heat flux. Generally as incident heat flux decreases, the time
required for ignition increases, and as the moisture content increases,

this time increases at the same incident heat flux. For example, a dry

sample ignited at 1.5 minutes at 1.9 W/cmz, whereas a sample with 27%
moisture content ignited at 16.5 minutes, for the same incident heat
flux. Thus, the presence of moisture in a sample would increase the
total energy needed to ignite the sample by increasing its ignition
time.

The higher heat flux data (more useful for fire spread purposes)

113

.oaae hogan soaufismu so ousueaox no uooumm ~.n enough

3 “.2: >53 29:29

—bb-—-b—.-— —--L-P—Lnb—bnn—~p-Lbn...-

09: 00m 00m DON. com com on:~ com com 00? o

 

PB“
V‘s—N

 

2::
*xxx

0:
O

 

 

11'1'I‘IU'IIIIIIIUIIITT'

 

m.—

0.
N

“2
(\l
(awe/M) Xn'ld 1V3H lNBGIONI

0.
1’)

O.
<-

114

shows also a large difference in ignition time. For example, a dry

sample ignited at 24 seconds at 3.5 W/cmz, whereas a sample with 27%
moisture content ignited at 90 seconds at the same incident heat flux,
suggesting that fire spread on wood containing moisture would be slower

than on dry wood.

5.4.2 Correlation of results

Ignition time data were correlated on the basis of a relationship
suggested by the solution of the one-dimensional heat diffusion
equation in the solid.

Consider a semi-infinite solid initially at a constant temperature

(To). At t > 0 the solid is exposed to a prescribed heat flux (F).

Assuming that the solid is inert and undergoes no thermal
decomposition. Since the heat transfer through an inert solid is by
conduction only, the following form of the one-dimensional heat

diffusion equation applies

2 2
pcg—f-xgfi- forx>0,t>0 (5.1)

where 0 - T - T”. The solid is assumed to be at the temperature of its

surroundings prior to the experiment, therefore

0 - 0 for t s 0, x > O
The boundary condition is obtained from the energy balance at the solid
surface (i.e., heat diffused into the solid - incident heat flux -
convective heat loss - radiative heat loss). Thus, the boundary

condition at the front surface becomes

115

a; _ _ _ a - a
- K 6x x_0- £8 (as, T0, t) e [P h as a((as + To) Tm)] (5.2)

where e - the emissivity (or absorptivity) of the surface,

5‘
I

heat transfer coefficient/e,

Stefan-Boltzmann constant,

e
I

F - the prescribed incident heat flux,

0 - Ts - Tdo where T8 is the surface temperature,
f - heat flux going into the semi-infinite solid at the front

surface,
and p,C, and K are, respectively the density, heat capacity, and

thermal conductivity of the solid.

Atreya (1983), obtained an analytical solution for the above problem by
the use of approximate integral methods. He solved for the ignition

time and gave it in the following form;

 

2
o (2so +r - Jfi)(r + J?) o (r + 230 )
t - M[-§§ + -§7§ 2n[ § ]- —§———1;—-JL], (5.3)
2fs p (25108 + r + /§)(r - J?) 5 £8
where M - g 2E% , r - -(h + 4 a T3), 5 - -2% a T: ,

2 AF d f* fig F o + 02
fl - (r - 3). an s - e - + r s s s

116

This equation was simplified by approximating the logarithm term, and

substituting the value of f: as the net heat flux diffused into the

solid [i.e. f: - (F - L)], where L, is the total heat loss by

convection and radiation. The final form of equation 5.3 is

 

2 .
- H 0 ]Q [J - ICEJZ . 2(r/0 + 23) (l-L/F)
t 2 -f5- 1 + 2 2 ---§* 2 (5.4)
fs 16s 08(1 - r /4Fs) 43(1 - r /4Fs)

The terms inside the bracket of equation 5.4, were examined closely,

and they were evaluated for a wide range of heat fluxes (P). Let

19 {J-IIE12 , 2(r/0 + 23) (1 . L/F)

e , e -

1 165 08(1 - r2/4Fs)2 2 as(1 - r2/4 Fs)
e3 - e1 + e2 and ea - e3+ 1. Figure 5.3 shows the value of each term
plotted against the incident heat flux. Note that at piloted ignition

the surface temperature as is approximately constant. Thus, the only

variable is F. For the range of heat fluxes studied (1.8 - 6 W/cm2
which are representative for igntion), two conclusions may be drawn

from Figure 5.3: (i) the variation in each of the terms e1 and e2 is

small compared to the dominant term which has a value of 1.0 (ii) the

value of all the terms in the bracket (ea) can be approximated by a

constant with an error of less than i 6% in the range of interest.

117

.xsam use: ucovaocH n> q.n coauosvm mo menus n.n shaman

ANEQE x3... .5: 5.362
m m. a. n N _

 

[Ll- - u — p p p H'- . . — p p p _ p p - n.0'
O O O ..
O OHoO O O O O m

mod

.wnd

. med

0 I on... o o M e m .

O O No 9 0 mum.O
my aw Aw mv aw .0 m

WNJ

.. mm;
.qo -- n
I ..

I I I I I I I ......m...

 

 

 

Pd

‘17'9 NOLLVOOB :10 SWHHl

118

By substituting the value of M given under Eq. (5.3), equation

(5.4) may be written as

64 pCK 0:
t um: . . (5.5)

Assuming that the front surface temperature at the time of ignition is
constant, so that it is no longer a correlating parameter, the heat

flux into the solid at the instant of ignition is given by

f -c*c - , (5.6)

/ pCK
where, C* - J 43:2

 

08 . (5.7)

Substitute the value-of f: back into equation 5.2, the following

relation between incident heat flux and time for ignition is obtained

F - 0* c’0'5 + L , (5.8)

where C* is given by equation 5.7, and L represents the heat loss by

convection and radiation. L, is given by the following equation

4

L - h as + 0((0s + I”)4 - Tm) . (5.9)

119

Equation 5.8 is the basic relationship for correlating the data of
ignition time as a function of incident heat flux. The ignition time
to the negative half power was plotted versus incident heat flux. The
data for the four moisture contents are shown in Figure 5.4; the
measured values were shown by points and the least square fit by the
straight lines. As shown in the figure each moisture data collapsed

onto a separate straight line, the intercept corresponds to L, and the
slope corresponds to 0* in equation 5.7.

*
The parameters L and C , were calculated for each value of the

moisture contents; L was determined from equation 5.9. For 08, the
value of the measured surface temperature at ignition was used. The

value of 6* was calculated from equation 5.7. The values for the
thermophysical properties (pCK) were obtained from the literature
[Parker (1988), Simms (1967)]. The oven-dry properties were given in
Parker (1988), who used the same type of wood. The properties for the
rest of the moisture contents were determined by using the formulas of

Simms (1967). A value of 1.75 was used for ca. This is the average
value of all the terms in the bracket of equation 5.4 from Figure 5.3.

