PLACE IN RETURN BOX to tomove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jl
' __JL__|[_
"7| II J]

MSU II An Afﬁrmdlvo ActloNEqual Opportunity lnstﬂmion

 

 

 

 

 

 

 

 

 

 

 

THERMAL DECOMPOSITION
OF CHARRING MATERIALS

By

Said Nurbakhsh

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Mechanical Engineering

1989

56‘30’77

ABSTRACT

THERMAL DECOMPOSITION
OF CHARRING MATERIALS

By

Said Nurbakhsh

In this work, the complex process of pyrolysis of charring solids was studied.
Experimental techniques and methods were developed to investigate the transient
process of wood pyrolysis under different levels of external radiation, moisture
content of the wood sample, and oxygen concentration of the ambient atmosphere. A
unique small-scale combustion-wind tunnel was constructed to conduct the pyrolysis
experiments and to obtain the time dependent gasiﬁcation mass ﬂux, surface and in-
depth temperatures, and evolved products of pyrolysis (C0, C02, H20, and total
hydrocarbons [THCD for thermally thick samples of Douglas-ﬁr. Experiments were
performed both in inert atmosphere (nitrogen), and in air at several different heat
ﬂuxes and three different moisture contents of wood Time dependent empirical
chemical composition, char yield, and the heat of combustion of the pyrolysis

products were determined.

The experimental results indicate that the presence of moisture reduces the
pyrolysis mass ﬂux and delays the occurrence of its maxima. Presence of oxygen
drastically increases the pyrolysis mass ﬂux but its effect specially at lower
temperatures depends on the experimental conditions such as the boundary layer
thickness over the wood surface. Char yield, chemical composition of the volatiles,
and the heat of combustion were not found to be constant, instead they vary during
the pyrolysis process and with changes in the environmental conditions and wood
moisture content. The results also show that the asymptotic fall-off of the pyrolysis

mass ﬂux is not proportional to the negative one-half power of time as predicted by

some simple models.

In the theoretical part of this work, the ’pyrolysis temperature’ assumption
often used for the simpliﬁed modeling of wood pyrolysis was examined in detail by
considering two otherwise identical models; one with inﬁnitely fast decomposition
kinetics and the Other with ﬁnite rate chemistry. The pyrolysis temperature used in
the pyrolysis temperature model was determined by enforcing conservation of mass
and energy in an integral sense between the two models. It was found that although
the surface and in-depth temperatures predicted by both the models were in
reasonable agreement, signiﬁcant differences were found in the mass evolution rate of
volatiles. Different pyrolysis temperatures were required to balance mass and energy
each time the heat loss boundary conditions or the incident heat ﬂux was altered. It
was concluded that the pyrolysis temperature is not a material property and different
pyrolysis temperatures are needed for every problem.

To the Memory of My Mother

iv

Acknowledgement

The author is greatly thankful to his advisor Professor Arvind Atreya for
introducing him to the area of wood pyrolysis and ﬁre research. His constant
guidance, encouragement and direct involvement in every step of the work was of
invaluable help in the completion of this work. I deeply appreciate his patience and
and his friendship. I am also very grateful to Professors John R. Lloyd, James V.
Beck, and David Yen, members of my PhD. Guidance Committee, for their helpful

comments, suggestions, and encouragements.

I would like to thank Mr. Leonard Eisele and Mr. Paul Faeth for construction
of the combustion-wind tunnel. I also wish to thank Mr. Robert Rose for his ever
generous helps in many details of the electrical components of my experimental
facility. Without their help this work would not have been possible. Many thanks to
my colleagues and friends; Mr. Kamel E1 Mekki for his help in conducting
experiments and Mr. Charles Gendrich for developing the data acquisition
system. Finally, I am thankful to my dear wife, Hengameh, for her patience during
this work.

This work was supported by the National Science Foundation under Grant #
CBT - 8415423.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................. x
LIST OF FIGURES ........................................................................................... xi
NOMENCLATURE ............................................................................................ xiv
Chapter One: Introduction and Literature Review .................................... 1
1.1 Description of the Physical Problem ............................................................... 2
Background and literature Review ........................................................... 4

1.2 Physical Properties and Chemical Components of Wood ........................ 4

1.3 Thermal Decomposition of Wood .......................................................... 5

1.4 Formation of Char and Its Effect on Pyrolysis Process ............................ 6

1.5 Energetics of Thermal Decomposition ..................................................... 8

1.6 Previous Work ................................................................................................ 9

1.6.1 Experimental work ............................................................................... 10

1.62 Theoretical Work ................................................................................. 15

1.7 This Work ...................................................................................................... 17

1.7.1 Experimental Work .............................................................................. 18

1.7.2 Theoretical Work ................................................................................. 19

1.8 Objectives ...................................................................................................... 20
Chapter Two: Experimental Apparatus ........................................................ 21
2.1 Purpose .......................................................................................................... 21

2.2 Experimental Facility ...................................................................................... 21
Small Scale Combustion—Wind Tunnel ........................................................ 21

2.2.1 Inlet Section and Accessories ............................................................... 26

2.2.2 Turbulence Manipulation Section ......................................................... 28

2.2.3 Test Section ........................................................................................ 28

vi

Ch;

2.2.3.1 Radiant Heaters (RH) ............................................................... 29

2.2.3.2 Top Tunnel .............................................................................. 30

2.2.3.3 Tunnel Mainframe .................................................................... 31

2.2.4 Exhaust Section ................................................................................... 32
2.2.5 Catalytic Combustion Tube .................................................................. 33

2.3 Data Acquisition Equipment ............................................................................ 35
2.4 Gas Analysis Equipment ................................................................................. 35
2.4.1 The Gas Analyzers .............................................................................. 35
Total Hydrocarbons [THC] .................................................................. 37

C0-C0; Analyzers .............................................................................. 37

H20 Meter ........................................................................................... 39

02 Analer ......................................................................................... 39

2.5 Sample Preparation ......................................................................................... 39
2.6 System Diagnostics Tests and Preliminary Experiments ................................... 41
2.6.1 Dynamics of the Flow in the Tunnel .................................................... 41

2.6.2 Temperature Distribution Along Top-Tunnel Glasses ............................ 43

2.6.3 Mixing of the Pyrolysis Products ......................................................... 44

2.7 Calibrations .................................................................................................... 44
2.7.1 Radiant Heat Flux ................................................................................ 44

2.7.2 Gas Analyzers ..................................................................................... 45

2.8 Experimental Procedure .................................................................................. 46
2.9 Data Reduction ............................................................................................... 46
2.10 Experimental Errors ...................................................................................... 49
2.10.1 Errors in Measurements ..................................................................... 50
2.10.2 Errors in Derived Quantities ............................................................... 52
Chapter Three: Pyrolysis Experiments ......................................................... 53
3.1 Preliminary Experiments ................................................................................. 54
3.2 ’Combustion Efﬁciency’ Experiment ............................................................... 55

vii

Ch

Ch

3.3 Pyrolysis Experiments ..................................................................................... 57

3.4 Experimental Conditions ................................................................................. 61
3.5 Derived Quantities .......................................................................................... 62
Experimental Results and Discussion .............................................................. 66
3.6 Pyrolysis Mass Flux ....................................................................................... 66
3.6.1 Effect of Extemal Heat Flux ................................................................ 71

3.6.2 Effect of Ambient Oxygen ................................................................... 72

3.7 Products of Pyrolysis ...................................................................................... 85
3.7.1 Permanent Gases ................................................................................. 88

3.7.2 Water Mass Flux and Effects of Moisture Content of Wood ................. 89

3.8 Time Integrated (Total) Mass of Pyrolysis Products ......................................... 92
3.9 Char Yield and Empirical Chemical Composition of Pyrolysis Products ........... 99
3.10 Heat of Combustion of Pyrolysis Products ..................................................... 103
3.11 sample Temperature Proﬁle ........................................................................... 104
3.12 Data Correlation ............................................................................................ 109

Chapter Four: The Effect of Thermal Decomposition Kinetics on

the Mass Evolution Rate of Charring Solids ................................................ 113
4.1 Model Formulation ......................................................................................... 114
4.1.1 The Pyrolysis Temperature Model ................................................................ 117
4.1.2 The Finite Rate Decomposition Model .......................................................... 118
4.2 Overall Energy and Mass Balance ................................................................... 119
4.3 Results and Discussion ................................................................................... 121

Chapter Five: Conclusion and Recommendations ..................................... 142
5.1 Experimental Work ......................................................................................... 143
5.2 Theoretical Work ............................................................................................ 144
5.3 Recommendations ........................................................................................... 145

Appendix A: Data Processing Procedure .................................................... 147

Appendix B: Finite Dzﬁ’erence Equations and Methods of

viii

Solution ......................................................................................... 157
Appendix C: Data From Pyrolysis Ecperiments ............................... 170

List of References ....................................................................................... 252

LIST OF TABLES

Table 3.1 Codes for pyrolysis experiments ...................................................................... 61
Table 3.2 Final char depth for Pyrolysis experiments at different conditions. .................... 95
Table 4.1 Constants used for the calculations. ................................................................. 126
Table 4.2 Percentage errors in the total energy balance within each numerical

scheme. ......................................................................................................... 127

Table 4.3 Percentage difference between the total input energies (Eh) for the
two models. This was used to dertermine the pyrolysis temperature Tp. ............ 128

Figure 1.1
Figure 1.2
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 2.9

Figure 2.10

Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5

Figure 3.6
Figure 3.7
Figure 3.8
Figure 3.9
Figure 3.10
Figure 3.11
Figure 3.12
Figure 3.13
Figure 3.14
Figure 3.15
Figure 3.16

LIST OF FIGURES

A qualitative diagram of surface temperature and mass ﬂux history. ..........

Two alternative routes for the decomposition of cellulose.

 

Schematic diagram of the experimental facility.

 

Pictorial view of the tunnel.

 

Cross sectional view of the tunnel.
Calibration of a sonic nozzle.

 

 

Heat distribution along the test section.

 

Schematic of the tunnel outlet and the catalytic combustor.

 

Gas analysis equipment.

 

Velocity proﬁle inside the tunnel.

 

Air ﬂow ﬂuctuations in the tunnel.

 

Schematic of the decomposition process occuring in the tunnel. ................

 

CaHs. C02, and CO in the combustion of C3113.
H20 and 02 in the combustion of C3Hg.

 

Number of carbon atoms calculated from products of C3Hg.

 

Total mass ﬂux of reactants and products of combustion of CgHg. .............

Pyrolysis mass ﬂux in three runs of an experiment on 8—9% moist wood at

 

2 W/cm2 in air (reproducibility test).

 

Pyrolysis mass ﬂux for experiments on dry wood in N2.
External heat ﬂux for experiments on dry wood in N2.

 

Pyrolysis mass ﬂux for experiments on 8-9% moist wood in N2. ................
External heat ﬂux for experiments on 8-9% moist wood in N2. ..................

Pyrolysis mass ﬂux for experiments on 17% moist wood in N2.
External heat ﬂux for experiments on 17% moist wood in N2. .....

Pyrolysis mass ﬂux for experiments on dry wood in air.

 

External heat ﬂux for experiments on dry wood in air.

 

Pyrolysis mass ﬂux for experiments on 8-9% wood in air.

 

External heat ﬂux for experiments on 8-9% wood in air.

 

Pyrolysis mass ﬂux for experiments on 17% moist wood in air.

xi

23
24
25
27
27
34
38
42
42
47
58
S8
59
59

68
76
76
77
77
78
78
79
79
8O
80

FL!
ﬁt
“3

Egg,

Flint

Figure 3.17
Figure 3.18
Figure 3.19
Figure 3.20
Figure 3.21

Figure 3.22
Figure 3.23

Figure 3.24-29

Figure 3.30
Figure 3.31
Figure 3.32
Figure 3.33
Figure 4.1

Figure 4.2a
Figure 4.2b

Figure 4.2c

Figure 4.2d
Figure 4.3a

Figure 4.3b
Figure 4.3c
Figure 4.3d
Figure 4.3e

Figure 4.3!
Figure 4.3g

Figure 4.4

External heat ﬂux for experiments on 17% wood in air.

 

Pyrolysis mass ﬂux for experiments on dry wood in N2.

 

Pyrolysis mass ﬂux for experiments on 8-9% moist wood in N2. ...................
Pyrolysis mass ﬂux for experiments on 17% moist wood in N2. ....................

 

Pyrolysis mass ﬂux for experiments on dry wood in air.

Pyrolysis mass ﬂux for experiments on 8-9% moist wood in air. ...................

Pyrolysis mass ﬂux for experiments on 17% moist wood in air. ....................

Time integrated mass of pyrolysis products as a function of incident heat
ﬂux.

 

Temperature vs. time at various locations inside wood; EXP. TM3N. ...........

Surface temperatures for experiments on dry wood in air.

 

Surface temperatures for experiments on 17% moist wood in air. ..................

Apparent latent heat of pyrolysis for N2 experiments.

 

Physical conﬁguration of thermal decomposition problem showing all
energy transfers.

 

From surface temperatm'es, with no heat losses at the front and back
Sin-faces.

 

Back sm'face temperatures, with no heat losses at the front and back
surfaces.

 

Total weight loss, with no heat losses at the front and back
surfaces.

 

Weight loss rate, with no heat losses at the front and back surfaces. .............

Surface temperatm'e, with different from surface heat losses and back
surface insulated

 

Back Stu-face temperature, with different front srn'face heat losses and back
surface insulated.

 

Total weight loss, with different front surface heat losses and back surface
insulated.

 

Weight loss rate, with different front surface heat losses and back surface
insulated.

 

Weight loss rate, with different front surface heat losses and back surface
insulated.

 

Weight loss rate at early times of pyrolysis.

 

In-depth temperature and density proﬁles in the wood slab with the

radiative heat loss at the front surface and the back surface insulated. ..........

Weight loss rates.

xii

81
82
82
83
83
84

107
108
108
112

129

130

131

132
133

134

135

136

137

138
139

Appendix A

Figure A.l
Figure A.2
Figure A.3
Figure A.4
Figure A.5

Appendix B

Figure 3.1
Figure 3.2

Appendix C

Figures Slxxx
Figures SZxxx
Figures PSxxx
Figures CHxxx
Figures YCxxx
Figures (3an
Figures 'I'Mxxx
Figures ERxxx
Figures HVxxx
Figures me

Polynomial spline ﬁt and rate of weight loss data.
Curve ﬁtting and response time correction of 02 data.
A well stirred system.

Response of C02 analyzer to a step input.

Recorded response to a piecewise linear input.

Schematic of ﬁnite-difference method for pyrolysis temperature model.
Schematic of ﬁnite-difference method for decomposition model.

Pyrolysis products (direct measurement).

Pyrolysis products after bruning in the catalytic combustor.
Products of pyrolysis as percent of total mass ﬂux.
Number of C and H atoms in C,H,0.

Char yield (gram char/gram wood).

Gram []/gram of QH,0.

Total mass balance for pyrolysis products.

Error in total mass balance for pyrolysis products.

Heat of combustion of pyrolysis products.

Temperature proﬁles.

xiii

 

 

NOMENCLATURE

>

= Pre-exponential factor, llsec

= Variable in equation (B-3)

= Variable in equation (B-3)

= Variable in equation (B-3)

= Speciﬁc Heat. J/kg.K

= Energy, W/cm2

= Activation Enegy, J/Kgmole

= Energy going into the solid, W/cm2

= Externally applied heat ﬂux, W/cm2

= Conductive heat transfer coefﬁcient, W/cm2 . K
= Heat of combustion per unit mass of 02, KJ/gm
= Heat of combustion of pyrolysis products, KJ/gm
= A1]

= Thermal conductivity, 1] rn . K . sec

= Length, m

= Number of nodal points

= Mass ﬂux, g/crnzsec

= Heat ﬂux, W/cm2

= Heat of pyrolysis. cal/gm

= At/A'q2

= Universal gas constant, Call gm mole K

= Non-dimensional distance
= Temperature, °C or K
= Time, sec
= Time constant, sec

= Nodal points adjacent to char-wood interface
= Distance, em

= Distance, em

= Variable in equation (B—12)

= Mole fraction
= Char yield

wgﬁumnympggg

B;Zl"‘?§

-
C

,9.

«yxnarr'amweo

J<

xiv

 

Greek

= Thermal diffusivity (Chapter 5)

= A constant (Chapter 4)

= More

= XIL

= m.

= Gamma function

= Density

= Nomdimensional time, equation (B-l)
= TIT sunoo

carton-educate

Subscripts

= Active material
= Char

= Final

= Fuel

= Incident

= Space coordinate
= Time coordinate
og = Out with gases
= Pyrolysis

= Pyrolysis

= Solid

= Wood

= Ambient

= Stored

~‘S'WWOQ

he

asemg'o

 

 

Chapter One

Introduction
and
Literature Review

The burning of wood and other cellulosic materials has long been recognized as
being of cenu'al importance in ﬁre research. These materials are very important sources
of fuel and the potential for energy and chemicals production from these renewable
sources has established great interest in their thermal behavior. A fundamental
understanding of the thermal degradation and combustion of cellulosic materials and
possible methods for controlling them is essential for protection and better utilization

of these materials.

The combustion of charring materials is a complex process. A major feature of
these materials is that the solid upon heating, undergoes a thermal degradation process

which generates the fuel gases to sustain a diffusion ﬂame and leaves behind a layer of

 

 

char. The formation and growth of this char layer complicates the ﬂaming combustion
of thick pieces of wood. The propagation of the pyrolysis front into the solid presents
a Stefan-type problem which introduces nonlinear coupling between the gas phase
combustion and the solid phase decomposition. For a realistic prediction of the
pyrolysis rate it is necessary to have data on the amount of char produced at any
instant (i.e. the char yield) and changes in the thermal properties of burning wood and

char.

In this work the transient thermal decomposition process of thick samples of
wood is studied under different levels of external radiation, moisture content, and
oxygen concentration of the ambient atmosphere. Experimental techniques and
methods were developed for this study. Numerous experiments were done on specially
prepared and instrumented samples of wood in an attempt to obtain reliable data for
the determination of the transient changes in the overall composition of the volatile

gases and the char yield.

1.1 Description of the Physical Problem

 

Consider a slab of wood initially at ambient temperature exposed to heat from
other burning objects in a furnace or a ﬁreplace. Upon heating of wood, ﬁrst the
adsorbed water is released (desorption), and then thermal degradation or pyrolysis of
wood provides a highly reactive carbonaceous char and a mixture of gaseous, volatile,
and tarry products. In the presence of oxygen, oxidation of char in the solid phase
produces glowing combustion, whereas the mixture of volatiles and tarry particles
(condensible hydrocarbons) that are carried out to the gas phase can produce ﬂaming
combustion if the mixture of fuel and oxygen is ignited Figure 1.1 shows a qualitative
diagram of the pyrolysis mass ﬂux history. During this process a pyrolysis zone will
be developed in the solid [corresponds to the initial rise of the mass ﬂux curve (Figure
1.1)] and begins to propagate into the material [corresponds to the gradual decay in

rd

wood i
of you
coup.
it are:

I

ant it

I. h ‘ I S ‘1‘
. . . .
‘ Q “a“ 0 0

AH.-. NS-ne\hvib Nudsﬂh '3‘!

the mass ﬂux curve (Figure 1.1)] leaving behind an insulating layer of char. This char
layer attenuates the fuel evolution rate, by acting as a barrier against conduction of
heat into the solid, and also radiates away part of the incident radiation to the
surface. If wood is heated in an inert atmosphere, this process continues until all of the

wood is pyrolyzed. In an oxygen containing atmosphere, non-ﬂaming gasiﬁcation rate
of wood will be increased by char oxidation.

In the burning of a charring solid, the gas phase ﬂaming combustion is directly
coupled with the solid phase decomposition process. For a complete understanding of
the overall combustion, one needs to study the phenomena in each phase. In this work

thermal decomposition of wood - the phenomenon occurring in the solid phase - will
be addressed.

1.25 L—

1.00 l-
0075 '-
0050 '-

O.25 -

HASB FLUX (mg/cm? sec)

 

 

Figure 1.1 Qualitative diagram of pyrolysis mass ﬂux.

H

2112

4

Background and Literature Review

1.2 Physical Properties and Chemical Components of Wood

Wood is a highly porous material made up of cells and ﬁbers that are vessels and
channels for transportation of water and gases. They also provide the mechanical
strength of wood. The porosity (ratio of the volume of pores to the volume occupied
by the cell wall) of the wood lies somewhere in the range of 40-75%. It is mainly
made up of two major chemical components carbohydrates (65-70%) and 1i gnin (18-
35%). Minor amounts of extraneous materials, mostly in the form of organic and

inorganic minerals (ash), are also present in wood.

The carbohydrate portion of wood comprises of cellulose and hemicellulose.
Cellulose (a polymer of glucosan (C6H1005),,) ranges from 40 to 45% of the dry
wood weight, and hemicellulose (a polysacharide producing wood sugars) ranges from
25 to 35%. The other major component of wood, i.e. lignin, is a multi ring compound

whose molecular weight exceeds one thousand.

Unless wood is dried, it usually contains some moisture. The moisture in wood
is in three forms: (1) Free water which is mechanically held - by surface tension - to
the walls of the cell cavities in the structure of wood. The energy needed to release
this water is only slightly over the latent heat of water. (2) Adsorbed or bound water
which is limited to approximately 30% of the oven dry weight of the wood. Here the
water molecules are attached to the cell walls via hydrogen bonds and it needs some
more energy to be released for lower moisture content. (3) The water of constitution
which is formed and liberated along with the other pyrolysis products of
wood. Primarily absorbed and adsorbed water are liberated within the early stages of
heating of wood. In general, it is not possible to distinguish between different kinds of
moisture being released. Practically, at any depth in wood, decomposition starts soon

after the free and bound water are evaporated. Furthermore, upon heating of wood,

a pal
been“

COTS

the release of water occurs in depth and the vapor need not all be driven to the surface
to escape. Some may move inward and recondense in cooler interior portions of the
solid. Thus even before decomposition starts the nature of the solid may be modiﬁed

by the heat.

To obtain "dry" wood, generally, it is not possible to remove all the moisture
content of wood without initiating any pyrolysis. However, theoretically when all non-
constitution water is evaporated the wood could be termed as "dry". This, of course,
will take a very long demoisturizing treatment of wood in a well controlled humidity
chamber to be achieved.

1.3 Thermal Decomposition of Wood

 

As it was mentioned earlier, cellulose is the major component of wood and other
cellulosic materials as well as the major source of combustible fuel for the ﬂaming
combustion. Hence, the chemisu'y of pyrolysis and formation of volatile products from
cellulose could well represent the corresponding processes occurring in wood. In this
section the pyrolysis of cellulose within thermally thick slabs of the solid is described
brieﬂy.

Pyrolysis of cellulose occurs in two distinct pathways [Shaﬁzadeh (1981)]. At
lower temperatures (200 - 280°C) dehydration of cellulose produces dehydrocellulose
and water. This process is slightly endothermic (except in the presence of oxygen) and
leads to the formation of char, water, and volatile gases such as C02, C0, and
hydrocarbons. The gases evolved in the dehydrocellulose route are primarily
noncombustible and the char which remains can oxidize through a surface (glowing)
combustion. At higher temperatures (280 - 340°C) endothermic depolymerization of
cellulose leads to the formation of tarry products which are highly condensible and
constitute the main gaseous fuel to support a gas-phase ﬂame. On further raising the

temperature, the tar-forming reactions accelerate rapidly and overshadow the

production of char and gases. The tarry products of primary reactions may undergo
further decomposition as they pass tluough the hot porous char. In fact, the analysis of
the pyrolysis products of cellulose show different yields of char and tar in vacuum and
in atmospheric pressure [Shaﬁzadeh (1981)]. This is due to the much lower residence
time of evolved gases when they travel through the hot char matrix in vacuum than in
atmospheric pressure. These pathways and observations were summarized by Atreya
(1983) in the following scheme which is consistent with the works of other
investigators [Madorsky (1975), Shaﬁzadeh (1981), Martin (1965), Panton et al.
(1971), and Broido (1976)].

1.4 Formation of Char and Its Effects on Pyrolysis Process

 

As shown in Figure 1.2, char may be produced by direct thermal degradation of
wood. It may also be produced from dehydration and condensation of the volatile
pyrolysis products (tar) in the active material. The porous char matrix is a highly
reactive substance which can have different compositions and properties at various
stages of charring. The reactivity of char depends on the interactions between the char
matrix and the volatile gases at different stages of pyrolysis [Shaﬁzadeh (1981)]. The
temperature history, the type of ambient atmosphere, and heat treatment conditions of

wood will change the rate of char yield and to some extent properties of char.

In the presence of oxygen, oxidation of char can lead to smoldering or glowing
combustion. The latter one, occurs in the presence of more oxygen and at higher
temperatures. Smoldering combustion occurs at lower temperatures and " incomplete "
oxidation of char leads to formation of C0 and C02 and generation of new reactive
cites. Oxidation of char causes the char surface to regress. However, the rate of char
consumption is not as fast as the rate of pyrolysis of the substrate [Shaﬁzadeh
(1981)].

 

Cellulose

 

 

 

[endothermic]
Scission of CO bonds
(high temperature)

{ Depolymerimtion

    
 

 

 

 

 

 

 

We water, gases

 

 

 

escape to annosphere
#5 C easier in the case
8 of small samples.

 

Pyrolysisoflevoglucosmsinee
itisrestninedtoeecepedueto
thehotciurmetrixoflergesemples.

 

 

 

(endothermic
except in the
presense of 07)

 

 

char, water, gases

 

 

 

Figure 1.2 Two alternative routes for the decomposition of cellulose.

.23

51

of

As the pyrolysis continues and the thickness of the char layer increases ﬁne
cracks begin to appear on the surface. The solid surface continuously shrinks and the
cracks grow in size. The direction and size of the cracks are different for different
types of wood and even for different samples of the same type of wood. The
appearance of these cracks may have a signiﬁcant effect on the pyrolysis of the
remaining material because they alter the course of heating of the solid and the motion
of volatiles within the char matrix by reducing heat exchange between the gases and
the solid [Atreya (1983)]. The shrinkage of the solid, which is more signiﬁcant at high
heat ﬂuxes, causes the actual temporal variations in the densities to be different from

those obtained based on the initial volume of the solid.

1.5 Energetics of Thermal Decomposition

 

The overall solid-phase decomposition process involves the following
phenomena: energy transport by conduction; kinetics of thermal decomposition;
outward migration of moisture and pyrolysis products and its interaction with the solid;
convective heat transfer and possibly‘ chemical reactions between the fuel volatiles
and the char matrix; inward migration, condensation, and regasiﬁcation of volatile fuel
gases; pressure build up in the interior of the solid due to the generation and migration
of these gases; radiative exchange within the porous cells of the char matrix; shrinkage
and cracking of the char which alters some of the above processes. The physical and
chemical properties of both char and wood also change as the temperature of the solid
changes.

 

"' The controversy over the occurrence of secondary exothermic reactions between the hot
slut and the fuel gases is not yet completely resolved. Atreya (1983), by recording the tem-
perature of the gas phase very close to the surface of the pyrolyzing wood, showed that the ex-
othermic reactions indeed occur not inside the solid but outside between oxygen and fuel
gases.’Ihisissuewillbeaddressedinthiswork.

Inclusion of all these parameters into a single analytical or even numerical model
will be extremely difﬁcult because of the nonlinearities associated with all the rate
processes. A detail theoretical model must include variations of thermal properties ( p,
Cp, K) of the solid and char yield in the pyrolysis of wood. For polymeric fuels
property variations associted with chemical decomposition are very signiﬁcant.
However, wood is a highly anisotropic and inhomogeneous porous solid and its
thermal properties are different in each direction. Hence, the transient process of wood
pyrolysis is very hard to be completely reproduced. Accurate reliable data on
temperature dependent thermal properties of wood and char, char yield and changes in

the empirical chemical composition of pyrolysis products are lacking.

1.6 Previous Work

 

Extensive experimental and theoretical work has been done on chemical and
physical processes occurring during pyrolysis of solid materials. A considerable
amount of this work have been focussed on "vaporizing" solids such as PMMA
(Polymethylmetacrylate). Many investigators studied pyrolysis of charring solids such
as wood. However, most of their attempts have been focussed on determination of
thermal and kinetic parameters (such as heat of pyrolysis, activation energy, and
frequency factor) that were incorporated in the models that describe pyrolysis
process. Several reviews have been published on the subject [Welker (1970), Roberts
(1970), Zardy and Pyle (1982)]. Since these works have been recently reviewed by
Atreya (1983), only relevant aspects are presented here. In Section 1.6.1 the history of
the previous experimental work is reviewed and work on pyrolysis of thick specimens
of wood is described in more detail. In Section 1.6.2 the past theoretical work is
discussed brieﬂy.

f‘ﬂ

'l
a.”

‘ A
NM:

Um

N“: l

M's“

5“; I"
b

10

1.6.1 Experimental Work

 

experiments on pyrolysis of wood and its components (cellulose, hemicellulose,
and lignin) may be divided into two major categories: (1) Data obtained on small
samples (i.e. thermally thin with almost no temperature gradient within the solid). (2)
Data obtained on large samples (i.e. thermally thick with an effective temperature
gradient inside the solid). These were reviewed by Atreya (1983) and Chan (1983).

In the ﬁrst category several techniques are used to determine the relevant
parameters related to the physics of pyrolysis (heat of pyrolysis), or chemistry of
decomposition (activation energy and frequency factor). These techniques are: (l)
Thermogravimetric Analysis (TGA) [Tang (1967), Rogers and Ohlemiller (1980),
Shaﬁzadeh (1968) and several other authors]; Differential Thermal Analysis (DTA)
[Tang and Neil (1964), Broido (1966) and several other authors]; and Differential
Scanning Calorimetery (DSC).

In the DTA method, the electrical power required to keep a wood test cylinder
and a compensation cylinder at the same surface temperature to assure equal heat
losses to the surrounding is measured. The difference in the power supplied to the test
cylinder as compared to the compensation cylinder is due to the sensible heat as well
as the heat of decomposition. This is called "energy capacity". By assuming the
sensible heat portion of the energy capacity to be proportional to the amount of solid
material left, the value of the energy capacity can be multiplied by the percentage of
weight remaining at a given temperature to estimate the sensible heat portion of the
energy capacity. Several authors have used this technique to measure heat of pyrolysis,

which is reported in a review by Welker (1970).

In differential scanning calorimetery energy differences rather than temperature
differences are measured. Both the sample and the reference cells are maintained at the
same temperature during the analysis by adjusting the electrical power dissipated in

each case in the required manner. Havens et a1. (1971) used this method to measure

11

heat of pyrolysis.

In thermal gravimetric studies of wood decomposition weight loss versus time is
measured at a series of constant temperatures over the desired range (isothermal TGA),
or weight loss is measured as a function of time while the sample is heated so as to
effect a linear temperature rise (dynamic TGA). From these data, the frequency factor
A and activation energy E are determined to ﬁnd a ﬁrst order Arrhenius reaction

equation for the weight loss (or density change) as a function of temperature

112-=Ame-ﬁ
dt

Several authors have used this method [Tang (1967), Shaﬁzadeh (1968), Rogers and
0hlemi11er(1980), and others].

In the second category, the pyrolysis of large samples of wood have been
studied. Large samples are thermally thick, resulting in a temperature gradient inside
the solid which plays a signiﬁcant role in the decomposition process, composition of
pyrolysis products, and the char yield. In this type of pyrolysis, wood gasiﬁcation is
conuolled by heat and mass transfer processes and is affected by other physical
variables such as moisture content, grain direction, and sample orientation. Only a few
studies on thick samples have been reported in the past and most of them have been
focussed on the determination of physical and kinetic parameters controlling the
pyrolysis process. Some of the most relevant work which has been done in the past is

brieﬂy reviewed in the following paragraphs.

Bamford et al. (1945) heated thick slabs of wood by luminous gas ﬂames and
measured the spontaneous ignition time and the temperature at the center of the solid
and calculated the heat of decomposition per unit mass of volatiles. Roberts and
Clough (1963) conducted experiments on wood cylinders under controlled atmosphere
in nitrogen to study the overall kinetics of the decomposition reactions. The rise in the

central temperature of the wood cylinder above the surface temperature was attributed

12

to the exothermic reactions within the solid. Their results showed that for experiments
in which the weight loss was high and the central temperature exceeded 350°C the
activation energy E was 15,000 cal/gm-mole and the frequency factor A was 9.0x104
1/min. For lower temperature and weight loss E=25,000 cal/gm and A=2.6x109
l/min. This was attributed to the difference between the early and later stages of the
reaction. From their experimental data they concluded that primary decomposition of
wood has low heat of reaction but when gases travel outwards through the hot char,
secondary decomposition of gases occurs, solid char acting as catalyst. Thus the
exothermic heat of reaction increases rapidly with distance from the axis and then

becomes approximately constant.

Kilzer and Broido (1965) analyzed data for pyrolysis of wood and suggested two
competing sequences of reactions. One a slightly endothermic dehydration to
"dehydrocellulose" followed by an endothermic sequence of reactions yielding char
and various gaseous products. The other one an endothermic depolymerization of
cellulose to produce volatile products. They concluded that under some conditions,
such as reduced pressure, the primary pyrolysis of wood can be endothermic. Under
normal conditions, at atmospheric pressure the primary pyrolysis of wood is
exothermic. They speculated that the difference may be caused by a change in the

reaction mechanism of cellulose due to catalytic or autocatalytic effects.

The wide discrepancy of reported values and even sign of the heat of
decomposition was considerably narrowed by the work of Kung and Kalelkar
(1973). They used a mathematical model describing the essential processes taking
place in wood pyrolysis, i.e., conduction, outward migration of volatiles, and a ﬁrst
order reaction kinetics to compare with the experimental data of Roberts and Clough
and deduce the values of heat of pyrolysis. The values of Q, A, and E which gave the
best ﬁt to all experimental data were -48.4 cal/gm (endothermic), 1.19x105 l/min and
15140 cal/mole. They found that for a pyrolyzing cylinder the central temperature

13

could rise above the surface temperature which might erroneously lead one to believe
that it is caused by exothermic reactions of wood pyrolysis. By expanding the energy
equation for the pyrolysing solid they showed that the effective local heat of
vaporization not only involves the endothermic heat of pyrolysis, but also includes the
energy carried away by the fuel volatiles and changes in the enthalpy of the solid due
to the thermal decomposition. These three components represent the instantaneous
local energy absorption associated with the generation of volatiles. The effecrive heat
of vaporization for wood decreases with temperature and in some cases it could even
become negative (exothermic) at high temperatures leading to the possibility of a
thermal runaway. They deduced that, this phenomena caused the temperature in the
central portion of the wooden pellets to be greater than the surface temperature at the
end of the pyrolysis process. Similar analysis was also adopted later by Atreya (1983)
who demonsuated that an exothermic heat of reaction could eventually lead to the

thermal explosion, a phenomenon that has never been observed for wood.

Lee et a1. (1976) studied pyrolysis of 2 cm diameter wood pellets heated by a
C02 laser between 3 and 8 W/cmz. They found that a pyrolysis zone (about 1 cm
thick) can be divided into three zones. (i) All endothermic primary decomposition zone
at T < 250°C. (ii) An exothermic partial char (active pyrolysis) zone at 250 < T <
340°C. (iii) An endothermic surface char zone 340 < T <520°C. The overall effective
heat of pyrolysis was determined to be -l46 cal/gm (endothermic). It was found that
the rate of pyrolysis depends on the heating rate and that the pyrolysis zone is thiner
and progresses faster if heat ﬂux is perpendicular to the grain direction. This could be
related to the samples’ much lower gas permeability in the perpendicular direction than
in the parallel grain direction relative to the heat ﬂux. The evidence was appearance of
more cracks in the parallel grain direction than the normal direction. In the latter case,
high pressures up to 0.3 psig are generated within the solid as it decomposes. This
pressrn‘e is much smaller in the case of parallel grain direction and gases are released

14

much more easier resulting in lower solid temperatures. Their data indicated that the
pyrolysis of wood is a rather complex process which is strongly inﬂuenced by the
endothermic and exothermic reactions, wood and char structural and thermal
properties, and the heating rate. Atreya (1983) suggested that the wide discrepancy in
the the reported values of heat of reaction could be due to either exothermic gas phase
reactions with oxygen, or to endothermic desorption of even small amount of
moisture. However, this question could be answered only when experiments were

repeated in an inert atmosphere, which eliminates the role of gas phase reactions.

Kanury and Blackshear (1967) reported temperature and density proﬁles in a
cellulosic pellet heated in air. They used the time-temperature history of pyrolyzing
cylinders to determine local heat source and sink strengths. They found that
decomposition is endothermic at 300-400°C, and is exothermic above 500°C. They
also found that gases produced by the endothermic pyrolysis of the solid in the interior
of the specimen undergoes exothermic pyrolysis in the hot char near the surface.

The effects of absorbed water on the pyrolysis and ignition of wood was
investigated by Lee and Diehl (1981). They soaked wood samples in the water to
obtain up to 60% water content and then exposed it to the ﬁre level radiation of 2
cal/cm2 heat ﬂux. They found that the absorbed water mainly delays the temperature

rise in the wet wood and the vaporized water convectively cools the pyrolyzing solid.

There are very few studies that have reported the measurements of products of
pyrolysis and until very recently no transient measurements of these products had been
reported. Chan (1983) studied the product distribution as a function of physical
variables such as wood density, length, grain direction, and substrate composition
(cellulose, hemicellulose, lignin). His results indicated that heat ﬂux has the strongest
inﬂuence on temperature and product disuibution. Although the products of different
components of wood behave differently as temperature changes, however, the chemical

composition of wood does not alter the general trends for increase or decrease of the

15

pyrolysis products as a function of temperature. He also found that the char yield
generally decreases with high heating rate and temperature, but increases with longer
residence time. The reverse is true for the gas yield. The tar yield increases with
higher temperature but may either increase or decrease with increasing heating rate
depending on the temperature. High temperatures, high heating rate and short gas
residence time would increase tar yield. Low temperature and low heating rate would
maximize char yield while high temperature, high heating rate and short gas residence
time would maximize gas yield. He quantitatively demonstrated the validity of these
ﬁndings.

The only study in which u'ansient measurements of major products of pyrolysis of
thick slabs of wood has been reported, is the work of Kashiwagi et a1. (1987). They
studied the effects of oxygen concentration of the ambient atmosphere and the external
heat ﬂux on the rate of pyrolysis and the evolution rate of evolved products, namely,
[TI-1C], C02, CO, H20 and tar. However, due to the difﬁculties in measurement of

H20, the study failed to quantitatively measure the water and tar evolution rates.

1.6.2 Theoretical Work

 

The process of formation and growth of a char layer which pr0tects the
decomposition zone and the virgin material involves many physical and chemical
parameters that makes it very difﬁcult to develop even a numerical solution which
contains all these effects. In the past several attempts have been made to develop
realistic models for this complex phenomena and compare with experimental
results. The investigators considered different kinds of simplifying assumptions in their
theoretical models. The simplest models are pure thermal models that only consider the
heat transfer within the pyrolyzing solid. Attempts to model the pyrolysis reaction rates
are mostly limited to a single ﬁrst order reaction scheme. Few authors have considered

higher or multiple ﬁrst order reactions [Shaﬁzadeh and Bradbury (1979)]. The

16

modeling of pyrolysis becomes more difﬁcult when large samples of wood are
considered. Often simultaneous solutions of mass, energy, and reaction kinetics that
involve highly nonlinear terms are involved. In the following paragraphs some the

most relevant theoretical work that has been done in the past is brieﬂy reviewed.