A comparison of the derived and the calculated values of L and C*
are presented in Table 5.1. Because surface temperature at ignition
was found to increase with moisture content (this will be discussed in
the next section) it would be expected that the heat loss increases.
The derived and the calculated values of L show an increase, as the
moisture content increases. The reason for a low value in the

experimental heat loss (L) at 27% moisture content is not known to the

120

 

.osua mud-n coauacuu uo soda-Houses <.n enema»

reabsv x3“. :m: 5.862.
m.» .onn. .mum.

cum 0.?

n p b p b L n

n n b n

\ .11.. 2 R nu

ma \ .1... 2 N t
\ 111 2 N Z
.11. #3

 

C) 0| ID II

 

 

(so-S) 31111 AV'EICI NOLLINOI

121

author; more data is needed to confirm this value. Because the

thermophysical properties of wood increase as moisture content
*
increase, it is expected that the value of C increases. The

experimental values of 0* have an average deviation of 10% from the
calculated values; this may be because of the error introduced by
approximating all terms in the bracket of equation 5.4 by a constant.
Such an approximation is expected to introduce ~ 12% error as discussed
earlier in this section. The sixth column in Table 5.1. was obtained
using a computer program (NLINA), provided by Professor J. Beck (1977).
This program can be used to estimate parameters for both linear and

non-linear cases. The calculated heat loss values along with the

ignition data were used to estimate 0*.

The ignition correlation reveals the strong influence of the heat
loss and moisture content upon ignition characteristics. It shows
that, as the time required for ignition tends to infinity, the heat
loss tends to the incident heat flux (see equation 5.8). Hence the
critical incident heat flux, below which ignition is impossible, is
given by L. This critical flux is, in effect, the rate at which heat
is lost from the surface at ignition temperature. Because surface
temperature has been found to increase with moisture content, it would
be expected that the critical incident heat flux would increase (see
Table 5.1). Thus the moisture has a significant effect on piloted
ignition; it increases the ignition delay time and more importantly
increases the critical incident heat flux.

Piloted ignition is closely related to fire spread. In fire

spread, the volatiles produced from the heated surface are ignited

122

either by the source of the heat itself or by the aid of external

sources .

Thus, fire spread may be regarded as a rapid succession of

piloted ignitions.

piloted ignition is a very important concept for fire spread.

This suggests that the critical heat flux for

Using

this value in building regulations will result in a safe separation

distances between housings, and this will reduce the risk of fire

spread to neighboring properties.

TABLE 5.1

*
Comparison of Calculated and Experimental Values of C and L

 

 

 

 

 

*
C
Moisture Experimental Using
Content Calculated Experi- Calculated Experi- Calculated Heat Loss
mental mental

Dry 1.01 1.13 8.6 9.78 10.73

11%. 1.05 1.20 13.2 15.53 14.84

17% 1.10 1.32 15.54 13.3 15.02

27% '1.30 1.14 20.86 22.85 20.20

 

 

 

 

 

 

5.4.3 Surface temperature

Surface temperature profiles for dry

and 11% moisture content at

2,3, and 4 W/cm2 are shown in Figure 5.5 and 5.6 respectively. These

Figures show that the rate of rise of surface temperature increases as

the incident heat flux increases.

Also the rate of rise increases as

moisture content decreases at the same incident heat flux as shown in

Figure 5.7.

123

At the time of ignition the flame appears in the gas phase, and this
causes a sudden increase in the surface temperature. The arrows in
Figures 5.5-5.7 indicate the time of ignition. These figures show that
ignition time increases as heat flux decreases (see Figures 5.5 and
5.6), and increases as moisture content increases at the same incident
heat flux (see Figure 5.7). Data for surface temperature at ignition
are summarized in Figure 5.8. There is some scatter in the data, but
the trends are clear. At any incident heat flux, the surface
temperature at ignition is higher for higher moisture content, and at
any moisture content, it increases as heat flux decreases. This
increase in ignition temperature is due to the slow decomposition of the
solid at the surface, and the resulting build up of the low thermal
conductivity char layer prior to ignition.

The sample moisture content also had a significant effect on surface
temperature at auto-ignition. In the decomposition experiments

discussed in Chapter 4, it was pointed out the auto-ignition occurs at 4

W/cm2 for the 3 moisture cases studied. The surface temperature of the
sample at ignition in these experiments were 510, 525, 550°C for dry,
9%, and 17% moisture content, respectively. These temperatures are much
higher than the piloted ignition temperatures under the same conditions.
This is because in the case of piloted ignition, the sample surface
temperature must rise only high enough to decompose the sample, provided
that decomposition is rapid enough to produce enough combustible gases
to form a combustible mixture with the surrounding air at the pilot
location. 0n the other hand auto-ignition, must be initiated by heating

either the sample surface to the point where it ignites the gases, or by

' 124

cowuavcoo man you ouauauoeaoa oueuusm so seam nee: acovaocu no uoouum n.n enough

3 m2:

000 005 000 000 00¢ [000 00m 009

uh. uh-

- n - —[Pf-|n - p

0

 

 

as: ZOHHHZUH

«an? 8 A

2

mafia. ZOHHHZGH

h

  

. _ o

h 8.
com
8...
8...
com
com
8s
8m
8m

 

-000?

(0.) BaniVHBdWElJ. 3011:1808

coauuvcoo acouaoo assuage: «as
new ensueuoaaos ooeuusm so seam use: ucoeqocu mo uooumu o.n «woman

3 m2:
com com 00¢ com com cos 0

 

125

 

b . . p L . _ . . . L [p . . _ p . p — . . b —. o
.. 00—.
BAHH. ZOHHHZUH h CON

h 00m.

~83 2s 00¢
«so; a; N5; 0 00

3sz 35:22 h

000
005
000
000

(0.) BaanHI-IdWEIl 30113803

 

- 000_.

126

«so; was Co 5.: use 0532:

 

 

me an ensueuoeaoa ooeuusm so acousoo assuage: mo uoomuu ~.n «woman
3 m2: ..
com com 00. o
n b n — n b — n p. — P L b
\

 

 

 

00 F
00m
000
00¢
000
000
00k.

- 000

(0.) 38101101381131 Bowans

127

0.¢

 

 

.uoquusmn ue eunuch—oases 3356 so acousoo assuage: 00 33.4.0 0.¢. can»:
Anemia x3: 2m: .2862.
an o.» 2 em as
b n b p b L p b L b n r h p — r b b L] b L p b P . .I 00”
won».
on .I l .....o wot».
Po . oo / / 0 m8...
1.. lllllll 0 ..
am.-----:: - -- ..-mn.J, I / womn
e -..?me / ...
III I [I/ WON¢
« woee
z e R .. .....I s
2 N t o .1... woo¢
2 a : . ...... -
..o 1.
>8 0 .ll m m
1 com.

 

 

 

(0.) NOLLINSI 1v BHMVHEIdWZ-ll 3011:1808

128

.Nao\3 o.“ u. ueoueoo assuage: «am new ooauouaqm omusueuoaaos oddeem o.n ouemam

3 m2:
com com com com com 00¢ com com 00, o

uh. DLf—-D.—bL- :b-b-nb D-PL-pP—D-Lo

 

111 208800

00 _.
DON
BE 5353
* con
Saga 00¢
com
000

00m.

 

 

 

- 000

(0.) 3801118381131 3181111115

129

heating the gases themselves to the ignition point. In either case,
this would require a higher surface temperature.

Sample surface and bottom temperatures during a typical test of
piloted ignition are shown in Figure 5.9. This Figure shows that the
bottom temperature increased only by 2-3°C before the time of ignition.
Thus the assumption that,the samples were thermally thick used in

section 5.4.2 is valid.

5.4.4 Sample Iass flux

Mass flux histories for dry and 17% moisture content are shown in
Figures 5.10 and 5.11 respectively. The arrows in these Figures
indicate the time of ignition. These Figures show that the mass flux
increased dramatically at the time of ignition and continue to increase.
Meanwhile a char layer on the sample surface forms. This char layer is
not completely inert but it is much more resistant to the evolved gases
than original wood. This is because char has lower thermal
conductivity, and the heat penetrating through this thick char would
decrease. This consequently decreases the volatile mass flux, which
continues to decrease until attaining a nearly steady state condition.

Numerical values of the mass flux at the time of ignition for the
different moisture contents at various incident heat fluxes were
determined from the mass flux histories. These values are presented in
Table 5.2. This table shows that the mass flux increases as the
incident heat flux increases, and at the same incident heat flux, it
increases as moisture content increases. This is because the initial
mass evolution from the samples is essentially desorption of water

followed by a large amount of tar. These quantities are evolved at

130

soauavsoo an: new
asap one: uoovcoaonuoaua so seam aeom ocueaosm mo uoouum o~.n enough

3 m2:

000. .muoxu 000 000 00¢ 000 00m 00w 0

b b p b n b n n n n n n b n -

MZHH :onHHZUH

a

 

      

 

‘1‘ .

 

mmzHa ZOHHHZUH

ao\3 0.N

 

N5; 34

 

 

 

       

 

 

(S ZwO/fiw) x013 ssvw

131

souuuvcoo ousuuwoz aha
you Mean and: unaccononloaua no Kean ueom.u=ov«ouu mo uouuum

3 m2:
000— com com com com com oov .omn

b.» —:—_-—- ——-—_—__~—— P

0

-CD

~«.n shaman

 

HZHH ZOHHHZUH—

       
   

N502» cm .H

Nau\3 H~.m

 

 

 

(s awe/5w) xmj ssvw

132

higher rates as moisture content and incident heat flux increases.

Table 5.2 shows that a mass flux of about 0.22 mg/cmzs is necessary for
piloted ignition to occur. Bamford et a1. (1945) proposed that piloted
igntion of vertical slab of wood is possible when the mass flux of the
decomposition products exceeds a critical value.I For deal, oak, and

pine wood, they gave this critical mass flux as approximatley 0.25
mg/cm2.s.

Table 5 . 2

Comparison of Mass Flux at Ignition For Different Moisture Content
At Various Heat Fluxes (F)

 

Moisture F I Mass Flux F I Mass Flux F I Mass Flux
Content W/cm2 I mg/cmzs S/cm2 I mg/cmzs W/cm2 I Mg/cmzs
| I I
Dry 1.85 I 0.22 2.6 I 0.28 3.48 I 0.48
11% 1.80 I 0.28 2.76 I 0.37 3.53 I 0.50
17% 1.84 0.30 2.36 I 0.44 3.21 0.42
27% 1.92 0.37 2.77 0.47 3.62 0.58

 

 

 

5.4.5 Products evolved

In this study oxygen depletion, production of C02, CO, total

hydrocarbons [THC], and water were measured before and after the time of
ignition. The procedure used for the analysis of these products is the

same as used for decomposition experiments, as described in section 4.3.
The data is presented as mass flux of permanent gases, and as percent of
mass flux. Plots for different moisture contents at various heat fluxes

are shown in Figures 5.12 - 5.21.

133

 

 

 

 

 

 

 

/-\

OD

N

E

(J

‘\t

(75

E

\_I

X

I33.

(I)

g 0.0-£1— fi-,--.,.‘-~~.-’
-O.2 UII'III'IIIIUITIIUI"

U 100~ 200 300 400 500
T1ME.(s)
. Figure 5.12 Mass Flux of Evolved Products at Incident Beet

PRODUCT AS PERCENT OF MASS FLUX

Figure 5.13

Flux 3.5 Vie-2 for Dry Condition

 

 

 

 

 

 

‘U U U U I I I I I U

200
TIME (3)

Products as Psrcent of Mass Flux at Incident Best

I I V l

300 400

500

Flux 3.5 H/cm for Dry Condition

PRODUCT AS PERCENT OF MASS FLUX

MASS FLUX (mg/cm“ 3)

134

 

 

 

 

: , ‘“~. CO A’ '
O 4- M’
° 0
I , o2 narrator:
0.2‘3 I ‘ s \e. TAR
O O- 3,... ' co and mc ~ . . , . -
. I I I2 I I I l I I v I ~ ‘
0 "JV 100 300 400 500 '

OOTIME (s)

Figure 5.14 Mass Flux of Evolved Products at Incident

Beat Flux 2. 6 “Ion2 for Dry Condition

 

 

 

 

 

Figure 5.15

 

Products as Percent of Mass Flux at Incident
Beat Flux 2.6 W/cm2 for dry Condition

135

 

'5 P- :83“
i}:

I“ 30

O
G
I

\ . .0
‘ ~\.}--“'- '

02 D PLETION

MASS FLUX (mg/cm2 3)

 

 

TIME (s)

Figure 5.16 Mass Flux of olved Products at Incident Beet
Flux 2.75 W/cI for 112 Moisture Content Condition

 

1 00.0

 

 

 

PRODUCT AS PERCENT OF MASS FLUX

TIME (s)

Figure 5.17 Proudcts as Percent of Mass Flux at Incident
Heat Flux 2.75 H/cm for 112 Moisture Content Condition

136

 

 
   
 

 

 

 

 

 

 

1.2 .

’“~ I

m 1.0-:-

N

5 0.8-3

E 2

E 0.6-; “2°

V - \k
. 0 DEPLETIO

x .. 2

0 0.4:

I: 0.2:

00 I

g 00-: ‘ * ' .mc, .