Tinney (1965) studied the pyrolysis of wooden dowels heated in air using the
Fourier conduction equation with a source term accounting for the exothermicity of
pyrolysis. Thermal properties were assumed constant and the convective heat transfer
of volatiles were neglected. Matsumoto et al. (1969) studied the thermal
decomposition of plastics which has some similarities with wood pyrolysis. They
included the temperature and degree of pyrolysis dependence of properties and
convection of fuel volatiles. The time-space region was divided into three distinct
regions; namely, char, pyrolysis, and virgin material zones. Kanury (1971) showed that
a pure transient conduction model (pyrolysis temperature model) overestimates the
burning rate and slightly underestimates its dependency on element thickness, i.e.
predicts less than experimental data. This is because of neglecting internal convection,
detailed pyrolysis kinetics, pyrolysis endothermicity, and thermal prOperty variations
with degree of charring. Kung (1972) assumed thermal properties of the solid to vary
continuously from their values for virgin wood to their values for char. Kung and
Kalelkar (1976) used this model to study the experimental data of Robert and Clough
(1963) and in a detailed analysis of the energies associated with pyrolysis process
explained the reasons that caused misinterpretations of those data in the past. This was

described in the previous section.

Kansa et al. (1977) included a momentum equation for the motion of gases
relative to the solid and obtained good agreement with Lee et al.’s (1976) experimental
data for low heat ﬂuxes. At high heat ﬂuxes, however, poor agreement was obtained
that was attributed to ignoring effects of structural changes (shrinkage and cracking)

and to the assumption of a single step pyrolysis reaction. All these models have used

17

a global ﬁrst order Arrhenius pyrolysis rate relation. The effect of evaporation and
inward propagation of moisture have often been ignored. Atreya (1983) included the
kinetics of moisture evaporation and migration in his numerical model. The effects of
cracking of char was simulated by shutting off the term representing the outward
convection of volatiles at a speciﬁed temperature of the solid when cracking was
expected to occur (650°K). A simple analytical model was developed by Delichatsios
and deRis (1983) which predicted the asymptotic pyrolysis rate for constant heat ﬂux

after pyrolysis is initiated at a pyrolysis temperature T1) at time t=tp. At large times
1
the pyrolysis decays proportional to (HP) 2 , that is negative half- power of

time. However, none of the available models, numerical or analytical, are detailed

enough to correCtly predict the fuel generation for wood.

1.7 This Work

 

Experiments on the decomposition of thermally thick samples of charring
polymers are usually conducted by exposing the sample to a known heat ﬂux and
measuring its weight loss. Uncertainties in the interpretation of these results arise due
to (i) spectral absorptance of the polymer surface relative to the spectral quality of the
incident radiation [Wesson et al. (1971)], (ii) changes in surface radiative properties
during decomposition, (iii) absorption of incident radiation by the decomposition
products [Kashiwagi (1981)], (iv) exothermic reactions in the gas phase [Atreya
(1983)], (v) endothermic moisture desorption effects in natural polymers [Atreya
(1983)], (vi) variations in char yield during decomposition [Atreya (1983), Broido and
Nelson (1975)], and (vii) lack of knowledge about changes in the physical properties

with temperature and decomposition.

The net effect of (i) to (iv) is that the boundary conditions at the polymer surface
are no longer precisely known. Uncertainties due to (i), (ii) and (iii) can be eliminated

by using surface temperature instead of the incident heat ﬂux. A method of

18

continuously measuring the surface temperature during decomposition has been
developed by Atreya (1983) and is described in Chapter 2. As mentioned before, (iv)
may be eliminated by conducting the experiments in an inert atmosphere. However,
items (v), (vi), and (vii) are processes occurring in the solid phase and need to be

understood and modeled.

In the studies in which chemistry of wood pyrolysis is investigated, usually
samples of the pyrolysis products are periodically extracted and analyzed [Chan
(1983)], or to prevent the condensation of heavy molecule hydrocarbons, they are
condensed in a cold trap and analyzed later. This approach preserves the chemical
composition of the fuel gases, but results in the loss of transient information. Thus an
alternative approach which allows the transient determination of carbon, hydrogen, and
oxygen atoms in the products of pyrolysis without preserving the chemical form is
adopted in this work. Experiments are performed in N2 atmosphere and in air. For N;
atmosphere, a known ﬂow rate of oxygen is introduced to the boundary layer ﬂow
downstream of the wood sample and a small sample of the resulting mixture is passed
through a high temperatrue catalytic combustor to produce water vapor and carbon

dioxide which are continuously measured.

-

1.7.1 Experimental Work

 

A small scale combustion-wind tunnel was designed and constructed to conduct
the pyrolysis experiments. Wind-aided and wind-opposed ﬂame spread experiments in
the ﬂoor as well as ceiling conﬁguration can also be performed in this facility. The
main features of this tunnel are: (1) It provides variable external radiation. (2) It allows
transient measurements of many physical and chemical parameters, such as
temperature of solid and gas, concentrations of evolved gases, and weight loss of the
wood sample. (3) It provides a well controlled atmosphere over the wood sample to

conduct the pyrolysis experiments. Detail description of the experimental facility is

ﬁrst-

I
“van-
Nirvi
u

 

 

19

given in Chapter 2.

The pyrolysis experiments were conducted to determine the evolved mass ﬂux as
a function of applied heat ﬂux. The objectives were: (i) to identify the dominant
chemical and physical processes that occur during the decomposition process, (ii) to
determine the amount of char formed at any instant under given environmental
condition, and (iii) to obtain time dependent empirical chemical composition of the
pyrolysis products.

The newly constructed tunnel was extensively tested prior to conducting any
pyrolysis experiments. The results of these tests are described in detail in Chapter
2. Also, a large number of experiments (>50) were conducted in order to identify
various factors that inﬂuence the experimental process and the accuracy of the
results. These factors were partly due to the nature of wood and partly due to the
ability of the instruments to do the measurements. For instance, accurate weight loss
and chemical species measurements were of utmost importance. For every experiment,
the accuracy of the results were determined from the temporal and overall mass
recovery in that experiment. Experimental procedure had to be improved and/or
modiﬁed in order to obtain reliable qualitative data. In these experiments, simultaneous
measurements of several physical and chemical quantities were made. The results of
the preliminary experiments and the ﬁnal procedures adopted to conduct the

experiments are described in Chapter 3.

1.7.2 Theoretical Work

 

As mentioned earlier, pyrolysis of thick samples of wood involves both the
physics of heat and mass transfer and the kinetics of thermal decomposition. In the
past, several attempts have been made to develop a simple model of this complex
phenomena with the objective of predicting the mass evolution rate of the

volatiles. The approach most often used to simplify the model equations is to assume

 

 

20

inﬁnitely fast decomposition chemisu'y at a speciﬁed temperature called the "pyrolysis
temperature." In this work the "pyrolysis temperature" assumption is examined in
detail by considering two otherwise identical models; one with inﬁnitely fast
decomposition kinetics and the other with ﬁnite rate chemistry. The pyrolysis
temperature used in the ﬁrst model is determined by enforcing conservation of mass
and energy in an integral sense between the two models. Chapter 4 describes the
formulation of the governing equations and results of the numerical
calculations. Derivation of ﬁnite-difference equations and the methods of solution are

presented in Appendix B.

1.8 Objectives

The purpose of this work is to identify the dominant chemical and physical
processes that occur during the thermal decomposition of wood. The speciﬁc
objectives are:

(i) To study the effects of external heat ﬂux, oxygen concentration of the ambient

atmosphere, and the moisttn'e content of the samples on wood pyrolysis.

(ii) To obtain transient data on the evolution mass ﬂux of pyrolysis products,

temperature proﬁles inside the solid, and the total pyrolysis mass ﬂux.
(iii) To determine instantaneous fraction of wood that is left behind as char.

(iv) To determine changes in the empirical chemical composition of the pyrolysis

products with the growth of the char layer; and
(v) To obtain the instantaneous heating value of the pyrolysis products.

To achieve these objectives, a series of systematic experiments w ere performed
on samples of Douglas-ﬁr with different moisture contents under different external
radiation levels in air and in niuogen aunospheres. The results of these experiments

are presented and discussed in Chapter 3.

Chapter Two
Experimental Apparatus

2.1 Purpose

 

The primary objective of the pyrolysis experiments is to understand and obtain
reliable data on relationship between evolved mass ﬂux and externally applied heat
ﬂux under different environmental conditions encountered in a furnace or a home ﬁre.
Upon heating by a neighboring ﬁre, wood undergoes thermal decomposition producing
fuel gases (volatiles) and char. In this work, external radiation from a ﬁre is simulated
by using electrical radiant heaters. In this chapter, the apparatus and the procedure
employed, the data processing methods, and the errors encountered in the experiments
are described in detail.

21

22

2.2 Experimental Facility

 

A schematic of all the components of the apparatus used in this work is shown in

Figure 2.1. This facility consists of three main units:

(i) A small scale combustion-wind tunnel which provides a well controlled
environment for performing the pyrolysis experiments and allows detail transient

measurements of physical and chemical variables; to the actual experiments.

(ii) Gas analysis equipment, to measure concentrations of chemical species in the

ﬂow;

(iii) Data acquisition equipment to collect, store and process the data. The
products of pyrolysis were measured once directly as they were produced and once
after they were burned inside a high temperature combustion tube that was attached to

the tunnel.

Small Scale Combustion-Wind Tunnel

The thermal decomposition experiments were performed in a small scale
combustion wind tunnel. This tunnel can also be used for ignition, ﬂame spread, and
extinction experiments. This facility was designed by the primary investigator of the
project and the technical drawings, modiﬁcations and construction of the tunnel was
carried on and supervised by the author. Figure 2.2 shows the pictorial view of the
tunnel which is rigidly mounted on a frame. The moving belt shown in this ﬁgure is
for the future ﬂame spread experiments and it was substituted with an adjustable table
and an electronic balance for the pyrolysis experiments. The cross-sectional view of

the tunnel is shown in Figure 2.3.

23

3:8. Eco—5.2.5 05 he Etna—e 3352.8 ﬁn Pin—h

mag—.5522 Elm-43:3 E
EU“: mac...— 5..—

382—3 ZOF<>¢mm¢O BO
.532. ANZZD... F

PE. Pat—39E B.

ESP 02:3:— .F:

mug 02:5:— 3:

«mm-255 02:92 02

mZO—h<—>—¢n< u—m—SLEOU

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:2: .558
. \zoEnSoo.‘ <5:
.5
maﬁa 96 236552 95
n8.— ..EF M2: 9.5:
:25. «z «o No gamma: :2: 5935 22.53.28
A _ 396 .54.. in: EEELSE
9338025.: :2: 55:55
6552:»... I 553 02:55?
95 :95 l t :26 B s\ . ezztmm
\ Jun—n ﬁa/ / . ﬂ
.. . i- . 53:5. Eases
_ .. B .
32 229. mm = . I a . e u .
/ :33 96 TN . . u
.+..T 3:3,. 02:08
I. «HEB: .2592.

 

 

 

 

 

 

B3553

5325—0 DEAF—hm I
223—. I: magma»: :0:—
io 20:8... 20:03 5:. 29.5%
b.32—

202228 :95

24

3520 5. 22 .365 2a mouse—Em :95sz 2: =u 82V
42:5. 2: no 32> .thoE ad 95”.:

 

 

:um 2:23..

 

 

 

e : ﬂ .5 LE .5 m.

 

 

.E.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\\Ms§

x

 

\

 

 

 

 

em

 

 

42:5. 2.. he 33> .338» $20 ﬂu «Suﬁ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. is
L es 2...». L..-
a I .. Boom.»
1 H. .H . -. .7 E
1.111 SHE
_ em 11. 69::
l _ shew t \xlzog E
v I. was 9:88
‘ s \
LE . --_H.H.=U\
-. m

 

 

33.39: . ,

 

an .95.
1.... g

 

 

 

 

 

 

 

 

 

\fﬁ
.2.
III
B
at
II
S
,/
E,—
_L ﬁe

 

 

.1—

26

This unique facility is comprised of several sections such that each section could
be detached and modiﬁed separately. These sections are: 1) inlet section and
accessories, 2) turbulence manipulation section, 3) test section, 4) radiant heaters, 5)

tunnel-top, and 6) exhaust section and accessories.

2.2.1 Inlet Section and Accessories

 

As illustrated in the schematic diagram of the experimental facility, Figure 2.1,
the inlet gases are provided by two separate routes to the tunnel, namely, compressed
air route and pressurized nitrogen and/or oxygen route. The compressed air which is
obtained from the main air pipe lines of the building ﬁrst enters a large pressure tank.
The mechanical oscillations of the air ﬂow initiated at the compressor are damped in
this tank. After passing through a set of globe and control valves, air enters a sonic
(critical) nozzle. The sonic oriﬁce maintains desired certain ﬂow rate if the inlet
(upstream) pressure controlled within the critical range. The temperature of the air is
measured by a thermocouple installed through a small well at the nozzle location. The
ﬂow rate of the air can be calculated knowing the diameter of the nozzle, upstream
pressure and the temperature of the air. The downstream pressure of the sonic oriﬁce is
nearly atmospheric and is controlled to obtain the desired pressure at the sample
location. The sonic nozzles were calibrated by employing a tracer gas technique. A
small known ﬂow rate of pure methane was inuoduced upstream of the oriﬁce and the
mole fraction of the methane in the resulting air-methane mixture was measured. The
discharge coefﬁcient of each nozzle was carefully determined from the ﬂow rate
measurements. Figure 2.4 shows calibration curve of a sonic oriﬁce. A similar

procedure was used for introducing cylinder nitrogen and/or oxygen gases.

The inlet air enters a distribution manifold and ﬂows into the settling chamber
through eight pipes which branch out of the manifold at equal angles. Within the
settling chamber, the high velocity incoming jets strike bafﬂes which are symmetrically

MASS FLOW RATE (lbm/ hr)

HEAT FLUX CAUSE
READING (nut

27

 

 

 

 

 

 

 

 

 

 

 

 

zso .,....,.-..,....,....,....
240- -
220‘ -
200- -‘
180- '-
1so-' -
140- ~
12°30' " 'Ss 3b is do 55 so
UPSTREAM STATIC PRESSURE (p810)
Figure 2-4 Calibration of a sonic oriﬁce.
‘T term-n or rue near-ens 4
LOCATION 0' THE HOOD SAMPLE
3‘ .L ..
L5€ L74 VT
"03: \\
”-1
”1 ‘§""t'o""t'e""z‘o""33"'it}

 

team or rue resr SECTION (16-)

Figure 2.5 Heat ﬂux distribution along the test section.

28

installed in the chamber opposing the air jets. The air looses its momentum and the
pressurized air flows into the tunnel through a set of canvas and metal screens. The
cylindrical settling chamber is made such that part of the cylinder can be detached and
slided out. The screens can then be cleaned or replaced if necessary. This
combination, therefore, creates a slightly pressurized ﬂow of gases while helps

damping out most of the oscillatory motions of the ﬂow.

2.2.2 Turbulence Manipulation Section

 

The turbulence manipulation section consists of a two-inch wide honeycomb and
three ﬁne wire screens installed two inches apart. The ﬁrst screen is attached to the
outlet plane of the honeycomb and the Others are installed 2 inches apart. Screens are
carefully made in a way not to disturb the ﬂow. They are tightly held in a metal frame
which is inserted inside a slot grooved on the side , bottom, and top of the tunnel.
The large scale turbulent eddies are damped in the honey comb section, and the
smaller scale eddies are dissipated by the wire screens. In addition, the two-inch wide
open space between the ﬁrst and the second screen is ﬁlled with 4 mm diameter glass
beads, which provides more uniformity for the ﬂow and clamps the oscillations even
further. According to Nagib’s (1976) turbulence reduction mechanisms, this
combination fulﬁlls the task of turbulence reduction and creating a uniform ﬂow which
is particularly signiﬁcant for ﬂame spread experiments as well as for weight loss

measurements in pyrolysis experiments.

2.2.3 Test Section

 

The test section of the tunnel consists of three modules: (1) The external radiation

source, (2) the tunnel-top, and (3) the main-frame of the tunnel.

29

2.2.3.1 Radiant Heaters (RH)

 

Two types of electrical radiant heaters were combined and employed to maintain
the external radiation for the pyrolysis and ﬂame spread experiments. Three high
temperature ( maximum ﬁllament temperature = 1230°K) ) quartz electrical heaters
were mounted in a metal frame [RI-I, Figure 2.2]. The three 10" x 10" square heaters
were conﬁgured at the top of a metal box to create a 10" x 30" infrared heat source. In
addition six U shaped Chromlox coil heaters ( Chromlox 3/8" dia. Incoloy sheath, type
UTU each 1.8 KW ) were added beneath the quartz heaters. Each group of the heaters
were controlled separately by a 3-phase 440 volt variable transformer. As shown in
Figure 2.3 a series of highly reﬂective aluminum sheets were installed as radiation
shields within the side walls of the heaters frame. The space within the walls of the
box were ﬁlled with ceramic ﬁber insulating material to minimize the heat loss to the

surrounding.

As shown in the cross-sectional view of the entire heating system [Figure 2.3],
the lO-inch width was reduced to about 6-1/2 inches by two reﬂective chromium
plated (mirror ﬁnish) stainless steel sheets. The resulting combination acted as a
furnace providing a highly directional radiative heat ﬂux towards the base of the tunnel
where the wood sample is located. The warm up time of the quartz heaters was
approximately 30 minutes. During this warm up time, a pair of water cooled blackened
aluminum plates [Figure 2.2] were used to intercept the radiation from the
heaters. These cooling plates slided out on several pairs of small ball-bearings which
were attached to the side walls of the heaters via two longitudinal L shape
holders. These cooling plates were used to expose the sample to external radiation. The
radiant heater module could be raised and lowered by two jacks which held the heater
box and were ﬁrmly attached to the main frame of the facility.

30

2.2.3 .2 Top-Tunnel

 

The top (ceiling) section of the tunnel was constructed as a separate module
which was hinged at the tunnel inlet. This water-cooled plate housed ﬁve 6" x 6" x
2mm pieces of infrared optical glass plates [Figure 2.2, KODAK IRTRAN II]. The
combination of the ﬁve pieces made a 6" x 30" window which was transparent to the
infrared radiation. The IRTRAN glasses were lined with ﬂexible silicon rubber to
allow for expansion. During the exposure of the glasses to external radiation a plane
jet of air was blown on the top, by a compressed air manifold, to convectively cool the
glasses. This high velocity air jet, however, caused entrainment of air into the heater
box and reduced the temperature of the heating rods. This was partially compensated
by increasing the electrical power to the heaters and raising the heater box slightly

above the air jet.

The entire tunnel top could be lifted at the downstream side of the tunnel by an
adjusting mechanism. This increases the cross-sectional area of the tunnel at the
exhaust end to compensate for acceleration of the gas core due to the boundary layer
growth and gas expansion from combustion heat release. By experimentally adjusting
the height of the top-tunnel it is possible to obtain a constant free stream velocity

01...).

To prevent leakage of the gases into or out of the tunnel through any possible gap
between tunnel-top and the main frame, two 1.5 inch wide 1/8 inch thick chromium
plated (mirror ﬁnish) bars were attached to the top tunnel next to the side walls of the
tunnel [Figure 2.3]. They were tightly in contact with the side walls and prevented
leakage when top tunnel was lifted. The mirror ﬁnish surface of these bars helped to

enhance radiation as well.

 

31

2.2.3.3 Tunnel Mainframe

 

The 48-inch long mainframe consisted of three sections: inlet, test, and outlet
sections which were 10, 32, and 6 inches long, respectively. The water-cooled side
walls were made of 1/8" brass plates and they house two observation windows and
three probe access slots on each side. Pyrex glass was used for these windows and it
was lined with a thin sheet of ﬂexible silicon rubber on the frame edges to allow for
expansion and prevent cracking. Some static pressure probes were also installed on the
side walls. The probe access slots were also lined with silicon rubber to prevent
leakage and allowing the probe to transverse up and down for velocity proﬁle
measurements. The base of the 10 inch long inlet section of the mainframe was also
water-cooled. This was done to ensure that the inlet air and the entering wood samples

(for future ﬂame spread experiments) were at the ambient temperature.

The cross-sectional area of the tunnel is shown in Figure 2.3. The conﬁguration
and the dimensions were selected to satisfy several objectives: (1) To obtain a uniform
heat ﬂux across (3" wide) and along (6" long) wood sample for pyrolysis and moisture
desorption experiments. (2) To prevent the interference of the side wall boundary
layers with the boundary layer over the charting solid at the presence of a ﬂame. (3)
To minimize the consumption rate of the inlet gases (nitrogen and oxygen). Two 1.5"
wide noncombustible blocks (Kawool insulating board from Babcok & Wilcox, type
3000 sprayed with a rigidizer and baked in oven at 400°C) mounted ﬂush with the
wood sample surface, were placed symmetrically on both sides of the sample. For
pyrolysis experiments all but the middle 6 inches of the 32-inch long sample space
was covered with the insulating board. In addition a 1/2" thick Kawool board was used
to cover entire tunnel base except for the sample location, and a 1/16" thick aluminum
plate with a sample size opening covered this board. This combination created a
laminar boundary layer over the ﬂat plate and brought the sample a little closer to the

heaters to increase the heat ﬂux. The radiative heat ﬂux was also slightly augmented

 

 

32

by the presence of the metal plate.

2.2.4 Exhaust Section

 

The main part of the exhaust section was the mixing chamber [MC, Figure 2.2].
This chamber was made as a separate detachable module. The primary purpose of this
chamber was to mix the stratiﬁed boundary layer ﬂow carrying the pyrolysis or
ﬂaming combustion gases before reaching the sampling probe. This was necessary in
order to obtain a well mixed and uniform sample of gases for chemical
measurements. In this chamber the gases were also heated to prevent any condensation

of the heavy hydrocarbons (tar) on their way to the gas analyzers.

Mixing and heating in the mixing chamber were achieved in several steps. The
incoming ﬂow passed through three bafﬂes installed at the inlet, middle, and outlet of
the mixing chamber. Large scale mixing of the stratiﬁed gases was accomplished when
they passed through these bafﬂes and a series of electrical tapes installed in staggered
position and up and down along the ﬂow. In addition, further downstream of the
mixing chamber a net of electrical resistance wire [Figure 2.6] was used to provide
mixing on a small scale. At the outlet of the mixing chamber a metal louver was

installed to ﬁnally assure mixing of the gasses.

The mixing chamber was made of stainless steel and was lined with oven treated
Kawool insulating boards. Quartz tubings reinforced with metal bars were used as the
holders for the heating tapes and wires. The input power to the heaters was controlled

by a small variable transformer.

A sample of the gases was extracted through a spatially integrating sampling
probe. This cross-shaped probe, with holes distributed along its four arms, further
assured a representative sampling of the product gasses for chemical analysis. A 12"
extension was attached to the MC which emptied the exhaust gases into a 4’ x 2’ x 2’

metalic chamber [Figure 2.1]. The gases were then sucked from the chamber by an

33

exhaust pump and were exhausted through the roof vent. The 16 ft3 exhaust tank was
used to suppress the mechanical oscillations which were created by the exhaust fan and
were reﬂected back upstream into the ﬂow and disturbed the boundary layer ﬂow and

the weight loss measurements.

The presence of all of the above mentioned components, i.e., heating tapes and
wires, bafﬂes, louver, and the exhaust chamber, substantially increased the pressure
drop. This resistance against the ﬂow was overcome by the exhaust fan. A motor
driven adjustable slot was installed on the exhaust side of this fan to vary the outlet
area. This outlet area was carefully adjusted until the pressure inside the tunnel at the

sample location was almost atmospheric.

The entire tunnel was rigidly secured by to a solid metal frame made of angle
iron bars. This allowed the tunnel to be used in the horizontal, vertical, and ceiling

conﬁgurations.

2.2.5 Catalytic Combustion Tube

 

A part of the sample gas from the tunnel ﬂow was passed through a high
temperature catalytic combustor (combustion tube) in which the fuel gases were burned
with oxygen and converted to C02 and H20. The combustion tube and the gas
sampling lines are schematically shown in Figure 2.6. A heating element (Nichrome
resistance wire and a piece of platinum wire attached in series) was made as a coil and
installed inside a ceramic tube. The inlet and outlet of the ceramic tube were tightly
sealed using several layers of high temperature ceramic adhesives. The ends of the
wire were passed tluough two small holes that were made close to the inlet and the
outlet of the ceramic tube. These holes were tightly sealed as well. The entire tube was
completely insulated from outside by ceramic ﬁber insulation. Temperatures of the
outside of the tube were measured at three different locations. The input power was

adjusted by a variable transformer to achieve the desired temperature. The premixed

34

zumz<zu uz_mhpmm

hm2<zxm

    

 

33388 2.».an on. use .256 .055. 05 he 2862.3 QN 0.53.,—

All > '11.?
_ _
a. : 85: 853:8 SEES

xtlruu- Egg ttttttt [It
. [V II:

 

 

mamm>m<z< m<o mom<3o~

, . N
_ muz_4 auh<m:

_ mpuzooxm m_m>4oz>m omzmamz:

 

 

 

mum—z oz_h<m: . mma<h uz_h<m:t

 

mmmz<zu c:_x_z
mm>2¢m uz_x_z

35

ﬂow of fuel and oxygen came into direct contact with the high temperature (>1000°C)

element while passing through the combustion tube.

2.3 Data Acquisition Equipment

 

All the analog signals from the thermocouples, gas analyzers, and the heat ﬂux
gauges were collected and stored with the help of a DEC PDPl 1-73 computer running
RSX-11M+ operating system. The analog signals were ﬁrst fed into a HP 3497A
controller digital voltmeter which was interfaced with the computer via an IEQ-ll
interface board so that the two machines could exchange information. The electronic
balance, was directly interfaced with the computer to record the transient weight loss
data.

Two separate data acquisition algorithms were developed, namely, a single
buffered (slow scan) option and a multi buffered (fast scan) option. The slow scan
acquired data at approximately eleven readings per second. The fast scan version could
take up to 20 readings per second (time constant of the system about 1/20
seconds). The accuracy of the acquired data was about 5-1/2 digits. The transient

weight loss data were obtained at a rate of about 5-6 data points per second.

A single computer program named ACQ invoked both the HP controller and the
digital balance. The system time was recorded at the start of both data to enable the
user to match the starting time of both weight loss and the other group of data

points. The data were stored on ﬂoppy diskettes.

2.4 Gas Analysis Equipment

 

Two sets of gas analyzers were used to continuously measure the concentrations
of the pyrolysis products. One set was used for direct measurement of these species as
they were produced and the second set measured the mole fraction of the gases after

they burned in the high temperature catalytic combustor. A small sample of the tunnel

36

ﬂow was extracted and passed through each set of the gas analyzers to continuously
measure the concentrations of the total hydrocarbons [THC], C02, CO, 02, and H20.
The sampling lines were all heated up to about 100°C via electrical heating tapes. The
transport times for the gases to reach the analyzers, (i.e., the lag time) were measured

for each equipment. A schematic of the gas analysis system is shown in Figure 2.7.

A high temperature vacuum pump (Metal Bellows model MBo601HT) was used
on side one (before combustion tube) to transport sampled gases to the gas analyzers
via 1/4 inch O.D. stainless steel and teﬂon tubings which were insulated all the way to
the analyzers. One port of the pump was used to feed the COz-HZO analyzer and the
other port fed the hydrocarbon, CO, and 02 meters. The hydrocarbon analyzer was
equiped with a built-in pressure regulator that maintained the desired pressure
throughout the system. Needle valves on the lines were used to control the ﬂow to all

analyzers and the ﬂow rates were monitored by rotameters.

On side two (after combustion tube), the selected sample was pumped into all
analyzers by a Metal Bellows vacuum pump (model P-40). The pressure on the system
was controlled by the 6" height of water inside a container that acted as a pressure
regulator. The sample gases supplied to the CO—C02 and Oz analyzers as well as the
CO analyzer of set number one passed through a cold trap and was cooled to -5°C and
dried. The cold bath was cooled by a Neslab U-cool immersion cooler. All sample
lines which fed the [THC] and H20 analyzers were electrically heated from their
entrance to the system to the analyzer entrances to prevent condensation of [THC] and

H20.

37

2.4.1 The Gas Analyzers

 

Total Hydrocarbons [THC]

 

In both sets the total hydrocarbons were measured by using a Flame Ionization
Detector (FID). On side one a Beckman hydrocarbon analyzer model 400 and on side
two a Shimadzu gas chromatograph GC-3BF was used as a [THC] detector. The
response time of both [THC] meters was very small (< 1 sec.). Both Fle were fueled
with a 40%H2-60%He mixture to minimize the effect of Oz in the sample ﬂow. The
response of the F11) for a constant concentration of [THC] is a weak function of 02
concentration in the sample stream. The FID response with 20% 02 in the sample gas
was 97% of the response for the same concentration of CH4 and no 02. However,
since the [THC] meters were calibrated using methane, the measured value was
considered as equivalent of CH4 concentration. The time constant of the Beckman

[THC] analyzer was 1.0 second and that of the Shimadzu analyzer was 0.86 seconds.

C0-C0: Analyzers

 

On side one a dual analyzer (AN ARD Series 600 nondispersive infrared) was
used to measure C02 and H20. It measured the concentration of a selected gas specie
in a multi-component gas mixture by measuring the infrared absorption of the selected
component in the mixture. The meter could measure C02 up to 5% and H20 up to
1%. The response time of the analyzer was very small. The sample ﬂow was ﬁltered
before entering this analyzer. A Beckman CO analyzer (model 3153 Infrared)was used

to measure CO on this side.
On side two an Infrared IR702 nondispersive dual gas analyzer was used for C02
(F.S. 0-20%, 0-6%) and CO (F.S. 0-12%, 0-3%) analysis. The time constant of this

analyzer was about 2.6 seconds at about 2 liters/sec ﬂow and the accuracy was il% of

the Full scale. The analyzer required a water ﬁ'ee sample.

38

:88qu £sz5 30 EN 953m

552588 02:33 as”? “N «Em

 

 

 

 

53:5 oh , .53..

v 3&5 ray:
v £5... 1E?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3082338 825

.5258 8 Arm!

 

 

_l

u— 2.5

 

 

 

Pun-(=5— O._. 7S]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

55,12...
.0 _
:
m5:
«Had—<5 ll
: a
.A
m
l” w“
.8358 ow
leY

 

 

 

 

 

 

39

H20 Meter

On side two a General Eastern 1200APS Condensation Dew-Point hygrometer
was used to measure the water. This equipment measured the dew point temperature of
the sample stream by optically detecting the condensation on a temperature-controlled
mirror surface. The instrument was accurate to :l:0.2°C when the mirror was clean (no
condensible contaminant). It was very sensitive to the presence of any condensible

other than water. The time constant was approximately 10 seconds at 1 liter/sec.

02 Analyzer

 

On side one, a Beckman oxygen analyzer Model 778 was used. This analyzer
measures the partial pressure of oxygen which diffuses into a layer of potassium
chloride gel through a Teﬂon membrane. The gel electrically connects the cathode of
the sensor to the anode. This analyzer had a lot of electrical noise. The response time
was about 2 seconds and the accuracy was il% full scale at constant temperature.

On side two, a Beckman OM-ll a polarographic analyzer, was used for 02
analysis. The time constant was about 1.5 seconds at 0.5 liters/sec ﬂow and the
instrument had a short term (10 minute) accuracy of 110.05% (i0.lS% long term). The

analyzer required a dry sample a 15-40°C.

2.5 Sample Preparation

 

The wood samples for pyrolysis experiments were cut from l’xl’x1.5" boards of
Douglas-ﬁr which was supplied by the USDA Forest Products Laboratory (Madison
Wisconsin). Samples of approximately 3" x 6" x 1.5" in dimensions were carefully
machined to obtain right angles and smooth surfaces. Prior to the experiments, the
samples were instrumented by very ﬁne thermocouples (type K, 0.002" Alumel-
Chromel) both on the top and the bottom surfaces. The surface thermocouples were
installed using the method developed and successfully used by Atreya (1983). In this

40

method, a very thin layer of wood surface was skinned off carefully by a sharp razor
blade. The thermocouple leads were then burried under this 1-2 mm wide skin and
glued to the surface by wood glue. It was then left to dry out under a heavy weight for
several hours. As a result, the ﬂatened thermocouple bead was well attached to the
surface. At high heat ﬂuxes and/or high moisture contents of wood, some
thermocouples were installed in depth of the sample at the designated locations. For
these thermocouples, a very ﬁne hole (0.8 mm diameter) was drilled to about 1 inch
depth on the side wall of the wood sample. A very ﬁne ceramic thermocouple insulator
tubing with two holes were used to secure the thermocouple leads inside wood and
maintain strength for thermocouple wires. The thermocouple leads were carefully
imbeded in ﬁne grooves made on the side walls of the wood sample and were glued to
stay tight in their position. This prevented the thermocouple wires to hang loosely

about the sample and disturb the weight loss measurement.

For moisture contents different ﬁom room condition the samples were
conditioned in a humidity control chamber to the desired moisture content. The
samples were ﬁrst heated in an air tight insulated electrical oven at a well controlled
temperature of about 104 °C for about two days. During this drying process a ﬂow of
pure nitrogen was maintained in the oven. This procedure resulted in drying the
samples in a very well controlled atmosphere with a known temperature and relative
humidity. The condition of the wood, therefore, was well known at the time of
removal from the oven. Samples treated in this manner were considered as "dry". For
cases other than "dry" a temperature-humidity test chamber (model TH-3 from BMA,
Inc.) was used. The controllable humidity range of this chamber was 20% to 95% and
the controllable temperature range was +2 to 85°C. This was limited by a :t2°C dew
point temperature. For humidiﬁcation, a vapor generator system incorporating a
stainless steel heater was used to introduce water vapor into the chamber. The

chamber was carefully tested and calibrated prior to its use. Also, before starting the

41

conditioning of the samples the history of the moisture gain with time was recorded. It
was observed that it was necessary to condition the samples for at least 4 weeks to
assure that they had reached the desired moisture content. However, the samples were
usually conditioned for about two months to obtain the best uniformity and

steadiness. The samples were usually conditioned at 20°C dry-bulb temperature.

For the pyrolysis and moisture desorption experiments all but the t0p surfaces of
the sample were covered with aluminum self-adhesive tape. This was necessary
because upon heating, some of the desorbed water and condensed volatiles tended to
move along the grain direction within the wood and drip out of the sample. These
products would remain unaccounted if they were not evaporated and carried away by
the tunnel ﬂow. It should be emphasized that any other method by which water and/or
fuel gases were collected and evaporated later, would have resulted in loss of transient
information about the products of pyrolysis of wood. Also, this procedure is necessary

to obtain a true l-D experiment, i.e. an inﬁnite slab.

2.6 System Diagnostics Tests and Preliminary Experiments

 

In order to ensure the validity and applicability of the experimental results, all
physical and dynamical aspects of the tunnel were examined in detail prior to the start
of the pyrolysis experiments. Every component of the tunnel was tested carefully and
necessary improvements or modiﬁcations were carried out. These tests are described in

this section.

2.6.1 Dynamics of the Flow in the Tunnel

 

The turbulence ﬂuctuation intensity of the ﬂow in the tunnel was monitored using
a HP-Real Time Analyzer. This device can differentiate the mechanical oscillations,
which may exist in the ﬂow, from the turbulence ﬂuctuations. Large number of tests

were performed and ﬂow ﬂuctuations were carefully detected in all possible

ll‘.‘

9.0.5..an0> DCCUIuUZSK
\nbcommcvckmﬁk 1!. C02

 

42

 

 

THEORY (BLASIUS)

a data points

 

r
a 1.0-l
3 .
3‘
'5 0.8-J
2
a) '1
>
3 0.6-
t:
.2 ‘
2
0 0.4-
.E. .
i”
c: 0.21
O
z 1
0.0
0.0
0.040

r T 1 I Y

4.0 816

Figure 2.8 Velocity proﬁle inside the tunnel.

1' 1 1
12.0 q-y‘lﬂg 16

 

0.030-

0.020-

H turbulence fluctuations

. H mechanical oscillations

H total fluctuations

‘r'i RMS

 

Non—Dnmensronol
Fluctuating Velocities

 

 

 

Star/U7

0.0104 c e a : er /
1>——-—'—""'

0.000w..s1....,.w.3....,....1.f.
0.0 0.5 1.0 1.5 2.0 2.5

2 inches
DISTANCE FROM CESVTER 1)O THE SIDE WALL

Figure 2.9 Air ﬂow ﬂuctuations in the tunnel.

 

3.0

 

i.O

 

cc thirst;
ﬂow, cxh
list he
asmm
ﬁgs: 2.5

or Blast.

' n1 ,
553$:le

Want A.
and; [7

ll] ‘1”:— c
““5“.“
v

Knox

absorbed f
0f ‘ht [R'
“M3 six:

2h: pem‘lg

in

43

combinations of all mechanical parts connected to the system (inlet pressure driven
ﬂow, exhaust fan, and roof fan). For instance, the ﬂow ﬂuctuations were recorded
when the exhaust and roof fans were off and the ﬂow was driven into the tunnel by its
upstream pressure. Then the exhaust fan was turned on and data were recorded.
Figure 2.8 compares the gas velocity proﬁle inside the tunnel with that obtained based
on Blasius theory [Potter and Foss (1975)]. Figure 2.9 shows that the typical

turbulence ﬂuctuations along the ﬂow direction (U10 are less than one

percent. According to the existing literature on the characteristics of small scale wind-
tunnels [Mitchel and Frobel (1988)]. this is a well reasonable nn-bulence level,

indicating that the ﬂow in the tunnel is laminar.

2.6.2 Temperature Distribution Along Top-Tunnel Glasses

 

Knowledge of the temperature distribution along and across the top-tunnel
infrared glasses is important to prevent them from cracking under high heat ﬂuxes.
Also the maximum temperature of these glasses must not exceed 200°C. Thus the top-
tunnel glasses were ﬁrst replaced by pyrex glasses of the same size and their
temperature distribution was measured while they were convectively cooled by a
planar jet of air. From the calculation of the comparison of the total radiation
absorbed by each type of the glass, it was determined that the maximum temperature
of the IRTRAN II glasses would not exceed the maximum permissible temperature
under similar conditions. The temperature gradient across the glasses was also within

the permissible range that was speciﬁed by the manufacturer.

 

 

44

2.6.3 Mixing of the Pyrolysis Products

 

Extraction of a well mixed representative sample of the gases produced during
pyrolysis is of great importance for chemical analysis. This was one of the most
difﬁcult tasks of this work. The ﬂow of the gases in the tunnel tended to be highly
stratiﬁed. Most of the product gases were carried by the free stream close to the
bottom of the tunnel. To quantify the mixing achieved in the mixing chamber, detailed
tests were performed by injecting a small known amount of pure methane after the
sonic nozzle into the air ﬂow. This methane was completely mixed with the air during
its travel in the pipe connecting the sonic nozzle to the tunnel inlet. The mole ﬁ'action
of the methane in a sample of this ﬂow was recorded by the gas chromatograph. Then
the same amount of methane was introduced at the sample location to be mixed with
the main ﬂow in the tunnel. The recorded mole fraction of methane in this ﬂow was
compared with the previous one which was considered as the reference point. The
difference between the two measured methane concentration was less than 5%. These
measurements were used to improve the mixing of the gases in the mixing chamber
(see Section 2.2.4). The ﬁnal result was satisfactory. The mixing was very good

specially at low ﬂow rates.

2.7 Calibrations

 

2.7.1 Radiant Heat Flux

 

The incident radiant heat ﬂux at the sample location was calibrated against the
incident radiation at the inlet of the test section (Figure 2.10). This was done both with
and without the top-tunnel infrared glasses. The calibration curves in both cases were
linear and were in good agreement. This indicated that the effect of geometry or
refraction of radiation on the heat ﬂux readings due to the presence of the glasses was

very small.

As she
hat ﬂux g3
of coarsest
as int-csog
is blowing

I I
&' mm?
“We:
‘7

2.72 Ga:
\

ill: at:
"in; Small
but ﬂuxes.
NEW 0}
at SW

at Condensi

mlil‘r‘xnv
m.

 

As shown in Figure 2.10, the thermal boundary layer starts after the water cooled
heat ﬂux gauge. Thus, this gauge measures only the incident radiation. The possibility
of convective cooling of the heat ﬂux sensor clue to the ﬂow of gases inside the tunnel
was investigated. The change in the heat ﬂux measured by this heat ﬂux gauge due to
the blowing of air, was negligible (maximum 1.5%.) This was accounted for when

determining the radiative heat ﬂux at the sample location.