-O.2. I I OI I 3 OI—I I '1

0 100 200 300 600

TIME (s)

Figure 5.18 Mess Flux of Exolved Products at Incident Heat
Flux 2. 77 His. for 271 Moisture Content Condition

 

 

 

 

 

PRODUCT AS PERCENT OF MASS FLUX

-2000 T I I I I I F! I t i I I I I I I I 1'1
0 1 00 200 300 400 500 600
TIME (s)

Figure 5.19 Products as Percent of Mass Flux at Incident Beat
Flux 2. 77 chmz for 272 Moisture Content Condition

MASS FLUX (mg/cm2 3)

137

 

 

 

 

 

 

1.0 .
0.6-3 kl
I ‘\*“-I
0.4”: ‘I 320
0.2; '- o n
. . II!
0.0-3 - """""”'"’."I' ,°° “‘1 m
"092 - I I l I I I I I I I I I I I 1 I l I I I '1
100 200 300 400 500 6 700
TIME (s)

Figure 5.20

MASS FLUX (mg/cm” 3)

Mass Flux of Evolved Products at Incident Beet Flux
1.85 via-2 for Dry Condition

 

 

 

1120
02 DEPLETION
"Vs

“\Joz ‘\

s v’

A L
' .I "‘ ‘ ‘0.

 

200
TIME (s)

Figure 5.21 Mass Flux of Evolved Products at Incident

Heat Flux 3.21 Vlcm2 for 172 Moisture
Content Condition.

138

Before ignition, the gases evolved are products of decomposition.
This was the subject of discussion of Chapter 4. These Figures show
that water is a major component of the evolved gases. The higher the
incident eat flux and moisture content, the higher is the water
production rate. As pointed out in Chapter 4, the presence of moisture
in the samples delayed and diluted the combustible gases. This is the
reason why ignition occurs at later times for samples with higher
moisture content.

After the time of ignition, the general trend of the measured species
is very similar for all moisture content cases at the various incident
heat fluxes; a sudden increase to a sharp peak, fall-off, and then
attainment of a steady state value. The sharp peak is caused by the
burning of the large amount of fuel available in the gas phase at the
time of ignition. Meanwhile a char layer starts to form which reduces
the rate of the evolved gases including the combustible gases. This
results in a reduction of the product rates, which eventually attain a
steady state.

Figures 5.12 - 5.21 show that the production of water after ignition
increases as the incident heat flux increases, and at the same incident
heat flux, the water production does not change significantly with the
increase in moisture content. This is because the combustion water is
much larger than the adsorbed water, and a large quantity of the
adsorbed water was evolved prior to the time of igntion. This is
supported by the fact that water production prior to ignition is higher
for higher moisture contents. These Figures also show that the

production of 002 decreases as moisture content increases at the same

139

incident heat flux, and decreases as incident heat flux decreases at the
same moisture content. Oxygen depletion follows exactly the same trend

as 002, as expected. This is because, as pointed out in Chapter 4, the

production rate of combustible gases decreases as the heat flux
decreases, and are diluted by the presence of moisture in the sample.
Figures 5.12-5.21 also show that the production of combustible gases is

almost zero, since these gases burn to produce C02 and water.

Figure 5.18 shows that at about 475 sec, as the flame gradually died

out, 002 production and 02 depletion gradually decreased, while the

production of combustible gases (C0 and THC), started to gradually
increase. After the flame died-out completely a steady state was
attained. They may now be considered as decomposition products.

The tar in some experiments shows a negative value (see Figure 5.12)
because it was determined by differences (as discussed in Section 4.3),

and any measurement error will show in the tar value.

CHAPTER 6

CONCLUSIONS

This chapter presents a summary of the important results for
different components of this work. Section 6.1 summarizes the
conclusions derived from the results of the droplet evaporation on the
hot porous and non-porous solids. Section 6.2 summarizes the
conclusions obtained from a set of systematic experiments performed to
investigate the effect of water on both thermal decomposition and
ignition. Some recommendations for future work are presented in

Section 6.3.
6.1 Droplet Evaporation Experilents

The transient surface and indepth temperature data obtained during
droplet evaporation reveal the thermal behavior of both the porous and
the non-porous solids. The following conclusions are apparent from

this investigation:

1. The only similarity in the thermal behavior of both porous and non-
porous solids is the agreement in the experimental contact temperature.
This is because it occurs during the initial contact and the draplet

sees both solids as a semi-infinite body.

140

141

2. The theoretical and experimental contact temperatures are in good
agreement up to the boiling point of water and then diverage with the
experimental contact temperature becoming roughly constant at a value

slightly greater than the boiling point.

3. During the droplet evaporation process, surface and indepth
temperatures for the non-porous solid remain nearly constant whereas
for the porous solid there was a continuous decrease in these

temperatures.

4. A thermocouple in the porous matrix at the same location as that of

the non-porous matrix cools faster under identical conditions.

5. Both, the recovery time and the envelope of droplet influence were
larger for the porous solid. These results confirm the indepth cooling

Iof the porous solid due to water penetration.

6. The evaporation time is longer for the non-porous solid than for
the porous solid for the same droplet diameter and under identical
conditions. This is because part of the droplet diffuses into the

porous solid matrix.

7. Smaller droplets are more efficient for cooling non-porous solid at
or below 100°C. This is because the evaporation process for larger
droplets attains a steady state and the extra fluid in the larger

droplets just increase the evaporation time. This phenomenon was not

142-

observed in the porous solid because evaporation process never quite

attains a steady value.

8. For the non-porous solid, the instantaneous evaporation rate, and
the instantaneous average evaporative heat fluxes were determined from
the transient measurements of surface and indepth solid temperatures,
and the droplet diameter on the solid surface. Although the total heat
transfer is more for large droplets the average evaporative heat flux
is higher for smaller droplets. This is because small droplets have

larger conductance due to their smaller thickness on the hot surface.

6.2 Thermal Decomposition and Piloted Ignition Experhments

The following conclusions are drawn from this investigation:

1. The presence of moisture in the wood sample has a significant
effect on the surface temperature and char oxidation. As the moisture
content increases, the rate of increase of surface temperature
decreases. Thus, the char oxidation occurs at later times for the same

incident heat flux.

2. The presence of moisture in the wood matrix delays the

decomposition process and dilutes the decomposition products.

3. The mass flux at the time of ignition increases as the moisture
content increases. This is because the initial mass evolution from the

sample is essentially desorbed water, and this component evolves at

143

higher rates as the moisture content increases. An absolute minimum

mass flux of about 0.22 mg/cm2.s is necessary for piloted ignition to

occur .

4. The presence of moisture in the wood matrix increases the total
energy needed to ignite the sample by increasing the ignition time. It
also increases the critical incident heat flux by increasing the heat

loss from the sample surface.

5. A single correlation was derived for all ignition data. This
correlation accounts for the moisture dependent thermal properties and

the heat loss from the sample.

6. The surface temperature at ignition increases as moisture content
increases. This is due to slow decomposition of the solid at the
surface and the resulting buildup of the low thermal conductivity char

layer prior to ignition.
6.3 Recommendations for Future work

Based on this research, it seems that the following areas need

further investigation in the future.

1. A study of transient cooling of the porous solid by droplet
evaporation of fluids other than pure water. This will hopefully yield

the optimum cooling agent.

144

2. A study of the effect of droplet release height on the cooling
process will yield the optimum strategy for delivering the coolant

agent.

3. A study of the droplet evaporation of common fuels such as methanol
and benzene at higher temperatures on solids with different porosity.
This may have direct application in the design of combustion chambers

for internal combustion engines.

4. Investigation of the effect of oxygen concentration on the ignition
process. This will allow determination the minimum oxygen

concentration necessary for ignition.

5. Investigation of the effect of air velocity on the ignition
process. Air velocity may significantly affect the ignition process,
for ignition to occur, the mixture of the products of decomposition
generated by the solid and the surrounding air should be within

flammability limits.

6. Develop a theoretical model for thermal decomposition of wood and

compare with the present experimental data.

APPENDIX A

APPENDIX A

METHOD OF ESTIMATING LOCATIONS 0F IN-DEPTH THERMOCOUPLES AND
THERMOPHYSICAL PROPERTIES OF THE CERAMIC

.In this appendix, the procedures used to determine the location of
in-depth thermocouples in the ceramic blocks and the thermophysical
properties of the ceramic are described. Some of the properties were
supplied from the Sandia National Laboratories [Taylor and Groot
(1985)]. However, since the properties of castable materials change
according to the procedure used in casting and drying them, it was
necessary to re-evaluate these properties. A program provided by
Professor J.V. Beck (1977) was used to estimate the thermal conductivity
and the density-specific heat product using transient temperature and

heat flux measurements.
A.1 Inrdepth.Thermocouples Locations

The locations of thermocouples, prior to casting, were measured to
within 2 0.1 mm. Due to the uncertainties introduced during the casting
process and shrinkage, thermocouple locations were evaluated by a least
square fit to the measured steady state temperatures. This criterion
yields a unique "best fit" line for the given set of data. Figure 1A
shows the least square fit to the steady state temperatures. The

evaluated thermocouple locations from the least square fit were close

145

146

to the measured values before casting. Table 1A shows the comparison
of the two values. Attempts to determine thermocouple locations

radiographically by using X-rays were not successful.
A.2 Porosity

Porosity is the fractional void volume of the ceramic. A simple and
straightforward method was used to determine the porosity. Several

samples were dried by a standard procedure in a temperature-controlled

chamber at 105 0C until constant weights were achieved. The samples
were then rinsed in distilled water until all the samples attained
saturated weights. Since there was no change in dimensions of the dry
and the wet samples (i.e, no water adsorption), the difference between
the weights of the saturated and dry samples was used to determine the
amount of water displaced. Then the volume fraction of the cell-wall
substance was calculated. The value of the porosity tabulated in Table

2A is an average of four tests.
A.3 Thermophysical Properties

The thermal conductivity, the product of density and specific heat
were estimated by a technique used by Professor J.V. Beck. Two thin
electrical circular heaters 3" in diameter were placed between two
identical ceramic solid-cylinders. Both solid cylinders were also 3"

in diameter. One of them was instrumented by the thermocouples. Two

147

aluminum solids were placed at both ends in order to obtain constant
temperature boundary conditions.

The solids were allowed to heat up until a steady state condition
was achieved and then an electrical step power input was applied to the
heaters. Finally, the power was turned off. During the experiment
temperature data was collected. The temperature history of the
thermocouples in a typical test is shown in Figure 2A. The time for
the step power input and the time when power was turned off are also
shown. Voltage and current inputs were also measured.

Program PROP, provided by Professor J.V. Beck, was used to estimate
the thermal conductivity and the density-specifitheat product. This
finite difference program solves a transient one-dimensional
homogenous partial differential equation with temperature-dependent
properties. The two boundary conditions are the heat flux at the upper
surface and the constant temperature at the lower surface of the solid.
The heat flux was calculated from the voltage and current measurements
and from the area of the heaters.

Results of the thermophysical properties at four different
temperatures are tabulated in Table 2A. From these results it may be
concluded that the thermophysical properties of the ceramic are not
strongly temperature-dependent. Table 2A also contains the

constituents of the refractory ceramic.

148

TABLE 1A

In-depth Thermocouples Locations

 

W L Warm-.1

0.0 | 0.0
1.3 | 1.69
2.9 | 3.06
5.5 | 6.63

. 6.3 | 7.33
10.5 | 12.27
26.0 | 26.56

(i)

(ii)

(iii)

149

TABLE 2A

Magnesium Oxide Ceramic-Refractory Grade

Constituents or chemical analysis by weight

 

 

_§emneune.. | Seanceitien...
Mg 0 | 96.11%
s1 02 I 2.2%
Fe2 03 0.36%
A12 03 0.19%
Cr2 03 0.06%
Ca 0 1.08%

Porosity 33.28%

Thermophysical Properties

 

 

Tempera- Density Specific Thermal Thermal
sure Heat Condustizisxm__nifussizi£x
L0G) (gm12m31_u_il£gm_.fil (uggm,k) (gmzzggg.)
65 2.053 1.00 0.02283 0.0111

100 2.053 1.076 0.02261 0.0102

150 2.053 1.307 0.02225 0.0083

240 2.053 1.307 0.02225 0.0083

 

 

 

 

 

150

.uousueuodsua uueum museum «:9 ca ugh museum ueuog <~ ouswum
AEEV 8:320
0N 0N mp

bIn—bL-Irb-bb

0p m 0
.L

b -

 

 

n b b p — n n b -
«E 96:00 «moo.- I N¢

20823302 dEo.» x

 

 

 

(0°) mmmsdwei

151

.xsdm use: no uuonuso use
oeeouosu doom < wcuusc mousueuoa509 :udovcu usuaeceus <~ shaman

Aoomv 08:.
on. 8. .2... .om. . 2.... an on 9 cm a

s. —.L p p n L p 0*

 

 

acnhnnzoo 529500 mags—ugh. 2.5200 7 v

 

 

 

..... . o . a 8030“—
9053 muhm

mushy—=0 .l O“

mac Guzman. mus?—

 

 

 

(30) SOJI‘IIOJOdwOIL q1dapug

APPENDIX B

APPENDIX B
CRITICAL NOZZLES

This Appendix presents a description and the calibration procedure
of the critical nozzles. These nozzles were used because they are a
very accurate way of controlling and measuring mass flow rates of gases
over a wide range of flow rates. Critical nozzles offer many advantages
over other measuring devices; they do not have any moving parts which
produce friction or wear [Aschenbrenner (1983)]; no pressure
differential measurement is needed [Arnberg (1962)); simple
construction, and simple flow equation that depends on pressure and
temperature of upstream.‘

The configuration of the nozzles was adapted from [Arnberg [1962]].
The general design is shown in Figure 13. A set of nozzles with

different diameters were manufactured to deliver a flowrate of air

anywhere in the range (175-32,000 ft3/hr). An identical set was also

manufactured to be used for either 02 or N2. A straight section of pipe

approximately 120" was placed in front of the nozzle in order to achieve

fully developed conditions at that location.
8.1 Mass Flowrate

The theoretical mass flowrate of a given nozzle is directly

proportional to upstream pressure and temperature. The static state is

152

153

given in ASME (1959)]. [Holman (1984)] presented the equation in the

following form:

2:.