2.7.2 Gas Analyzers

 

The amount of gases produced by thermal decomposition of wood samples were
very small compared to the main ﬂow in the tunnel (less than 5%), specially at low
heat ﬂuxes. Thus, measurement of each chemical component was very sensitive to the
accuracy of calibration of the gas analyzers. Furthermore, the measurement of H20
needed special care. The dew-point hygrometer was extremely sensitive to the presence
of condensible hydrocarbons and to the temperature of its surroundings. This analyzer
was used after the combustion tube where all the fuel was burned and converted to
002 and H20.

Prior to the start of experiments, ﬁrst [Tl-1C], C02. CO, and 02 analyzers were
calibrated using a ’zero gas’ (21.3% oxygen and the balance nitrogen) and a
calibration gas (0.2% CH4, 2.01% C02, 0.195% CO, and the rest N2). Then for the
experiments conducted in the nitrogen atmosphere, ﬁrst nitrogen was ﬂown into the
tunnel while an inert block (the size of the wood sample) was placed at the sample
location. The dew point and oxygen concentration of the ﬂow was monitored until it
reached a steady value. The steady value of the dew point was taken as the reference
point (zero line) for the moisture produced later during wood pyrolysis. The steady
value of oxygen recorded during this time was used to determine the leakage of the air
into the tunnel. The amount of air leakage (assumed to remain constant during the

experiment) was used in the calculation of volatile mass ﬂuxes. All the recorded data

46

during pyrolysis were corrected for this leakage.

2.8 Experimental Procedure

 

As the dew point and oxygen readings reached steady values, indicating that the
initial conditions for the experiment were all known, the wood sample was placed and
positioned properly such that it did not touch the tunnel. This was done as quickly as
possible to: (i) minimize further leakage of the air into the tunnel and, (ii) to
minimize the surface desorption (or adsorption) of moisture out of (or into) the wood
sample. Then, as soon as the analyzers reached steady readings again, the heating of
the sample and the data collection was started. A schematic diagram of the thermal
decomposition process occurring in the tunnel is shown in Figure 2.10. Here, it should
be emphasized that, if a uniform external heat ﬂux [Figure 2.7] is maintained over the
entire surface area of the wood sample, the mass ﬂux of the products evolving from
wood would also be spatially uniform [Figure 2.10]. As shown, the velocity boundary
layer starts considerably ahead of the entrance edge of the sample and the thermal
boundary layer starts after the heat ﬂux sensor which is water-cooled. Thus, the
convective heat transfer is nearly constant along the length of the sample and the
system is convectively well deﬁned. Also, the heat ﬂux gauge is not convectively
affected by the incoming ﬂow which is at the same temperature as the gauge, thus it
records only the incident radiation.

2.9 Data Reduction

 

The data collection by the computer was started a few seconds before exposing
the wood sample to the external radiation. This was done to record the initial
conditions (reference points) of the data recorded by the gas analyzers during the
experiments. The data for weight loss was recorded in one data ﬁle and all other data

(temperatures, heat ﬂux, and chemical measurements) were stored in another ﬁle. As a

Jr W «or W

ZA v~.—.<-A ~<V~ t~<ZV~m§Px-

 

42:5. 05 59.5280 $80.5 5.233800% 2: Lo ouansom aﬁn 953k

  

vUOAm awn—n50 amt—000 MESS
E mOmme X34”. Ream—I

 

 

 

 

 

 

 

 

a
M
ZOF<JDmZ~ \
«:2, . \ Hanna}
\\
Q \ V»
t E \\ g \
NZ. \\\.&JA$¢ \
\\ \ .J \ \ l 00:
\\\\\ a - it \.\ /
\ \\ \\ \I it .13 >.:oo.._m>

\ w W X

ZOF<~Q<M A<Z¢Evnm

 

48

ﬁrst step, these data were converted into a tabulated form (value vs. time) and were
plotted for a preliminary analysis. From these curves the offset (base line) of each
chemical measurement was determined. This was used for the ﬁnal data processing and

calculations which were performed in several steps.

There were two types of disturbances in the experimental data which had to be
distinguished from the actual pyrolysis data. The ﬁrst type, was the electrical noise
that existed in most of the gas analer outputs. The second type, were the
disturbances caused partially by the ﬂuctuations in the tunnel ﬂow and partially by the
motion of the ambient air due to the exhaust fan and air circulation in the room. The
latter mostly aﬁ‘ected the weight loss measurements. The output signals, therefore, had
to be smoothed in such a way so as to preserve the overall characteristics of the
phenomena while ﬁltering out the noise. As stated by Atreya [1983], to do this on a
computer, we must use a procedure ﬂexible enough to keep the important trends, but

not too ﬂexible so as to include the noises.

The smoothing was also necessary to obtain the weight loss rate (pyrolysis mass
ﬂux) and other derived quantities [Chapter 3]. In addition, the data is never recorded
simultaneously for all channels, and yet, for calculation procedure the value of all
signals at speciﬁed time intervals is needed. This was done by interpolating the data by
ﬁtting appropriate polynomials through them. The procedure used for lag and response
time corrections, smoothing, and curve ﬁtting is the one used by Atreya [1983] and is
described in Appendix A.

The ﬁnal output of this data processing procedure was a tabulated form of all
measured quantities corrected for offsets, time lags, and response times at speciﬁed
time intervals. Further calculations were done to obtain derived quantities like
empirical chemical compositions of the pyrolysis products and the char yield. These
are described in Chapter 3. Finally, these corrected data and derived quantities were

plotted using a plotting routine called PlotlT furnished by the College of

Ezglnt

WE“
ﬁbmr u

r—l

r
be

49

Engineering. The data processing and plots were carried out on a Micro VAX

computer.

2.10 Experimental Errors

 

The difference between the true value (unknown) and the result of measurements
(best experimentally determined value) is the error. Generally, two types of
experimental error may be recognized; systematic errors and random
errors. Systematic errors are repetitive and usually of ﬁxed value such as calibration
errors that occur consistently every time the measurement is made. These errors are
not susceptible to statistical analysis. Random errors are of inconsistent nature and may
change ﬁom one experiment to another, such as some errors due to the nature of the
material being investigated and environment and human errors. For example, electrical
noise in the output signal of gas analyzers or disturbances in the weight loss
measurements caused by the motion of the air, are of this type. These noises can be

removed by a data processing routine.

There are several sources of experimental errors in the pyrolysis
experiments. Different types of error may assume different signiﬁcance during the
course of wood pyrolysis. For instance, at early times, wood decomposition occurs
very rapidly and small errors in the weight loss measurement (caused by air ﬂow
disturbances) may translate to large discrepancies in the pyrolysis mass ﬂux that is
numerically calculated ﬁ'om the weight loss data. At larger times, shrinkage and
cracking of the solid changes the effective surface area of the sample, which in turn
affects the calculated total mass ﬂux. In addition, the derived quantities of interest,
such as char yield, are often determined using several measured values. They

incorporated all errors existing in those measured quantities.

The errors encountered in the experimental results can be categorized in three

broad groups [Atreya (1983)] as:

/‘N

The

Wet:

Sum

50

(i) Errors in the measurements of basic quantities;
(ii) Errors in the derived quantities;

(iii) Errors due to the response time of the instruments.

2.10.1 Errors in Measurements

 

In every experiment three types of data were obtained. (i) Weight loss data; (ii)

Chemical measurements; (iii) Temperature measurements.

I . Errors in Weight Measurement

 

Weight loss measurements were disrupted by disturbances induced by the ﬂow of
gases inside the tunnel and the motion of air in the room. These ﬂuctuations caused a
random error in weight loss measurement. The electronic balance was able to measure
the weight up to 0.01 grams with an accuracy of $0.005. The typical disturbances were
in the range of i002 grams. The noise in the weight loss measurements was highly
magniﬁed in the calculation of the weight loss rate which was obtained by numerically
differentiating the weight loss data. The data smoothing scheme (Appendix A)
maintains the continuity of the data but it is not possible to fully maintain the
continuity of their derivative with respect to time. Thus, sudden changes in the
measured quantities may cause discontinuities in the mass ﬂux curves. The maximum

error in the calculated weight loss rate was less than 120%.

2. Instrument Errors

 

Several sources of error existed in the measurements of the chemical species.
The errors due to the electrical noises, calibration, and the accuracy of each analyzer
were specially important at the times when the measured quantities were very

small. For instance the lower range of the C02 analyzer was 06% while the gas

CUE

WC}

I.

:3
re

51

concentrations ranged ﬁom zero to 2%. At low heat ﬂuxes the concentration of gases
were usually very small and became comparable with the accuracy of the instrument
(i196 full scale). In the measurement of total hydrocarbons [THC] another error was
introduced due to the use of the ﬂame ionization detector (FID). This instrument
basically responds to the presense of C atoms in the sample stream and is usually
calibrated with pure methane. The response of FID per carbon atom is altered if
oxygen atoms are present in the sample gas. For the calculations of the mass ﬂow rates

of [THC] the measured concentrations were assumed to be equivalent to that of CH4.

3. Errors Due to the Nature of Experiments

 

Some systematic errors were caused due to the nature of wood and pyrolysis
products. For instance, the condensation of the heavy molecule hydrocarbons (tar) in
the tunnel and in the tubes, slightly changed the concentrations of other gases reaching
the analyzers. However, as explained in Chapter 3, since the pyrolysis products were
highly diluted with the carrier gas in the tunnel, the contribution of this error was
negligible. Also, internal pressrues generated inside the wood and the char matrix
(Chapter 3) may cause random local variations in the weight loss data as the gases
evolved from the wood surface. These variations could not be distinguished from

weight losses due to the evolution of water or fuel volatiles.

Other errors in the measured chemical quantities were due to the imperfections of
the experimental apparatus. A possible source of error was the imperfect mixing of the
gases before reaching the entrance of the sampling probe. The error due to the
incomplete mixing of the gases was not more than :l:5%. Another type of error was
introduced due to the leakage of the air into the tunnel specially at higher heat ﬂuxes
when the top-tunnel glasses had to be removed. The ﬂow rate of this incorniug air was
determined using the tracer gas technique and was accounted for in the

calculations. The accuracy of the results depended on the accuracy of the measurement

such a
"ﬂies.
a bios
“’Oiudc

comP03

A?an

52

of the leakage. In the experiments conducted in the nitrogen atmosphere, since the
amount of air leakage was very small compared to the main ﬂow (less than 5%), its
effect on the calculated mass ﬂuxes was very small. In experiments performed in air
this effect was more signiﬁcant because more leakage had to be allowed into the
tunnel (up to 20%).

4. Errors In Temperature Measurements

 

During pyrolysis, as the wood chars it cracks and shrinks and its surface recedes
gradually. The location of the surface and in-depth thermocouples, therefore, changes
as the pyrolysis continues. This effect is more at higher external radiations and also is
larger in air than in nitrogen atmosphere due to char oxidation. The position of
thermocouples can be measured at the end of experiment. A linear correction may be
assumed for the intermediate times. However, the change in the heat ﬂux at the wood
surface due to surface regression was not signiﬁcant because the regression was not

more than a few millimeters even at high heat ﬂuxes.

2.10.2 Errors in the Derived Quantities

 

From the equations for the char yield, chemical composition of volatiles, and heat
of combustion of pyrolysis volatiles (Chapter 3), it is seen that the errors in the
measurement of the pyrolysis mass ﬂux and the gas concentrations were accumulated.
The systematic errors due to the assumptions made in derivation of these quantities,
such as a constant compositions of char and wood, also affected the determined
values. The accuracy of these calculation was experimentally determined by injecting
a known fuel (propane) into the tunnel at the sample location which simulated the
evolution of volatile gases from the wood. The total mass balance and chemical
composition of the products of combustion was calculated. The results are presented in
Appendix C [Figures ERxx] and are discussed in chapter 3.

Chapter Three

Pyrolysis of Wood

In this chapter the pyrolysis experiments are described and the results are
presented and discussed in detail. In sections 3.1 through 3.5 preliminary experiments,
combustion efﬁciency test, difﬁculties in the pyrolysis experiments, experimental
conditions and derived quantities are described. Section 3.6 analyzes the data on
pyrolysis mass ﬂux and examines the effects of all variable parameters in
detail. Products of pyrolysis are discussed in section 3.7, and section 3.8 describes the
time integrated mass of these products. Char yield and empirical chemical composition
of pyrolysis products are examined in section 3.9. In section 3.10 the results of heat of
combustion of pyrolysis products are presented, and ﬁnally the sample temperature

proﬁle is discused in section 3.11.

53

"NF ]
EU“

"0'
bag

54

3.1 Preliminary Experiments

 

As mentioned in Chapter 2, a large number of experiments were performed in
order to obtain reliable and accurate weight loss and chemical species
measurements. The main focus was on obtaining a reasonably acceptable overall mass
balance within every single experiment. That is, the sum of the mass ﬂux of all
volatiles leaving the wood sample had to be equal to the pyrolysis mass ﬂux measured
by the electronic balance at any time. The major difﬁculty encountered was the
condensation of the heavy molecular weight hydrocarbons (tar) that left the wood upon
thermal decomposition. The importance of this parameter and its effect on other
chemical measurements were not originally known to the investigator. The presence
of condensible tar in the sample stream signiﬁcantly affected water
measurement. Initially, water content of the pyrolysis products was measured by a
dew-point hygrometer. This analyzer measures the dew-point temperature of water in a
stream of gases by letting the water condense on a small ﬂat gold mirror (sensor) that
is cooled (or heated) by a heat pump. The meter, however, can not distinguish between
water and any other condensible molecules. As a result, presence of even minor
amounts of condensibles in the sample stream would considerably alter the water
measurements. The readings were usually higher than that was expec:ed. Even
ﬁlteration of the sample stream did not eliminate this problem. Kashiwagi et al.(l987)
have also encountered similar difﬁculties. After extensive investigations, a new water
meter 'was purchased and employed. Although the new system was also somewhat
sensitive to the presense of high concentrations of contaminants, however, by proper
ﬁltering of the sample gas stream (which removed larger particles without affecting the
response time of the analyzer), the water measurement data were much more
quantitatively accurate. The system is described in Chapter 2. Thus transient
quantitative measurement of permanent gases {[THC], C02, CO] , H20 and 02

depletion were successfully obtained before and after catalytic combustion tube

55

(Section 3.2). The mass ﬂux of condensible portion of the pyrolysis products was then
calculated by difference, i.e.,

that=rhp—(m. m1Q+rhC02+thC0+IhH20) (3'1)

3.2 ’Combustion Efﬁciency’ Experiment

 

A sample of the tunnel gases was extracted and passed through the high
temperature catalytic combustion tube where it was burned with oxygen producing
C02 and H20 and trace amounts of CO and [THC]. The catalytic combustion tube is
described in Chapter 2. As will be discussed later, from the measurements of products
after catalytic combustion, the char yield and empirical chemical composition of the
pyrolysis gases were calculated. The relevant question was "how complete is the
combustion process in the tube", and whether all reactants were oxidized or not. Since
the pyrolysis products contain many unknown hydrocarbons, this question could be
answered only by using a gaseous fuel with known chemical composition. The
calculated chemical composition of the fuel was compared with the actual composition
to obtain a measure of the efﬁciency of the combustion process in the catalytic

combustor. Propane ( C3H8 ) was found to be a very suitable fuel for this purpose.

A known ﬂow rate of propane was introduced into the tunnel at the wood sample
location. This closely simulated the evolution of the volatile products ﬁ'om the wood
surface during pyrolysis experiments including any effects of imperfect mixing of the
tunnel ﬂow before reaching the sampling probe. The data were collected for a few
minutes without heating the combustion tube and then power was suddenly applied to
the heating element resulting in the combustion of the propane and producrion of C02
and H20 and depletion of 02. This corresponds to t=150 sec in Figure 3.1. As

shown, during the warm up time of the heating element, fuel concentration gradually

" ".
bUh

mil

”"’

ugh

CUE

a]; ‘

56

decreases while concentration of C02 increases. The corresponding increase in
concentration of H20 and depletion of 02 is shown in Figure 3.2. Initially, temperature
is not high enough for complete oxidation of propane, thus small amount of CO is
generated which diminishes as temperature rises. The steady burning of propane
continues until the power is cut-off at t==570 sec where the combustion process stops
and fuel concentration increases. In this experiment due to the lack of equipments (a
very sensitive pressure regulator or control valve) it was very difﬁcult to maintain a
constant ﬂow rate of the fuel. The ﬂow rate of propane slightly decreased during the
test (see Figure 3.1). This is the reason for the difference between the initial and ﬁnal
concentrations of [THC] in Figure 3.1. The data shown in Figures 3.1 and 3.2 are the
recorded gas concentrations and are not corrected for the zero shift, lag and response
time (Appendix A). Assuming the chemical composition of the fuel (treated as an
unknown) to be CXH (normalized with respect to the number of H atoms), then the
combustion process within the catalytic combustor can be represented by

CXH + 02 -> C02 + C0 + [TI-1C] + H20 + 02. (3-2)

This expression is merely a symbolic representation of the reaction occurring in the
combustor and does not represent an actual chemical reaction equation. Here, [THC]
represents the unburned portion of propane. However, the total hydrocarbon analyzer
(FID) was calibrated with CH4 and the resulting data indicates the mole fraction of
the leftover fuel "equivalent" to CH4. From the measured mole ﬁactions the mass ﬂux
of the products (taco, moo, mung] and mm] were calculated. These were used to

determine x from carbon balance as follows

57

_1_ _1_ 1 .

44 ””2“"— 28 mm“ '1? mm“

x = l . (3-3)
[inf-12(—1-m¢02 +_ 218 mco +_ 16 1111.120”

 

Here m, represents the mass ﬂux of propane. The calculated values of x (multiply by
8) are shown in Figure 3.3. The sudden discontinuities appearing at the beginning and
end of the combustion process, i.e. times around 150 and 570 seconds, are caused by
the very sharp rise (or fall) of the gas concentrations due to suddenly applying (or
cutting-off) of the electrical power. This results in small over-prediction of gas
concentrations at these "cut-off" times (after correcting for lag and response time - not
shown here). These errors in the mass ﬂow rates of chemical species accumulate in the
calculated value of x through equation (33) and cause large discrepancies at the cut-
off at times in x values (Figure 3.3). Beyond that, the predicted numbers of carbons
are very close to the expected constant value of 3 in C3H8 and the overall agreement is
quite satisfactory. The total mass balance, i.e. sum of the mass ﬂuxes of C02, CO,
H20, and [THC] is also plotted against input mass ﬂow rate of propane (Figure
3.4). The difference is less than 10%. This demonsu'ates that the overall performance
of the catalytic combustion tube is satisfactory and the values calculated from the

measured quantities are accurate within the errors of experiments (i.e. i10%).

3.3 Pyrolysis Experiments

 

The main parameters of the pyrolysis experiments were: external radiation,
oxygen content of the ambient atmosphere, and the moisture content of
wood. Experiments were conducted in two major groups: (1) in inert atmosphere , and
(2) in air. In each group three different moisture contents of wood, (i.e., oven-dry
(considered as containing 0% adsorbed moisture), wood at "room" condition (8-9%
moisture content), and "moist" (about 17 % moisture content) were examined under 4

different external radiation levels of 1, 2, 3, and 4 W/cmz.

.~.\a..,

\/-.9-

-\

I.

-\ I!

i

If\\

GAS CONCENTRATIONS 3:

H20 concentration

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l .O r I I I I I I ' I
- — THC
-- C02 i
I “““““““““““ ~ co
0.8- f \\ 4
i I
I I
0.6- I -
l
I
0.4- -
l
l
0.21 l ‘
I. I
0.0 "I 3,. """ :"","“'L' ‘; """""" .
O 100 200 300 400 500 600 700 80C
TIME (sec)
Fig. 3.1 C3H3, C02, and CO in the combustion of C3H3.
1.6 . , . . I p a . , . , . , . ,
‘ 11 ,’ -““"'“‘ -21.2
1.4" / “ ..
" I, ‘ ”21.0
"2" ' ~ lzo a O
. f .‘ _ ’ N
1.0- : I :20.6 8
i l t -20.4 (a,
0.8“ l l ’ t0
. F I —20.2 3.
\ - '1
0-5- \ -2o.0 2.:
. I, ‘ \ ' O
0.4- I J \4 _-19.a :1
‘ O -19.6
0.2- - _ __ - " 2 -
0.0 I ' V I I l U I I I I l I I
O 1 OO ZOO 300 400 500 600 700 BOO
TIME (sec)

Fig. 3.2 H20 and 02 in the combustion of C3H8.

7,— _

a..2av-<

7: 1 _Z<C

IIV

3;

4‘

1.

A1....d\?....v

A g

4:}; >55. ~ \

.3133

:31:

ﬁg 3.4

O)

59

 

U"
1

4s
1

,1; OF CARBON ATOMS IN CXHB
'1’

 

U
l
B

 

 

 

I .
100

I I
200

I r

300 I 400 l 500
TIME (sec)

. I ,
600 700 800

Fig. 3.3 Number of carbon atoms calculated from products of combustion of C3H3.

40

 

(A
O
l

1

TOTAL MASS FLOW RATE (mg/cm’)
N
o
l

 

 

I r If I T

 

 

 

10— I", _
(P cm
W
O T I I r I I r I m I I F ' r *
O 100 200 300 400 500 600 700 800
TIME (sec)

Fig. 3.4 Total mass ﬂux of reactants and products of combustion of C3H3.

exp

III

60

The maximum external heat ﬂux was limited due to the capability of the
experimental facility. Moreover, although the heaters were warmed up for a long time
(4 to 8 hours) before the start of the experiments, it was very difﬁcult to maintain a
very constant heat ﬂux and, more importantly, closely reproduce the same level of
external radiation for similar experiments. This was because some experiments had to
be conducted under special condition i.e. without the tunnel-top infrared glasses (which
will be discussed later), and possibly there were variations in the current of the main
power supply to the laboratory (from time to time). However, for the sake of
simplicity and convenience, the experiments are nominally coded for their
corresponding heat ﬂux levels at each moisture content of wood in either air or
nitrogen atmospheres. For example, the experiment conducted on dry wood at 2
W/cm2 in nitrogen is called experiment ’DZN’. The experiment conducted on wood
conditioned in "room" moisture condition at 4 W/cm2 in air is termed as experiment
’R4A’. The experiment on wood with 17% moisture content (herein termed as "moist

wood") at l W/cm2 in nitrogen is considered as experiment ’MlN’.

In addition, to improve the weight loss measurements, most of the experiments
were conducted with only front and back surface thermocouples installed. Separate
experiments with samples instrumented with in-depth‘ thermocouples were carried out
for in-depth temperature measurements. These experiments are coded with an
additional letter ’T’. For example, TD3A represents an experiment on dry wood at 3
W/cm2 in air with only temperature measurements. These nominal classiﬁcation of the
experiments are summarized in Table 3.1. Due to the large amount of experimental
data, most of the results are presented in Appendix C. In the following sections
wherever a reference is made to the Figures from Appendix C it is explicitly referred
to this Appendix. Otherwise the Figures are shown in this chapter.

61

 

 

 

 

 

 

 

Table 3.1
CODES FOR PYROLYSIS EXPERIMENTS
Atmosphere Nitrogen Air
Moisture Content Dry 8-9% 17% Dry 8-9% 17%
g 1 DlN RlN MlN DlA RlA MlA
E
548‘ 2D2N R2N M2N D2A R2A WA
3 a
a
g E 3 D3N R3N M3N D3A R3A M3A
£5
4 D4N R4N M4N ‘ D4A R4A M4A

 

 

 

 

 

 

 

 

 

 

 

3.4 Experimental Conditions

 

Unfortunately, even after adding extra heating coils to the external radiation
source I RH in Figure 3.2], it was not possible to obtain 4 W/cm2 at the sample
surface in the presence of the infrared glasses. Thus, for experiments conducted at
more than 2 W/cm2 external radiation the top-tunnel infrared glasses were removed.
This allowed us to obtain more than 4 W/cm2 heat ﬂux on the wood sample. In this
condition, the 1" gap between the top-tunnel and the cooling plate [Figure 3.3] was
sealed with insulating boards, and the tunnel ﬂow rate and the exhaust suction were
adjusted to eliminate leakages of volatile pyrolysis products out of the tunnel. The
boundary layer over the sample surface was not signiﬁcantly altered in this condition.
Small leakages of the ambient air into the tunnel were measured by tracing gas
technique prior to the start of every experiment and were accounted for in the

subsequent calculations.

V. a.
be

3X?

10‘?

f
I

SA!
'1-

I

GOT“.

\‘Uc

62

In the absence of the top-tunnel glasses, however, the inside of the radiant heater
module was likely to be ﬁlled with smoke, specially at the start of the pyrolysis. This
caused partial blockage of radiation and reduced the magnitude of heat ﬂux from its
desired constant values. Figures I-IFxx show the measured heat ﬂux values during the
pyrolysis experiments. The maximum reduction in the heat ﬂux was about 15% at the
early seconds of the experiments performed at 3 and 4 W/cmz. The condition
improved as the pyrolysis continued and as the tarry contents of the pyrolysis products
were reduced. The external radiation was much more constant at l and 2 W/cm2
experiments as can be observed from Figures HFxx (maximum change less than
10%). The slight fall off of heat ﬂux in some of these cases was due to the cooling of
the heating rods [Figure 3.3] by the air jets used to cool the inﬁared glasses.

3.5 Derived Quantities

 

The measured output of the continuous gas analyzers were corrected for lag time
and instrument response time [Appendix A] to obtain mole fractions of the measured
species as a function of time when they were produced. Since the instruments used to
measure 02 , C02, and CO after catalytic combustor and CO on the direct
measurement line required a dry sample stream, the measured values were on a dry
basis. To obtain the real mole fractions, they were converted to a wet basis using the

measured mole fraction of water, i.e.,

( mole ﬁaction on wet basis ) = ( measured mole fraction on dry basis )

x ( 1.0 - mole fraction of water removed by the drier ). (3-4)

The corrected mole fractions of all the species (major constituents) were then

converted into mass production rates by using the exhaust mass ﬂow rate [mass ﬂow

talc C
Incl:

one

O“

Io'rscr
cornbt
s cor
amour.

porno

Where,
LT
13° iii-r
film ,., .
a,
may be

63

rate of nitrogen (or air) determined from the sonic nozzle relations + mass ﬂow rate of
pyrolysis products determined ﬁ'om weight loss measurements + mass ﬂow rate of

oxygen (for experiments in nitrogen) + mass ﬂow rate of air leaking into the tunnel].

Methods of determining the empirical formula of the pyrolysis products [Susott
(1982), Parker (1985)] are in principle similar to the method adopted in this work
(described below). In these methods, which are used to determine the heat of
combustion of volatiles and the kinetic parameters as well, a small ﬂow of the volatiles
is completely oxidized in a catalytic converter to carbon dioxide and water. The
amount of oxygen consumed in this process is measured and used to calculate the

potential rate of heat release and the heat of combustion.

In this work, the difference between the input mass ﬂow rate of 02 before
combustion tube and mass ﬂux of 02 after combustion tube is the amount of oxygen
consumed by the complete combustion of the pyrolysis products. This difference
multiplied by the nearly constant heat of combustion per unit mass of 02 [Hugget
(1980)], determines the heat of combustion of the pyrolysis products as a function of

time

AH, = onz . AHc 66)

where, All, is the heat of combustion of fuels per unit mass of oxygen.

Instantaneous char yield and the time varying empirical chemical composition of
the pyrolysis products are calculated as follows. Let CuHVO represent the empirical
formula for wood, and CquO represent the empirical chemical formula for char (these
may be obtained from elemental analysis), then the thermal decomposition of a gram

of wood can be written as

64

lgramof(CquO)—>chramof(CquO)+(1-Yc)gramofoH,O (3-6)

where, Yc is the char yield (grams of char formed per gram of wood) and CxHyO is
the empirical chemical composition of the pyrolysis products.

These pyrolysis products are later burned to C02 and H20 and trace amounts of
CO, (neglecting [THC] which was practically zero after the combustion tube). This

reaction is represented by

CXI‘IYO + 02 —) C02 + C0 + H20. (3'7)

The two unknowns x and y in expression (3-7) are easily obtained from the two

equations for carbon and hydrogen balance in terms of the species mass production

 

 

rateas:
4 . 4.
—mco,+"mco
11 7
x= ; (3-8)
dip-lithe +-3—mco+iu°usi
11 02 7 9 2°
and
_12.
9 mm
y= . (3-9)

. 3 . 3 . l.
nip-[ﬁmcoﬁ7mcoi'3mmo]

The instantaneous char yield can now be obtained from equation (3-6). Since the
composition of wood is unknown and is assumed to be invariant, u and v are known
constants. Hence we are left with three unknowns Yc , p , and q and only two
equations (for the carbon and hydrogen balance; oxygen does not yield an independent

equation because the empirical formulas are normalized with respect to

65

oxygen). However, previous investigations [Atreya (1983)] shows that the char
composition is very nearly constant during the decomposition process and the only
variable effectively changing the composition of the wood volatiles is the char
yield. Thus, p and q can be determined from elemental analysis and assumed to be

constants. Hence char yield can be obtained from the expression

_ (12p+q+16) 12(x-u)+(y—v) _
Y°— [(12u+v+16)]°[12(x-p)+(y-q) (310)

From the calculated empirical composition of the pyrolysis products the amount of
oxygen needed for complete combustion in the combustion tube can be determined

from the reaction

CxHyO+(x+-§---%-)Oz—>XCOZ+%H20, (3-11)

32(x+%--;T)
“1‘50“: (12x+y+16) mp.

 

(3-12)

This serves as a criterion for checking the accuracy of gas composition measurements.

66

Experimental Results and Discussion

3.6 Pyrolysis Mass Flux

 

The mass ﬂux curves represent the single most important characteristic of
pyrolysis of wood, because the ignition, gas phase ﬂame propagation, steady burning,
and extinction phenomena are all primarily controlled by the rate at which the solid
can provide fuel to the gas phase. Also physical and kinetic quantities such as heat of
pyrolysis, activation energy, and reaction rates are obtained ﬁ'om models which best ﬁt
these curves. Experimental data of pyrolysis mass ﬂux, specially the ones that
represent short and long term behavior of the solid, are often lacking. Before
discussing the effects of each parameter on the pyrolysis mass ﬂux the general

characteristics of these curves are described.

The pyrolysis mass ﬂux data was obtained by numerically differentiating the
weight loss data vs. time and dividing it by the original surface area of wood that was
exposed to the external radiation. This was a difﬁcult task because disturbances were
caused in weight loss measurements due to both the nature of the wood and the
experimental conditions. Two sources of disturbances existed in the weight loss
measurements. The ﬁrst included the mechanical ﬂuctuations and small level of
turbulence that exist in any tunnel ﬂow, and are harder to eliminate in small scale
tunnel ﬂows [Michel and Froibel (1988)]. These "noises" were mostly smoothed out by
the data ﬁtting technique described in Appendix A. Care was taken to smooth the data
in such a way that the fundamental characters of the phenomenon was preserved. The
second source of ﬂuctuations in the weight loss data originated from the nature of the
wood and the physics of pyrolysis. It has been shown [Lee et al. (1976), Fredlund
(1988)] that high pressures are generated inside wood as the pyrolysis zone propagates

67

within the solid. Moisture and fuel volatiles travel outward through the pores of wood
and char, and burst out of the surface in an irregular fashion. At higher temperatures
occasionally cracks appear on the char surface and fuel gases are released through
these cracks more freely. Wood is also an inhomogeneous material. Grain direction
and density growth rings [Atreya (1983)] have signiﬁcant effect on how easily the
pyrolysis products are liberated from the front surface. These two groups of
parameters caused some local variations within the pyrolysis mass ﬂux curves (Figures
3.6-3. 17). It is not possible to distinguish between the variations caused by each of
these factors at any time. No further attempts were taken to smooth these curves any

more.

Another common problem with the mass loss curves is that in most experiments
the ﬁrst few seconds of data are not generally reliable. This is due to practical reasons
that have existed in other investigations as well [Atreya (1983)]. For example, in the
preliminary experiments, it was found that a sudden exposure of the tunnel to external
radiation increased temperature and pressure of the gases inside the tunnel, thus
disturbing the weight loss measurements. This situation was signiﬁcantly improved by
modifying the procedure for starting the experiments. A thin ﬂat aluminum plate
slightly larger than the wood sample size was laid down over the sample location such
that it covered the sample without touching it. The cooling plates were removed and
the tunnel was exposed to radiation moments before the start of the experiments. After
a few seconds the ﬂat plate covering the wood surface was removed without contacting
with the sample. This was done successfully in most experiments. However, due to
the extremely rapid pyrolysis and water desorption, disturbances at the initial times,
could not be totally eliminated. To demonstrate the reproducibility of pyrolysis
experiments, three similar experiments were performed on three wood samples. In
Figure 3.5 the pyrolysis mass ﬂuxes of these experiments (wood heated in air at 2

W/cmz) are compared. Considering all experimental problems associated with wood

68

.98. 3:96:33: he E «8&3 N 3 coo? EOE Ham :0
.58:qu so mo 2.: 025 E 3:350 .3: 39: £3.93 Wm .uE

 

A83 m2:
coop com com com com com cow com com 00— o
l? — n — p — . — . — p _ n — . _ . _ 0.0
— E: o
. N E: b n-
n E: x .

 

 

 

 

(s-zuuo/Buu) xms ssvw SISA‘IOHAd

69

pyrolysis, the agreement is encouraging.

In order to capture both the small time and the large time characteristics of the
pyrolysis process, the majority of experiments were continued up to 30 minutes. This
is about twice the duration of similar experiments conducted by Other investigators
[Kashiwagi et al. (1987)]. The mass ﬂux curves for each group of conditions are
shown in Figures 3.6-3.17. These data are also plotted on log-log coordinates in order
to observe some of their characteristics more clearly. The most striking observation
about all mass ﬂux curves conditions is that the gradual decline of the pyrolysis mass
ﬂux following the rapid initial rise, does not show a negative one-half power
dependence on time as is assumed by many investigators and is predicted by simple
theoretical models (Chapter 4). Such dependence is observed rarely and for short
periods such as in Figures 3.18-3.23. For the time duration of most of these
experiments the mass ﬂux, after the rapid initial rise and the subsequent gradual fall-
off, decreases very slowly and in some cases tends to increase and becomes almost
constant. The sudden rise and gradual fall-off type behavior of mass loss curves was
previously observed [Lee et al. (1976), Chan & Kreiger (1980), Kashiwagi et al.
(1987)] and is predicted by numerical models of wood pyrolysis (see Chapter
4). However, in all the experiments performed in this work, the fall off in mass ﬂux
curves does not appear to be proportional to the negative one-half power of time for
the entire duration of the experiment. on Douglas-ﬁr. Although other types of wood
and different grain direction relative to the incident radiation may cause quantitative
variations in the mass ﬂux, it seems unlikely that the characteristic behavior will be
signiﬁcantly different. Atreya (1983) examined many types of wood and this trend,
i.e. decreasing the pyrolysis mass ﬂux proportional to the negative one-half power of
time, was not observed. Kashiwagi et al. (1987) reported the pyrolysis mass ﬂux data
for similar experiments performed on white pine. In their data it appears that the fall-

off of the pyrolysis mass ﬂux is closely related to the negative one-half power of

70

time. For this reason, while for similar experiments (e.g. experiment at 4 W/cm2 in N7)
the mass ﬂux values in the initial times are in reasonable agreement with our data
(considering the two different woods), the large time values are consistently
smaller. Although these data and Kashiwagi’s data (some of the original data were
available) were examined in detail, the main reasons for this difference are not yet
fully known to the author. However, it appears that the ﬂow geometry may be a major
source for this different behavior in the rate of gasiﬁcation. As it will be discussed in
more detail in section 3.6.1.2 (for experiments in air), in their experiments the
boundary layer over the wood sample starts from the sample edge resulting in highly
non-uniform convective heat transfer over the wood sample. The convecrive heat
transfer coefﬁcient along the length of their sample (3.8 cm) is very high and the
sample is convectively cooled. In fact, even at the early times of heating, the results
show that the mass ﬂux in our experiments typically rises faster than their
experiments. As the surface temperature rises, which corresponds to the chaning
regime of pyrolysis process, the effect of this non-uniform convective cooling becomes
more signiﬁcant. Thus, the mass ﬂux falls-off more rapidly than in our experiments
where the boundary layer is almost fully developed and is uniform along the entire
length of the sample. This ﬂow geometry has even a more signiﬁcant effect for
experiments conducted in air, where the char oxidation occurs. This will be discussed
later. A second reason for higher mass ﬂuxes in our experiments could be the
oxidative effect of minor amount of oxygen (typically less than 1%) into the tunnel in
experiments conducted in N2. Of course, no char oxidation was visually observed even
at high heat ﬂuxes and the effect of this small amount of 02 can not be
signiﬁcant. But, possibly some oxygen may have diffused into the hot char pores and

reacted with it. The extent of this effect, if any, can not be quantiﬁed.

71

3.6.1.1 Effect of External Heat Flux

 

Obviously, the rate of pyrolysis (for the same ambient conditions and moisture
content of wood) increases as the external heat ﬂux increases. As shown in Figures
3.5-3.17, the maximum in the mass ﬂux curves occurs earlier for higher heat

ﬂuxes. This is also predicted by theoretical models.

A. Dry Wood

 

For oven-dry wood heated in inert atmosphere, the mass ﬂux rises rapidly and
only one sharp mass ﬂux peak appears (Figure 3.6). This peak corresponds to
approximately 375°C (range of 350-400°C). The initial early peak appearing in mass
ﬂux curve at 1 W/cm2 is not believed to accurately represent mass ﬂux values. This
could be partly due to the desorption of small amount of moisture that might have
diffused into the surface layer of wood during the short period of time that the sample
was made ready for experiment. It could also be caused in part by the previously
mentioned initial disturbances that are signiﬁcant compared with the mass ﬂux values
at low heat ﬂuxes. This problem generally exists at low external heat ﬂux experiments
where the mass loss rate is so low that the experimental errors become comparable

with the actual data.

In air, the experiments performed at 4 W/cm2 heat ﬂux show a very rapid rise in
mass ﬂux followed by a sharp increase in surface temperature (Figure 3.31). The
sample spontaneously ignited after about 82 seconds. This corresponds to the surface
temperature of about 510°C which is followed by a sharp temperature rise of up to
about 700°C. Similar experiments on wood with 8-9% and 17% moisture content also
ignited at surface temperature of about 500 - 540°C. The ignition was delayed due to
the desorption of water. The sample with 8-9% moisture content ignited after 127
seconds, and the one with 17% moisture ignited 150 seconds after the start of heating.

After ignition, the pulsating motions of the ﬂame caused the mass ﬂux to be very

72

oscillatory. These ﬂaming conditions are not of interest here because they can not be
compared with the non-ﬂaming conditions; since, the heat ﬂux impinging upon the

surface of wood changes drastically in the presence of the ﬂame.

B. Moist Wood

 

For moist wood at 4 W/cm2 also only one peak was observed in both air and
nitrogen atmospheres. The ﬁrst small hump in experiment R4A [Figure 3.14] at about
120 seconds is likely to be caused by disturbances and does not represent actual mass
loss. However, for 17% moisture content, the successive steps of drying, thermal
decomposition (tar evolution), and char production can be distinguished [Figures 3.10
and 3.16]. For example, in Figure 3.16 for experiment M4A the ﬁrst peak corresponds
to the maximum moisture evolution rate, whereas the second one reprr sents the
maximum pyrolysis mass ﬂux and the formation of char. The effects of moisture

content of wood are discussed in more detail in section 3.7.2.

3.6.1.2 Effect of Ambient Oxygen

 

In oxygen containing atmosphere char oxidation and exothermic gas phase
reactions between volatile products and oxygen at the vicinity of the surface contribute
to the energy which enters the solid. This energy supplements the external heat ﬂux
so that the mass ﬂux increase is due both to increased pyrolysis and char
gasiﬁcation. Like the external heat ﬂux, the char oxidation heat in conﬁned to a thin
surface zone. However, the char does not oxidize as fast as the inward moving
pyrolysis wave produces char, thus the pyrolysis mass ﬂux falls-off with time in much

the same manner as in pure nitrogen atmosphere.

At lower heat ﬂuxes, under non-ﬂaming condition in air [Figures 3.12-3.17], an
early hump appears in the mass ﬂux curves followed by less rapid rise to a second

larger peak which is followed by very gradual decline of mass ﬂux rates as the

73

pyrolysis continues. This is a very interesting phenomenon. To explain this behavior,
we should note that in these experiments the velocity and thermal boundary layer starts
ahead of the edge of the sample. Thus, the boundary layer is thick and almost fully
developed over the sample. In this situation less 02 can diffuse into the surface of the
hot char compared with a thin boundary layer. Then, let’s compare experiments D2N
and D2A as well as experiments R2N and R2A with each other. In these experiments
the external heat ﬂuxes are almost equal to 2 W/cm2 and are nearly constant during
entire duration of experiment [Figures 3.7 and 3.13]. In both experiments D2N and
D2A it took approximately 100 seconds before the mass ﬂux reached the ﬁrst peak. In
both cases, this corresponds to the surface temperature of about 375-400°C. The ﬁrst
peaks of experiments in air, however, are smaller than those in nitrogen. This is likely
in part due to the chemisorption of oxygen into the surface of freshly produced
reactive hot char at lower temperatures, and in part due to the less complete char
oxidation in the absence of enough 02. Notice that at earlier times, e.g. at =50 sec.,
when the surface temperature is almost equal in both cases (= 300°C) [Figmes TD2N
and TDZA, (Appendix C)] the magnitude of mass ﬂux is nearly equal. As the
temperature of porous char, which is highly reactive, increases and before signiﬁcant
oxidation of char is initiated, oxygen molecules are absorbed into the char surface and
consequently the weight loss rate at this point is relatively lower than the similar
situation in inert atmosphere. Upon formation of char, in experiment D2N the pyrolysis
mass ﬂux decreases rapidly after the peak appears. In experiment D2A the mass ﬂux
slightly decreases (due to formation of char), but begins to increase again with a
slower rate than the early times. This increase continues until about 560 seconds where
the mass ﬂux starts to decline due to the growing char thickness. This peak is slightly

higher than that of D2N experiment but occurs at a much later time.

This phenomenon is closely related to the smoldering combustion and involves

the formation of reactive char by pyrolysis, chemisorption of 02, evolution of CO,

74

C02, and generation of new reactive cites [Shaﬁzadeh (1980)]. This process, which is
accompanied by evolution of incompletely oxidized volatile pyrolysis products at
relatively low temperatures; is usually distinguished from the more rapid and
incandescent combustion of char (glowing combustion) at higher temperatures and in

the presence of more Oz.

This phenomenon does not appear in the work of Kashiwagi et a1. (1987). The
mass ﬂux curves in air do not show such early shorter humps and the sharp peaks of
the curves are always higher and occur faster in oxygen containing atmospheres than
in nitrogen. The reasons are not yet fully known to the author. However, different
experimental conditions are likely to have considerable effect on the pyrolysis
process. In their experiments the boundary layer starts at the lower edge of the sample
which is heated vertically. The boundary layer is thin over the length of the sample
and is even thinner over part of the sample before its thickness grows. Thus, the
amount of oxygen diffusing into the surface of the char from the free stream ﬂow is
higher than our condition resulting in the rapid oxidation of char. The pyrolysis mass
ﬂux, therefore, rises quickly and reaches its maximum. Chan and Kreiger (1980) also
obtained different gas evolution patterns than those of Kashiwagi et al. (1987). This

problem is not yet well understood and needs more experimental veriﬁcation.

The second peak in these mass ﬂux curves {Figures 3.12-3.17] corresponds to the
"thick char" regime [Wichman and Atreya (1986)], in that further decomposition is
hampered by the formation and growrh of an insulating layer of char. The consumption
of the char in the glowing combustion, however, continues. Therefore, after the second
peak in the mass loss curves the rate of weight loss is much higher than the similar
case in nitrogen atmosphere. The surface temperature of the solid [Figure 3.31] shows
an inﬂection point after the start of char oxidation which results in much higher
temperatures in air than in nitrogen. The effects of oxygen on the surface temperature

will be discussed later in more details. Similar phenomena are observed in

75

experiments R2N and R2A, i.e., experiments on 8-9% moist wood at 2 W/cm2 in
nitrogen and air respectively [Figures 3.6-3.11 and 3.12-3.17]. In these experiments
the very early humm are due to the desorption of adsorbed water upon heating of the
sample. The second peak is the time when rapid decomposition leads to the formation
of char, after which the one in nitrogen decreases monotonically. The magnitude of
this peak in experiment R2A is smaller than that of experiment R2N (due to
chemisorption of 07). The experiment in air experiences a very small decline for a
short time after the ﬁrst pyrolysis peak, followed by gradual increase leading to the
second peak. This phenomenon is also observed in experiments conducted under 3
W/cm2 external heat ﬂux in air as well. The second peak is very clear in the
experiment D3A and occurs earlier than in experiment DZA, as expected. The
experiment R3A is not very clear due to the excessive disturbances in the weight loss
data caused perhaps by moisture evolution. The magnitude of the highest peak of the
mass loss rate curves for the same moisture content is much higher in air than the
maximum mass ﬂux in nitrogen atmosphere. The difference is larger at higher heat
ﬂuxes. Thus, the pyrolysis mass ﬂux for wood samples degrading in air is considerably
larger than those decomposing in inert atmosphere, even when the char layer becomes
thick.

To summarize, the pyrolysis mass ﬂux increases with an increase in ambient
oxygen concentration. The chemisorption of the oxygen on the reactive char surface,
smoldering and then glowing combustion signiﬁcantly change the physics of pyrolysis
and the weight loss rate. The actual energy going into the solid in oxygen containing
atmosphere is the external heat ﬂux plus the exothermic energy released by the char

oxidation and exothermic chemical reactions in the gas phase above the wood surface.

76

 

 

 

. “Q . EXP.D3N -
A: -

EXP. DIN

 

 

Pyrolysis Moss Flux (mg/"'cmas)

"‘

. I V I Y I I I T I T 1' I ‘7
0 400 800 1200 1600 2000

 

 

 

Time (sec)
Fig. 3-6 Pyrolysis mass ﬂux for experiments on dry wood in N2.

 

 

 

 

 

 

5 I 1" V l I I f r T 1 I. I I r T I I' ﬁr
- a
mom

”a? 47 -
N.
E u q
\ 34 510.03»: _
3 /
V
x rampart '
3 2-4
u i
1— J 4
ﬂ armour
I 1-i ..

01

 

 

'I‘r‘r'r'rrr'r'rrlr

O 200 400 600 800 10001200140016001800 2000
TIME (see)

Fig. 3.7 External heat ﬂux for experiments on dry wood in N2.

77

 

 

 

 

 

 

 

 

 

 

 

1'0 I l l T T T f r
”v? .,
”é
\u 0.8- _
5 a.
33, ."-.
0.5%. _
_S w
[_1_ .
m M EXP.R4N
8 0.4-. ’\ -
2 r' ' E rum:
.2.‘ 0.2-0: : T— v "
8 I -"
‘ -' EXPRIN
O>_‘ K ﬂ
O-O'F’jI‘T'f'TfT'r'I‘I'r‘
0 200 400 600 300 1000 1200 1400 1600 1800 2000
TlME (sec)
Fig. 3.8 Pyrolysis ma: ; ﬂux for experiments on 8-9% moist wood in N2.
5 l r I ' Iﬁ 1 r I ' r l ' l
I; 40 __ mam _
N ' /
E . ,
ﬂ 3 m m
g .
v /
>< .
3 2-
0. exam
*- 1
Ff .-
0

 

 

 

Fig. 3.9

TIME (sec)

'Trf'l‘l‘l'rﬁﬁ'lll'
0 200 400 600 300 10001200140016001800 2000

External hear ﬂux for experiments on 8-9% moist wood in N2.

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 1.0 . I I I I I I I I I
m. —
NE 0.9- .
o ‘ u—d
\ 0.8- .
O .
E, 0.7—
X _
0.6-
533 ‘ EXPM4N _
0.5-i. f/W
% ":- v L l _ EXPMBN __
< 0.4‘5: '. w \
2 03;. —
{-0- . .' __ EXP.M2N
m 0.2-»'\/"‘\—~—H_, \_ _\ _
>_’I .. - —;
00C (DJ-"k — EXPMlN
)- «D
0" 0-0 I ' ' ' I ' I ' I ' I ' I ' I '
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. 3.10 Pyrolysis mass ﬂux for experiments on 17% moist wood in N2.
5 ' j T l ' I ‘ I TI r I ' I ' 0 I I
BIPJMN
0? 4~ ._ _
“. /"""" exams:
E ‘ a
3
v
X i
3 2“ exam .1
LI. M
5 m *
I 1— .MIN ..
.I
0

 

 

Fig. 3.11

rIrFfI‘rfIrﬁrI'r'Tr
0 200 400 600 800 10001200140016001800 2000

 

TIME (see)
External heat ﬂux for experiments on 17% moist wood in N2.

79

 

 

    

 

 

 

 

 

 

 

 

 

 

”a? 1-5 I r I r
m ° 4 -
E :. .
a
or 1.2- . . ‘
E - , .
v < '. '1
X ° '-.
3 0.9— ,' \x . J
d ' . 'o q :‘e 0:0 ..o J
'. ° ’ . 'A :1;
(fl . V: ’3'. '1’ V 1.: \ZvEXP‘D‘tA 1
(<12 0-5‘ ' EXP.D3A ,‘ "
2 - :. .
., M a o 0.
En. -' ‘ . '. ' ,
(>0- ' EXPDZA “
—J -I
0
g EXPDIA -
0" I I I I f r T I I I #‘ T r
0 500 1000 1500 2000
TIME (sec)
Fig. 3.12 Pyrolysis mass ﬂux for experiments on dry wood in air.
5 . I I . I I I I I ' j T I f ' I
EXP. D4A d
A 4-
N //
E .
ﬂ
34 momma _
E
X 4/ .I
:1 EXP. DZA
Ll. 2 _
'—
f5 - ..
EXP. DlA
:I: 1 q
.I .1
0

 

 

 

Fig. 3.13

‘ITIIIITIIﬁﬁrrIIII'
O 200 400 600 800 10001200140016001800 2000

TIME (sec)
External heat ﬂux for experiments on dry wood in air.

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 1.5 I I I f T
U)
or. 1
. E .
o
\ 1.2- ‘
CJ‘1 .1
E
V 1
X o -*
D 0.9“ "
rd
R4A
(f) ‘\ {\Ei-VV EXPRBA
‘2 I. "
2 :A 1. :k"
(I) 1/ __ A EXPRZA'?‘
Q I v i
_l _ .
g3: ." EXPRIA
)—
0. ,,..,,......,..
500 1000 1 500 2000
TIME (see)
Fig. 3.14 Pyrolysis mass ﬂux for experiments on 8-9% moist wood in air.
5'I'I'I'I'I‘I'I'I'Ir
EXP.R4A
A 4‘ d
N
E d/ ‘
0 EXPJUA
\ _
v
x . ‘
D
d 2_¥ exam -
l.—
Lfr ‘ EXPR
I . 1A _
1..
0

 

 

. . . . . , . , . , . , .
0 200 400 at?) 800 10'001200140016001800 2000
TIME (see)

Fig. 3.15 External heat ﬂux for experiments on 8-9% moist wood in air.

1.5

1.2

PYROLYSIS MASS FLUX (mg/cm2.s)

0.0

81

 

 

 

 

 

 

 

' ‘ ' ' ' ' I ' I ' I r I '
CT 200 400 600 800 10‘001200140016001800 2000

 

 

 

 

 

 

 

TIME (see)

Fig. 3.16 Pyrolysis mass ﬂux for experiments on 17% moist wood in air.

5 ' I l ' l ' l ' I ' l ' l ' l ' l
g“ 4- m. M4A —
E ‘/-—l— .-
\ EXP- Me:
3 3‘//’( d
V
X - .
3 EXP. m
LI... 2-—‘ -
E] . .
I 1 _f EXP. MIA

0

 

 

 

0 ' 2007 400 ' 500 t 800 r10100T12[00t14:00T16I00'18‘00T2000

Fig. 3.17

TIME (see)
External heat ﬂux for experiments on 17% moist wood in air.

82

 

1.000? ‘2‘

0.100-g.

0.010“:

 

LLllll

PYROLYSIS MASS FLUX (mg/"cm2.s)

 

 

0.001 . .,,kﬂ,l , ,.,,,,ﬁ
10 50 100 500 1000

TIME (sec)
Pyrolysis mass ﬂux for experiments on dry wood in N2.

 

Fig. 3.18

 

T I I I—Frf T ' ‘7 T ' ' ' '[
1.0001

0 o 9 0....“ EXP.R4N

: . . . a o 0...... § :‘ EXPR3N

EXP.R2N

l llllll

0.100

1 LLIIIIL

1

0.010

L1

1 l

LLlllllLL

PYROLYSIS MASS FLU}: (mg,/cm2.s)

 

 

0.001 . r .l...‘ .
10 50 100 500 1000
TIME (sec)
Fig. 3.19 Pyrolysis mass ﬂux for experiments on 8-9% moisr wood in No.

 

83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

rm? T V T I I I j
0 3. EXPM4N 3
‘\ L" . : 3 a a I Iran-IIIMA -_A‘:-
0‘ - EXP.M3N ‘ _ ‘
e . /~ M,
EXP.M2N .

>5 0.100—“Ii “M W
_I : EXP.M1N 1
LL 1 1
U7 - .
m -1

< a
2 0 010

(fl I .11
>- .1 1
_J a .
O 4

Q: .4
>_

0- 0.001 . ﬁ.,.ﬁ.., . -,l.,,,r

10 50 100 500 1000
' TIME (sec)
Fig. 3.20 Pyrolysis mass ﬂux for experiments on 17% moiSt wood in N2.
’0‘? r T If r r I r I 1 ﬁ ﬁ I—T I I I TI I
Né 1.00021 ...,-.° 0.00% .3!
$9 3 exam». : 3
E a . EXP.D2A -'
’3‘ 0.100 ' . _ - ' E
_l . _ " .r
L . ' EXPDIA j
m ' I AVA“
m / J
<1: v, -
E 0 010 w —:
<2 '
9” 3
r I
_l ..'
O 1
CE I
[y !
Q- 0.001 _"'T‘ '1" V frrrrr ‘—T_ f I ff r FT—r i
- 10 50 100 500 1000

TIME (see)
Fig. 3.21 Pyrolysis mass ﬂux for experiments on dry wood in air.

 

 

 

 

 

 

”('0‘ I I I r I r I I
N . 1.0001 EXPR4A 1
E : ..W- 3
\ . EXPRBA d ." "'
0'1 . . nail-"'u _M —
é ‘ ' 3 . ___,..---"" EXPRZA ~
>< _ ' _
3 0.100 5 .....\ EXPRIA 3
u. ' ““‘i
U) a ‘
If) .
<3: 1
E
(L) 0.010"E ‘1
9 3 1
_I I
O
D:
>_
0- 0.001 r flrrrrr r v [vrvrr

'10 50 100 500 1000

TIME (sec)

Fig. 3.22 Pyrolysis mass ﬂux for experiments on 8-9% moist wood in air.

 

 

 

 

 

 

 

’5; ﬂ f I I . I ' ' ' ‘ I
"E 1.000-5 EXPM4A . '3
U : . . . . . . . °"000.oeo°..’2 M - :1
‘ EXP.M3A ..." q

E . . exam». _— :“_
E . ..r

v 3’

X

3 0.100-:- EXP.M1A 1
_I 1 A 1
L‘— 11 ~‘ _I
m . i
(f)

<

2 0 010

(g ' 3 —:
U) I 2
>- .1 -1
_l . I
O

11’. ‘ .
>-

O- 0.001 .,.,.., f .l....,

10 50 100 500 1000
TIME (sec)
Fig. 3.23 Pyrolysis mass ﬂux for experiments on 17% moist wood in air.

85

3.7 Products of Pyrolysis

 

Data from transient measurements of the products of pyrolysis of wood are
plotted in Figures Slxxx (Appendix C). These data are also plotted as a percentage of
total mass ﬂux in Figures PSxxx (Appendix C). In theses Figures, xxx stands for every
experiment and is consistent with the nomenclature described previously. As
mentioned earlier, the mass ﬂux of total organic condensibles (tar) was not measured
direCtly. It was calculated by subtracting the sum of the mass ﬂuxes of [THC], C02,
CO, and H20, from the pyrolysis mass ﬂux. This is exactly true for the experiments
in nitrogen atmosphere. In air, however, due to the char oxidation the total mass ﬂux
of the produced gases is more than the pyrolysis mass ﬂux and the mass recovery is
more than 100%. At higher heat ﬂuxes part of CO and C02 are the products of char
oxidation rather than thermal decomposition of wood. Thus, generally for the
experiments in air, mass ﬂux of tar is not shown except at low heat ﬂuxes where the
surface oxidation is not expected to be signiﬁcant, tar is plotted for comparison with
experiments in inert atmosphere. It is important to note that for data obtained in
nitrogen atmosphere, the only errors in the products mass ﬂuxes and their trend are the
experimental (measurement) errors. The sample weight loss is directly measured and
the mass of tar is whatever that has not been measured by the gas analysis
equipment. Thus, these curves represent quantitative transient data for the pyrolysis of
wood in inert atmosphere. In a recent study [Kashiwagi et al. (1987)], water and
therefore tar data, were only qualitative. In most cases, the magnitudes and sometimes

even the trends were not closely conﬁrmed by this study.

In these measurements, all the sample lines were electrically heated at least up to
100°C. Nevertheless, it was not possible to completely prevent the condensation of tar
even inside the tunnel where it was exposed to the external heat flux and the
temperature was even higher. However, since the total mass ﬂux of the pyrolysis

products was never more than 5% of the main ﬂow in the tunnel (nitrogen or air), the

86

changes in the concentrations of the permanent gases due to any tar condensation
before they reached the analyzers were negligible. Thus, the recorded mole fractions
of C02, CO, H20, and [TI-1C] very closely represent the actual mole fractions in the

mixture of the tunnel ﬂow.

In some experiments specially those at high heat ﬂuxes, the water meter reading
on the direct measurement side was slightly shifted upwards. Thus, at the beginning of
every experiment, the line ﬁlter was replaced and the meter was set back to zero. At
the end of the test, the reading of the meter for a "zero gas" was recorded and the

entire curve was linearly corrected for this shift.

For "dry" wood a large portion of the volatile pyrolysis products consisted of
organic condensibles (tar) [Figures SleN]. Although the wood is termed as "dry" the
initial mass evolved (for the ﬁrst few seconds) is water followed by a rapid and large
outﬂow of fuel gases. Thus, most of this water is water of constitution which is a
product of direct pyrolysis of wood and is possibly produced as fuel volatiles pass
through the hot char matrix and are cracked down to lower molecular weight
hydrocarbons, C02, CO and H20. The peak of these curves obviously corresponds
closely to the peak of pyrolysis mass ﬂux curves and occur at temperature about
375°C. In all cases the tar mass ﬂux falls off rapidly and is responsible for the rapid
fall-off of the pyrolysis mass ﬂux after the peak.

In nitrogen atmosphere [Figures SlxxN (Appendix C)] the interesting feature is
the downward turn in the tar evolution which is steeper than the corresponding mass
ﬂux at larger times. For example, from experiment D3N one gets about 0.175
mg/cmzsec tar at 400 sec while the pyrolysis mass ﬂux at this time is about 0.35
mg/cm2.sec (Figure 3.6). This difference is much larger as the moisture content of
wood is higher. For instance, the corresponding proportions in experiments R3N and
M3N are 0.075 to 0.3 and 0.1 to 0.4 respectively. This type of behavior was also
observed by Kashiwagi et a1. [1987]. This is consistent with the phenomenon

87

explained in Chapter 1, that the chemical composition of the pyrolytic products is
altered as they pass through the growing hot char layer by cracking into small
hydrocarbons or by polymerization to char. Thus, as heating continues, more
permanent gases (C02, CO, and noncondensible hydrocarbons) and water are produced
while production of tars generally decreases. An exception was the experiment D4N
(dry wood at 4 Wm2 in nitrogen) as is shown in Figure 3.6. In this experiment the
pyrolysis mass ﬂux and the corresponding tar [Figure SlD4N (Appendix C)] increases
again after the rapid fall-off and a second peak appears at about 1000 sec. An opposite
trend has occurred in the H20 curve (same Figure). This experiment was conducted
twice and both times the second rise in mass ﬂux was observed. It is likely to be due
to the appearance of huge cracks on the char surface at high heat ﬂux that allows more
pyrolytic tar to escape without contact with the hot char. Less condensible molecules
were cracked down and also more radiation penetrated further inside the solid. Thus,
more tar and less water were produced. Later as thickness of char increases, the rate of
tar production reduces again. The same trend, although to a lesser extent, is observed
in Figure SZD4N, which shows the products after the combustion tube. Here, a
second peak similar to that of the mass ﬂux curve is observed in both C02 and H20
curves. This indicates that indeed more fuel has been generated resulting in more C02
and water on this route. Of course, not all of the produced hydrocarbons entered the
catalytic combustor due to the condensation of tar. Thus, the second up and down
trend in these curves are less steep than in Figure SlD4N. This type of trend was not
observed in the work of Kashiwagi et al. (1987). In their experiments the heating
period was 12-15 minutes and they used different type of wood and the heat ﬂux was
parallel to the grain direction (in this work it was always perpendicular.) Of course,
different woods have different densities and rosin contents, and other properties may
differ as well. This observation signiﬁes the importance of the effect of the cracking

of char specially at high heat ﬂuxes. Therefore, cracking of char not only alters the

88

heat transfer between the outgoing volatiles and the hot char, but it also could change
the mass ﬂux as well as the chemical composition of the pyrolysis products. For a
detailed theoretical model of wood pyrolysis, therefore, it will be important to include
the effect of char cracking.

3.7.1 Permanent Gases

 

As the heat ﬂux increases the evolved mass ﬂux of permanent gases ([THC],
C02, CO) generally increases [Figures S2xxN and SZxxA (Appendix C)]. When the
sample is heated in air the evolution rate of all permanent gases is larger than those in
nitrogen atmosphere. The most signiﬁcant increase is in production of C02. Since the
wood samples in experiments at 4 W/cm2 in air all ignited, they can not be compared
with the similar experiments in nitrogen atmosphere. Figure SlR3N and SlR3A
(Appendix C) show such experiments at 3 W/cm2 in N2 and in air on wood with 8-9%
moisture content. At about 900 seconds mass ﬂux of C02 in nitrogen and air are 0.035
and 0.15 mg Icm2.sec respectively. These correspond to 10% and 35% of their
pyrolysis mass ﬂuxes, respectively [Figures PSR3N and PSR3A Appendix C)]. Those
at 2 W/em2 heat ﬂux [Figures SlR2N and SlR2A (Appendix C)] are 0.045 and 0.01
mg/cm2.sec, respectively. These correspond to 2.5% and 14% of their pyrolysis mass
ﬂux respectively [Figures PSRZN and PSR2A (Appendix C)]. Production rate of
[THC] increases only slightly in air and part of it is possibly consumed in endothermic
gas phase reactions occurrng at the vicinity of the char surface. Mass ﬂux of the
evolved [THC] in experiments R2N and R2A at 900 seconds is 0.02 and 0.035
mg/cm2 (corresponding to 6% and 10% of the pyrolysis mass ﬂux), respectively
[Figures SIRZN, SlR2A, PSR2N and PSR2A]. Since the differences in percentage
values are within the ranges- of experimental error, distinct conclusions may not be
derived from these values. CO also shows increase in oxygen containing atmosphere. It

is about 5% and 15% in R2N and R2A experiments, and 9% and 18% in R3N and

89

R3A experiments respectively. In most of these experiments the production rates of
permanent gases (Figures SlxxA) remained almost constant after initial rise. Thus,
since the pyrolysis mass ﬂux (total) was gradually decreasing, the mass fraction rates
(percentages) of production of these components show a gradual increase (Figures
PSxxA). The continuous increase in the production rate of C02 with time that was
reported for wood with 5% moisture content by Kashiwagi et al. (1987), was not
observed in these results. As it was discussed in section 3.6.1.2, in their experiments
the boundary layer started from the edge of the wood sample, whereas, in these
experiments the sample was located in the middle of the boundary layer and far away
from its leading edge. As a result, for similar heat ﬂuxes, the wood surface was
exposed to different amount of oxygen in each case, resulting in different amounts of
products. This shows that when the boundary layer starts from the sample edge, the
process in not convectively (heat and mass transfer) well deﬁned, and the magnitude

and history of the pyrolysis mass ﬂux will be affected.

3.7.2 Water Mass Flux and Effects of Moisture Content of Wood

 

In this section the moisture desorption rate and the effects of moisture content on
energetics of pyrolysis and combustion of the solid is discussed. Although, adsorbed
moisture plays an important role in pyrolysis and in ignition of wood, it has often been
ignored in the theoretical models. Atreya (1983) showed that unless the wood is
completely "dry" the terms related to moisture in the model equations for the pyrolysis
process must be retained. Considerable improvement in the theoretical prediction of
pyrolysis mass ﬂux were obtained by including moisture desorption in his model.

In this work, samples with three different moisture contents, ("dry" == 0%, room
=8-9%, and moist =17%) were examined. All the water measured in nitrogen
atmosphere experiments (direct measurements) represents the water produced during

pyrolysis. This consists of adsorbed moisture and water of constitution. For

90

experiments in oxygen containing atmosphere, a part of the evolved water vapor,
specially at higher heat ﬂuxes, is generated by char oxidation and exothermic gas
phase reactions between fuel volatiles and oxygen above the char surface. For
example, in experiment R3N (8-9% moist wood in N2) the water mass flux at t=400
see is 0.14 mg/cm2.sec (Figure SlR3N), while that of experiment R3A (in air) is 0.2
mg/cm2.sec (Figure SlR3A), an increase of about 40%. Part of this extra water is
produced in the gas phase reactions. Other part is generated ﬁ'om drying and pyrolysis
of the solid due to the heat of char oxidation conducted into the solid. The
corresponding values for experiments R2N and R2A at t=1000 see [Figures SlR2N
and SlR2A Appendix C] are 0.035 and 0.075 mg/cmzsec, respectively.

As mentioned earlier, almost all of the water appearing in experiments DxN (dry
wood in nitrogen) is the water of constitution which is a product of thermal
degradation of wood. At higher moisture contents, as wood is heated, ﬁrst the desorbed
water is evaporated and then pyrolysis starts. The mass ﬂux curve of evolved moisture
is very similar to the pyrolysis mass ﬂux curve indicating that the desorption of water
is also affected by the formation and growth of the char. The adsorbed moisture is a
few molecules thick layer of water molecules that are attached to the cell walls within

the micro-pores of wood via hydrogen bonds.

Moisture content of wood primarily changes physics of pyrolysis and ignition and
its effect on the chemistry of pyrolysis is not signiﬁcant. Some of the effects of
moisture on the energetics of the thermal decomposition and ignition of wood are
discussed in the following paragraphs. Water (depending on the magnitude that is
adsorbed or absorbed) delays the process of pyrolysis and dilutes the products of
pyrolysis of wood. For example, we can compare products of pyrolysis for experiments
conducted on samples with three different moisture contents at -2 W/cm2 [Figures
SlD2N, SlR2N, and SlM2N]. These data indicate that while the trend of production

of each species at different moisture contents are alike, suggesting that they follow the

91

same type of reaction kinetics, the mass ﬂux and time of occurrence of their peak is
different. In these experiments the magnitude and time of maximum mass ﬂux of tar
for dry, 8-9% moist, and 17% moist wood are 0.28 mg/cm2.sec (at t=100 sec), 0.135
mg/cm2.sec (at t=180 sec), and 0.08 mg/cm2.sec (at 210 sec), respectively. That is, the
evolution of tar is reduced and delayed as moisture content of wood increases. An
opposite trend in the production of water can be observed in these Figures. Similar
trend can be observed in other experiments as well as for experiments conducted in
air. As a result, the ﬂammibility of wood is decreased as more water is adsorbed (or
absorbed) in it. As mentioned previously, in experiments conducted in air at 4 W/cmz,
the samples ignited and the ignition was delayed as moisture content of wood
increased.

Comparison of pyrolysis mass ﬂux curves [Figure 3.6-3.11] also shows that at
each external heat ﬂux, the magnitude of the mass ﬂux reduces as the moisture content
of wood increases, while the time of occurrence of the pyrolysis peak (i.e. tar
evolution ) increases. Only at very low heat ﬂux, where thermal deco: I Iposition of
wood is very small, mass ﬂux increases with moisture content of wood. For example,
for experiments at l W/cmz, the weight loss rate increased as wood contained more
moisture indicating that most of the weight loss was due to the moisture evaporation
[Figures SxDlN, SleN, SleN (Appendix C)]. In fact, while for dry wood in
nitrogen atmosphere [experiment DlN, Figures SxDlN], some degree of pyrolysis has
occrured and sorrre tar and small amounts of other pyrolysis products have been
generated, for 8-9% moist wood [ Experiment RlN] mostly water and small amounts
of fuel gases were produced. For 17% moist wood [Experiment MIN], almost all of
the weight loss of the sample was due to the water desorption and very little tar and
consequently small amount of char were produced (Figure SlMlN). The same effects
is observed in experiments in air [Figure SlMlA].

92

The effect of moisture is basically cooling the solid by absorbing energy from it
for its desorption and during its outward migration. A comparison of the surface
temperature of the pyrolyzing solid can demonstrate this effect. Figures TD1N , TR4N
and 'I'I‘M4N (Appendix C) represent the surface and in-depth temperature history of
samples heated in N2 at 4 W/cmz. At t=200 sec the surface temperature of dry, 13-996,
and 17% moist wood are 620, 580, and 460°C, respectively. For similar experiments
conducted at 2 wlcm2 in N2 [Figures TD2N, TR2N, and 'l'I'MZN (Appendix C)], the
surface temperatures at t=200 see are 420, 390, and 360°C, respectively. The presence
of moisture changes thermal capacity of the solid and therefore changes the heat
transfer and hence the temperature rise history of the solid. The thermal conductivity,
speciﬁc heat, and density of dry wood are normally less than those of moist wood.
Hence the surface temperature, which is proportional to (ka)‘°'5 [Carslaw and Jaeger
(1959)], will be higher for dry wood than for moist wood. This will cause delay in the
pyrolysis process (and ignition) of moist wood. Finally, heat is transferred directly by
inward and outward migration of water. Forward migration and condensation of the
moisture ahead of the arrival of the pyrolysis zone, changes wood properties even

before decomposition starts.

3.8 Time Integrated (Total) Mass of Pyrolysis Products

 

Since the pyrolysis mass ﬂux was calculated numerically and tar mass ﬂux was
obtained by difference, the data on tar involves both errors in the measurement of
permanent gas concentrations and errors in pyrolysis mass ﬂux as well. An alternative
way of studying the effects of each parameter on the mass evolution rate of pyrolysis
products is the comparison of the total amounts of each product after a speciﬁed
duration of heating time. The total amounts of each product was obtained by
integrating their mass evolution rate. Here, since some experiments were continued for

only 15 minutes, although majority of experiments were carried out up to 30 minutes,

93

all evolved rates were integrated for only 15 minutes. This also allows us to compare
these results. with the data on other types of woods [e.g. with Ohlemiller et al.
(1987)]. Figures 3.24-3.29 show the results for nitrogen and air. The heat ﬂuxes shown

on these curves represent the average values obtained from Figures 3.7-3.17.

The results indicate that for experiments in nitrogen atmosphere the amount of all
major products of pyrolysis increases with the heat ﬂux, while their corresponding
percentages relative to the total mass ﬂux do not show a similar trend. For dry wood
[Figure 3.24], the percentage of tar shows a slight decrease with increase in heat ﬂux
while more water in generated except for experiment at 4 W/cmz. For experiments at
1, 2 and 3 W/cmz, this is due to cracking of heavy molecule hydrocarbons as they pass
through the hot char. For experiment D4N, as it was described earlier, f0rmation of
huge cracks is likely to be responsible for escape of more tar without contacting the
hot char. As a result, less total amount water is also observed. Increase in the total
percentage of CO is not signiﬁcant, but more C02 is generated at higher heat
ﬂuxes. For moist wood, water accounts for most of the products at low heat ﬂuxes and
it decreases at higher ﬂuxes. For moist wood, production of tar generally increases
with heat ﬂux, except at the highest heat ﬂux [Figures 3.25 and 3.26]. It is because for
moist wood specially at lower heat ﬂuxes, the desorption of moisture is the dominant
process and as heating continues more fuel volatiles are generated. At 4 W/cmz,
however, due to the rapid rise in the surface temperature, the drying process occurs
very rapidly and thermal decomposition of wood continues more similar to that of dry
wood. Again, the total amount of C02 increases at higher ﬂuxes and increase in the
production of CO is not signiﬁcant, except for 17% moist wood in which more CO is
produced at 4 W/cmz. This is because water dilutes the mixture of products and
reduces the solid temperature altering the trend of cracking of tar while passing

through the char matrix.

94

For samples heated in air since all samples at 4 W/cm2 ignited, obviously almost
all of the products are C02 and H20. Therefore, basically only experiments at heat
ﬂuxes lower than 4 Wm2 are compared In all of these experiments because of the
char oxidation the mass recovery is more than 100 percent. Moreover, the gas phase
reactions with oxygen produce some products that are not products of
pyrolysis. Production of CO and C02 show considerable increase compared with
nitrogen atmosphere. Also, generally increase in the production of C02 in air is more
signiﬁcant than changes in 00 yield indicating that CO is mostly a product of
pyrolysis not incomplete oxidation.

The ﬁnal char depths in the pyrolysis experiments are shown in Table 3.2. These
were found by cutting the sample in half in the middle and measuring the char depth
from the original surface of the wood sample to the sharp black char-wood
interface. Often, specially at higher temperatures, gases evolved from the sample even
after removing the external radiation. Thus the values represent the approximate char
depths at the ﬁnal instant. The numbers are correct within :I:0.5 mm. In some cases,
due to the structure of that particular piece of wood, huge longitudinal cracks appeared
which signiﬁcantly altered the normal heating rate and the ﬁnal char depths were much
higher. The general trend is higher char depth at higher heat ﬂuxes. Also more char is

produced in air than in niuogen at similar external radiations.

95
Table 3.2 Final Char Depth For Pyrolysis Experiments at Different Conditions.

(in mm}

(For 30 minutes heating time unless speciﬁed in parenthesis)

 

Atmosphere Nitrogen Air

 

Moisture Content Dry 8-9% 17% Dry 8-9% 17%"I

 

External 2 10.0 5.0 4.0 15.0 15.0 16.0
(32) (20)

Radiation 3 5.0 7.0 6.4 19.5 18.0 19.5
(20) (5)

Mom2 4 18.0 7.3 10.5 24.0 19.5 21.5
(1)

 

 

 

 

* Large longitudinal cracks appeared.

 

 

 

 

 

 

500 ' I I l E I I
g" . Dry wood heated in a
E nitrogen for 15 rain.
0
E 400‘ (ntunbers shown next to '—
5 each bar represent the
V percentage of each product) ,II’ ‘
Z! / ’
g: 500- ,x _
_ol TAR
/ .
s . / ”
Q /
L 200‘ // L13 _
O 3 / ” AL
9 a , ’ ’ '5’ H50 ‘
U /r> / ..
D /
o 100‘ / _ ~ / Ali -'
O / 2 ,/ g /" C0
m 0‘ / v I / / ,0
°' ‘ “/ x“ —---a, ’ ,I’PS co, 1
0 '34 4:2 =39 “r: =7 .4;- " "“ THC
r I '7 r I h ! I
0.0 1 .0 2.0 3.0 4.0 5.0

HEAT FLUX (W/em’s)
Fig. 3.24 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.

 

 

 

 

 

 

 

 

00 t I y I I I I I I
8-92 moist wood heated .
"E in nitrogen for 15 min. .
0 u
E 400‘ (numbers shown next to
E each bar represent the ..
V - percentage of each product) /4\
(n
— / ‘
g3 300- / {g TAR
3 / A
/ a
95 ‘ /
a- //’ //
LI. 200‘ ,/ g / '-
C at” ’ 5 H 0
m J / ”’ 2 d
8 //BAF”
8 100‘ //// o ///
/ / O\ z/ ’
3.: 1 '53 f” o ...- ?” 1:: C02 1
0 Mat-"‘3‘ 5—'4
r ' I ' '
0.0 1 .0 2.0 3.0 4.0 5.0

HEAT FLUX (W/em’s)
Fig. 3.25 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.

 

 
  

 

 

 

 

 

 

 

500 . I . r . I . I
g—\ . 177. moist wood heated .
E in nitrogen for 15 min.
0
E 400— (numbers shown next to ,A\ q
E each bar represent the x/ N TAP
I; percentage of each product) P/ ’ u ‘ ‘
a _l —r
5 300 //{g //,/‘
8 / / /
>" / /r)’
a. / /
LI. 200‘" / / U1 Ch H90 “
O / / \J N .'.
m ’1‘ //
d / cl
8 / / 52/
D , I ’
8 100‘ u : / ’ 83 CO A
t: m _"/
0- ‘ ~° ,AP- —' " "“ co ‘
“ —:___o~¢t._-——m-o 2
0 _:——u¢)-————u-#7mc
T I I ‘ I'
0.0 1.0 2.0 3.0 4.0 5.0

HEAT FLUX (W/cm=.s)

Fig. 3.26 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.
1000 . l .

 

 

 

 

 

 

 

 

l ' I ' r j—
g‘ . Dry wood heated in -
E air for 15 min. MDTAR
o / ..
E 800- (numbers shown next to // \
E each bar represent the // b 320
V ‘ percentage of each product) // U“ '
(D
5 / ... co
0 / x /
c: " / ‘
>- / / /
0.. / //
u_ 400‘ / / -'
o //
, I

g / / é — co .
fﬂ / o / :3 2
0 // u I // a
0 /# I ’ N /

/ C ’ a H
g:- ' / /, ’ N " ’ "9 1
i===Zr ’-—“ :T‘“~
O ' é ' I H r I w 4'74: '
0.0 1.0 2.0 3.0 4.0 5.0

HEAT FLUX (W/cm2.s)
Fig. 3.27 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.

98

 

 

 

 

 

 

 

 

1000 . x . r ' I ' '
g" . 8-92 moist wood heated
E in air for 15 min.
0 d
E 800 (numbers shown next to
E '1 each bar represent the -
V percentage of each product)
In a
'g': 600- ”fun ‘
.01 / /3 .
. /'/ ‘5
E / / H20
/ / _
.,_ 400- 3A0 ... o
0 ’ / //
/ // .
g ‘ // /6' ’/
u / N
D I // I’l/ a ' "'
8 / ‘0 / [I’m
a 'I I /’ ’ a /’/// I 8
G 0 :—-_—'_€ :¢""~dr_--_N C u
0.0 1.0 2.0 3-0 4-0 5°°

HEAT FLUX (W/cm’s)
Fig. 3.28 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.

 

 

 

 

 

 

 

1000 v I W T ' F ﬁ ' ﬂ

5‘ . 172 moist wood heated ‘
g in air for 15 min.
\ 300.. ‘
CED (numbers shown next to
V . each bar represent the j
0, percentage of each product)
ﬁ ..
:3 600 ///= TAR d
0 q / / .
E x ’
0. AI / “ F‘20
O / a / N

" / U. l / d
:2 / , ///
3 II A! /’ U'
o 200‘ /t':; //// N .-
05 y / ’ °° ’ I: ’ 2 .
l m I "I- "" ..U l l ’ df

o ' Ijii‘ r ‘ I h V
0.0 1 .0 2.0 3.0 4.0 5.0

HEAT FLUX (W/cm’.s)
Fig. 3.29 Time integrated mass of pyrolysis products as a function of incident heat ﬂux.