-_ _1_
m A9. P1 RTl 1+1[(1+1) ' (3'1)

.2.
1'” .,
where m - mass flowrate (lbm/sec),

A* - area of the nozzle (in.2).

P1 - inlet static pressure (Psia),

g - gravitational conversion factor (32.2 lbm ft/lb secz),

R - gas constant (ft lbf/lbm R),

T1 - inlet temperature (°R),

1 - ratio of specific heats of the gas (Cp/Cv)°

B.2 Calibration

A trace gas technique was used for calibration of the nozzles. A
known flowrate of methane (measured by a flowmeter) was-introduced
downstream of the critical nozzle. A sample of this mixture passed
through the gas analyzer, where the percentage of methane was recorded
by the gas chromatograph. The experiment was recorded by the gas
chromatograph. The experiment was repeated several times for different

upstream pressures at the known fixed amount of methane.

154

The experimental gas flowrate was calculated by writing a mass

balance of methane.

_flmetuf_methene_
mole fraction of methane total flowrate of gases . (B.2)

The mole fraction of methane was determined from the chromatograph
reading (A), the flowrate of methane is the measured value of the

flowmeter (Q1), the total flowrate is the flowrate of air and methane

(Qa + Q1). Thus the flowrate of air (Qa) is given by
Q -Q
Q J—J-. (3.3)

The experimental and the theoretical flowrates are compared in
Figure 28. A discharge coefficient of 0.964 is listed in the literature
for a similar critical nozzle design. The discharge coefficient is the
ratio of the experimental to the theoretical value. Figure 23 shows an
error of ~ 4%. The discharge coefficient used here is 0.97. The error
is probably due to the combined error in the measurements of the nozzle

diameter, flowrate of methane, and the reading of the chromatograph.

155

0.5“
I‘—~~

L.--—

 

 

 

 

 

 

 

 

—---

:1 4

 

 

 

Figure 18 Configuration of the Critical Nozzles

156

cm

..noac.o nouoauan uo.oauuoz muoauauo
a amsouna a: mo .33— 305 acoauouoofi. was #353:an am can»;

Agmav 8:381 3282

 

 

 

no on as oe mm on
L.Pb.PLb—-Lpb—-ppph-Lbb pr-L omF
a
100m
a
ro¢u
r
IomN
n
8205590 0 town
30:83:... ..I .

 

 

(Jq/yo) sing MOU

APPENDIX C

APPENDIX C
THE GAS ANALYSIS SYSTEM

This Appendix describes the gas analysis system. This system is
mainly composed of the individual instruments used to make continuous

measurements of H20, C02, C0, 02, and total hydrocarbons (THC) in a

sample stream. A schematic of the instruments along with the
accessories needed to feed the desired sample stream into each analyzer
is shown in Figure 2.4.

The gas sample was pumped by a metal bellows pump model MB 601 HT
driven by 0.75 HP motor. This pump was originally designed as a two
stage pump. The two stages were used as separate pumps; one side was

used as a suction pump for H20 and C02 analyzers, the other side was

used to pump through the rest of the analyzers. A pressure regulator
originally built in the THC analyzer was used to regulate the pressure

before the CO and the 02 analyzers. A heating tape was used to heat the
sample line in order to prevent condensation of HC and H20. Needle

valves were used to control the flow through each analyzer which was
monitored by rotameters. The sample gas supplied to the C0 analyzer was
dried by cooling it to -5°C, using a Neslab U-cool immersion cooler as
required by the instrument. The sample lines were conStructed from

teflon tubing and stainless steel swagelock fittings.

157

158

C.1 H20 - CO Analyzer

2

An infrared AR-600 nondispersive infrared dual gas, dual range

analyzer was used for H20 (Full scale 0 - 5.0%) and C02 (full scale 0 -

1.0%) analysis. The analyzer has a 100 ms response time and an accuracy
of i 1% of full scale (F.S.). The sample gas is drawn through a 0.5
micron inline filter, and into the analyzer by pump suction. This
arrangement yields the fastest response time. I

This analyzer has certain unique features: (1) the measurements
are not affected by the presence of HC in the sample gas, (2) it is
equipped with a sophisticated differential signal processing techniques
which eliminate zero drifts, (3) it is instrumented with an

interconnection network between the C02 and H20 measurements which

provides a spectral balance adjustment for the spectral interference
range of the two gases, and finally (4) the analyzer is equipped with an
internal calibration reference for setting the span control whenever a

calibration gas is not available.
6.2 CO Analyzer

A Beckman 3153 infrared analyzer was used to continuously analyze
the concentration of C0 in the sample gas. The analysis is based on a
differential measurement of the absorption of infrared energy. The
analyzer has an adjustable full scale sensitivity, 90% response time in

0.5 second at 400 cc/min, and an accuracy of i 1% of full scale.

159

C.3 0 Analyzer

A Beckman 778, a polaragraphic analyzer was used for 02 analysis.
The concentration of 02 is determined by measuring the partial pressure

of oxygen. Because of this characteristic, it is important that the
sample gas mixture be kept under the same total pressure as when the
instrument was calibrated. The analyzer has three measuring ranges (0-
5, 0-25, 0-100%), 90% response in 10 seconds, and an accuracy of i 1%

F.S.

6.4 Total Hydrocarbon (THC) Analyzer

A Beckman 400 was used for analysis of non-condensible
hydrocarbons. The analyzer utilizes the flame ionization method. The
sensor is a burner; a regulated flow of sample gas passes through a

flame sustained by regulated flows of a fuel gas (40% H2 - 60 He) and

air. The concentration of total hydrocarbons in the original sample is
essentially measured by determining the rate at which carbon atoms enter
the burner.

The analyzer has an adjustable range from 1 ppm CH4 to 2% CH4,
response time for 90% CH4 of 0.5 second with sample by-pass flow of 3

liters/minute, and an accuracy of i 1% F.S.

LIST 01“ REFERENCES

LIST OF REFERENCES

Abu-Zaid, M., and Atreya, A. [1988a], ”Transient Cooling of Hot Porous
and Non-Porous Ceramic Solids by Droplet Evaporation," Submitted to the
J. of Heat Transfer.