99

3.9 Char Yield and Empirical Chemical Composition of
Pyrolysis Products

 

 

Char yield (Y c) was calculated using equation (3-12) and the results are shown in
Figures YCxxx (Appendix C). The values of x and y in CxI-IyO were determined from
relations (3-8) and (3-9) and are plotted in Figures Cl-lxxx (Appendix C). Figm'es
SZxxx show the mass ﬂuxes of the products of combustion of pyrolysis volatiles in the
catalytic combustion tube. These products are mainly C02 and H20, and the amount of
CO is usually negligible indicating nearly complete combustion. Only at the start of
the combustion for a short period of time higher amounts of C0 are detected. This was
because the mixture of fuel gases and oxygen was not yet fully ignited. The most
signiﬁcant source of error in these data is the condensation of the heavy hydrocarbons
(tar) before the sample gas enters the combustion tube. As described in Chapter 2,
considerable care was taken to minimize the condensation by heating all the
components of the tunnel and the connecting lines. Nevertheless, some tar always
condensed on the interior surfaces of the tunnel and the sample lines and never
reached the combustion tube. To determine the extent of this partial loss of tar, the
tatal mass balance was calculated from the measurements of C02, CO, and H20 after
the combustion tube. For this calculation the amount of oxygen consumed in the
combustion process was determined from the difference between the mass ﬂuxes of
oxygen before and after the combustor. Figures TMxxx (Appendix C), represent
comparison between the measured pyrolysis mass ﬂux and the computed total mass
ﬂux of the gases [( tin-:0z + mm + mm) - mm; after combustion. Figures ERxxx
(Appendix C) show the percentage of the difference between the two curves. As
expected, the highest degree of error in the total mass balance occurs at the early times
of the pyrolysis process where most of the tar is generated (see Figures Slxxx,

Appendix C). As the char layer thickens, less condensibles are liberated and the total

 

100

mass balance considerably improves. This "lost" tar is not obviously accounted for in
determining the empirical chemical composition of the pyrolysis products and the char
yield.

To analyze the data for Y6, and number of C and H atoms in CXHYO, consider the
processes that occurs during wood pyrolysis. As described in Chapter 1, the pyrolysis
of cellulose (major component of wood) starts with the dehydration process at low
temperatures which leads to the formation of char, water, and fuel gases. At higher
temperatures depolymerization of cellulose produces levoglucosan which is further
pyrolyzed in large samples, while passing through the hot char matrix. In these
processes, the rate and composition of the pyrolysis products changes with the heating
rate and temperature. As the pyrolysis takes place deeper inside the solid, the products
of pyrolysis contain less ’carbon’ and ’hydrogen’ atoms relative to the number of
oxygen atoms, due to the cracking of heavy molecules while passing through the hot
char. These products are reported by several authors [Goos(1952), Madorsky (1975),
Hileman et al. (1976), Chan (1980), Kashiwagi et al. (1987)]. The average empirical
chemical composition, char yield, and heat of combustion of volatiles of various woods
are reported in the literature. These were summerized by Atreya (1983) in a review of
pyrolysis of wood. The average formula reported for pyrolysis products of Douglas-ﬁr
is CLMHugO. The average char yield is 0.36 and the lower heat of complete

combustion of CL04H2330 is determined to be 15.33 KJ/gm (same reference).

In our data [Figures CHxxx and chxx (Appendix C)], the complete prediction of
time variations of char yield Y0 and number of ’C’ and ’H’ atoms during the heating
period of wood, was hampered by the condensation of tar. Since the heaviest organic
volatiles, containing the largest number of carbon atoms, escape during the early stages
of pyrolysis and they are partially condensed, error in the calculated values of x, y, Yc
and the heat of combustion is higher at the early times. This error is consistent with

the trend of errors in the total mass balance [Figures ERxxx (Appendix C)]. The

101

difference between the two curves in Figures TMxxx (Appendix C), specially around
the peak of the mass ﬂux curves, represents mainly the amount of tar which has been

lost by condensation and was not accounted for.

Experimental data [Broido and Nelson (1975), Shaﬁzadeh (1981)] shows that char
yield increases with decrease in temperature and also the extent of the sample exposure
to lower temperatures. Under different experimental conditions char yield on heating of

cellulose for one hour at 370°C varied by a factor of ﬁve [Broido and Nelson

(1975)]. According to Shaﬁzadeh’s (1981) data for cellulose, (-pp—f) (ﬁnal density of

W

char/density of wood) varies from 21% at 300°C to 3% at 500°C. From equations (4-8
through 4—10) it is evident that the variations in the chemical composition of the
pyrolysis products is closely related to the variations in the char yield. The calculated
values of ’C’, ’H’, and Ye in these experiments at larger times generally become closer
to the values reported in the literature and was mentioned above. Figures CHD4N and
YCD4N (Appendix C) are good example of variations of these parameters during an
experiment with relatively low total mass balance error. Except for early stages of
pyrolysis, for this high heat ﬂux experiment on dry wood, the average char yield is
about 0.35 which is very close to the average value reported in the literature. Almost
similar values (average) are obtained for 4 W/cm2 experiment performed on dry wood
in air. In air, of course, due to the char oxidation char yield is expected to be
lower. The discrepancy is due to higher experimental error for experiments conducted

in air (because of more air leakage).

Theoretically, the char yield right at the start of heating (i.e. at time-=0), must be
unity because no carbon has left the wood as yet. This point can not be obtained
experimentally due to the uncertainties in all quantities involved at the ﬁrst few
seconds of each experiment. Then, char yield increases monotonically during the rapid

rise and fall-off of the pyrolysis mass ﬂux [Figures 3.6-3.17], because as char thickens

102

more ’carbon’ atoms are caught in the char matrix as the pyrolysis products travel
outwards. However, the trend observed specially at the early stages of char yield and
’C’ and ’H’ data are opposite to the expected trend. This is obviously the consequence
of the condensation of that portion of tar that never entered the combustion process,
and seemed as if it remained in the solid and was converted to char. The magnitude of
this error is different in different experiments depending on how much tar has been
condensed. However, generally at lower heat ﬂuxes and higher moisture contents, the
values of ’C’ and ’H’ and Yc appear to improve signiﬁcantly [Figures CHxlx, YCxlx
(Appendix C]. The results for each group of experiments show that for the same
ambient condition and moisture content, char yield increases with decrease in heat ﬂux
(and consequently with decrease in the solid temperature). Moreover, the results
indicate that the char yield increases by moisture content of wood". For example, a
comparison of the ’average’ char yield in experiments at 4 W/cm2 in nitrogen [Figures
YCx4N (Appendix C)] shows that char yield increases with increase in the moisture
content of wood. In low heat ﬂux experiments on 17% moist wood in nitrogen, where
small amounts of tar are generated and mostly water is released [Figures YCM2N and
YCMlN], the char yield is almost invariable and is about 0.65, indicating that of and
pw basically represents densities of dry and wet wood. Notice that CxHyO, empirically
represents those gases which left the wood. Thus, the char yield, as calculated from
relation (3-4), is a measure of the ’carbon’ (and ’H’ and ’0’) atoms which stayed in

the solid.

The values of gram C/gram of CxHyO and gram I-I/gram of CxHyO are also
plotted in Figures GGxxx (Appendix C). From the average reported empirical formula

of Douglas-ﬁr volatiles, i.e. €1.04H2380, the average values of

_sa£., _saLH_, and _sm9_
gm CxHyo gm Cxﬂyo gm CXHYO

* Deviations from these general observations are due to the experimental errors in any par-
ticular experiment.

are 0.404, 0.077, and 0.519

 

103

respectively. The results show good agreement and the errors follow the same trend as

in the other data.

To summerize, although due to the condensation of tar the overall time variations
of Ye, ’C’, and ’H’ for the entire period of heating was not completely obtained, the
results conﬁrm the fact that char yield and chemical composition of pyrolysis products
are not constant and they vary with time, temperature, and heating rate. Moisture

content of wood also dilutes the pyrolysis products and increases the char yield

3.10 Heat of Combustion of Pyrolysis Products

 

For the determination of combustion efﬁciency we need to know the lower heat
of complete combustion. Hugget (1980) found that for most hydrocarbons, which
included polymers, organic liquids, and natural materials, the amount of heat released
per unit mass of oxygen consumed for their combustion is almost constant. A value of
13.1 KJ heat released per gram of oxygen consumed was a close
approximation. Deviations from this average value were typically on the order of
i5%. In this work, the value 13.5 was considered which is the constant heat of
combustion of fuel per unit mass of 02 required for the complete oxidation of a unit
mass of fuel obtained from An'eya (1983). The heat of combustion of pyrolysis
products is determined from the relation

mm,

AH,=13.5 ,
mp

 

in which AH, (KJ/g) is the heat of combustion of volatile products, onz
(mg/cmzsec) is mass ﬂow rate of oxygen consumed for the combustion of pyrolysis
products, and mp is the pyrolysis mass ﬂux (mg/cm2.sec).

.PP Figures HVxxx (Appendix C) show the calculated values of lower heat of

combustion of pyrolysis products. These data include the error due to the condensation

104

of that part of the heavy hydrocarbons which did not enter the catalytic
combustor. However, average large time values are mostly within the ranges reported
in the literature. The average reported lower heat of combustion of Douglas-ﬁr
volatiles is 15.33 KJ/g [Atreya (1983)] which may be determined from complete
combustion of Cl.O4H2.380 (average empirical formula for products of pyrolysis of
Douglas-ﬁr). The results show that for higher moisture content of wood, the heat of
combustion of volatiles is lower. This is due to the presence of more water in the
evolved products which dilutes the mixture. Higher heat ﬂuxes also yield higher heat
of combustion of volatiles which is due to the presence of more heavy hydrocarbons in

the pyrolysis products.

3.11 Sample Temperature Proﬁle

 

In most of the experiments, temperature at speciﬁed locations in the samples
along with the front and back surface temperatures were measured. The distance
between the in-depth thermocouples was greater as they were located further away
from the heated surface. This was done because these measurements provide data for
the calculation of thermal properties of the pyrolyzing solid and their variations with
time. Sensitivity of these calculations to the changes in the temperatures larger for
higher temperature gradients closer to the hot surface than to the back surface. The
surface thermocouple usually remained in place for most of the duration of
experiments in nitrogen atmosphere. In air, however, due to the oxidation and more
frequent cracking of the char, surface thermocouples came off the solid much
faster. Moreover, the location of the in-depth thermocouples could not be maintained
the same in all samples because the growth rings of the wood samples made it
impossible to drill a straight ﬁne hole for the insertion of the thermocouples. For this
reason direct comparison of the in-depth temperatures is not made. However, the exact

location of the thermocouple beads were determined and recorded. The temperature

105

versus time measurements are shown in Figures Txxx in Appendix C.

From the in-depth temperature-time proﬁles, it is clear that the temperature
histories of the solid are profoundly altered by some nonconductive processes. For
moist wood there is a characteristic plateau at about 100°C. Such plateau does not
appear in the temperature proﬁles of dry wood [Figure TD4N, Appendix C]. The time
of occurrence and duration of these plateaux increase as the distance from the heated
surface increases. These plateaux are distinctly visible in experiments on wood with
17% moisture content [Figure 3.30 and Figures TMxN Appendix C]. Such an
endothermic effect is quite characteristic in pyrolyzing organics and is due to the

diffusion, condensation, and re-evaporation of adsorbed moisture.

A second plateau, although not very distinctly clear, can be observed at
temperature range of 350-400°C [Figures Txxx (Appendix C)]. As the wood is heated
and the surface of the sample is pyrolyzed, while most of the volatile products of
pyrolysis move into the the boundary layer outside of the solid, some of them are
driven (and diffuse) into the cooler interior of the solid. These gases carry energy
from regions of higher temperature to the relatively cooler inner layers and they
condense where the temperature is as low as their condensation temperature. The
energy released in this process heats up the intact solid prior to the arrival of the
pyrolyzing zone. As heating continues, energy is conducted into these regions which

causes re-evaporation of condensed pyrolysis volatiles.

The in-depth temperatures continuously increase with time while the rate of
temperature increase reduces as the char layer thickens. The surface temperature,
however, ﬁrst shows a rapid increase and then the rate of increase reduces drastically
and becomes almost constant. This is because although the absorptivity of the char is
higher than wood and its conductivity is low, both of which result in increasing the
surface temperature, however, the surface emissivity increases as well and much of the

incident heat ﬂux is reradiated from the surface. These two effects offset each other

106

and the surface temperature tends to remain constant and even occasionally decreases.

Heat ﬂux has the most signiﬁcant effect on the temperature distribution along the
sample thickness. Figures Txxx (Appendix C) show that as heat ﬂux increases
temperature increases. Appearance of cracks may alter the surface and in-depth
temperatures. In general, temperature decreases with increasing sample depth and so

does the temperature gradient.

Ambient oxygen signiﬁcantly increases the surface as well as rhe in-depth
temperatures at any given time [Kashiwagi et al. (1987) obtained data for interior
temperatures measured at equal distances in three different ambient oxygen
concentrations. They also observed these effects]. The most important effect of

ambient oxygen is on the surface temperature.

Surface temperature considerably increases in the presence of oxygen. This
appears at all levels of external heat ﬂuxes and the effect is higher at higher heat
ﬂuxes [Figures 3.13-3.14 and Figures Txxx (Appendix C)]. The surface temperature of
the solid decomposing in air, however, shows an inﬂection point followed by sharp
increase before it gradually becomes ﬂat (Figures 3.31-3.32). This is due to the
occurrence of exothermic gas phase reactions between the fuel gases and oxygen
adjacent to the surface [Atreya (1983)] followed by oxidation of. char. Other
investigators [Shivadev (1974), Quintiere et al. (1982)] have also observed such
inﬂection points in their surface temperature data. Atreya’s (1983) data showed that
the inﬂection point in the surface temperature corresponds to the deviation in the
required heat ﬂux (calculated from the measured surface temperature) from that
measured by the heat ﬂux sensor. Further oxidation of the char surface and the
glowing combustion increases the surface temperature drastically. In all experiments at

4 W/cm2 heat ﬂux in air, the wood was ignited and ﬂaming combustion continued.

The increase in the interior temperatures in the presence of oxygen is mainly due

to increased heat conduction through the sample from a higher surface temperature

107

.235. .58 ”too? 2: 028% 3238. £5.53 .a 25. .m> BEER—Eva.

83 m2:

DCON comp 00.0— 00%; oo_N— 00m: chum own 0a.; 0mm 0

cad .uE

 

 

 

 

mo<mm=m xu<m

ES ~.q~

msm Hzomm

 

 

 

 

(0.) Eanrvaadwar

TEMPERATURE (°C)

TEMPERATURE (°c)

800

108

 

7007
500-

500+

 

400-

200
100

o 1

T F ' I T l T I 7 T I I I I I ‘
EXP. D4A

acres». ‘I Is.

EXP. D2A

EXP. DlA

 

 

800

o 1 260 . 460 ' 600 I 800 01 0100‘12r00‘14I00'16I00'181002000

TIME (see)

Fig. 3.31 Surface temperatures for experiments on dry wood in air.

 

700—
600;
5003
too;
300;

200]

 

100

 

I ' I I ' r . ' r ' TT ' T - 1

EXP. M4A .1

 

 

0 r

O

 

. I ' z ' r T I ' 1 Y I ~ I r I ‘ . i
200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. 3.32 Surface temperatures for experiments on 17% moist wood in air.

109

caused by the oxidation of char. It may also be partially caused by the increased
penetration of external radiation inside the solid through cracks. Oxidative degradation
of wood due to the diffusion of oxygen into the solid does not seem likely to be
signiﬁcant. Because the penetration of oxygen into the solid is mostly overcome by
the outward ﬂow of the volatiles and it is also mostly consumed by the char oxidation
reactions. Indeed, the upward trends in the interior temperatures follow the same route
as the surface temperature when the conduction wave reaches those thermocouples,
indicating that no signiﬁcant ’extra’ source of heat (due to the exothermic reactions
between char and 07) has been present at those locations. This is also observed in the
data obtained by Kashiwagi et al. (1987). Thus, the exothermic secondary reactions

occur outside not inside the solid.

3.12 DATA CORRELATION

 

In an actual ﬁre environment, it is very helpful to be able to even approximately
predict the burning rate of the objects from the surface temperature information,
without resorting to a detailed theoretical calculation. In an attempt to ﬁnd an
empirical correlation for the experimental results, the solution to the problem of 1-D
conduction of heat in a semi-inﬁnite inert solid initially at a ﬁxed temperature is
considered. The conduction of heat in the solid at time t > 0, initially at a constant

temperature of T... is described by the equation

as _ aze ,
a: _kax2 ,x>0 t>0, (3-13)

where

110

and a generalized heat ﬂux boundary condition is given by

—k-g-xe-=f,(0,,T...t)=e[F-h0,—o((0,+T.,)4-T..4)]. (3-14)

where F is the externally applied heat ﬂux, and f, is the heat ﬂux going into the solid
at the surface. Other symbols are deﬁned in the nomenclature. Atreya (1983)
obtained an approximate integral solution for equation (3-13) with the generalized
boundary condition (3-14) using Duhamel’s theorem and assuming that f, = a t“,
where a and or are real constants. For a constant heat ﬂux boundary condition, a = 0,
whereas for a constant temperature boundary condition or = -1/2. The details of the
derivations are given by Atreya (1983) and will not be repeated here. From the

resulting equations, it was found that

r: AzpCpk(-t:£2 (3-15)

where A=l" «1%)/l" (0t+ l), in which I‘ () is the gamma function. The total

energy stored per unit area in the semi-inﬁnite solid, E", is given by

 

n t Azpcpk 932
E .. {modi- “+1 (78—). (3-16)

Even though these equations were derived for a semi-inﬁnite inert solid, they collapse
the pyrolysis data very well. Figure 3.33 shows the resulting correlation. In this ﬁgure,

the heat ﬂux f, is the net heat actually going into the solid and is given by

111

f, = e[ F - 0 (T54 — 11,4) - h (1“. - T..)E(b)] — cpgrimr, - T..). (3-17)

where F is the prescribed incident heat ﬂux, T3 is the surface temperature, h is the
convective heat transfer coefﬁcient, and Cpg is the speciﬁc heat of evolved gases. In
this equation, E(b) and Cpgm”(T, - T.) , respectively, are the required corrections due
to the blowing and the enthalpy carried by the pyrolysis gases; E(b) [deRis (1978)] is
given by

E(b) = —-b—— and b = EEC-Pi (3-18)

(eh—1)

The abscissa of this ﬁgure, TIME(--), is

e, 2 f, 2
TIME(—)=(7r—)/(-I;)

where 0 =T — T... and T... is the ambient temperature.

The pyrolysis mass ﬂuxes of 9 experiments performed in nitrogen at 2, 3, and 4
W/cm2 on wood samples with three different moisture contents are included in this
Figure. Due to the large number of data points all the graphs are plotted with the same
symbol. The mass ﬂux changes from being proportional to Oszlfs, which may be
interpreted as the instantaneous energy content of the solid , to being proportional to
0,1'65/f,1'12. The correlation provides a very convenient tool for approximate prediction
of the burning rate in an actual ﬁre given the incident heat ﬂux and the surface

temperature of the solid.

112

ANZ cm macgtomxmv mam—8.?— uo Sum “:23 2.33.? mm... .wE

Tums:

 

 

 

5.0

. opd

O
O.
P

o
O.
o
P

50.50.. oo+.uo.— no+.mo.— 3.7m...— O.Poop OJ... 0.“: amp —.0

. .. /
/
.. .. / .
... h o. / .-
. .. / ... w
n . r / N... a u
u .1. / N E ..
n. ..v. / . mm . I
....J/I. / mo—

. O 1
. n
m. w.
_. a
r.
n a
D. D D -II D D -ID D I -PP P D DP. P D -PD I I I-DI P h -b#P D

 

 

 

(tub/m) xnlr seem/Sr

Chapter Four

The Effect of
Thermal Decomposition Kinetics
on the Mass Evolution Rate of
Charring Materials

As it was discussed in Chapter 3, previous theoretical studies can generally; be
divided into two categories. The ﬁrst category contains detailed numerical studies
[Kansa etal (1977), Kung (1972), Atreya (1983), Chan and Kreiger (1984)] that
attempt to provide comprehensive description of wood pyrolysis by including as many
heat and mass transfer processes as possible. Single-step overall apparent
decomposition cherrristry is often used to describe the complex decomposition
kinetics. The second category contains simpliﬁed analytical studies [Kanury (1972),
Delichstios and de Ris (1983), Wichman and Atreya (1987)] that attempt to develop
practical and useful formulas for important quantities such as the evolved fuel mass
ﬂux and surface temperature. Such models are purely thermal and treat thermal

decomposition as a phase change process, in which wood abruptly converts to char at

113

114

a speciﬁed pyrolysis temperature. Recently, an attempt has also been made to obtain
an analytical solution to a simpliﬁed set of equations that include Arrhenius
decomposition kinetics [W ichman and Atreya (1987)]. Formulas for volatile mass
efﬂux were derived for the initial stages ( kinetically controlled regime) of pyrolysis.
Also, as expected, it was found that for the later stages (diffusion controlled regime)
the results were similar to those obtained by the pyrolysis temperature (thermal)
models [Delichaetios and de Ris (1983)]. However, obtaining a uniﬁed analytical
description of the entire process was not possible and a pyrolysis temperature had to
be assumed in the diffusion controlled regime. Pyrolysis temperature models are often
used when seeking simpliﬁed formulas for predicting the evolved fuel mass ﬂux and
the sample surface temperature which are needed for solving complex problems such
as ﬁre growth in buildings. This chapter examines their validity by considering two
otherwise identical models; one with inﬁnitely fast decomposition kinetics (i.e.
assuming a pyrolysis temperature) and the other with ﬁnite rate decomposition
kinetics. The dependence of pyrolysis temperature on various parameters is also
examined. For simplicity, the solid is assumed free of moisture and changes in

thermophysical properties with temperature are ignored for both the models.

4.1 Model Formulation

 

Consider a "dry" inﬁnite slab of wood of thickness L and initially at ambient
temperature, T... At times t > 0, let one face of this slab be subjected to a constant
heat ﬂux, 61in”, and let the other face be well insulated. Thus, as heating proceeds, the
temperature throughout the solid gradually increases. At some time a pyrolysis zone
begins to develop at the surface and then propagates slowly into the interior of the

solid leaving behind a thermally insulating layer of char.

The thickness of the pyrolysis zone depends upon the decomposition kinetics.

For inﬁnitely fast decomposition reactions, this thickness tends to zero and a pyrolysis

115

front propagates through the solid. This sharp front is associated with a temperature
(T1,), at which wood abruptly converts into char. For ﬁnite rate decomposition
kinetics, the density of the solid in the pyrolyzing zone continuously changes ﬁ-om the
initial density of wood (ow) to the ﬁnal density of char (pf). At any instant, a partially
pyrolyzed element of wood in this zone may be considered to consist of char
distributed through the unpyrolyzed active material. Since zero shrinkage is
assumed, all densities used in this work are based on the original volume of the wood
element. Thus, ps (x,t) = P. (x,t) + pc (x,t) , where pa and pc are the densities of the

active wood material and char respectively. Also, at t = 0, p8 = p,l (x,0) = pW and at t

= co, p8 = pc (x,oo) = pf and p8 (x,oo) = 0. For constant char yield (i.e. Bef- = constant),

W

ps is linearly related to p, according to:
Pf
P; (x,t) = (1 - p—) p; (xvt) + pf. (4'1)
W

Similarly, a linear variation with density for the thermal conductivity (ks ) of the

solid in the pyrolyzing zone is assumed;

._. 2:. a
., Mam.

Here, lcw and kc are the thermal conductivities of unpyrolyzed wood and char
respectively. These are assumed constant and independent of temperature. Also, for
temperature independent thermal properties, the enthalpy per unit volume of an

element in the pyrolysis zone is given by;

116

pshs = p: CpuT = (Pprw + pc Cpc )T. (4.3)

where, CW and Cpc are the speciﬁc heats of unpyrolyzed wood and char respectively.

Assuming no heat transfer between the pyrolysis gases and the char matrix, the

energy equation describing the process becomes;
a _ .9. a:
'5; (9311:)- ax [k8 ax ] + Q‘ (4'4)

where Q is the rate at which energy is released per unit volume during the
decomposition process. For a single ﬁrst-order overall decomposition reaction, Q is

given by

 

as
Q=Qp-;T=-AQp(p.-Pf)CXP ). (4.5)

E
(RT

whereas, for an inﬁnitely fast decomposition reaction (i.e. a propagating pyrolysis

front) Q is given by

Q=-Q,<pw-pf)%5<x-x,). (46)

Here, Qp is the heat of pyrolysis per unit mass of volatiles (positive when
endothermic) and x1) is the location of the pyrolysis front. The other symbols are

deﬁned in the nomenclature.

117

The associated boundary and initial conditions are:

at t = O:
T(X.0) = Ta; pc(X.0) =0; p.(X.0) = pw = p. (X.0); (4-7)
Mx=Qt>0:
8T ° II 4
- k,-a—x- ,. = o = qin - h [T(0.t) — T..] - 80 WW) — T... 1; (4—8)

andatx=L,t>0:

3T _
-k’$ x=L"’O' (4-9)

4.1.1 The Pyrolysis Temperature Model

 

For inﬁnitely fast decomposition kinetics, equations (4-4) and (46) reduce to the
conventional formulation of the phase change problem. Note that for
0<x<xp(i.e. inthe char),Q=0,p,h,=prchandk,=k¢.Forxp<x<L(i.e. in
unpyrolyzed wood), Q = 0, p,h8 = pW CWT and k, = kw. Thus the energy conservation
equations for wood and char with temperature independent thermal properties become:

 

an, 321',
Pw pr—g't" = kw 31?; (4-10)
and
are 82'1"c
pccpc—aT=kc axz- (4-11)

118

At the wood-char interface (i.e. x = xp), the temperature is continuous and is

equal to the pyrolysis temperature, Tp, viz.,

x=x1,;TW=T¢=Tp (4-12)

The energy balance at the interface may be obtained by integrating equations
(44) and (4—6) from x = xp(t) - e to x = xp(t) + e and letting 8—90. This procedure
yields,

‘3?

c
x = x,+£ ax

 

- Qp (pw - p.) d7? (443)

X=Xp-€—

 

Equations (4-10) through (4-13) along with the boundary and initial conditions given
by equations (4-7) through (4-9) were solved numerically by using the Crank-Nicolson
ﬁnite difference method. Special, completely implicit ﬁnite difference equations were
used for the start and end of charting of the boundary nodes and the nodes adjacent to
the char-wood interface. The numerical procedure is similar to that described by
Elhrich [1958]. The numerical code was veriﬁed against Neumann’s exact solution
[Carslaw and Jaeger (1959)]. The results were in excellent agreement. The ﬁnite
difference equations and methods of solution are described in Appendix B.

4.1.2 The Finite Rate Decomposition Model

 

As noted earlier, for ﬁnite rate decomposition kinetics, the density and therefore
the thermal properties change gradually from their values for wood to their values for
char inside the pyrolyzing zone. Equations (4-1), (4-2) and (4~3) show the linear

variation assumed. Thus, equation (4-4) may be rewritten as:

119

3T 3 3T
(pacpw+pccpc)7t'='a: [ 5"]

8p, (Cpc - PGCc)
(pw - p!) (T )1 ( )

Here, the last term is the energy required to produce a unit mass of volatiles minus the
energy carried away by this unit mass of volatiles. In the pyrolysis temperature model
this energetic process occurs solely at the wood-char interface.

Equations (4-14) and (4-5) along with the boundary and initial conditions given
by the equations (4-7), (4-8) and (4-9) were numerically solved. A ﬁnite difference
form of these equations was obtained by using the Crank-Nicolson method to express
the derivatives. A simpliﬁed version of the computer program developed by Atreya
[1983] was used for this purpose. This program has been previously tested for

accuracy against the analytical calculations of the same reference.

4.2 Overall Energy and Mass Balance

 

Overall energy and mass balances were used as criteria to check the accuracy of
both models and to enable comparison between them. Various terms in the energy
balance are schematically indicated in Figure 1. Let t: be the time it takes to
completely pyrolyze a slab of wood of thickness, L. Then, the total mass of volatile
gases produced during the time tf and the sum of all energies lost, stored and absorbed
during the pyrolysis process must be the same as the total input energy and also the

same for both the models.

For constant values of initial (pw) and ﬁnal (pf) densities, the total mass evolved
per unit area [(pW - pf)L] is automatically balanced. However, for the energy balance

various terms need to be considered separately. These are:

120

(i) Total input energy per unit area (J/cmz)
Ein = 9m” (tr "' t0)-

(ii) Enthalpy remaining in the system at the end of charring (J/cmz)
L
E... = pf c,. I mas) - T.) dx.
0

(iii) Energy consumed during thermal decomposition (J/cmz)
Epy=Qp(pW—pf)L'

(iv) Convective and radiative heat losses from the surface (J/cmz)

E10,, = [so (T,4(t) - T...“ ) + h ('I‘,(t) -— 1.)] dt.

6'":

(v) Energy carried out by the products of pyrolysis (J/cmz)

(from equation (14));

E03:-

L
as w w-
j—‘i— (p :3 _ 5:”) mm) — T.) dx at.

6'5“

(4-15)

(4-16)

(4-17)

(4—18)

(4-19)

121

a
For the pyrolysis temperature model 325 may be calculated from Equations (45)

and (4-6), and using T(xp,t) = Tp, whereby

Bog = (pwcpc - prpc) (Tp ' To.) L. (4'20)

Thus for the overall energy balance the sum of the terms given by equations

(4—16) through (4-19) must be equal equation (4—15). For the pyrolysis temperature
model, equation (4-20) must be used instead of equation (4-19). All of these terms
were calculated for both the models for the purpose of comparison.

4.3 Results and Discussion

 

Property values used in the calculations are listed in Table 4.1 for both the kinetic
and pyrolysis temperature models. Calculations were performed with both the models
for four different incident heat ﬂuxes and four different heat loss boundary conditions
on the front surface. The boundary conditions used on the front surface were; (i) no
heat losses, (ii) only convective heat losses, (iii) only radiative heat losses, and (iv)
combined convective and radiative heat losses. For purposes of comparison, the case
with no heat losses is taken as the base case. An appropriate pyrolysis temperature
was determined by enforcing the overall energy and mass balances between the
pyrolysis temperature and the kinetic models. The inﬂuence of the heat of pyrolysis

was also investigated

Figures 4a-d show the history of total weight loss, front and back surface
temperatures and the weight loss rate for the cases with no heat losses at the front
surface. This is the base case. The solid lines are the results of calculations
performed by using the kinetic model, whereas the dotted lines are for the pyrolysis

temperature model. The pyrolysis temperature used in these calculations was

122

determined by matching the energy and mass balances between the two models for the
case with 3 Mom2 incident heat ﬂux, giving 'rp = 425°C. In these cases, the only
variable is the incident heat ﬂux. The front and back surface temperatures predicted by
both the models are in reasonable agreement. The curves deviate only slightly as the
incident heat ﬂux is varied. This is expected because the boundary conditions and the
thermal properties used in the two models are identical. However, the total weight
loss (which is directly proportional to char depth) and thus the volatile mass ﬂux
predicted by the two models differ signiﬁcantly with the incident heat ﬂux. The
deviations in the mass ﬂux curves calculated by two methods are more pronounced at
the early times (corresponding to ignition) and the later times (corresponding to

extinction).

Next, more realistic cases with surface heat losses were considered. The accuracy
of both the numerical schemes was determined by performing the overall energy
balance described in Section 4.2. Table 4.2 shows the percentage errors in the total
energy balance within each numerical scheme. The pyrolysis temperature model has
more computational error, because of the abrupt conversion of wood to char. This
error is introduced during the time step in which the pyrolysis front crosses a nodal
boundary. In this time step, the location of the pyrolysis front is iteratively determined
by satisfying Equation (4-13). If the pyrolysis front is within a small prescribed
distance from the nodal boundary, its location is taken as the nodal boundary. Thus,
an error is introduced in the energy balance. This error may be reduced by reducing

the time step. However, it was considered acceptable for the present purpose.

Table 4.3 shows the percentage difference between the total input energies for the
two models. This difference was minimized to determine the energy-balanced pyrolysis
temperature for each of the cases listed in the table. The pyrolysis temperature was
found by uial and error by enforcing the condition that the total energy for the front
model be equal to that of the kinetic model. Also shown in Table 4.3 (in the last

123

column) is the percentage difference in the total input energy between the two models
when the pyrolysis temperature was held constant (T1, = 425°C). As can be seen, for
the most realistic case with convective and radiative heat losses, the error in the energy
balance is 33%. In order to reduce this error to 6.9% the pyrolysis temperature had to
be lowered by 75°C to 350°C. It is possible to obtain even better agreement between
the two models by further lowering the pyrolysis temperature. As a general rule, higher

the net heat ﬂux into the solid (i.e. k(%E-)) at x = o the higher is the pyrolysis

temperature required for balancing the energy. Thus, the pyrolysis temperature is
clearly not a material property. Instead, it varies considerably with the net heat ﬂux
into the solid, which may be altered either by changing the heat losses from the

sample surface or by changing the incident heat ﬂux to the sample surface.

Figures 4-3a to 4»3e show the time histories of front and back surface
temperatures, total weight loss and the weight loss rate, as calculated by the two
models for different front surface heat losses. Once again, from and back surface
temperatures predicted by both the models are in good agreement when identical
boundary conditions are used. The effect of decomposition kinetics on the total weight
loss and the weight loss rate can be seen from Figures 4-3c, 4-3d and 4—3e. The
agreement between the decomposition kinetics model and the pyrolysis front model is
excellent for Case 1, i.e. with no heat losses. However , it deteriorates when larger
heat losses are permitted at the front surface; even for the cases where the energy-
balanced Tp was used. For the case with nonlinear radiative heat losses, the error in
the mass balance (Figure 4-3c) at 500 seconds is 27% when the energy-balanced Tp is
used (Case 3) and is 51% when TI, is kept ﬁxed at 425°C (Case 6). This error is
clearly unacceptable for predicting ﬁre growth in practical problems. Furthermore,
even with a T1, that satisﬁes the energy balance requirements, it is not possible to
predict the correct pyrolysis mass ﬂux (as evidenced by Figures 4-3d and 4-3e)
specially in the initial and ﬁnal stages. Thus, pyrolysis temperature models cannot be

124

used to predict ignition and extinction phenomena. Figure 4—3f shows the difference
between the mass ﬂux predicted by the pyrolysis temperature model and that predicted
by the decomposition kinetics model at early times using a TI) that satisﬁes the energy
balance requirements. The pyrolysis temperature model predicts that both piloted and
spontaneous ignition will occur at 30 seconds when the mass ﬂux sharply rises to 0.8
mg/cm2.sec and that the surface temperature at that instant is equal to the pyrolysis
temperature. These predictions do not agree even qualitatively with the experimental
observations. Figure 4—3g shows the in-depth temperature and density proﬁles for the
case with radiative heat losses at the front surface and with the back surface
insulated. Three different calculations are compared at two instances; 300 and 600
seconds. These calculations correspond to: (1) decomposition kinetics model, (ii)
pyrolysis temperature model with energy-balanced Tp = 375 °C, and (iii) pyrolysis
temperature model with ﬁxed Tp = 425 °C. As can be seen, although the front and
back surface temperatures are reasonably close the temperature and density proﬁles
inside the solid are substantially diffemnt This difference accounts for the large

differences observed in the volatile mass ﬂux shown in Figures 43d and 4-3e.

The effect of the controversial heat of pyrolysis on volatile mass ﬂux is shown in
Figure 44. Three different values of the endothermic heat of pyrolysis were used. For
a ﬁxed pyrolysis temperature ('I‘p = 425 °C), the total energy between the two models
balanced to within 3%. The volatile mass ﬂux predicted by the two models (Figure
44) also shows good agreement except at the initial and ﬁnal times. Thus, heat of
pyrolysis does not have a signiﬁcant effect on the pyrolysis temperature. The front
and back surface temperature histories presented in the Figures 4-2a, 4-2b, 4-3a and
4-3b show that there is relatively small difference between the predictions of the two
models. This is because identical boundary conditions are used. Thus, the heat loss
term [Equation (18)] in the overall energy balance is essentially identical for both
kinetic and pyrolysis temperature models. Energy consumed during thermal

125

decomposition [Equation (4-17)] is also the same for both the models. The enthalpy
remaining in the system at the end of charring [Equation (16)] is very small compared
with the other terms in the energy balance; hence, any differences in energy between
the two models may be ignored. Thus, the primary difference in the energy balance is
the energy carried out by the products of pyrolysis, i.e. Equations (4-19) and (4-20).
Thus, for a given problem, an approximate energy-balanced pyrolysis temperature may
be found from the terms (4-19) and (4-20), viz,

 

top = A‘ [A p's amﬁmndt‘ (4.21)
where
(T - T...) x , t , (p,(x.t) - pt) E t
=———, =—, =—, a: , =——,A=A
e T. n L I t: (on - Pt) 4) RT. If

and to is taken as zero. Clearly, the non-dimensional 9? depends upon the history of
the pyrolysis process. The higher the average temperature, 9, at which pyrolysis
occurs (i.e. higher incident heat ﬂuxes), the higher is the value of 9p. Equation (4-
21) has to be solved numerically since it requires the knowledge of p', (x,t), and
9(x,t), which must satisfy Equations (44) and (4-5).

126

Table 4.1 Constants used for the calculations.

 

 

A = 250 x10°

Cpc = 712
CW = 1465.1

E = 125.58 x 108
Kc = 0.0712

K, = 0.1675
L =1
h = 11.4

Qp = 30
T, = 25
e = 0.95
o = 5.67 x 10-8

pw = 676

pc = 162.24

 

 

 

 

 

 

 

127

Table 4.2 Percentage errors in the total energy balance within each
numerical scheme.

 

 

 

 

 

 

 

 

CONDITIONS:
External Heat Flux = 3 W/enn2
Qp = 30 cal/mg
Insulated back surface.
- — E + E + + E
Percentage error = 53—3—35- 1” E” 3‘) x 100
Bin
PYROLYSIS
KINETIC TEMPERATURE
MODEL MODEL
1) no surface heat losses 0.002 1.45
('I‘p = 425°C)
2) convective heat losses 0.06 3.00
('I‘p = 405°C)
3) radiative heat losses 0.9 2.20
(Tp = 375°C)
4) convective & radiative 1.64 1.95
heat losses (T1, = 350°C)

 

 

 

128

Table 4.3: Percnetage difference between the total input energies (Bin) for the two
models. This was used to determine the pyrolysis temperature Tp.

CONDITIONS:

 

External Heat Flux = 3 W/cm2
Q = 30 cal/mg

Insulgted back surface.

 

Total input energy (Joules)

Percentage difference

 

 

 

 

Pyrolysis Pyrolysis
Kinetic Temperature Temperature (1) & (2) (1) & (3)
(1) (2) (3)

1) no surface 6.100E+6 5.985E+6 5.985E+6 1.87 1.87
heat losses (Tp = 425°C) (T1) = 425°C)

2) convective 8.640E+6 8.500E+6 8.820E+6 1.62 2.08
heat losses (Tp = 405°C) (T1) = 425°C)

3) radiative 19.800E+6 20.000E+6 24.600E+6 1.01 24.24
heat losses (T p = 375°C) (Tp = 425°C)

4) convective
& radiative 22.679E+6 21.110E+6 30.200E+6 6.90 33.05
heat losses ('I‘p = 350°C) (Tp = 425°C)

 

 

129

£035.: .335 .3
9.327. 80363 :eEmquoRe—u 38.3... .«e soup—amazon. .3325 ﬁe «...-awn

 

33.35
Aux - 1 _
_
629.0 r _
L))\L EM ’ _ 3.255
n a .
so I .
no . / —
4 cos ._. _
_

 

ell ..F /..m

8 I «I
Asylum: £"ss>¢8~.y Snow AV Elﬁn—ovum I: know

130

Junta moan

ecu Eat 0:. 5 name. 32. e: .23 .mouBEuQEo. 38:.» 2.8”. «N... «...—”E

 

 

Aooov ”.5:
com com 00¢ com com 00 — o
p a . . . . . _ . _ . _ o
S
n
a.
11111 . V
\\\\\ \ s I l—
2; xx t 3
“.225. ..II xx x w
3322.53 33.85 ---- xxx d
\E -m E
x .. V
a- m
NE W ”a a
Nae; m .3 E 3
~65 e - : ..N— \l/_
t S
5:8 cm ... no - /
9 one as t 3 3 ml.
mummop camp. 0: "mmmu mmmm t (

 

(O
P

 

131

com

.mooatam x93
93 Bob 2.. 8 momma. 32. e: 55 .8232.an conga 325 and one»;

66$ 22;
can con oo.

com 00*
. . . _ t _

p .