Abu-Zaid, M., and Atreya, A. [1988b], "Heat Transfer During Evaporation
of A Water Droplet on A Heated Non-Porous Ceramic Solid," accepted for
presentation at the Eastern Session of the Combustion Institute,
Florida.

Alvares, N, Blackshear, P., and Murty, K. [1969], "The Influence of Free
Convection on the Ignition of Vertical Cellulosic Panels by Thermal
Radiation," Central States Section, Combustion Institute Meeting,
University of Minnesota.

American Society of Mechanical Engineers Research Committee on Fluid
Meters [1959], "Fluid Meters, Their Theory and Application,” 5th Ed.,
Published by ASME, New York.

Arnberg, B. [1962], "Review of Critical Flowmeters for Gas Flow
Measurements," J. of Basic Engineering, Vol. 84, 447.

Aschenbrenner, A. [1983], "The Influence of Humidity on the Flowrate of
Air Through Critical Flow Nozzles," Proceeding of Flomeko 1983 Imeko
Conference on Flow Measurement. Budapest, Hungary.

Atreya, A. [1983], "Pyrolysis, Ignition and Fire Spread on Horizontal
Surfaces of Wood, ” Ph.D. Thesis, Harvard University, Cambridge, MA.

Atreya, A. [1984], "Fire Growth on Horizontal Surfaces of Wood,"
Combustion Science and Technology, Vol. 39, 163.

Atreya, A. [1985], "Effect of Water on Ignition of Cellulosic
Materials,” Research Proposal, submitted to the National Bureau of
Standards, Washington, D.C.

Atreya, A. Carpentier, C., and Hankleroad, M. [1986], ”Effects of Sample
Orientation on Piloted Ignition and Flame Spread," First International
Symposium on Fire Safety Science, Vol. 1, 97.

Atreya, A., and Wichman, I. [1987], "Heat and Mass Transfer During

Piloted Ignition of Cellulosic Solids,” 2nd ASME-JSME Thermal
Engineering Joint Conference, Vol. 1, 433.

160

161

Avedisian, C., and Koplik, J. [1987], ”Leidenfrost Boiling of Methanol
Droplets on Hot Porous/Ceramic Surfaces," Int. J. Heat Mass Transfer,
Vol. 30, No. 2, 379.

Bamford, C., Crank, J., and Malan, D. [1946], "The Combustion of Wood,"
Proc. Cambridge Phil. Soc., Vol. 42, 166.

Baumeister, H., and Simon, F. [1973], "Leidenfrost Temperature - Its
Correlation for Liquid Metals, Crygens, Hydrocarbons, and Water," J. of
Heat Transfer, Vol. 95, 166.

Beck, J., and Arnold, K. [1977], ”Parameter Estimation In Engineering
and Science," John Wiley and Sons, New York.

Beck, J., Blackwell, B., and St. Clair, Jr., C. [1985], "Inverse Heat
Conduction Ill-posed Problems,” A Wiley-Interscience Publication, New
York. '

Bonancian, C., Comini, G., and Del Giudice, S. [1975], "Evaporation of
Atomized Liquids on Hot Surfaces," Letters in Heat and Mass Transfer,
Vol. 2, 401.

Bonacina, C., Del Giudice, 8., and Comini, G., [1979], "Dropwise
Evaporation,” J. of Heat Transfer, Vol. 101, 441.

Butler, C., Martin, S., and Lai, W. [1956], "Thermal Radiation Damage to
Cellulosic Materials - Part II - Ignition of Alpha-Cellulose by Square
Wave Exposure,“ U.S. Naval Radiological Defense Lab. Tech. USNRDL-TR-
135, NS 081-001, AF SWP-906

Carslaw, H., and Jaeger, J. [1959], ”Conduction of Heat in Solids,"
Oxford University Press, Oxford, 2nd Ed., p. 88.

Chapra, 8., and Canale, R. {1985], ”Numerical Methods for Engineers,"
McGraw-Hill Book Company, New York.

di Marzo, M., and Trehan, A. [1986], "Transient Cooling of A Hot Surface
by Droplets Evaporation,” National Bureau of Standards Interaency
Report, NBS-GCR-86-516.

di Marzo, M., Wang, 2., and Meng, W. [1987], ”Transient Cooling of A Hot
Surface by Droplets Evaporation," National Bureau of Standards Report,
NBS-GCR-87-534.

Emmons, H. [1973], "Heat Transfer in A Fire," J. of Heat Transfer, Vol.
99, 145.

Evans, 0., and di Marzo, M. [1986], ”Evaporation of'A Water Droplet
Deposited on a Hot High Thermal Conductivity Solid Surface, National
Bureau of Standards Report, NBSIR 86-3384.

Garden, R. [1953], "Temperature Attained in Wood Exposed to High
Intensity Thermal Radiation, "Fuel Research Lab., MIT, Tech. Report No.
3.

162

0005, A. [1952], "The Thermal Decomposition of Wood,” Wood Chemistry,
Vol. 2, ACS Nomograph Series No. 97, 2nd Ed., Reinhold Pub. Co., New
York.

Gottfried, B., and Bell, K. [1961], "The Evaporation of Small Drops on A
Flat Plate in the Film-Boiling Regime," Paper Press. at 54th Meeting of
the A.I.Ch.E.

Gottfried, B., Lee, C. and Bell, K. [1966], ”The Leidenfrost Phenomenon:
Film Boiling of Liquid Droplets on A Flat Plate, Int. J. Heat Mass
Transfer, 9, 1167.

Gradshteyn, I., and Ryzhik, I. [1980], "Table of Integrals, Series, and
Products, Academic Press. Inc., Orlando.

Holman, J. [1984, "Experimental Methods For Engineers," McGraw-Hill Book
Company, 4th Ed.

Holman, J. [1981], ”Heat Transfer," McGraw-Hill Book Company, 5th Ed.,
New York.

Incropera, F., and Dewitt, D. [1981], "Fundamentals of Heat Transfer,"
John Wiley and Sons, New York.

Kanury, A. [1972], ”Ignition of Cellulosic Solids - A Review," Fire-
Research Abstracts and Reviews, Vol. 14, 24.

Kashiwagi, T. [1981], "Radiative Ignition Mechanism of Solid Fuels,"
Fire Safety Journal, Vol. 3, 185.

Kashiwagi, T., Ohlemiller, J. and Werner, K. [1987], "Effects of
External Radiant Flux and Ambient Oxygen Concentration on Nonflaming
Gasification Rates and Evolved Products of White Pine,” J. Combustion
and Flame, Vol. 69, 331.

Keltner, N. [1973], "Transient Heat Flow in Half-Space Due to an
Isothermal Disk on the Surface," J. of,Heat Transfer, Vol. 95, 412.

Lee, C., Chaiken, R., and Singer, J. [1976], ”Charring Pyrolysis of Wood
in Fires by Laser Simulation," Sixteenth Symposium (International) on
Combustion, 1459.

Lee, C., and Diehl, R. [1981], "Combustion of Erradiated Dry and Wet
Oak," Combustion and Flame, Vol. 42, 123.

Lee, L., Chen, J., and Nelson, R. [1985], "LiquidsSolid Contact
Measurements Using A Surface Thermocouple Temperature Probe in
Atmospheric Pool Boiling Water," Int. J. Heat Mass Transfer, Vol. 28,
1415. -

Lee, T., Loftus, J., and Gross, D. [1974], "Effect of Moisture on
Surface Flammability of Coated and Uncoated Cellulosic Materials," ASTM
Special Technical Publication, No. 385.

163

Makino, K., and Michiyoshi, I. [1979], ”Effects of the Initial Size of
Water Droplet on Its Evaporation on Heated Surfaces," Int. J. Heat Mass
Transfer, Vol. 22, 979.

Martin, S. [1965], "Diffusion-Controlled Ignition of Cellulosic
Materials, by Intensive Radiant Energy," Tenth Symposium (International)
on Combustion, 877.

Martin, 8., Lincoln, R., and Ramstad, R. [1958], "Thermal Radiation
Damage to Cellulosic Materials, Part IV. Influence of the Moisture
Content and the Radiant Absorptivity of Cellulosic Materials on their
Ignition Behavior,” Report No. USNRDL-TR-295, AFSWP-1117. Naval
Radiological Defense Laboratory, San Francisco.

Michiyoshi, I., and Makino, K. [1978], "Heat Transfer Characteristics of
Evaporation of A Liquid Droplet on Heated Surfaces,” Int. J. Heat Mass
Transfer, Vol. 21, 605.

Mikkola, E., and Wichman, I. [1988], "The Thermal Ignition of
Combustible Materials," Accepted for presentation at the Eastern Session
of the Combustion Institute, Florida.

Ohlemiller, T., Kashiwagi, T., and Werner, R., ”Wood Gasification at
Fire Level Heat Fluxes,” Combustion and Flame, Vol. 69, 155.

Parker, W. [1988], ”Prediction of the Heat Release Rate of Wood," Ph.D.
Thesis, George Washington University.

Pedersen, C. [1970], ”An Experimental study of the Dynamic Behavior and
Heat Transfer Characteristics of Water Droplets Impinging Upon A Heated
Surface,” Int. J. Heat Mass Transfer, Vol. 13, 369.

Quintiere, J. [1981a], "A Simplified Theory for Generalizing Results
from Radiant Panel Rate of Flame Spread Apparatus," Fire and Materials,
Vol. 5, 51.

Quintiere, J. [1988], "The Application of Flame Spread Theory to Predict
Material Performance,” J. Research of National Bureau of Standards, Vol.
93, 61.

Renkisizbulut, M., and Yuen, M. [1983], ”Experimental Study of Droplet

Evaporation in A High-Temperature Air Stream," ASME J. of Heat Transfer,
Vol. 103, 384.

Rizza, J. [1981], "A Numerical Solution to Dropwise Evaporation," J. of
Heat Transfer, Vol. 103, 501.

Sauer, F. [1956], "The Charring of Wood During Exposure to Thermal
Radiation: Correlation Analysis for Semi-infinite Solids," Intrim, T.
AFSWP-868 U.S.D.A. Forest Service.

164

Schwenker, R., and Beck, L. [1963], “Study of the Pyrolytic
Decomposition of Cellulose by Gas Chromatography," J. Polymer Sci. Part
C, 1, No. 2, 331.

Seki, M., Kawamura, H., and Sanokawa, K. [1978], ”Transient Temperature
Profile of a Hot Wall Due to an Impinging Liquid Droplets,” J. of Heat
Transfer, Vol. 100, 167.

Simms, D [1960], "Ignition of Cellulosic Materials by Radiation,"
Combustion and Flame, Vol. 4, 293.

Simms, D. [1963], ”On the Pilot Ignition of Wood by Radiation,"
Combustion and Flame, 7, 253.

Simms, D., and Law, m. [1967], ”The Ignition of Wet and Dry Wood by
Radiation,“ Combustion and Flame, Vol. 11, 377

Smith W., and King, J. [1970], "Surface Temperature of Materials During
Radiant Heating to Ignition," J. Fire and Flammability, Vol. 1, 272.

Spiegel, M. [1968], "Mathematical Handbook of Formulas and Tables,"
Schaum's Outline Series, McGraw-Hill Book Company, New York.

Stamm, A. [1964], "Wood and Cellulose Science," The Ronald Press
Company, New York.

Steward, F. [1974], "Ignition Characteristics of Cellulosic Materials,"
Heat Transfer in Fires, Scripta Book Company, 379.

Taylor, R., and Groot, H. [1985] "Thermophysical Properties of Cast
MgO," A report to Sandia National Laboratories, Purdue University.

Thomson, H., and Drysdale, D. [1988], "An Experimental Evaluation of
Critical Surface Temperature as A Criterion for Piloted Ignition of
Solid Fuels," Fire Safety Journal, Vol. 13, 185.

Toda, S. [1972], "A Study of Mist Cooling (lst Report: Investigation of
Mist Cooling),” Heat Transfer Japanese Research, Vol. 1 (3), 39.

Tzeng, L., Wichman, I., and Atreya, A. [1988], ”A Theoretical Study of
One-Dimensional Piloted Ignition," Submitted to Combustion and Flame.

Vovelle, C., Akrich, R., and Delfau, J. [1984], ”Thermal Degradation of
Solid materials Under A Variable Radiant Heat Flux," Twentieth Symposium
(International) on Combustion, 1647.

Wachters, L., Bonne, H., and Van Nouhuis, H. [1966], "The Heat Transfer
from A Hot Horizontal Plate to Sessible Water Drops in the Spherodial
State," Chemical Engineering Science, Vol. 21, 923.

165

Wachters, L., and Westerling, N. [1966], "The Heat Transfer from A Hot
Wall to Impinging Water Drops in the Spheroidal State," Chemical
Engineering Science, Vol. 21, 1047.

Welker, J. [1970], "The Pyrolysis and Ignition of Cellulosic materials:
A Literature Review," J. of Fire and Flammability, Vol. 1, 12.

Wesson, H., Welker, J. and Sliepcevich, C. [1971], "The Piloted Ignition
of Wood by Thermal Radiation," Combustion and Flame, Vol. 16, 303.

Wichman, I. [1983], "Flame Spread in an Opposed Flow with a Linear
Velocity Gradient,” Combustion and Flame, Vol. 50, 287.

Wichman, 1., and Atreya, A. [1987], "A Simplified.Mode1 for the
Pyrolysis of Charring Materials," Combustion and Flame, Vol. 68, 231.

Yuen, M., and Chen, L. [1978], ”Heat-Transfer Measurements of
Evaporating Liquid Droplets," Int. J. Heat Mass Transfer, Vol. 21, 537.