 

 

33:3. III
mezuoemasma mpm>_oexg tttt

~59}. _ tMv
Nani: N t m
~33: m tMNW
NEU\3 v t —
53—3 on u ac
9 one mNe u on

momma. use; o: "mmou omen

\ E E E E

 

mugw

 

(°°J./ ql) aanlvaawal NOVB

132

.38.“ :3 x23 98 EB“ 05 3 momma. 82. o: .23 .32 Emma? .85. on... 0.53.”—

 

 

 

 

 

Aoonv m2:
com com on:V ocn com 00 — o
_ . _ . b . u . t. . _ . to
52.3. II x

mg “8.853 9539.3. ---- xx t
loop m
l..—
. W
..oom we
. mu.
m.
loom m
S
- S
. Mg M “a ,8... m
.4 use} n -3 .. /
e ~53 e - _ m
x :e loom {u

5:8 on n no
u use awe t eh .
mammo— umws O: "mmnu Ummﬂ
com

 

 

133

 

.3035.“ “.25 98 Eat 2.. 3 momma. .3: o: .33 .22 82 £30? an... «._.—ar—

 

 

 

 

 

 

 

 

Aoomv u}:
000 can 9.3. own CON 00 P o
“1 u w _
. .
. \\r
. \
_ _
. .
. .
. .
. xx .
E“ x u
5 0:2; II "
mgzumgmasmu mpm>_o;>m ---- u N _
- T
E r. u.
1 n. Tm
NEU\= _ -qu .n “
~53 N - m .. .. ,.
Nag; m - ~ .. u.
NE; e -3 a; ..
.u ..m
5.58 an n no u. E
u 0% we " A: ... 1
mummO— uww: O: nwmmu wmmm

 

op

 

(8 gum/5w) xrm ssvw TILLV'IOA

 

1594

62285 coatam 323
95 332 an: coatam Eat 583.5. .23 .28E383 coatsm an... 95w?—

 

 

 

 

 

Aommv m2:
coo. 0mm 2.5 cm; . cam . m. p
3 Ex: I! ..N
”.29.; II.
.....zc..m:.._3 339;; ---- l l L: 1 a-
.... I .1...“ Iﬂml \ivv- i -n
N: EN: 2; .2
é
.u not nuuuno_-a mm
.3 33a. 322...: a onng” a .35 2. a
.x. ,Mwﬁism M :2.
A. g x. momma. m>_.e en; a u»..u~>cu q
i “a. Maﬁa : jun...“ \ é
A. n a. nonao_ a: \\
NE; n u ...... ...: a S. \ 2.

 

 

 

OVJEHS

EFLLVBBdWEJ. 3

-!
q
d

i)

(“i/

 

135

.3235 Sara x23 Ea
moms— .ao: coatsm :8: 2.80%? 55 .032quan couta xoum and «...—mr—

 

hoomv ”.5:
coo. 0mm 2% 2.: cam
h b n n -
ah tmx__ 1.1:
u_ac:_g III:
ogzgmgmasmu m_mx_c;>; ----
G 8.. a. . at \
.3. momma. 0,.3..g.g a ~,_.u~,ou NW .muwnu_un \
mac 0) a as
AMWV mac—_u>.”wm”nou Wm. p aux—b \“V

Ab ¢ x. momma. o>.aa.ceg a o>.uuo>oU e
.m a x. “no. o>.uu_uug n cause—an \
Am ¢ x. «we. o>.uuo>ou mwm xagocu \\\
A; a x. mammo— o: q . \\
:.
Nauxx n u so ‘x\“v
. \;§Av\
“
v \\\\\NV
lllll |. ‘\ \.\.\u\HV\
xxx
E s \ \\ \
.... l. w. ll...r...... .... \ E E a. A:

 

 

0N

 

(”i/“Daamvaadwai BOVJHDS xcva

136

62235 coatsm v.25
can 382 82. 3&5.“ 89» Each? .33 .32 £303 .58.

and 9.53,.—

 

Aoomv mic.
OOO— owm 0&0
ah umx_u I...
Uuumc—x III...

.baaagoaaaa m_mx_ogxm ----

 

at 3. .1

 

 

:oop

100m

1000

soov

icon

 

 

Dom

(am/5w) ssm .LHSIBM “mm

137

62285 coats...”
“.03 EB «88— .3. coatsm 59c Bahama .53 .08.. $2 2&0? um... 9:33—

aoom» m2...
com com omN CON on — 00— on

L-nhl_b‘-P#—hbbb Dunn—b-hn—nbhn—nh

 

 

 

3 85 ll
U:25. ll

mgaumgqumu mpm_ogxa ---- e

 

(s znus/Em) Xﬂ'L-l ssvw 31LLV'IOA

 

 

138

622.55 08.35
x03 can 332 “no: coatzm 89c Eouobmu .23 .28 $2 Ego? and «._.—um..—

 

Gomv us:
CON 0
00m: . o m . 2mm . Om; . . _ .J 10.0
. 3 thll _ _ \
n 0.59.3— I _ .ﬁ w
u m sumgmasw» mwmx_og>¢ ---- _ u
.
.

 

I;

V:::I:I“““’”

max

/ l
/ ~-
gulf—......
,i’

(8 znus/fun) XfT‘L-i ssvw 311mm

 

 

 

 

139

63523 .«o 88: 2:8 8 28 was. 390.3 and 0.53...—

 

Aoomv m2:

con 0mm cow an P 00 P on o

p n P b P b b n n — b p I? k b h n b n n p b h b — p 0.0
.
u_um:mx will. . :
wgaamgmasmu mpmx_ogxa ----

1N6
#8

mummo_ m>_uum>:ou a m>_um_umg
umu=m_mn xmgocm

u awn own ah
Em\_mu am no

NEU\3 m "z”w “Aev ammo

 

 

 

 

(3 sz/ﬁw) xrm ssvw 3111mm

140

.82235
each; :03 2.. 28 copra Eat 05 a $2 30: 03368 o...
5; 32m coo? on. E 8an bacon Ea 888382 53?.—

:\xv mmwzxoi» 9% 8o;

an... «Sur—

 

 

 

 

 

 

 

oé md 66 To Nd 0.0
n b p - L til b — D P b - n b b h b b mop
Noe-J o o o o o 0’?” m
.I\\\\\.r£ _ _ m -. y.
C 1 A3 A H W
m. _ n N.
Ste]. .............. “X H a
w .... ..... ..v..l.I/.M.. 3
d .. u {1.1.1. ma.
% . _ H z 0
....— O 01 um... comp.» an 3 “ 3v H U
m 8. SE a. 3 _ H m
n _ . )
L H I.
. _ .
m o, _ f. .M
a _ . . (
_ umx... “ A.V
. 33:3: Ill .

mgsuagmqsmu mpm>_cha_ ----

 

 

 

_ H

 

Q

..c::: &C «5.»: ...: ucekkm

 

 

 

 

 

 

 

 

 

 

141

.33.. m8— EmmoB v... 0.5»:—

 

803 m2:
omw cam cm. 2.: cm
W _
_ _
_ ..I .l ..IJW .\
Amv_ \ a uwumcmx Ill
mg: mgm Em m m o; I...
r. \\ Am «Am wgzwcgmasmw mwmxuogxm ----

 

 

 

. _‘
__ , Em\_.u so u as An.
_ aa\_.u m_ n ac ANV
_ 53—3 on a QC AS
AN. _. ~50\3 m u .a

mammo_ you; up

 

Ago. m.m»_ogxa yo anus .. “a...“

 

 

(3 gum/5w) Xn'L-l ssvw 3‘lliV'10A

Chapter Five

Conclusions
and
Recommendations

In this work, pyrolysis of thick samples of wood was experimentally
investigated. Pyrolysis experiments under different levels of external heat ﬂux and
moisture content of wood were conducted in air and in nitrogen atmospheres. A unique
small scale combustion-wind tunnel was constructed for these experiments which is
useful for obtaining simultaneous transient measurements of several physical and
chemical variables. Transient quantitative data on mass ﬂux and pyrolysis products
were obtained. In addition, a new approach in the chemical measurements was
adopted. For this purpose a simple, but very efﬁcient, catalytic combustor was
constructed. A sample of pyrolysis products was burned in this tube. From the
products of combustion after the tube, char yield, empirical chemical composition of

pyrolysis volatiles, and their heat of combustion were determined. These are new

142

143

ﬁndings that provide understanding about changes in the chemical and physical
parameters dming pyrolysis. In the theoretical part, the concept of ’pyrolysis

temperature’ was investigated.

5.1 Experimental Work

 

The physical and chemical data obtained during pyrolysis experiments indicate
that:

1. External heat ﬂux has the strongest inﬂuence on the rate of gasiﬁcation and
temperature distribution. Rate of production of fuel volatiles increases and the char

yield decreases as heat ﬂux is increased.

2. Ambient oxygen signiﬁcantly increases the pyrolysis rate and char oxidation
provides additional energy which supplements the external heat ﬂux. Surface and in-
depth temperatures, char depth and pyrolysis products show increase in oxygen

containing atmosphere.

3. The pyrolysis mass ﬂux does not show a negative one-half power dependency
on time. This is conclusive for experiments conducted in air. In the experiments
conducted in nitrogen atmosphere in this work the surface temperature (and visual
observation) do not show any evidence of char oxidation due to leakage of small
amount of air (less than 5%) into the tunnel. However, the comparison of the results
with the results of similar experiments in the literature indicates that presence of even

small amount of 02 in the atmosphere may affect the gasiﬁcation rate.

4. Flow geometery, i.e., position of the sample in the boundary layer ﬂow, can
have signiﬁcant effect on the gasiﬁcation rate and time of occurrence of its
maxima. When the boundary layer starts from the sample edge, it is not convectively

well deﬁned and heat and mass transfer are not uniform along the sample.

144

5. It was found that at lower temperatures in air, chemisorption of oxygen into
the surface of freshly produced hot char may occur which reduces the pyrolysis mass
flux. This is followed by smoldering and glowing combustion as temperature rises and
char oxidation initiates. Therefore, extreme care is needed to provide well deﬁned

conditions for experimental study of wood pyrolysis.

6. Moisture content of wood primarily effects the physics of pyrolysis and
ignition and its effect on the chemistry of pyrolysis is not signiﬁcant. It cools the
solid as it desorbs and reduces the tar evolution rate and delays ignition time in air. At

lower heat ﬂuxes for moist wood, mainly water is desorbed.

7. Char yield in pyrolysis of wood is not constant. It changes during pyrolysis as
well as with changes in the external heat ﬂux and moisture content of wood. Lower

heat ﬂux and higher moisture contents result in higher char yield.

8. Chemical composition of the pyrolysis products changes during the course of
heating as well as with changes in the heating rate and moisture content. Most of the
heavy hydrocarbons (tar) evolve at the early times of pyrolysis and more ’carbon’

stays inside the solid as char thickens.

5.2 Theoretical work

 

1. The gasiﬁcation rate of thermally thick charting materials was
investigated. Two models were developed for this purpose; one with inﬁnitely fast
decomposition kinetics (pyrolysis temperature model) and the other with ﬁnite rate

decomposition chemistry (kinetic model).

2. The pyrolysis temperature, which is often treated as a known constant is
neither known nor is it a constant. It depends on the net heat ﬂux entering the solid,
which in turn depends upon the boundary conditions and the thermal properties of the
solid. Each new problem requires a new value of the pyrolysis temperature. An

improper choice of this temperature will produce serious errors in the predicted fuel

145

mass evolution rate.

3. For a given problem, a proper pyrolysis temperature may be chosen by

enforcing the overall energy and mass balance between the two models.

4. Despite large differences in the fuel mass evolution rate, the front and back
surface temperatures calculated by the two models are in reasonable agreement. This is
because the same boundary conditions and heat of pyrolysis are used for both the
models. However, signiﬁcant differences exist in the in-depth temperature and density

proﬁles.

5. The effect of the heat of pyrolysis on both models was similar and no
signiﬁcant differences in the predicted fuel mass ﬂux was observed.

6. Although the pyrolysis temperature assumption considerably simpliﬁes the
solution of the governing equations, the results are incomplete because the pyrolysis

temperature is unknown.

5.3 Recommendations

 

There are several topics that can be recommended to extend the present study in

the future.

1. Experiments may be conducted in nitrogen atmosphere with only weight loss
and surface temperature measurements. The pressure inside the tunnel should be
allowed to be slightly above atmosphere in order to prevent any leakages of air into
the tunnel and to eliminate any possibility of char oxidation. The results can be
compared with the literature to accurately conﬁrm the trend of fall-off of mass loss

rate.
2. Similar experiments may be performed in air and in nitrogen with gas phase

thermocouple installed very close to the surface of the solid without concerning about

weight measurement. This data will show the exothermic gas phase reactions at the

146

present of 02 and will help to clerify the question of exothermicity in wood pyrolysis.

3. From the temperature measurements obtained in this study, changes in the

thermal properties of the solid can be obtained using an appropriate model.

4. To interpret the experimental results, and for prediction purposes, a detailed
numerical model for wood pyrolysis should be developed. Such a model must include
moisture desorption and cracking of char. Experimentally obtainad char yield and
thermal properties of active material can be incorporated in this model. This model can

be extended to include char oxidation as well.

5. It would be interesting to extend the "pyrolysis temperature" model to contain
desorption of moisture as a second from traveling inside the solid. Such a model can
be compared with a decomposition kinetic model which contains a ﬁnite rate kinetics

term for moisture desorption.

Appendix A

Appendix A

Data Processing Procedure

As mentioned in section (2.9), the data acquisition system in the fast mode was
able to collect data for more than 20 channels in every second. In a typical experiment
period of 20 to 30 minutes between 1200 to 1800 data points were collected for more
than 20 channels. In many cases 23 channels (12 thermocouples and 11 instruments)
were used. During the same period the number of data points for weight loss was 5 to
6 times as much. These data were processed separately for each channel. The data
processing procedure consisted of two major steps, namely, smoothing (curve ﬁtting)
and lag and response time correction. The FORTRAN code was the one developed by
Atreya et. al. (1983). This code was used with minor modiﬁcations.

A.l Smoothing

 

The raw experimental data curve showed different degrees of ﬂuctuations
primarily caused by the electrical disturbances of the equipments. In the case of weight
loss measurement in addition to the electronic noise, small ﬂuctuations were caused by
the ﬂow within the tunnel and motion of the air within the room. The experimental
data had to be smoothed in such a way that while the main characteristics of data were

preserved the unwanted disturbances were damped out. For weight loss data this task

147

148 Appendix A

was much more critical because the rate of weight loss, which is equal to the pyrolysis
mass ﬂux, had to be numerically calculated. Any small discontinuities in the weight
loss data are highly magniﬁed in the weight loss rate curves. The curve ﬁtting scheme
must ensure continuity of value and continuity of slope as much as possible. The data
ﬁtting program ﬁts piece-wise polynomials to a given data set. A small segment“ of
data points was read into the computer and ﬁtted with an orthogonal polynomial using
the least square criterion. For each segment the highest polynomial ﬁt (between 10th
degree and a cubic) with no more than one inﬂection point in the ﬁtted range was
chosen. This was done because not more than one inﬂection point was expected to

occur in the physical phenomena.
In order to preserve the continuity in value and slope, a segment was divided into

7 sub-segments and "§7' of the neighboring splines were made to overlap. The curve

ﬁtted data points were then extracted only from the central -;- portion of the

splines. Since the recorded number of data points were too large, the smoothed data
were stored every 5 seconds. The results show that this scheme was quite satisfactory
and no signiﬁcant discontinuity was observed in the smoothed data. Only if there was
a very sudden change (almost step-wise) in the data, small discontinuities were
observed because continuity was not enforced mathematically. Figure A.1 shows a
weight loss curve, smoothed data curve, and its derivative which is equivalent to the
pyrolysis mass ﬂux (weight loss rate of wood samples). Figure A.2 shows the original
(raw) and curve ﬁtted data points for a typical C02 measurement. The smoothed data

was then corrected for lag and response time.

 

‘ThenumberofdaupoinuinamanaPOMcmndbevarriedtoobnindredesireddegreeof
«mom.Typ'cally35to70numberofdaupoinuwereukenforallsigmls.ForweightabthBOdatapoinuwere
used. SpecialanmtionwupaidtotheselectiouofLPOMinordertopmemsharprulwhiehwereusuallyexperi-
eneed in weight loss rates and chemical species coneartntiom measurements.

 

 

 

 

 

 

 

 

 

 

 

 

149
400 r I f I T ' l I r 1 r 7 r 1 .0
J I
6; +0.8
0 3004 5 _
\ . '.
O J .0
E L06
V . HA
(I) —I . l'
m 200 \\
3 .. . r—O.4
i.—
I I-
O o
m 1 00-4 - Weight loss data L-O.2
; x polynomial spline ﬁt
.4 . . Derivative of weight loss curve I
l
0d - 1 t r I ' I ' l f ﬂ T I T ﬂ
0 1 00 200 300 400 500 600 700 800
TIM E (sec)
Fig. A.l Polynomial spline ﬁt and rate calculation of weightloss data.
22 r I r I ' “‘“I’ ' T ' I T I T—j
A .'
,3, . l
— Recorded data I
Z . x Polynomial spline ﬁt I
O 21 -I .. - Data corrected for .1 ,
-- o lag and .
:E 1 . response tune ,
m «I . J.
I— '. f
E 20 g
g l.
O .
O 4
Z I
LIJ 1 9 '1 —
O
>_
X .
O l
1 3 r r r I ' r ' r r T I
O 1 00 200 300 400 500 600 700
TIME (see)

Fig. A.2 Curve ﬁtting and response time correction of 02 data.

(saws/53w) xma SSVW

150 Appendix A

A.2 Lag Time and Response Time Corrections

 

During the experiment some measurements such as weight loss, temperature, heat
ﬂux and pressure are fast enough to be considered instantaneousFor others, such as
gas analysis equipment it takes some time for the information to reach the
instrument. This time lag can be measured and simply subtracted from the recorded
time. Further, the instrument has a response time which it takes before it can react to
the information. The recorded data, therefore, has to be corrected for this response

time. The response time of each instrument has to be determined experimentally.

The correction method described below and the computer program was adapted
from Atreya (1983).It utilizes the response of the measuring system to a unit step. A
typical recorded output of an instrument to a unit step input is typically shown in
Figure AA. The instrument response is the same as the recorded output corrected for

the lag time. This is modeled as a transient process within the instrument.

Consider a gas analyzer where the gases are completely mixed while passing
through the it. The outlet concentration, therefore, will be the same as concentration of
the gases within the instrument at any instant. This is similar to a lumped heat
capacitance system in which it is assumed that the instantanuous temperature T of
every point within the body is the same at any time. Figure A.3 schematically

represents such a well stirred system.

INSTRUMENT

dy
y: in "q" 71% J“ yg out

Figure A.3 A well stirred system.

 

NORMALIZED RESPONSE Ys(t)

GAS CONCENTRATION

 

 

 

 

1.0— ..-
Swp innnX.(t)
0.8-I
0.61
* I
0.44 I
. I
-hummnuicxlznsnmwe
(LZA * Eqummﬂﬂﬂmﬂbnﬁn
hﬂodanmﬂﬂwﬂuhw
. constant tc = 2.61 sec
Lagthne Deunsmmneth]
0.0 —T I I r T ﬁ- r ﬁ— I a? ‘r _l_j_ T _1_ I

r r T
O 10 20 30 4O 50 60 . 7O 80

t I TIME (sec) I
Fig. AA Response of C02 analyzer to a step input.

 

 

X(t)
xl, II...”
I /’
I /
I //
X3 /’|’
33 1 c—-"'---
33 ' ' I]...
I ./
X2 ’AI/ , Y(t)
I ' ,-—-— “"ﬂ——
| . 0 I,
a
X ’a'.” . .
I I . .—
Y4 ...—"—-
a ' I '1 I’.’
1 I ' l" I
’ I / I I
c—J' Y3 p .
‘ A ' —————
yl I ,L—v’d
...—’7 I l ‘ _.
t
0 t1 t2 t3 t4 t5

TIME (see)
Fig. A.5 Recorded response to a peice-wise linear input.

152 Appendix A
Species mass balance for the stirred reactor of volume V is given by
v dY
ﬁc-afI = Y... - Y... . (A-I)

in which p is the density of the gas, rir is the mass ﬂux and Y, is the mole fraction of

the specie in the mixture..ps 12 Using Y,” = Y, and letting Lv = tc we will get:
In

Y8 = ngn . (A'Z)

 

[1+tc-élt-

This equation represents the instrument response Y, to the input signal Ym.

Some instruments are better modeled as containing two well stirred reactors in

series. The governing equation for such systems is:

2
[ﬁg-ntc-EIH Y, = Y,” (A-3)

dtz

 

The response of the ﬁrst order system modeled by equation (A-l) to a unit step input

is

rl»

Y, (t) = l - e , (A-4)
and the response of the second order system modeled by equation (A-2) is given by

Y,(t)=1—(l+-;:-)exp(l-i). (A-S)

153 Appendix A

The only unknown in these equations is the time constant, t, which is
experimentally determined. For a more complicated time dependent input data X(t), the
instrument response (for a sufﬁciently linear instrument) can be found by adding up

the responses of the instrument to step inputs according to the Duhamel’s superposition

integral

l
Y(t) = I! Y,(M) %dr (A-6)

Here, Y(t) is the measured output and Y,(t) is given by equation (A-4) or (A-S) for
ﬁrst or second order systems. Y(t) consists of a set of measured values Y, ( t,) at ﬁxed
time intervals At (= t,“ —t,), and we want to ﬁnd out a corresponding set of input
values X, (t, ).

Assuming that the input X(t) to be piece-wise linear [Figure A.5], then the
corresponding recorded output, Y(t), can be reconstructed by superposition of the ramp
inputs. Each linear input ramp will be X, = a 1:5 + b. The response of the analer to a

ramp input is obtained from integrating equation (A-6) with Y, from (A-S):

{-4 {-1

‘I ._.
YRi(t)=ajI (1—e t‘ -£-_-'-t-e l‘)dt
‘I—1 t‘

3:. l‘ x
= a,- (At + art-2n) e “ -(tj+1-t-2t-2t,) e ‘= ) (A-7)
and for j=i and t= t;
.A‘.
1.

YR, (t) = a, At — 2tc - (—At-2t,) e (A-8)

154 Appendix A

The recorded output Yi(ti) at any instant t,, [Figure A.5], is the sum of responses
of ramp inputs at previous times (tH, with j=0 to i) plus response of ramp input at
time i [Eq. (A-4)]. Therefore,

i—l

j-O

.r-Ie

i—l
= Z YRi(t,) + a, At - 2tc+(At + 2n) e (A-9)
i=0

In this equation the only unknown is a;. Once a, is known the input (piece-wise
linear) will be Xian = XH + aim. The time constant of each instrument was found by
introducing a known step input and using a least square method to match

approximately with a ﬁrst or second order system.

A calculation scheme where a, (i=1,2,...,n) and hence Xi (i=1,2,...,n) were
successfully determined from the knowledge of tc, ti, and Y, (i=0,l,...,n) was used for
the response time corrections of C0, C02, 0;, H20, and [THC] analyzers. The response
of the dew point hygrometer was better modeled by a ﬁrst order system and, except
for the change in equation (A-l), an identical procedure was used.

The error due to the response time of instruments is basically due to the
uncertainty of predicting the slope of the piece-wise linear input from the recorded
output (Appendix A). Atreya (1983) discussed the error in the prediction of these
values. He showed that for stability of the numerical criterion with a given
measurement error, it may be necessary to use a ﬁxed constant time correction factor
tab. and in each time t, shoot for the current time (t-tcb) instead of the time Step At. The
values of td, were determined such that the prediction of the input signals be

stable. Values of t,., were typically between 10 to 40 percent of the instruments time

155 Appendix A

constants. Since both the instruments and the data reduction procedure were mainly the

same as used by Atreya (1983), the discussion will not be repeated here.

Appendix B

Appendix B

Finite Diﬁ‘erence Equations

and
Methods of Solution

In this appendix the numerical methods for solving the "pyrolysis temperature"
and the "decomposition kinetics" models which were formulated in chapter 4 are

described and the ﬁnite-difference equations are presented.

B.l Pyrolysis Temperature Model

 

The numerical method for solving this problem is similar to that of Ehrlich
(1954) with several modiﬁcations to improve the accuracy of the results and the
stability of the numerical scheme. There exist several different methods in the
literature for solving this Stefan-type problem. Most of these methods, however,
primarily address the prediction of temperature proﬁles and the time dependent
location of the phase change interface, corresponding to a "charring front" in this
problem. These methods fail to correctly predict the rate of the propagation of the

interface at the early times of phase-change process. Physically, the phase change rate

157

158

Appendix B

may not start at a ﬁnite value, but it has to begin ﬁ'om zero and increase

subsequently. The method adopted in this work successfully predicts the charring rate

of wood (as a phase-change problem) at the starting times as well as the large times.

13.2 Nondirnensioalization

 

The governing equations and boundary conditions are normalized by deﬁning:

 

r kw! a.
e=—s n=£91= 2tB=—9
7'. L prpr a,
7”": ,=A“__._.JL
kw' n2 122

Phase I . Inert heating of wood slab

 

Before onset of pyrolysis; for t > 0 and T, < T,

 

as, 329,,
GE. — =
at 8112
Boundary conditions:

39,,
1) — Wm) = ( b1 +122 +b 3 ) — e,( 1.29,3 + b;,)

2) ae"(lo-o
an 9
where
“if“
'_ oeLT,,3

kw

03-1)

03-2)

(33)

(3-4)

159

Initial condition:

e (“90) = 1

Phase II. Charring of Wood

 

Governing Equations:

I) On the wood side:

30W 326W

31=an2

 

II) On the char side:

29.: - .1. .32:
3‘! - [5 3'12
Boundary conditions:

39,.
1) - 3140.1) = % (b1 +122 +b 3 ) - 91(b26,3 + 1),)

2) - 33150.1) = 0

At the interface of char and wood, i.e., at x=xp

TIP“) =(W-1) All + S (1) AT] = [(W-1) + S (1) h]

and

80, 30,,

ds
—-——=bh-—
Van an 4 d1

Appendix B

03-5)

03-6)

(B-7)

03-8)

03-9)

(B-lO)

(B-ll)

160 Appendix B

B.3 The Finite Difference Equations

 

The basic method is Crank-Nicolson with a linear temperature proﬁle within a
nodal control volume. fully implicit scheme is used for some "special" cases to
improve the stability of the numerical scheme. The set of ﬁnite difference equations
for the phase I are as follows:

Node 1 : (implicit)

[3+4rh(-}1;+b2x,+b3)] or“ + (1+4r) 61/“

 

 

 

= [Hrh(%+b2xl+b,)] 011' + (1+4r)e{ + 8rh(b1+b2+b3) (B-12)
in which
e{+1+ei 3
X1 3 (91M)3 = -_T;
Interior Nodes:
l-4r . 1+1 j+1 _1_-ﬂ, _ j+l
[6+8r] 9" + 6‘ + [6+8.]°‘“
= 1+4' . ,- . ,- 6-3' i -
[6+8I‘J( 9,.1 + 9H,] )+ [6+8r]e, (B 13)

Last node (insulated); (implicit)

< 71,--r W” + < %+r)e~.r'” = < %—r>e~.li + ( ﬁmw‘ (B-14)

161 Appendix B

B.4 Phase II: Charring of Wood

 

1. Node 1 at the start of charring (implicit):

'ZP-Sz ”1+(-Y—+P%h-)Sjﬂ+l=
£52141 91j+ 6,, + 2%(b1‘l’bﬁb3) (3'15)
2. First node when the ﬁrst slice is all charred; (implicit):
[BWI‘M -Z-+b,x,+b,,]e,i+‘ + (By-4rh)9{”
= [spy-4r“ -Z-+b2X2+b3)]91j+ (Wrap; +8rh (b1+b2+b3) (B-16)

3. "Regular" nodes on the char side:

[ﬂ]iﬂ 95-1 +9!” + [ﬂ] 9min

6(5-0-8 613-0-8r
=[-6%+ﬂ8—] (es-149141,) + [ﬁ]0,j (3'17)

4. Node w adjacent to the interface on the char side; (implicit):

 

21'5“ . . . 2
[14'5“ l]ew-l ( st-I-l)ew Sﬁlﬂew 1+5“ 1 p ( )

5. Node w+1 adjacent to the interface on the wood side; (implicit):

162

(2-sj+1)(1+2r—s}+1)9,,+1j*1 - 2K1-3ﬁ1)ew2i+l =

(2—sj+1)(1-sj+1)e,,+li + m,

6. "Regular" nodes on the wood side; (Crank-Nicolson):

 

1-4r . '1 ,4] 1-47 . J41:
[6-1'87]e‘-11+ +9, + [64-87] 914-1

 

 

1+4r . j _ ' 6—8r
[$8,] (91-1 Wm") + [6+8r]

7. Last node when charring: (implicit)
[(l-Sm)2+2']91v+1j+l = 2’9N+1j + (1‘5ﬁ1)29N+1j

ﬁnite difference equations for char-wood interface:

a. At the start of charting: at n = O,

is. _ 31.
8n v
, 36¢ ,

expanding ?“- about (0, 'c) and usmg

36c _ 1 BZGC

a: " B 3112
we get

39. 1

an _' 7A2

Appendix B

(B- 19)

03-20)

03-21)

03-22)

(B-23)

163

where

This equation will be used until there are at least two nodes on the char side.

b.l As the charting continues

3%. _1_
at h
where
S 1 l 1+S'1 .1 1.231
A2 _-—_ _L6 1‘" ._Lewﬂ' ...—LOP
1+3,” Sf.” Sj+1(1+5j+1
and
30... 1
_ = —A
at h ‘
where

23- 1

.. . -—s- .
A1: _ﬂewlﬂl-_£ﬂ.ew2)+1_

l-S 124.1 2-3,“

c. when sjﬂe O or ﬁnal-91 , we will have

1-2(1-S 1+1)

 

(l-SmXZ-Sju) p

36: _ 1

at ’ MA“
where

M’ = [<1-2s,-.1)e...2"*‘-4<1- mealﬂ‘ua-zsjawf‘]
and

BGw-IA
at" 1

2h

1

Appendix B

03-24)

03-25)

03-26)

(B-27)

(B-28)

(B-29)

B-30)

(B-31)

164 Appendix B

where

41’ = [- (543,41) amt}+1 4' 4(2-Sﬁ1)9w+zm - (3-23j+1) Gleam] (3'32)

B.5 The Method of Solution

 

The above equations when, considered for the related cases, provide a system of
N+l equations that have to be solved simultaneously. The N+1 unknown temperatures
and the location of char-wood interface or alternatively 53+! being the other unknown
for which the interface equation provides the extra equation needed. For the

approximation of the new location of the interface we will use:

s,» = £44 7 A2 - A: > + s,- (3-33)
such that
If “H < 1.25 use A2’ in (B-33)
If 0.75< Si.“ (0.25 use A1, in (3'33)
If O< Si.” <0.25 “86 W12” in A

Otherwise use () as is.

For the initial approximation of SN value at the onset of charting, the exact

solution to the melting of a metal slab with ablation [Landau (1949)] is used to

determine % as follows:
. C .1.
is- = 2M arctan 45-1 2 (B-34)
dt 1: kapo 3,

in which 1,, is the time when pyrolysis begins. The initial approximation for s,“ will be

“7+1 = ‘2—131' For every new approximation of SN the tridiagonal system of N+1

165 Appendix B

equations are solved iteratively for temperatures using the recursion solution given by
Carnahan et. al. (1969). Once the solution converges for temperatures a new value for
3H is approximated and the procedure continues until 3}” is also converged. To
accelerate the convergence of the iteration process every new guess for SM is found by
using an underrelaxation or an interpolation between the old value of 3M and its new
value. The underrelaxation factor a) was determined by trial and error such that the

best convergence was achieved.

The process continues until the char interface comes very close to the next node
at the boundary of two adjacent slices. when | 3&1'1 I is less than a pro-speciﬁed error
criterion, charring of that slice is completed. Here, to avoid passage of the interface
from across a nodal boundary before convergence, the time step is cut shorter
successively until it converges. The validity of the results was examined by comparing
the numerical solution for phase change problems at constant surface temperature with
the exact solution [Carslaw and Jaeger (1959)] for freezing of water and the results

were in very close agreement.

166 Appendix B

3.6 The Decomposition Kinetics Model

 

Using the Crank-Nicolson scheme, the ﬁnite-difference form of the governing

 

 

 

equations will be:
. . .4 . . .41
TI“ - T! Tm 2-2T. 2+?“ 2
[534(3pr cCp.] T Ax = 1‘: Ax
”.1. . 1
1+-
_ Ax [33?] 2 [Q,+h,]i 2 (B-35)
in which
a e e
[6%] ' = —A p"‘—-—e£ exp E 1 (3'35)
' l - & RT.”7
Pi ‘
For the interface nodes 1<i<N+1 we’ll get
m.- M“ + u.- T!“ +f.- 2s: = b.- (13.37)
where
1‘,
mi = — E (B'38)
Ax k.
“i = [(pacp.+pccp‘)'A_t + E] (3'39)
f.-=- {:3— (3-40)
It: - k: - k.
b; = (TA?) 1.5.1] + [(DCCP) - 3]]? + E + D; (3’41)
and
. , ' .9; 3»;
2'
0,. = .. MP; . Q, + (ha-he) 9" (B-42)
r 1-3L
Pw .~

 

 

. l

#3 _
ha = Cam " T.)

. 1

1+- _
h...- 2 = can — T...)

if” +sz

r.= 2

For the surface Node i=1

urTrm'i'fr T2i+l=b1 +01

where
u1= la) C
2 a Pa+
1‘:
f1 - "‘21:
1
bl [2( pa
and

=cj,-,,+hT...,+ec:T..,4

For i = N+1 (insulated)

p‘C”)At2Ax2

167

+-’3-+

 

wma p)(ch WP)[

1;:jTM]Ax

[ 13
11.2]
-86T

2

Appendix B

03-43)

03-44)

03-45)

(1346)

(B-47)

(B—48)

(B-49)

(B-SO)

(B-51)

168 Appendix B

1 i

re ,7 ...;- . 1

= k. " - 3%] (Q + ha?! (B-52)
‘ N+1

 

Ax 2

B.7 Method of Solution

 

The method of solution of the decomposition ldnetic model is described in detail
by Atreya (1983). The system of (N+1) linear equations are solved to determine the

. 1
(NH) unknown temperatures (Tr-2.). In this case, the density of the solid is

continuously changing as the wood is heated. Changes in ’p,’ are calculated from

. 1
equation (B-36) and the mean density during a time step (puff) is obtained. Then the

. 1
1+—
system of N+1 linear equations is solved to determine new values of T,- 2 , i=1,...,

N+l . The convergence is checked based on the difference between the new predicted
value and the previous one. The iteration continues until it is converged. The next time

step is evaluated and the procedure is repeated for the new time step.

169

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pyrolysis Front
__ CHAR :‘r: WOOD
5 rm, “7" 1
3°- ' l/ 6.7. |.¢.V. CN-
/{ l//' '
I 14 2 3 r [w [L I; 3.1 N |
a. w ¢I V? i/Q i M»!
/| V W
2 / / I ¢¢l P¢ :
0' a r/iéuﬂd 9" ’
._.: 3?le (lde d——o-
KP
L r

 

 

 

Enemy — (5+ ‘ O i
. 14- 1
balance I I

SHCOS

 

 

 

 

--——------ — -d

{ i-N+l

 

Figure B.2 Schematic of ﬁnite-difference method for decomposition model.

Appendix C

Appendix C

Data From Pyrolysis Experiments

In this appendix data obtained from the pyrolysis experiments are presented in
the following order:

I. Experiments conducted in nitrogen atmosphere:
I.1 Exeriments conducted on dry wood.
1.2 Experiments conducted on 8-9% moist wood.

1.3 Experiments conducted on 17% moist wood.

II. Experiments conducted in air:
11.1 Experiments conducted on dry wood.
11.2 Experiments conducted on 8-9% moist wood.

[1.3 Experiments conducted on 17% moist wood.
In each set of experiments the results are presented for 4, 3, 2, and 1 W/cm2 external
heat ﬂuxes respectively. For each experiment in nitrogen atmosphere the measured

and derived chemical quantities are given in the following order:

Fig. 8 lxxx Pyrolysis products (direct measurement).

170

171 Appendix C

Fig. SZxxx Pyrolysis products after burning in the catalytic combustor.
Fig. PSxxx Products of pyrolysis as percent of total mass ﬂux.

Fig. CHxxx Number of C and H atoms in CXHYO.

Fig. YCxxx Char yield (gram char/gram wood).

Fig. GGxxx Gram[ ]/gram of C,‘I*iy0.

Fig. '1qu Total mass balance for pyrolysis products.

Fig. ERxxx Error in total mass balance for pyrolysis products.

Fig. HVxxx Heat of combustion of pyrolysis products.
For experiments in air only the ﬁrst six sets of data are presented. This order for
every two consecutive experiments may have been altered to save space. Some of the
low heat ﬂux data curves are not presented due to high order of
scateredness. Selected data on temperature measurements are given following the
above data in the same order. In these notations xxx stands for every experiment and

are consistent with notations summarized in table 3.1

171

MASS FLUX (mg/cm2.s)

MASS FLUX (mg/cm2.s)

 

172
O 4"" I I I r r l T f I i I i I i r i I I I
—— HC
4 A —- C02 _
.. co
. x H20
0.3-* A TAR

‘\ " .
0.2 — ‘JA‘JI‘. “ _.

 
  

 
 
 

......... ---.-"-------

__ ‘A‘

-C Vﬁohx-wﬁﬁ”'

 

 

 

0.0—r

 

 

 

 

I I . I - I ' I . I r I . I r I . I r
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
ﬁg. SlD4N Pyrolysis products (direct measurement); EXP. D4N
O 8— I r r I f l I i r i r l r i 1— r I- r I
0.7— ._. —
o 5 1 3:. A
. -‘ .5. ... . C0 “
~ ‘. ...,f°‘.~ 2 My, "‘
0 5- /‘ 'M‘w-ﬁw WM _
‘ WM -1
0 4e ' H20 _
0.3~ -—
0.2~. _
J. .
0.1—9 _
A, \\‘ CO .—
O.O—r r I 7"?"5'7"?‘7":"F'TT‘T"57"?"’5' I ,
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

"ig. SZD4N Pyrolysis products (after catalytic combustor): EXP. D4N

PRODUCTS OF PYROLYSIS
AS PERCENT OF TOTAL MASS FLUX

AppendixC

 

173
100 T I T I I l T T F I I r FT 1 Fr l T FT 7 T FT T 1
0 tar
_I x H20 ‘
— THC
80- -*

60-

4o—

_1_
0 o
'I

 

 

 

 

 

 

204 ‘I
4 s
O TYII—TjIrTTIIlIﬁTITIrITF frllifrrilierU
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig.1’SD4N Prodmtsofpymlysisaspaecmofmnlmassﬂmm.D4N
O>‘5 I I I I I I r I I I I I I
I
ox "\Hydrogen "
z 4 d .35; ." vi‘A‘Xt. -, “E’W'w‘ﬁ' r
— ;. #- 'Nw 3‘5""
(f) .4 ; NH!" .
2 ,1; ..
O 3_ i '9'
I— _
<
I s _
=8 2- a —
0
LI. "1 Carbon 4
O
m 1" . :v’zhwm _.
LrJ
m . v .- .
2 .
3 O . .
Z I I'I‘I‘I'I'I'I'I‘
0 200 400 600 800 1000 1200 1400 1600 1800 2000
-TIME (sec)
Fig. CHD4N Number of C and H atoms in the products of pyrolysis: EXP. D4N

 

 

 

CHAR YIELD

Grams [ ]/Crom

Ammmﬁxc

 

 

 

 

 

 

 

 

 

174

1.0 I I I I F T I T I I
0.9-I q
0.8~ —
0.7— ..
0.6-a _

u ..‘ -I
0.5— I ‘. -

'1 ’ z -I
0.4-1 3.; . —

.1 ‘0 '1 , MM” .1

. * ‘VQ\. -‘ C

0.3“ .1 .' . 0",” 0 “Mg _
0.2- d
0.1- ..
(10 I I I I I I I’I I I I* I I I I I I F I

0 200 400 600 800 1000 1200 1400 1600 1800 2000

TIME (sec)
F&;YCDQN (ﬂmrﬁddoymncMmbmmIwnmnEXRJlﬂl

1-0 I I I I ‘I I I T I I I r I I I I

J T, 1
(19‘ -

.. .I
0.8—I _.
0f7- A

1 3‘ L
06-1 : ._. Oxygen .
CLS“ ‘JQHE_ .Ua‘Nn*i-"yﬂﬂﬂua-“"”“¥u~u—nnni -a

.1: ._. H d _
0'4 d '1 :. ,*,%W q
0.3-1 .30", _J
0'2? Carbon j
OO1-‘ﬁW‘V—* ﬂ -—4
(10 ' . I r’I I T f I I r - I I I - I I I .4

0 200 400 600 800 1000 1200 1400 1600 1800 2000

TIME (sec)

Fig. GGD4N Gram [l/gmm of CxHyO; EXP. D4N

MASS FLUX (mg/cm2.5)

0%)

h

ERROR IN TOTAL MASS BALANC

175

 

 

 

"PP . . .

.ﬂlcoz'i-sz
I T I T I r . I . I I T I I - I I T
200 400 500 800 1000 1200 1400 1600 18r00 2000

 

 

 

 

TIME (sec)
FﬁﬁﬂﬂMN TUﬂnm3EMM=ﬁrmmmMsmuMwnEniDﬂi

50 I u I I I I I I F f I I I I I I I
401 " -
1
304 ' a

b
I:' ' a
204 . -
q :1": . . ’ -1
10“ - . f: o T . . ‘ O

-I ‘ ’ M'o'. o;°. . '. ... ..., . .2 I." .4
0.. ............. ...}.-i.‘---o..2--.ot--f'-'.".-..;..¢<....g‘\.-:-! ....... _.
. ' . ' ‘- ‘ ' - ‘. '1'; .
_10_. ~ 1' a _1
‘ 1
-20- _
-30-. 4
I I
—404 ~
.50I I

 

o 260 400 600‘ 860 10001200 1400 1600 Iéoorzooo
TIME (sec)
Fig. ERD4N Error in total mass balance for pyrolysis products; EXP. D4N

HEAT OF COMBUSTION (KI/g)

J/Q)

\’

HEAT OF COMBUSTION (

Appendix C
176

 

O

 

 

 

I I IIIIIIIIIFIIIII

o 260 460 600 800 100012001400160018002000
TIME (sec)

mm Heatofcornbnsﬁonofpymlysismdmm.D4N

 

30 I I I 1 r I T T T I I r j r I I I I ﬁ
24-I -
8 .' , ~15 ..
1 “I a h... 0 Q. ‘~’.' _
I. I o. .. O . . ..... O . . u
3.4“}: ' .— l 0:55,?QTo 3'. " " 0' -I
'v «93(0' .c. 9‘ '5‘. . : '
4 o .0. ' W 0‘ .- .1
V: -
I 2 - ~
6—I -—<
O

 

 

 

IIIIIIIIIIIIIIIIIII

O 200 400 600 800 1000 1200 1400 1600 1800 2000
THAE (sec)

Fig. HVDSN Heat of combustion of pyrolysis products: EXP. D3N

Am..-cco\m:tv X3,: mm<§ ~m....-.:u¢;.:CO X3 : haw/\E

6» RI

 

MASS FLUX (mg/cm2.s)

MASS FLUX (mg/cm2.5)

Appendix C

 

 

 

 

 

 

 

 

 

 

 

177
0.4
— HC
J “ C02
CO
><-—>< H20
O.J— 9. A TAR
_ f ‘. TAR
0.2 . «K. M
‘ \/ I
‘ QP
« . n“. :
NMJ gi‘M-ﬁ‘yv
O.l~: H20 w—w4\
O-O'TH‘: I_‘I-TI-iI-TTi-IT'IT-vaiw-TIT—I I- I I ﬁ I I I
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. SIDBN Pyrolysis products (direct man); EXP. D3N
0-8 IIIIIIII ﬁIIIIﬁIIr
O.7~ _
I
0.6— —
0.54 —Il
0.4-I ~
‘ H20
0.3“ 5" ' ‘3’. _.
l 5 C02 4
0.2—if _
0.1—j _
d]: CO ...
0.0/77‘1-T—r—T-1—‘__—E_r_1——,v¥117 _ |
O 200 400 600 800 1000 1200 1400 1600 1800 2000

TI M E ( s e c )
Fig. SZD3N Pyrolysis products (after catalytic combustor); EXP. D3N

PRODUCTS OF PYROLYSIS

Appendixc

 

 

 

178
1IOO P f I f j I I ﬂ ‘ I T w'

. O tor I
>< - X H20 ‘
3 ” — THC I
L‘ 80— — C07 ~
(I) u
(I) . -- CO
3 '.
4: .
‘J 60" ono‘ _
E ./ 1‘. TAR

. ‘r’\ ﬁva
I; ‘3. VMWW'
O .I HO _-

40“ x. 2

52 d' ‘ lI‘a~r‘\k~unf‘saﬁ"‘h"_\\Jﬂ\\J“n
L5 "' k \x
83 .
o. 20— _
(f) ‘ .. ...... 4
< --.-u ,,,,,,,,,, .. ~.. ..

L," _____ __

O

 

I I IIIIIIIIITIIﬁII
o 250 460 600 800 100012001400160018002000
TIME (sec)

Fig. PSD3N Products of pyrolysis as percent of mu mass ﬂux; EXP. D3N

 

§Z\D I Y I T i T 7 ‘ I I I T I - I
l
x q d
O ‘ . :
I f -
Z T T 4x -’ :6 A; _d
— ‘ I: ' '- T It s
5 f a It“ 5“ I,
I . f

< .4. . 'r
I .1 . 1 ...: J

‘ I

i

J
0'6 2 A_ q
C.)
LL T Carbon 1
0
Cr 1 '_" {92.21} ""s hw¢.O.-\_\-*~‘ivv: ._ ,-.,’..‘=*W‘NO:.. _1
LL] {‘4- ‘ ~w Av‘
(I) . .-‘"'V"
2— ' g 4
D _.
Z O

 

 

 

I I I I I I I I I I I I I I I I I T I
O 200 400 600 800 1000 1200 1400 1600 IBIOO 2000
TIME (sec)
Fig. CHDSN Number of C and H atoms in the products of pyrolysis; EXP. D3N

CHAR YIELD

Grams [ ]/Cr0m

1.0

Aqnnndhzc

 

d

0.9d

.4

0.8—

.1

0.7—.
0.6~'-

0.5—
1

O.4-

0.3-
-I
0.2-
I
0.1 d

4

l

l 1 l 1 1 l l

—

-I

 

 

0.0

1.0

TIME (sec)
Fig. chn

Char yield (gram char/gram wood); EXP. D3N

IIIIIIIIIFIIIIIIII+T
O 200 400 600 800 10001200140016001800 2000

 

 

0.9-

0.8-1

q

0.7—
1
0.5—
I
0.57
0.4-

0.3-I

0.2-m

d

0.1-

d

I ‘ I ‘ T ' T ' I ' I

K

k

X

X
’k
a

I!“ Oxygen

WW

Hydrogen 55 ;
‘0’ M”! “W
“J“‘Iw‘ w“

I

Carbon

‘ ‘ ... w

p

 

0.0 I

T

L_r

llllllllLJll

l

J

l

-1

 

TIME (sec)
Fig. GGD3N Gram [ I/gram of CxHyO; EXP. D3N

IIII-IIIIIIIIIIIﬁ
0 200 400 600 800 IOOO120014‘001610018002000

 

Appendix C
180

 

 

 

 

 

 

I O T T ﬁ— l T F j T
0.9— ..
772‘ 0.8—I 4
NE .I
\U 0.7-: ..
51 0.6— _.
é . I
X
D _
-J I
Ll.
U) _
U) .I
< _
2 .
.: ‘Pv . . .
.' . mcoz‘I’moo‘I'U‘HZO‘mw,
0-0 I T I I I I I I I I I I I I I T I I ~
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. M3N Total mus balance for pyrolysis products; EXP. D3N
50 I l I I I I . T H I I I j I I I I
E ‘1 ‘
:3 30* ~ ...: '4
2 q 1"? .
2.5 ‘o‘ ‘0
CD 4' .' '. ." "
m IO— .... ;.. ,. . I
m 00:;- 0. '1 . '
g ...-as l. . ...;s. .3;
z T ....... t- . ;--".:. ;'\' ..... :.‘ ............................ _-
_J .. ..0‘ I .. ... .5. 1‘1... . .... ... . .
f.- —IO-I ".1"- .- -,'.",'-"..'.":' " -
O o .0 . ' '0'. 0" 0 ﬂ "
I— °'-~ '
z ‘ ."I' ‘
g —301 " q
CZ
0:
Lu .
—50"

 

 

 

O " ZOO 400 BOO BOO 10:00 1200 I400 1600 18.00 2000

THAE (sec)
Fig. ERD3N Error in total mass balance for pyrolysis products; EXP. D3N

MASSFlUX Gog/ahas)

MASSI%lWZOng/cwﬁs)

Appendix C

 

181
0-4 II I I I I II I I I III
—— THC
>< C02
-- co
V x H20
On)“ a TAR
A
0.2—I‘ A

0.1-‘

 

 

 

(10

 

 

 

THAE (sec)

Fig. SlDZN Pyrolysis products (direct measunnent); EXP. D2N

 

I I I If I I 7 I I I I I
O ZOO 4OO GOO BOO 100012001400160018002000

 

 

 

 

 

 

 

 

O'BT'T‘ITV'I'TTI'I I'II
. ‘i
I
06* -I
i
04« j
I
NCOZ“
0.2-I .3 #x L ‘h.
. 5/ H20 —~ *2 ‘
JZE‘ co
0.0 . "Tr ———————————————

 

THME (sec)

Fig. SZDZN Pyrolysis products (after catalytic combustor);EXP. DZN

 

PRODUCTS OF PYROLYS|S_
AS PERCENT OF TOTAL MASS rLUX

 

 

 

 

 

 

 

 

 

 

182
100 f I I I I T" I _fT I I . I T
A tcr
~ X H20 ‘
. —— THC -
804 :’\ -- C02 ‘I
i‘ \. W0 1
‘ ‘3‘ TAR :
60- 3W 4
40~ —
4' .4
. H 0 .
20-1 - A
IKJ/ nm 4_ I
“I ,_~__-;":*;. "‘J—‘=_:;__;_: :
o‘r’T‘T'TT'rww ”I“
O 200 400 600 800 1000 1200 1400 1600 1800 200(
TIME (sec)
Fig.PSD2N Prodmtsofpyrolysisaspereentoftotalmassﬂux;£XP.D2N
O 5 I l T T I I I I I I T I I f I
>~
I .
ox ‘ -
z 4" . $1, :1 0:“: {\b m
m .J ;?:. HYd to g €13, i=1-5" 1" : u"
2 f .5 ’f“ 5‘ V?-
- a :
S J" :‘ﬁ: 5“ \-
< u‘. : "
I '1. 1;:
68 2_. ' ..
O .4
n.
O ._4 Carbon 4?.- :
tr ' ' .ﬁ” ad *#~“*J. T
ES «5" l
2 .
D 0 j . T I ’ 1- ' I ' . ‘ . I ‘ z 1 .
Z O 200 400 600 800 1000 1200 1400 1600 1800 2000

17!?le (sec)
Fig. CHDZN Number of C and H atoms in the p'oducts of pyrolysis: EXP.

CHAR YIELD

HEAT OF COMBUSTION (KJ/g)

183 AmnmExC

 

1-0 ' I’I I I I . I . - I I I . I II.
O.8~

0.6-I \

 

 

 

 

 

 

K“ .M I
M. Aﬂr
0.4— \/ -
CLZ— -
(IO I I I T I I I’I I I’I I I I I I I I I
O 200 400 600 BOO 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. YCDZN Char yield (gram char/grain wood); EXP. DZN
‘50 1 I ' I ' I ' r ' I ' I r T ' I ' T ‘1
24- _I
18‘ —i
. .. ‘ i
.. ’:\...’. I ~ g ..
. . § .{§ I; 3 \E 3
q .- IO’XM?‘ 31-?“ ..p' w :ﬁ.‘~’5.~.“ Ii", 4
12_ .: :- o o "5 '. —4
.. .... .f .4
. . ..r q
6" —I
I .
O I I I I II I I I I I I I I I I I I I I
O ZOO 4OO BOO 800 1000 1200 1400 1600 1800

TIME (sec)
Fig. HVDZN Heat of combustion of pyrolysis products; EXP. D2N

 

2000

Appendix C
1 84

 

0.1-

MASSIqID<(nKp%xn?S)

 

 

 

 

(10

50

 

‘ z ‘ T I ‘ ‘ I ‘

800 1200 1600 2000
TIME (sec)

Fig. TMDZN Total mass balance for pyrolysis producrs; EXP. D2N

' 4OO

 

30-

IOd

-10-

-301

ERROR IN TOTAL MASS BALANCE (7:)

 

—50

.l r rﬁr I r F I T r r v 1 If I 1— f r 1
C
C
O
C
I
O
'.’-
'3'; U 9"... g’:
J‘. O “ g ‘ ‘ . .“ I
'6 e. .‘~.~ f ' ' .5 ... 3 _a
'1‘” '- s a- ~.'. '91. -. ‘
0’. " .0 .. I. .
'. 0 ° . ?ﬁ.' 1
""""""" ""I"" ‘f"" .I.-- "'""‘°""""""‘~
I . 2. ’1.
.I.' o O
o. .
I. d

 

 

'IOOO ‘ISCO 12000

TI M E (sec)
Fig. ERDZN Error in total mass balance for pyrolysis products; EXP. D2N

SOO r

\

.SI

MASS FLUX (mg/cm"-

MASS FLUX (mg/cmzs)

Appendix C

 

 

 

 

 

 

 

 

 

 

 

2000

 

185
0.10 f a
—— HC
-- C02
CO
0.08 x H20
I 5 TAP
c105—-' -
0.04— -I
002 TAR
. _ “/V M
d A H20 f \n-A
0.00— . I I"; . , “7":“ﬁ-‘r‘; , . .
0 200 400 600 800 1000 1200 1400 1600 1800
TIME (sec)
Fig. SlDlN Pyrolysis products (direct measm'ment); EXP. DIN
0.05 W
0.04—
0.03“ H20 “
IDIW
0.02- -
.C02 ,
(TOT-I _rw””' co - A
..I"//d’
0.00 ’

 

 

 

O 2OO '4OO' 6OO‘ 800 1000 IzoO'Iioo 1600 1800

TI m e ( S e C )
Fig. SZDIN Pyrolysis products (after catalytic combustor); EXP. DIN

2000

MASS FLUX (mg/cmRS)

)

av
O

ERROR If! TOT—“L MASS BALANCE (

Appendix C
186

 

 

 

 

 

T T T

. T . rT a. . r .ﬁ
800 1200 1600
TIME (sec)
Fig. TMDIN Total mass balance for pyrolysis products; EXP. DIN

460

 

 

50 fkt r r T I 1 r . I r ‘ I T j ﬂ
.4 q
30-+ -
.0
’ \
1 ' '
0‘ .
0 .~
0 O
1.0—J ‘~ ' <
\
'1 ------------------------- I ---------------- 3 ----------------- —1
O
1 '. . . .
._.— . O—J . §\‘. ——4
, - . °. ~ .
. J ~ '
ﬂ .0. - —4
0 '~"

..30_

O ’ ..
\

 

-50

 

' 10063. ' 1500
TIME (Sec)

Fig. ERDIN Error in total mass balance for pyrolysis products; EXP.

500 '

DIN

 

2000

187

Appendix C

 

 

 

 

 

 

 

 

(WI ' :I 3.... I. -
\J . '
‘ - .. \ .
g. 1
E "-.
~.-/ 0.2“ \~_/\'\ ‘
jg ’. H O
:3 ' ‘- 2
._: ~ 2."-
L“ 2; 3"... TAR
3 Q ‘ ...‘I'Q" '
.1 . I _ 0 . 0“ . T“
< V v J’N"/.ﬁ. \« K‘: C02
2 23!?
I THC .
.. M
4'/-/ __________ C0
,/ """""""""""
0.0 . f I I 1 1 ‘ I . ‘ I ‘ I ‘ I ' I A
0 200 400 600 800 1000 I200 I400 1600 I800 _"000
TIME (sec)
Fig. SIR4N Pyrolysis products (direct meannement); EXP. R4N
0-8 ' I ' I ' r T I z
I
0.7—
/"‘"‘\
m ‘
m; 0.6— —
.~ 1 r; I“
O“ \— u-I . éw H20 —
‘V' 0.4". “v, R
if ‘. . .
'3 4 WWW» (:02
,e’ 0.3-4 a
m “
(,3 '12— c
: q .I
0 1..., —I
I co —
:3 O A‘--—A——.— --------- . . , , . . ~-
0 200 400 600 800 7000 1200 7400 1.600 180C .2000

ec)

Fig. SZR4N Pyrolysis products (after catalytic combustor); EXP. R4N

 

PRODUCTS OF PYROLYSIS
AS PERCENT OF PYROLYSIS MASS

I

FLUX

NUMBER OF 0 8c H ATOMS IN CxHyO

Appendix C

 

 
 

 

 

 

 

188
00 I '1 I I I I I I I I I.
80—2 ..
60-2 __
., “29.
40—1 “nusx’ _
‘I 3‘ TAR ‘
INTI ‘ «ﬁyﬁﬁ‘ —
’bov' 'Jﬁﬁyg
2 THC
____________ co ‘
O I I I I I I I I I I I i I I I I I I I
O 200 400 600 800 1000 1200 1400 1600 1800 2000
THAE (sec)
Fig. PSR4N Products of pyrolysis a percent of total mass ﬂux; EXP. R4N
5 I I I I I I j I I I I I r I I— I
4-. __
.JE
‘ MW 'W'J'K,‘ Hydrogen
'I‘Uﬁ.~ ' .
')_I

Q

0’?

._. ... W... Carbon

 

 

 

1,

0 200 400 500 800 I000 I200 I400 :

THJE (sec)

C)

O
(I)
(D
O

I.)
I

Fig. CHR4N Number of C and H atoms in the products of pyrolysis; EXP.

l3

Appendix C
189

 

IIEID
SJ

1 1

L1

 

CHAR \

 

 

0.0 A

0 200 I 400 I 600 ‘ 800 I000 1200 14001600 1800 2000

 

TIME (sec)
Fig. YCR4N Char yield (gram char/gram wood); EXP. R4N
1.0 . I . , . I

 

I ' I E l ' I ' I

0.6ﬁ
0.5-

0.4—

q

CromS [ ],’Cr

0 . 3 -
HYDROGEN ‘

S 071% OXYGEN —
W

0.2-I_' .
3‘ 1A CARBON -
0.1—W —

..I
0.0

 

 

 

I I I I I I I I . I I I - : I . I , .
O 200 400 600 800 1000 1200 1400 16100 1800 2000
TIME (Sec)
Fig. GGR4N Gram I I/gram of CxHyO; EXP. R4N

I:I‘I«I/~'III".:.I

l 'I II

f H

0".

.IA

‘0.

:I‘Z)

li/II Al ICI.

‘(T‘
J.)

[O I‘.-‘-I.

ICRROI’.

a [A
I h

v t V

A ,.
' I he.

“ I
-\ A
‘ _
A A

Appendix C
190

 

— -—.-.—‘._-—.._._._

 

 

 

200 400 600 800 1000 1200 12.00 I600 IBCC

TII ': \
IIIVII. \SGC/

Fig. M4N Total mas balance for pyrolysis products; EXP. R4N

 

 

 

 

 

ﬁf I f r
I
" 'I
«...-I O
a"
.".
e I a.
' '
< .. I . . -<
0.. . ..
OI ...! .
-I e .0. .' 0' —t
.- C ... .' .. ..
.0 ' ~' 0
O
s 3' ‘ ."
.0. --------- J- .- nd-c-l ----------------------------------------
' . ', l.-
O
' 0...”: I
' I. *
- .'. _.
.0
I
-‘ —I
A. “A,“ qﬁﬂ f‘ﬁ ‘i‘jr‘ 1’qﬁf-‘I aﬁ’q.a‘ 4 .f‘lﬂh AAA‘A‘ 1 F‘- 4‘,an
I.— _UL - iL SLIIV CUL' .'_JL‘— ..l.\p\.. ;“‘~b‘ Cui— ~8v— —\—'-\.¢
*9} I: :alq‘tk
IIII— wv

\J ~

Fig. ERR4N Error in total mass balance for pyrolysis products; EXP. R4N

HEAT OF COMBUSTEOII (1 J (3)

HEAT OF COMBUSTION (NJ/g)

Appendix C

 

 

 

 

 

 

 

 

191
50 r
78—
‘ :-
1 2 _ . . ‘ ‘:\\. 'f“. 0.)" ::;$‘.. _
iﬁ'lgo:# I ' 5 " ..
6—J
O ‘ I ' T ' . * t T . 1 T ' . ' I ‘
0 200 400 600 800 1000 1.200 132.00 1600 1800 2000
TIME (sec)
Fig. HVR4N Heat of combustion of pyrolysis products; EXP. R4N
30 f 1 ' l I r F ' l ' l ' T
24— .-
18— _
. J
I. V
12~ ' z 'j —
."!o .'."-:a p ‘1' .~ ~
.. ’1. 4*. ,o \ «0.3: ‘ .J's
'._' A” ' 4
w .
6~ . ' E
J" .
Q- r I Y a r 7 . ‘ . , I .
0 200 400 600 800 1000 1.200 1400 1600 18002000

Fig. HVR3N

T! M E ( s e c )
Heat of combustion of pyrolysis products; EXP. R3N

MASSFIUX Uhg/mhas)

Appendix C

 

 

0.0*

 

192
' ' —— Em k
- - C02 .1
- CO
X H20

 

 

0

.0
4.

”T I f I r I ’ I ' T ‘ I r I ' I ‘ I '
200 400 600 800 1000 1200 I400 1600 1800 2000

TMJE (sec)
Fig. SlR3N Pyrolysis products (direct measurement); EXP. R3N

 

 

C)
l

 

O
O

I ' I ' I ' I ' I ’ T ' I

‘~_—_———

 

 

200 400 500 800 7000 1200 1400163007800 2000

Il-

.IIE (sec)

Fig. SZRBN Pyrolysis products (after catalytic combustor); EXP. R3N

Appendix C

 

 

 

 

 

 

 

 

193
700 I a . ,
5.: 0 KY
:3: I x H20
'~'- ' — Tau
I233 80". -- C02 '
mfé -- CO
:3“ “-1 ‘
OQ -
Egg/3 OO—I :\ _—
id 2% .': =2". :59 H20
0%: d v... wa'ﬁ/‘l ‘I. f “
‘. d \
ELL «1.0—I OV“ —
UO .' 3
3+— . 14%“ r I ‘
33$ .; "M‘Av. 3"": TAR
01% 2o—, 3;: 3,, _
LIJ ...’ -
:3.
(f7 1 r” V. V w_’:' T\[;;: d
< w ...,
O . I I r I - T I f I I I 7 I I I T I I .
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. PSR3N Products of pyrolysis as percent of total mass ﬂux; EXP. R3N
O 5 f I f I T r T1 r
>‘
I . l
X
0
z 4" I
(2 ,.
O 3 ﬁ , Hdro gen - r. ,4
P u
< V -~WJ".I~'- .I‘f'" 3’
I ._g-ﬂ-y’ '
0'6 2— 5" -I
0 " 5"
LL .__.
g H _
w I. Cari?“ .. .
m ........................................................................................................
2 ......
D 0 . y _
Z o 200 400 600 800 3000

TIME (sec)
Fig. CHRBN Number of C and H atoms in the products of pyrolysis; EXP. R3N

CHAR ‘I’IEII

Grams [ Ix/Grom

Appendix C

 

 

 

 

 

 

194
1.0
3.8-~ ~
it"K
3.6“ . ‘w.’ .\ ..
WW‘A‘A‘ {’3
.I ;. f:
0.4- ~
1
0.2—I _.
" 'I
0‘0 I I I I ' ‘ ' , ‘ T . i I I I— . I
O 200 400 600 800 I000 IZOO I400 1.600 1.800 2000
TIIIIE (sec)
Fig. YCRSN Char yield (gram char/gram wood); EXP. R3N
1-0 I I I I I I I I . I I I I I I I I
0.9-I .e
-I ..
0.8-.1“ - -—I
_, ‘ _s- K I
O. / _‘ ‘W.’ -—4
q L- ' 'I- Oxygen '4
0.6“d 3:7 g: _
0.5— _
0.4d —I
0.3~ ;. , . Hydrogen ﬂ
‘ I‘\. ..‘o‘: _.
0.2- * ,ﬂ‘" = “V" " _
«V 9' Carbon -
O-I‘V _____N _
r'
0.0

 

 

 

I I . I I I . I . I . l . I I l I .
O 200 400 500 800 1000 1200 1400 1600 IBIOO 2000
TIME (sec)
Fig. GGR3N Gram [Mgr-am of CxHyO; EXP. R3N

SI

.5

'Cnf.

k

MASS FLUX (mg,

ERROR III TOTAL- MASS BALANCE (3:)

Appendix C
195

 

 

‘\ 2‘
I.
3 4 I ; -
I I
2 I' -r
I "X
, .
I I
OHS— ‘
I 5‘. I ’ I."
o n 0'. I /V ‘V—
r’ V
02—3. I
0.1— T
.4 - {PP . . .
. mcozmchmnzo-mao,

 

 

 

013.
0

026010460. sooxaoo'Iooo'Iéoo'I4oo 1600 I800 2000

TIME (sec)

Fig. TMRSN Toral mass balance for pyrolysis products; EXP. R3N

50

 

30- :3

 

—I
.I
.0
0.
I
I
.I . -—¢
:0-
0 0" '
' I
I
---------- gfau7-n-----k-—---------------------------------4
I
0' o
I
.I
I
I
o O . —I
v
. I
I
Q
0.
d ,"
o —1

ITIIIIIIIIITII~II.

 

 

200 400 600 800 1000 I200 I400 16001800 2000

TIME (sect)
Fig. ERRJN Error in total mass balance for pyrolysis products; EXP. R3N

Appendix C

 

 

 

 

 

 

 

 

 

 

196
0.4 ﬁ fr T i r r {
—-- HC
4 _— C02 1
A CO
(I) . H0
N. 0.3— 2
E J :. MP.
‘0
0‘; -I
E
\‘1 0.2— -~
5
._J
Li.
«A
a OI ’ 3 -
<1: ° . 'RAR
:5 150 I K
’—\\V\”_W’\:—f\%———~
O-O‘L-é'fﬁI F—T--I--T-‘TU-:--Iv-‘::l I VI f I I I
0 200 400 600 800 100012001400160018002000
TIME (sec)
Fig. SlR2N Pyrolysis products (direct measurement); EXP. RZN
O‘BTI‘I‘TﬁI‘V‘IT'ﬁ'T
1:;
N 0.6- —
E
u
.V _I 4
C)"
E
"J 0.44
><
3 H20 ‘
L
5’2 CO ..
it: 2
CO
0.0 ﬂ

 

 

T I I I I . -
0 200 400 600 800 1000 1200 I400 7600 7800

TIME (sec)

Fig. SZRZN Pyrolysis products (after catalytic combustor); EXP. R2N

 

PRODUCTS OF PYROLYSIS .
AS PERCENT OF PYROLYSIS MASS FLUX

“I 0

F1

IUI=.'IBER OF C & H A'I'OZIS III Cv'

I

Appendix C

 

 

 

 

 

 

 

 

 

 

TIME (sec)
Fig. CHRZN Number of C and H atoms in the products of pyrolysis; EXP.

197
I00 T . ;
a. .1. tor I
-I ' [I/ H20
—— m,
. ”pvt-u" —— CO
I ”\r. .,’ 1
f, ". .‘W'
60— 2 ‘0' _
"N"
.‘V‘.
. TAR
. A .
\. I" .
20—1 K ”we. _.
w \Ww-m,
\: —\ v d W ~ W “
O‘I TV I I I f I I I I I f I - I I I . I
0 200 400 600 800 1000 I200 I400 1600 I800 2000
TMAE (SEC)
Fig. PSRZN Products of pyrolysisasperceutof total mass ﬂux; EXP. RZN
5 r" I I I I I I I I I I I I I I I I m
‘ I
3- a. ,- -I
:F w ‘1 " \VA'VM‘ﬁVN-gv Hydro gen
.I ' ._ “V": .I
2~ ' 5' s
1"”
'I .
I ~ _
:2 Carbon
4.: am
Ir
0 I r : I I I I R . I I I I I - I
0 200 400 600 800 I000 I200 I400 I600 I800 2000

 

 

I'IE'LD

CHAR

Grams [ L’Grom

AppendixC

 

 

 

 

 

 

 

 

 

 

198

7.0 E r

. I
0.9-I .

4 I
0.8~ _.

~ 'I
0.7— ~

_: :M. 4
06 -4 " ' ‘~.’/\~v U H" I
05* a
0.4— -—
0.3— _.
02* —
on! S
0-0 I I I i I 7 I I I I I I I I . I I I I

0 200 400 600 800 I000 I200 I400 I600 I800 2000

TIME (sec)
Fig. YCRZN Char yield (gram char/gram wood); EXP. R2N

1-0 I I r I I t I . I I I I . I . I I I

I -I
0.9- J
0.8~ -.
0.7%0‘ :M‘o‘ ‘—
0 6; "_ _' WVW Oxygen :
0,50 -' s
0.4- " ...
0.3—4°. Hydrogen :
02‘ ”VJ ! A Carbon :
0.1-‘5': ‘W'v _—
0-0 ° I I I I I I I I I I I ‘. . : . I . I .

0 200 400 600 800 1000 1200 I400 1600 I800 2000

TIME (sec)

Fig. GGRZN Gram [ ]/gram of CXHYO; EXP. RZN

“
. ..J

(mg, "(:m‘

FLU}:

5.1 ASS

ERROR HITOTALIMASS BALANCE 0:)

Appendix C

 

 

 

 

 

 

 

 

 

199
-." f R I
I
3;_ I
N I
C 2+ "
I: .I
0.1“ I —-4
I
‘/.oo .1
— , , . I
O O {,2- . u‘nao;mcoIIIrnmo-ﬂloo2 j
. ‘ I I I I I. f I I . ' I i ‘ . ’ 7 '
0 200 400 600 800 I000 1200 I400 1600 I800 2000
TIME (sec)
Fig. TMRZN Total mm balance for pyrolysis products: EXP. RZN
50 I . I I j I T I I I I I I I . I I I I
30~ . 3‘ .
" -:
10d .. ‘ , ’ o
-I --------- ;%:-;:-‘- ---------------------------------------- -I
-IO— ' 3:, . —
—3o— ”:2? Z _
0 200 400 600 800 I000 I200 I400 I600 I800 2000
TIIJ’IE (Sec‘)

Fig. ERRZN Emr in tom] mass balance for pyrolysis products: EXP. R2N

Appendix C

 

 

 

 

 

 

 

 

 

200
30 T I
/'\ J .
\U‘ .
6
I: I8—
U7
3
co ' .. ,.
0 ”...“.5 .‘o .o‘ I: . .0 '
CE) 12 .. : . 5.3.! ...:5' 1...: §;§h:f,;-;:5;&{:’x
Q yf... o ..l. a 0
L...
O
I— 6—
<
59
OII'IIII.IIII;Iwa.
0 200 400 600 800 I000 I200 I400 I600 I800 2000
TIME (sec)
Fig. HVRZN Heat of combustion of pyrolysis prodm; EXP. RZN
1OO'—mi I I rrI‘Tfr'l
7. \ o tar
X I H \ X H20
3 . ..g H O — THC
g3: 30-_ x 2 a”.— co:
W ' ‘. I. f f 1. co
*m o '5 . ﬁ
2!; ,: ¥I§KWVINWM e
E3 “‘25
“_o .-
o" ‘ :'
LI. .‘
8° 4o~".
l- -.
Dz . I
on: . TAR
. . .0 o o . ’0 .
“E 20-. - I” .I. g!" '-.-‘I “*N
‘2 . - ”UH .-
Np/“"s\~§_s"s» Q‘Q.‘ I -
O-rl*‘l'T—r TI'T'W‘xﬁ'I‘
0 200 400 600 800 1000 1200 1400 1600 1800 2000

 

TIME (sec)

Fig. PSRlN Products of pyrolysis as percent of total mass ﬂux; EXP. RlN

 

 

MASS FLUX (mg/cm’s)

MASS FLUX (mg/cm’a)

Appendix C

 

201
0010 U I I I V I I I l | f r Y I T I I I l’
. ._ HC .
0.09- - - 002 -
cl .--- CO cl
0.08“ x H20 -‘
0.07.‘ o TAR

111111111111

 

 

 

 

o 200 400 600 800 10001200140016230'18'00'2000

TIME (sec)
Fig. SlRlN Pyrolysis products (direct memurement); EXP. RlN
o.1o...,.,.,.,.,.l

 

I ' l T
0.09-I ‘
0.08-

q

0.07%

0.06-
.I

0.054

llLllilll

0.04-
0.03-I
.I

0.02-I
0.01-

1 I l l 1 l

 

-

.pr'" '
...:- co

(3.00-“u?" ,— 1 I h r‘n—T—T‘l—T—ﬁ—r—I—f I , I r
0 200 400 500 800 1000 1200 1400 1600 1800 2000

a!”

‘

 

 

TIM E (sec)
Fig. SZRIN Pyrolysis products (after catalytic combustor); EXP. RlN

 

Appendix C

 

202
0-4 r* I I I . I I ' I I I I I I I I ’ 7
M THC
- C07 i
”A. CO .
U). _, \: H90
I" 0.3-I “

A TAR

O
[\J
l
3 ’
_1_..- -4." a.-__i A---

 

 

III/188 FLU >1 (mg, ’c m

 

 

 

 

 

 

 

I I
O O 4"” I
~ I' I I I I r r I I I T I I I I I I I I I I
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. Sill/MN Pyrolysis products (direCt measurement); EXP. M4N
0.8 I I I I I I I I Iﬁ I I I .I I I
I J
O "_. ..
13 « .
II; 3.6— HO -
C V
.1; /’ W’ -I
r n ;_ . ..
g \Jov
I: ~ 1
v 0.4— _
.: 0.3— -
gq ‘
m 3.2— -I
<1
:5 ‘ CO2 «
C e_ A * i __ ‘fff 4
/:'\ I, --------- ”r---
F: ~~~~~~ ”’ C0 +
0.0

 

 

 

0 200 ‘ 400 600 800 7000 “12000 74001600 ‘800 2000

“ME (sec)
Fig. SZM4N Pyrolysis products (after catalytic combustor); EXP. M4N

PRODUCTS 0F PYROLYSIS
AS PERCENT OF TOTAL MASS FLUX

 

 

 

 

 

 

203
1001
o tor
‘< I420 -
_ ——- THC
80*. -— C02 _
H20 -- C0 1
60“ \fw _.
40- a
r 9"
E \$ TAR
20" \‘Mm —
P,—:'5:-::NL:‘ ;;;;;
O'¥¥rT—’Ii F’ r V‘ I L I I F’ I I I I I I I I I
0 200 400 600 800 100012001400160018002000

TIME (sec)
Fig. PSM4N Products of pyrolysis as percent of total mass ﬂux; EXP. M4N

 

I ' I ' I ' I r I ' Y ' I I ' I

Hydrogen ‘

Carbon

II

*
r

A

NUMBER OF C & H

 

O I

 

 

‘ I T I ' I ' I ' I . ' I I T ‘ . ‘
0 200 400 600 800 1.000 I200 1400 1600 1:800 2000
TIME (sec)
Fig. CHM4N Number of C and H atoms in the products of pyrolysis; EXP. M4N

IIELL)

CHAR

Grams [ ]/Grom

204

AmnmﬁxC

 

(18—

(16—

(14—

(12—

 

(10

 

- -..L... ..-...L-_.

L

 

-____L___J___.L

 

TIME (sec)

FQIYCMWN lamrﬁddemnmmﬂthwuanXleﬂl

I ' ' T I f I r I r I I I I I r
o 260 460 600 800 1000 1200 I400 1600 1800 2000

 

L0
(19-
C18—
0f7+
(16;
0.5;
0.4!

(13-

02-,
0.1 —.

-I

 

0.0 I

Hydrogen Carbon

 

 

TI M E (sec)
Fig. GGM4N Gram I I/gram of CxHyO; EXP. M4N

I I I I I I I I . I F’ I I I r’ I I : f’*.
0 200 400 600 800 1000 1200 1400 1600 1800 2000

MASS FLUX (mg/cmzs)

)

0"
O

ERROR IN TOTAL MASS BAIAHCE (

Appendix C
205

 

 

53
m
l

-— [hp "
‘ .mcoz:lilm+mnzo-mAOY2 -I
DO ’ r T If I r f 7—7' R T . .

O 200 400 500 800 I000 1200 1.400 1600 1800 2000

TIM E (sec)
Fig. TMM4N Total mass balance fa pyrolysis products; EXP. M4N

 

 

 

 

 

50 f I Y- T I I t 1‘ 1* T T— .r T— v j— T T f T
'n :. I
-.l'.
30~ x. . 3‘3» 4
' o
z ‘0.
.4 I .0.
. ‘ ‘
'0 a. :5; I. .SLA '0‘ :00 A... 0..
1O—I ". "ﬂ-f , :o'.o ' "'3' I? -«
' c . 0". ." ~q '
' e n o.
o ’
J». .......................................................... —l
-10- S
-I ..
I
-50 i

 

 

C3 200 ' 400 600 800 ‘Io'oo'Izoo 14001600 1800 2000

TI M. E ( s e c I
Fig. ERM4N Error in total mass balance for pyrolysis products: EXP. M4N

HEAT OF COMBUSTION (VJ/g)

HEAT OF COMBUSTION (KJ/g)

AppendixC

 

 

 

206
3 O I I T ﬂ. I r F
24—
1 ..
I 8 d T
4 a
_4 -4
I 2 —I '—
‘ ctt‘fm . g
r. o . ‘ . . . ‘M _
'I W‘ Wﬂﬁ'ﬁlﬁﬂ$ﬁ{~a ti 3"!va ' °
ES-I T
-I- a
0

TIME (sec)
Fig. HVM4N Heat of combustion of pyrolysis products; EXP. M4N

IIII IIIII'IIII
o 250 450 600 800 1050 1200 1400 1600 1800 2000

 

 

I ' I

. . . ‘
.‘It. ° h' H 'M'b'ipﬁ‘t

I

I T

:w.«..-n.r~.~s~;w.~tx-\~x

I

r I

 

-

 

 

0

r

200

T

IIIIII
400600800

T

r
1000

r

TI M E ( s e c )
Fig. HVMBN Heat of combustion of pyrolysis products; EXP. M3N

r

T r I I T « F
12001400160018002000

 

MASS FLUX (mg/cm2.s)

MASS FLUX (mg' "cm2.s)

Appendix C

 

207
04 T I I I I r If I ' I t I
— THC
+ - 002 4
C0
. X H20 _J
0.3“? TAR
J H20 -1
0.2-I 3‘s, —
. I
. In; T...
0.14 ‘V 4

     

-m-~--
”OWOdOQ”
... q. ”CV.---o-
v -—’ \ -v'
I -

 

 

0.0

II I IIIIIIIIIIIFI
o 260 460 600 800 10001200140016001800 2000

 

TIME (sec)
Fig. SIM3N Pyrolysis prodmts (direct measurement); EXP. MBN
0.8 I I I I I n I I I I I I I I I I I
I
0.74 a
0.6“ -—I
0.5-4

 

 

p-—__------------—-—-—---—--—--

 

 

0.0

 

IIIIIIIIIIIIIIIIIII
0 200 400 600 800 100012001400160018002000

TIME (sec)
Fig. SZMBN Pyrolysis products (after catalytic combustor); EXP. M3N

 

NUMBER OF C & H ATOMS II‘I CXHyO

 

 

 

 

208
10C) I’ I I II I I’ T I I r I I
I o tcr

>: I —- H20 I

3 I" ----- THC
92:; 80% ——- coz I
In -- O
g3” . O .

‘- H
‘33—: 60— 2 I
ELIE
59

LL
(DO A
5: 4o. “....
C3361 4 “A TAR y
CDL)
“E
0'0. 20— s

m

< ‘ WOUN—

/\—»~’““‘ . V“ ....... 7: :37??? ....................................... x
O "'Iﬂ'I IfI I I I I rﬁ II ﬁIrﬁrI
O 2'30 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. PSM3N Products of pyrolysis as percent of total mass ﬂux; EXP. M3N

UT

 

I I I I U I l I U I I I I I r l I I

L

4-I I

.- Hdro en
_ A Y 8

.3 :ﬂg" a'I-ﬂ-xh¢af5nuh(“qdﬁf‘ﬂ'—‘~—N~F-’~“’-’
#V

-_b [\J
I E . L

L

Carbon

‘_‘

O

 

 

 

' I I I I I I I I I I - I I I I - I -
O 200 400 600 800 1000 1200 I400 IBIOO 1800 2000
TIME (sec)
Fig. CHMBN Number of C and H atoms in the products of pyrolysis; EXP. M3N

 

 

 

 

CHAR YIELD

 

 

 

0.0 .r.,.,.,.,.,.r,f,lf
0 200 400 600 800 100012001400160018002000
TIME (sec)

Fig- YCM3N Char yield (mm chm/gun wooa); EXP. MBN

1.0 y—r . . I 1 1 f1 fr I I 1 ﬁ' j I

 

09:: Oxygen -
0.8—IMM"? — v -
0.7-:
_I
0.64

.J

0.5-

.I

0.4-I

I....L_--I__.I-_-

1 L__.I._-

111]

Grams [ ]/Grom

0.3-I

.1

0.2 - Hydrogen

d

O . I —I-\--:;‘w —“v ,

A

 

LIJILI

.: _ Carbon

0'0 . I T r f ' I I F ' I I I r I ﬁ— I I I I

O 200 400 500 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. GGMBN Gram [l/gram of CxHyO; EXP. M3N

 

 

 

MASS FLUIIZ (Ihg/ICFTIQLS)

ERROR III TOTAL MASS BALANCE (3:)

210

Appendix C

 

 

 

 

 

 

 

 

 

0.2- —
I — .
o 200 400 600 860 1000 1200 1400 1500 1800 2000
TIME (sec)
Fig. TMM3N Total mass balance for pyrolysis products; En! M3N
50 I I I I I I I I I I I I Iﬁ I I I I
30d 3’23 _
0".
. . . 'z...” ...! ...: .{:$
IO -' 0. . a'.&‘ .‘ 5“. .. \I‘). a o _I
I- .................. E: 5“?“ ETI’IT'If.f;"::~"“1u
-IO-« .
-30-. s
“‘50 I E I I I . . I T I I I I - I I , ;
O 200 400 600 800 1000 1200 I400 1600 1800 2000

TIME (sec)
Fig. ERMBN Error in total mass balance for pyrolysis products; EXP. M3N

 

MASS FLUX (mg,"cm2.s)

MASS FLUX (mg/'cmzs)

 

211 Appendix C

 

 

  

 

 

0.5 I I - I . , . I . ‘ I '
--~— THC
.I —- 002
C0
0 4- .\ H20
.3 TAP?
(13—
(12- ~
' H20
0.1 - _
TAR
.I.
O O —.—.__~.c.-mun-"J?".72.-.-.-.---m--m---ﬁ--a—- ---:~-r/; ﬂ r ‘
O 200 400 600 800 1000 I200 1400 1600 1800 2000

TIME (sec)
Fig. SIMZN Pyrolysis products (direct measurement); EXP. MZN

 

 

 

 

 

 

 

0.8 ﬂ 7 ‘ T V 1: ‘ I r . x ' v
— C02 q
‘7 "" CO
OJ ‘ x H20 ‘
(16— -
(15- S
0.4— ~
0.3-1 H20 -.
‘M - — ~
(12— r
0.1— _.
0.0 T "‘ ': """""

 

 

‘ z - Y ' T ' I ' I ' I ‘ I ' l . *
O 200 400 500 800 1000 1200 1400 1600 1800 2000
TIME (sec)
“8° SZMZN PYTOIYSiS products (after catalytic combustor); EXP. MZN

PRODUCTS OF PYROLYSIS
AS PERCENT OF TOTAL MASS FLUX

IIIIMBER OF C &: H II! C.v

Appendix C

 

 

 

 

 

 

 

 

 

 

212
‘100 . f c : I c
"-1... o tcr
. I.‘ .. H20 -
80 " N’_,/a{—mw-,'HC
n20 ... 1“; 552 “
1"! sf \f/Vﬂ ‘
. /
60a 4" —
404 -
3 TAR
4 Fa
: “.N'AI' ,
’) ' b ' n"
...O-I :.’ V‘IH/‘NNH‘ _I
-I -' -
.j' w T-?.IT._’I..‘1W_T-~:-:*.¥—=:’::lq‘"""R
o—ﬁzr’" e r f T 2+! . T I
O 200 400 600 800 I000 1.200 1400 1600 I800 2000
TIME (sec)
Fig. PSMZN Products of pyrolysis as percent of total mass ﬂux; EXP. MZN
5 I I I I : I I I r I I I I
. 4~.
.I .J
p," Hydrogen
3-—I :I’ .' ‘/ ”V --4
. V/‘ V
24
I— _
1 Carbon
0 . T : I I T Y . ' T ‘ ; - '
O 200 400 600 800 I000 I200 I400 1600 I800 2000
TIME (sec)

Fig. CHMZN Number of C and H atoms in the products of pyrolysis; EXP. MZN

 

 

CHAR YIELD

Grams [ ] "(3mm

AppendixC

 

 

 

 

 

 

 

 

 

 

 

213
1.0 + I . r I
0.8-I -
~ -I
W
(Is—II —
I5
0.4-3' .
0.2-
0-0 I I I I I I I I I I I I I r I I. I T i
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. YCMZN Char yield (gram char/gram wood); EXP. MZN
1 O I I I T I T T l
..I
O.9~ 4
IBM _ — I
‘ f,‘_ w V .
0.7—. -_
0.6ﬂ _I
0.5-I ._.I
0.4- -—
O.3—P —«
O . 2 + - Hydrogen ‘
4A; __ v‘d‘.‘ A ﬂ 1
O.I~ _ _.
. Carbon .
0-0 I r R I I I I I I T r I I r T r f 3 r
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. GGMZN Gram [ ]/gram of CxHyO; EXP. M2N

 

 

 

MASS FLUX (mg/cm2.s)

ERROR IN TOTAL MASS BALANCE (75)

0.3

214

 

 

T I Y I T r r

. WW”:

 

 

0.0

50

. , . . T f , , . . ,
o 260 460 660 860 1030 12‘00 14Too 161001800 2000

 

TIME (sec)

Fig. TMMIN Total mass balance fa pyrolysis products; EXP. MIN

 

I I ' I ' I f I ' 1 j I ' I F r I
30.. —I
‘0 ‘ . q
o . . 1 . 2 O
10.- o I o O o 0...:. o .. 0" .... ......o: .. .. -I
P g... . .50 1 ‘ '- ..«.:‘ 0'
.° 0 ‘ . .0 . ... . 0 . .o.".
1.00,.3’ ......... W---;,;~.’-"-q‘-d‘.-*;-...‘t-.’--=“ ....... «—
o o .‘ '0' “O . ‘ ‘.... . o . ...
. . O . . .é" .. .‘ .. O ‘. . ‘ . O 0 .
-10- ... ...0 .0 V. . \.o. . . .0 d
‘ I .. 0 $ 0
o 3'
. \ o . . u
‘ . o . o
p
-50

 

 

Y I - V r I I' Y I V ‘— I
o 260 450 600 860 1600 who 14‘00 1600 who 2000

 

TIME (sec)

Fig. ERMIN Error in total mass balance for pyrolysis products; EXP. RlN

 

 

MASS FLUX (mg/cm’.s)

MASS FLUX (mg/cm2.s)

AppendixC

 

 

 

 

 

215
0'20 ' I I r I I I I r r I ' I
— THC
- C02
I -- c0
— H20
0.15" O TAR
HO
0.10 2
0.05-
TAR
0.00 . ,ﬁ—FJ‘,’ . M ,
O 200 400 600 800 10001200140016001800 2000
TIME (sec)

Fig. SlMlN Pyrolysis products (direct Went); EXP. MIN
0'3'll'l'l'l'l'l'l'l
0.2-4

. H20
0.1-
CO
........ I CI?
0.0- --------------------

 

Fig. SZMlN

 

 

 

 

W, I I l ' , r
0 200 400 600 800 1000 1200 1400 1600 1800 2000

TIM E (sec)
Pyrolysis products (after catalytic combustor); EXP. MIN

 

Appendix C

 

 

 

 

 

 

216
100‘l‘l'.l'l'l‘l'l'l'-l‘
o tor \ H0 .-
X " “2° \ﬁwﬁw ‘
3 —- THC
A _-
(IILL 80‘ - C02
03‘" -- 00
).(D
30;;
._l —I _
E5- 60
35 <
:32
88
0:01
a: 20" TAR "'
(D
“ VII 1'
{J "' W‘-
0- .
0 200 400 600 800 10001200140016001800 2000
TIME (sec)
Fig.PSMlN Products of pyrolysisaspercent of totalmassﬂmt; EXP. MIN
O>‘5 lrl'lﬁﬁl'r‘ff'lr
I
x .
O
z 4- -
g . .
I'O' 3‘ .\ Hydrogen -
< . i‘Jl-M
I 1.. o _4
«I 2- J _
0
LL ..
O
o: 1- _
E
2 -. Carbon
3
Z 01

 

 

 

*T‘I‘I‘Trl‘l‘l‘l'
0 200 400 600 800 10001200140016001800 2000

TIME (sec)

Fig. CHMIN Number of C and H atoms in the products of pyrolysis; EXP. MIN

CHAR YIELD

HEAT OF COMBUSTION (KJ/g)

217 Appendix C

 

 

 

1.0.,.I.,.,s,,,,,,I,IE

.. -I
0.8-I -
0.6 .J
0.4 1
0.2- q

-I .
0.0

 

 

 

0 I 250 r400 ' 600 I 800 ‘1OIOOI12‘00'14100'16100‘18230'2000
TIME (sec)
Fig. YCMIN Char yield (gram char/gram wood); EXP. MIN

 

 

 

 

30 1 I ' I ' I ' I j I ' T I r ' T ' I
J -I
J -1

24J -
.I

18d 4
-I 4

12~ -
-I . -
-I ' . ~ .

5‘ . . 3' ,. , t . . : ,, .41.“! 39:9. {Aggfw -
1:" ...". ‘5. “‘9'. =°- 'I«1°‘I‘:'-.°5~"I;<'°~‘:I'J"- :0 °.- -.-'= : . ' x" .
‘ ..‘ﬁ...°'.;"o.f .55 .g ' ‘ . ' '° 0 o

0 .° . 1

0 I I I I .f I . I I I I I I I I I I r I

O 200 400 600 800 1000 1200 1400 1600 1800 2000

TIME (sec)
Fig. HMIN Hat of combustion of pyrolysis products; wood; EXP. MIN

2.23)

em

f.
I

.ASS FLUX (mg,

I

ERROR IN TOTAL MASS BALANCE (7%)

Appendix C
218

 

 

 

 

0.3
s
-I .0
°-.:.'
. /‘-° :, ‘ ‘ﬂx
." | a: o We'r‘1?’ ‘
p.24 i ,é. .... :. ‘1... . .‘0.°:’;. .. ~" :I \i A\
\1/ '0'"...
s.
.I
0.1 d
4i
qmp .
0.0

 

O

50

I ' I I ' a . . ' . ' ' . I
200 400 600 800 I000 1200 I400 1600 1800 2000

TIME (sec)

Fig. TMMZN Total mass balance for pyrolysis products; EXP. MZN

 

SO-I

‘-

I r I ' I ' . ' I ' I ‘ I ' I ' I

4:
ﬂ‘. 0 O Q ’
I I o O : o o. .
o. '0 :0. . ' g ’0 g ‘ o '0 O .
.00 '0 . : . . .. ’ 0 0.. . $ :~ C... to J.°f‘ Ito... .. o" a
ﬁ ‘ ~o . 00 .0 .0. \\‘¢' ‘9 0. o 0 0'.
0 . ' 0..-:23- ' \4 -~ ' f ' "
. --. ...... %.‘.'.. . . a 0‘ ‘ ‘: ........ '-%—...-'-'.. ..........
o O 0 ’g '0 . o ‘
0 0.00 c 0‘. .l o n
' a... I o. o 0 ‘

 

 

0 l 200 ' 400 ' 600 I 800 I000 I200 1400 16001800 2000

TIM E (s e c)
Fig. ERMZN Error in t0tal mass balance for pyrolysis products; EXP. RZN

 

 

 

 

MASS FLUX (mg,'cm2.s)

.8)

‘-

IIIASS FLUX (mg/hm

I

 

 

 

219
213 I I LL
~ — - HC
~ __ C02 I
1 - 00
5‘ , H20 ’
4 I“ A TAR
‘ ‘1
-I H -I
II
I.2~ '\ a
I \
—I I \ J
_+ I I _1
I g x \I\ r"‘\ ’h CO d
0.8 _‘ :ll ‘\ A, l y 4P\\Il 2 -I
-I I -I
‘ :I II
‘ 'I H O ‘
0'4‘: “w 2 “
". I: 'I
. I." .I
O L
4 -I
0.0%: I

 

*IIIﬁI
0 200 400 600 800 1000 I200 1400 1600 1800 2000

TIM E (sec)
Fig. SID4A Pyrolysis products (direct measurement): EXP. D4A

 

 

 

200 I I I I r I I I I I r I r I I ' I I
1.6% E
-I \ _‘
.I
I
. I
I .2 ‘I - -I
.I ‘ ._. C02
I ”I‘Vawo'vﬁ‘v' I
(la-I , E
I KM; H20
0.4-I { _
45; -
If I co .
<30 : “ -------

 

IIIIII'IIIII-IIIII

O 200 400 600 BOO IOOO 1200 I400 I600 1800 2000
Time (sec )

Fig. 8204A Pyrolysis products (after catalytic combustor); EXP. D4A

 

 

 

NUMBER OF c & H ATOMS m CXHyO

CHAR YIELD

 

 

 

 

220 Appendix C

5 ' O J a l l l '

4 . O a. _.
'1 1
" '1
..J .1

3 O A _ _
‘ 1" f -‘ 52"." _ .' Hydrogen ‘
J . : ' v "- "1 ' d
4 . I I ' I ' ¥ ..

IZO—J' f? V a _
J - ‘ 1
: '-.' .. '
q 5: . g .‘ Carbon .

1 O- ' 3'". "- .3 ”v 5 4
d . o.~o h. :v‘ .v. \J~ q
J.‘ v .
4 J .

013‘ I

 

 

‘I‘I‘f'rtlfwrrrl'IT
O 200 400 600 800 100012001400160018002000
TIME (sec)

Fig. CHINA Number of C and H atoms in the products of pyrolysis; EXP. D4A

 

 

LOrlrﬁ‘I'f‘I'I'rﬁr
.4
0.8—
0.6—' _
:9:
, 4'-
.4
_ 4', e. . . 1
... .o ...c
0°44 "‘:.' -I.'.h.. i
o “... -o"’..
o 011‘
(12+ -
0.0

 

 

o 200 400 600 860 1000 1200 1400‘1600'1800 2000
T! M E (sec)
Fig. YCD4A Char yield (gram char/gram wood); EXP. D4A

MASS FLU}.

PRODUCTS OF PYROLYSIS

AS PERCENT OF TOTAL

Grams [ L-‘Grom

 

Appendix C

 

 

 

 

 

 

 

221
1OD —. _ _ ‘ . . 7 . . . ., v
T '-.- ‘ C0 0 H20
T 2 -- THC
CO
80-1 -- C02
60— :‘ --
A P "‘-'-A‘ J". : H20
’1' V 1.; '-.-' a;
40— . '
20— ° -
. V§\ ..
O‘P-Fﬂﬁa—T‘ r r r r . r . r . l . u '
C 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
‘ig.PSD4A Products of pyrolysisaspcrccntoftotalmassﬂux; EXP. D4A
1-0 I ~ I ' l ' l ' T f I ﬁr r I ' T
0.9
0.81
0.7—4 ' ——f
4p . *1
0.6-4l 7,4 : . —3
.1 a _ a". ‘0‘." . . '5 t ﬁns. Oxygen 4
05.19: 3:5“ :_ -«
4:. . s ’4 ‘
0.44 _c, .1 3°35 :3 Carbon ‘
‘1! ... tzt‘ ‘9‘ .‘J. ‘4‘...“ "
0.34 =~ _
0.2—i ..
f Hydrogen ‘
O 1—§ ... ; -
O-O‘; ' I . l r l r I 1 y 1 .r . r ' 1 F r '
T 200 400 600 800 1000 1200 1400 1600 1.800 2000

TIME (sec)

ig. GGD4A Gram [ ]/gram of CXHYO; EXP. D4A

 

 

MASS FLUX (mg/"cm2.s)

MASS FLUX (mg/cm2.s)

Appendix C

 

222

0.2. , . T . , .

——- HC

-- C02 4

co

/ H O
0.3— 2 ~
0.2- -—

0.0 "

 

 

C
C
d I _
. ﬂ
- \
P _ ’ .
I
I

 

 

 

 

. . . . I r T ‘ T ' I r l T

o 260 460 660 800 1000 1200 1400 1600 1800
TIME (sec)

Fig. 8103’“ PleSiS W8 (direct measmement); EXP. D3A

 

l I 1 l l r I r I r j T r 1 I

 

 

ﬁlrl A

‘ T ' I ' l T T T E T l T f.
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
F ig. 5203A Pyrolysis products (after catalytic combustor); EXP. D3A

 

PRODUCTS OF PYROLISIS
AS PERCENT OF TOTAL MASS FLUX

IOO

223

Appendix C

 

80— _

60-

0 I r I I I I r F ‘ T F F ' I I I ' I ‘
O 200 400 600 BOO 1000 1200 1400 1600 1800 2000
TIME (sec)
Fug. PSDBA Products of pyrolysis as percent of total mm ﬂux; EXP. D3A
1.0 I I I I I I I I I I I I I I I I
0.9— I
0.8— I
0.7— a j
. -' 'x_ T
0.6— " -'- F
4.. \x_ Oxygen . 5‘ .' '1 T
0.5‘” . ky-uM'ﬁjzso: 33.51 :17 .:
O A .1 . ~’\' E a; :5 3i . T
_4 ’. O ’ .d‘ I“: :‘J ‘0 no .- _
1 ' ’4’ J Cum.” ‘3 '14" ‘ z
o 3— - .e/ \’ T 1
. "v’ T
0.2— .c
0.1 _‘1 A Hydrogen ‘W -
xv— ﬁ T
0-0 I I . I I I I I I I I I I I I I - I I
O ZOO 400 600 800 1000 1200 1400 1600 1800 2000

 

THC

J

 

 

 

 

 

 

 

 

 

TI III E (sec)
Fig. GGD3A Gram [ l/gram of CxHyO; EXP. D3A

 

 

 

Appendix C

 

 

 

 

 

r.‘
L‘ t
....

III

‘/I

CIIAR

224
O 5
>
I
x I .I
O .
z 4* "
(é, .
E 3— :m.\v“-:a- 5‘: Hydrogen _
< I; y" .
.t ’5‘ 4
I _J' x": I"
‘5 2~ / -
U ”I t 3 Carbon
.- .-I ' r‘
LI. 1; '\ . \ 3
O lh .s “a l.‘ 10::
a: 1‘3 ‘."/.~ ' "“‘€ if T 3; 4
LU \/ 's’
a: 7.'
2 I
3 o
z I I I I I I I I I T I I I I I I I I I
O 200 400 600 800 1000 I200 1400 1600 1800 2000
THAE (sec)
Fig. CHD3A Number of C and H atoms in the products of pyrolysis; EXP. D3A
1-0 I I I I I I t - I I . I 1 I I - I
0.84 2
0.6“ ‘4‘ ._t
. \. I
d 6‘
\ .
(l4~ K3££. 5 ~
"' '-. . .. .'-.-*:'.
- ”...-...: I .0:- 1’ Q“ _
0.2.. _.
1 .
0.0 .° . 7 . .. i , - w . , . . . -
O 200 400 600 800 1000 I200 1400 1600 1800 2000

 

 

 

TINT:

It‘—

(sec)

Fig. YCD3A Char yield (gram char/gram wood); EXP. D3A

 

Appendix C

 

 

 

 

 

 

 

225
'3.4 IT ; V : f 1

Ex

"I -..j- ._
E

U

o?

E

T" 0.2d ..
5

_J J

LI.

(fl

(fl _
<1:

2

I ,—«“ ’
O-Q‘IIIIIIIIIIIIIIIIIII

O 200 400 600 800 100012001400160018002000

TIME (sec)

Fig. SIDZA Pyrolysis ploducts (direct measm'ement): EXP. D2A
I-OIIIIIIIIIIIIIIIﬁIII
0.9— _.
0.8; “
or! —~

 

MASS FLUX (mg/ ' cm2.s)

 

 

 

 

CO2 _‘
{MM 2
H20 ‘
CO T‘
1 .- ___________________
0'0 r T 1 ‘Iﬁ ' I F T T I I I T I I r 7
O ZOO 4OO GOO BOO 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. 8202A Pyrolysis products (after catalytic combustor);EXP. D2A

CHAR ‘I‘IELD

Appendix C

 

 

 

226
O) 5 I I T t n T I I I ‘ T
I I
x s
O
— Hydrogen
(D ‘t
IC2 3— Of '
< 2&1”ka J
I I '-. ....I
‘8 9—4 'tv/ ‘
0 ' Carbon .
L5 -4 c
E TJ “Raf/W
(D .5 ° -
5 I
3 o

 

‘ ' T ' I I I I I I T I I I
O ZOO 4OO BOO BOO IOIOO 1200 1400 1600 1800 2000
TIME (sec)

Fig. CHDZA Number of C and H atoms in the products of pyrolysis; EXP. D2A
I .O . ,

 

I ' I ' T 7 I . I

0.8—
0.64.

04-2: «I In.

0 2 . WJ\(\'.¢W€?.OE

 

0.0

 

 

IIIIIIIIIIIrIIIIItI
O 200 400 600 800 1000 IZOD 1400 1600 1800 2000
TIME (sec)

Fig. YCDZA Char yield (gram char/gram wood); EXP. D2A

PYROLYSIS PRODUCTS

AS PERCENT OF' TOTAI. MASS FLU?

Groms [ ]/"Crom

100

Appendix C
227

 

 

 

O

1 J 4

1

 

 

 

 

 

- I I T I I I I I r I I I I I t I T I
O 200 400 600 800 1000 1200 1400 1600 1.800 2000
THAE (sec)
Fig. PSDZA Products of pyrolysis as percent of total mass ﬂux; EXP. D2A
1.0 I I I - I I I I f I I I ‘ I ‘ I
0.9-
0.8-'_ -
0.74: _.
O.6~- . -
‘ '-.:'r\'\ Oxygen J
0.5- " ‘
O.4— :‘t ‘
. I ". Carbon 1
03-: _
0.2—j -
O.1-' \ * deroggn _
./ _—
0.0 f r I

 

 

 

.1. I.
O 200 400 600 80

. I I I I I I I I I I

1000 1200 I400 1600 1800 2000
TIME (sec)

Fig. GGDZA Gram [ ]/gram of CxHyO; EXP. D2A

MASS FLU): (mg,‘cm2.s;)

MASS FLUX (mg,/'cm2.s)

Appendix C

 

 

 

 

 

 

 

 

 

 

228
O..-
".j—I
0.2—
0.1— 5 . -
..- Mﬁ/
OO/VIIIIIFIIII'IIIIII
O 200 400 BOO 800 1000 1.200 1400 1600 1800 2000
THAE (sec)
Fig. SlR3A Pyrolysis products (direct measurement); EXP. R3A
LO I I I I I I I I I I I I I - l
0.9- -
I
0.8-I
_I
0.7-I 1
0.6-1 .5." _.
O 5: {f 1' C02 ”'3’“ _:
. d ... ‘m~'.,:'. :" ..w'stwA-Mw:~v -I
.' #-
O.4— ;' —
0.3: I}! H20 _d.
o.2~ ;_' _
I 2' co ‘
0.1ﬁ.‘ ’/.—r‘~\‘/-—_”~’_.‘ ______ -
IC/ — ‘ — -I
O'OJ‘I'ITI’T'E'I'I'I‘T
O ZOO 400 600 800 1000 1200 1400 1600 1800 2000

THME (sec)
Fig. 82R3A Pyrolysis products (after catalytic combustor); EXP. R3A

CHAR YIELD

Appendix C
229

 

 

C? S

‘If‘

LE

_ .I-

U7 .

E): "—I
o

<E

I I

% 2_‘

C.)

LL. .4

O

o: 1“

LL]

G3 .

‘5

:: O-ﬁr

' I
5 Hydrogen a
I: ‘1. g." ‘ .1
' ' - It 4" "A
I . ~ '0 A “ I T .
I . ‘ .-a c. e.’ 0~§ II Q I. ‘H‘ :3
- ; «...: '- . -. I . .
A n a L I - N. I
‘ I. .' i e- f ' I —I
'5..- ‘V \- '._ -
‘I
- Carbon ’
. A .0
d ‘t ' o
a g, [xx-1.; 'w-"~"WNJ-J W \090‘ °

 

 

IIIIIIIIIFIIIIIIIII

200 400 GOO 800 1000 1200 14001600 1800 2000
TIME (sec)

Fig. CHR3A Number of C and H atoms in the products of pyrolysis; EXP. R3A

 

1.0
0.9-
O.8~

0.7- t

0.61.
0.5;
0.4.2
0.3L
0.2!
0.1.3

 

0.0

$
. _
. -—I
Q
$ \ ..
o~5 ~
Q‘- 0‘ I $.éa. ‘1. :'
“C . . . —I
~ ' o. o ' “0‘ 5 a” s O. ’ °
0 ‘ .f O F . ’3’:’ ’ .0” ‘. ...“ . ‘ 0. d
O O . . .
...o '0': ... .. '0‘. ~‘ %' c o :
.0 ‘ " .VJ

[ . , . U I I I r r . I r

 

 

Iobo Isoo

Tl M E (s e c)
Fig. YCR3A Char yield (gram char/gram wood); EXP. R3A

560‘

Appendix C
230

 

 

 

 

5 .
o’lT- 80— _
""(f) ..
93 In
LJ<J§ ‘ T
E". ‘ H o -
if: 60‘; 2 I
tho 4 .W 1
C"- .
WELD - ft- 7'le
+- 40— s E;
0+— _/ .-I /I s.- I co ~
3: T o' . -\ m2 A / ..
DUJ _1 _' v’ \ w / \J m A A I I.
00 n ‘ \‘ V V" \ﬂ.‘
0:01 4 ' '“ CO :‘I .
Q'LLJ 20 '- l/ / \ /\d\ s \ /\ - “ /- ‘ ‘J'

C— -a: / /\\/ / ‘~’\T/HC TTTTTTTTT T ‘1 \V" _

’3 ".le I - I "

(2T: _I ... 37’ Nﬂvww/ Wq/

-I 4/ .4
o-é7
O

500 1000 ' IISbO' 2000

TIME (sec)
Fig. ‘PSRBA Products of pyrolysis as percent of total mass ﬂux; ED. R3A

 

 

 

 

 

1'0 ’ T ' T I I I I I I I T I I I I r I r I I
O.9-‘ _
0.8“ -. _
g or7~ 3 _
r ' Oxygen ‘ g
L. $ .-
‘(D O ' 6 d ~AVM.‘¢ 3“ Inﬁll" - ‘51-“ . a» I; T —
F: .. '5‘}: "VH’ 1’ 5 ..pu‘q. 'W' ‘0’” o-v' '2: _‘
._g <15- - , r
I .
E 0.44 (3 ._. . Carbon ...f ~
8 -9 d ’5..- s‘V’ a", ‘Eo¢‘:.~;’ ..., ...; “K “...”,‘fkﬂ's' \... ' o :.
CD on)“ ..I—t' a; a d
0.2- : 4
J . Hydrogen ‘
O '1-if ‘: _‘v AW— i; v. . ._.:
0.0rrITIVIrvTTrIe..--T.
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

 

Fig. GGRBA Gram [ ]/mm of CxHyO; EXP. R3A

 

MASS FLUX (mg/cmzs)

MASS FLUX (mg/cm2.3)

Appendix C

 

 

 

 

 

 

 

 

 

 

 

231
o... I
0.3‘1 _
0.0-Ir . I . , r T I ,
O 500 1000 1500
TIME (sec)
Fig. SIRZA Pyrolysis products (direct measurement); EXP. R2A
1.0 r I I I I I I I I “I
0.8- 5
0.65 -*
0.4— —
0.0- , T r , r T , r ,
O 500 1000 1500 2000
TIME (sec)

Fig. 82R2A Pyrolysis products (after catalytic combustor); EXP. RZA

100

Appendix C
232

 

PYROLYSIS PRODUCTS
AS PERCENT OF TOTAL MASS FLUX

 

 

 

 

 

 

TIME (sec)

Fig. PSR2A Products of pyrolysis as percent of total mass ﬂux; EXP. RZA

 

 

 

r Y I T ‘ . I I I T I Y T Y z T I

 

 

 

 

E
O
L
$3
2 0.5:- L
m .
O.4-— _
CE) - Carbon
(5' 0.3: #5.. A__*_;-":"
0.2- _
_ Hydrogen
0.1-J : A ‘v
“"F
0.0-F} . I rﬁ . : ' I ' I - I ' I f I . I .
O 200 400 600 800 1000 1200 I400 1600 1800 2000
TIME (sec)

Fig. GGRZA Gram [ I/gmm of CxHyO; EXP. R2A

 

CHAR YIELD

NUMBER OF C 85 H ATOMS IN CxHyO

 

 

 

 

 

 

 

233
5 r T T I I Y I r I I r r I
4- _
‘ I I
3- _
1 Hydrogen q
2‘ -"\.n-"‘.-a.-~v""""’"T:
.pz'-’-”
1" I .—
.k Carbon
IV 4. ..
4' va “ﬁ— ..
l
l
0 If t r t m 1 f ' r t 1 I y I
0 500 1000 1500 2000
TIME (sec)
Fig. CHRZA Number of C and H atoms in the products of pyrolysis; Exp. RZA
1.0 I I 1 r I T ﬂ 1 r f f 1— r , I I r ﬁ r j
-I
J
j
—4

 

 

.rr .ﬁ.
1000 1500

TI M E (s e c)
Fig. YCRZA Char yield (gram char/gram wood); EXP. R2A

.1.
500

_L.1_.J_ -2 _. _

2000

MASS FLUX (mg/cm2.s)

MASS FLUX (mg/cm’e)

234 Appendix C

0.10 ...r.,

 

0.08-

0.06-

 

0.04

0.02~

TAR

 

0.00 “ _...____..._. ’ '

0 200 400 600 800 1000 1200 1400 1600 1800 2000

TIME (sec)
Fig. SlRlA Pyrolysis products (direct measurement); EXP. RIA

0.10 v I ' I I I ' l

 

 

 

 

 

l ' I ' r u I I

0.084

0.06 -1 H20

0.04 -

0.02

C0 - a a. a“ l; :-
4..::'.-."'u-:--:="'::"”U0-2
0.00

 

TIME (sec)

Fig. SZRIA Pyrolysis products (after catalytic combustor); EXP. RlA

 

«I

 

 

0 200 ' 400 ' 600 i 800 '10'00'12'00'1400'16'00' 18'00'2000

 

 

NUMBER OF C 8: H ATOMS IN CxHyO

CHAR YIELD

 

 

 

 

 

235
5 r . . 1 I . 1 . , 1 1 . I j
4- d
3- ‘
Hydrogen
z-vtﬁf‘d; VA_;__ A‘ -
"I
1- A
' Carbon
0 ...................

 

0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. CHRIA Number of C and H atoms in the poducts of pyrolysis; EXP. RIA

 

 

 

 

 

1.0-......,...,.,...,
0.8- -
0.6 v V -
0.4 -
0.2 _
0.0

 

0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. YCRlA Char yield (gram char/gram wood); EXP. RlA

 

Appendix C

 

 

 

 

 

 

 

 

 

 

236
100 . .
x ‘
3 .
L‘- ao-I d
mm ‘
I—U’ ‘ ..
gs . .
604 -I
a: . .
0.0 .
h d
m d
as ‘
>- 40- -
A d
05 .
gm J «I
m8 .
m d
a II
9 .
. I . . I .
O 500 1 000 1 500 2000
TIME (see)
Fig. PSRlA Products of pyrolysis as percent of total mass ﬂux; EXP. RlA
100 I 7 I F r I I I r r f I y I r , I— 1 1,
H20 I
O 2 3
d 804- __ CO _‘I
In“) . a CO I
I—m I ‘- ﬂ"\ 2 4.
< , . .
8: KW “s W r'/\
8"] 50‘ \‘m/ \j\ ‘ -;
EE ,. H20 , ;
(£2: 4 J‘WV «W’ j
90 404 _*
_JI__ 1.
OZ :
[ILL] .' .4
>-o I - I
(LE P", ;
0. ZO-I 5" E;
m _f I
< .Il \\ .1l
0 A “““““““““““““““““““ - I
I f I I r ' I . f I I , T r I T I I I . s
O 200 400 600 800 1000 1200 1400 1600 1800 2000

TIME (sec)

Fig. PSM4A Products of pyrolysis as percent of total mass ﬂux; EXP. M4A

 

 

 

MASS FLUX (mg/amass)

MASS FLUX (mg/’cm23)

Appendix C

 

 

 

 

 

 

 

 

 

 

 

237
2.0 . .
. . HC
. -— CO
_ - CO
7.6“ H20
«.2.
.4 |I\\
0.8— I I
I I
" I I
I \ CO2
04— : I~x’“\--/I ____________ __
/I H20 \W
0.0 J'AI - I - : I I ImI I .I_ I I In"? mm """'_""" "" .
O 200 400 600 800 1000 I200 1400 1600 1800 20
TIME (sec)
Fig. SIM‘“ PYTOIYSiS products (direct measurement); EXP. M4A
2-0 I I I I I I I I I I I I I I I 7: I I I
d 'I
I..- .. _I
: ‘~' I
I 2; - '° 4
.4 . i. .I
. ’1': CO2 _I
q WIN ‘
0.8“ o s' {N d
.. {0? .‘ﬁ -I
+ ' .' H 0 ..
O.4~ '1‘ i 2 .q
4: u \ 1
. C0 ________ -
0.0;???71-I‘I-F7T. I I I I r I I .
O 200 400 600 800 1000 1200 1400 1600 1800 2000

TI M E (s e c)
Fig. SZM4A Pyrolysis products (after catalytic combustor); EXP. M4A

OO

 

 

CHAR YIELD

Appendix C
238

(J

 

L

L:
I
L

Hydrogen
'7 A» I,"
A {fig-"rd \u': ﬂ'\- j/x\ "‘
I 5%,: Carbon M \g.‘ *

.‘

.11": / -ﬁ

.'T‘I‘IEI'T'ITITT'T‘

O 200 400 600 800 1000 I200 1400 1600 1800 2000
TIME (sec)

Fig. CHM4A Number of C and H atoms in the products of pyrolysis; EXP. M4A

l

L

NUMBER OF c 8: H ATOMS IN CxHyO

 

O

 

 

 

 

 

I -O I I - l I I I T I I "I I I F u I . I
o . 8 I

-I ...: q

's
O . 6 _‘ \3, ‘
‘e
f
'4 o f ’ﬁ" -I
V:\ .A" . “V“ o.
' _I a O .. / ‘ ‘d'. / 0‘ N f

O * ’ .f ‘P' ‘5‘ ..s\ o’ 1

-< -I
O . 2 ..

.I .I
0.0 I S

 

.'IﬁITtTT‘TTST.‘. :-
O 200 400 600 800 1000 1200 I-IOO 1600 1800 2000
TIME (sec)

Fig. YCM4A Char yield (gram char/gram wood); EXP. M4A

 

 

 

MASS FLUX (mg/cmzs)

MASS FLUX (mg, 'cmzs)

Appendix C

 

 

cup-....
I
\
a
\
s
‘ }
|
I
ILJJLL I ll

1111111111

111

1114.1

11111111

1 l

‘-

_ .-----O-----‘
0---- ' ‘--' -\ ‘
-- .---.

 

 

 

 

 

 

0'00 ' F ' I ' I ‘ I T l ' T T I ' T ' I ' 1
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. SlM3A Pyrolysis products (direct measurement); EXP. MBA
1.0 I I I I r I . I r I r I x I r I I I r
0.8-I CO2 ‘I
1' 70.5“ n g -I
' . at" 4N
. clan-f
O.6-I 3. - _.
.’ ". .3 H20
,- u
o 4. 5' :W _.
m I
0.2- 3; ..
.5; co
9...: __________
o o ’5 ----- T _. .-
- I my I I I I T - I I T I h I I I n I
O 200 400 600 800 IOOO 1200 1400 1600 1800 2000

TIME (sec)

Fig. SZM3A Pyrolysis products (after catalytic combustor); EXP. M3A

CHAR YIELD

Appendix C
240

 

0‘
<1

W
[II-V

_- U Carbon

.|_ . _J
g: N

1": 1

NUMBER OF c 8: H ATOMS IN CxHyO

 

 

 

I I I I I r u I I I f r I T I I I T I
0 200 400 500 800 1000 1200 1400 1600 1800 2000
TIME (sec)

Fig. CHMBA Number of C and H atoms in the products of pyrolysis; EXP. MBA
I.O . I T I

 

 

 

I - I I I I I I I I I I F
0.8- ._
f‘.
0.6“ z‘A‘ i. _
q ‘A
w -I
.0 4“
s-r‘
0.4— ‘wg” N —
V
I
0.2— ._.
0.0 r r

 

I ' I I l I ' I ‘ I ' f r . 1 T 7 .
O 200 400 600 800 1000 1.200 1400 1600 1800 2000
TIME (sec)
Fig. YCMSA Char yield (gram char/gram wood); EXP. M3A

 

 

PYROLYSIS PRODUCTS
AS PERCENT OF TOTAL MASS FLUX

PRODUCTS OF PYROLYSIS
AS PERCENT OF TOTAL MASS FLUX

Appendix C

 

241
100 T I I I I I I T I I l
ao-I -‘
eo- -
40- ..
1 H20 d

100

 

 

 

 

IIrIIIIIIIIIIIIIII
O 200 400 600 800 10001200140016001800 2000

TIME (sec)

Fig. PSMSA Products of pyrolysis as percent of total mass ﬂux; EXP. MBA

 

 

 

I U V I U I I U U I r V I I T I

 

l l l l I I l l A l l l l

 

 

 

co .......... ..
-' ’f—nu—g; J
. I . I . . , . . , . , , . . , .
O 300 500 900 1 200 1500 1 800
TIME (sec)

Fig. PSMZA Products of pyrolysis as percent of total mass ﬂux; EXP. M2A

MASS FLUX (mg/cm’.s)

MASS FLUX (mg/cm’.s)

 

Appendix C
242

0.4

 

0.31

0.2-I

 

 

‘ISbo' ' zooo

 

 

 

TIME (sec)
Fig. SlMZA Pyrolysis products (direct measurement); EXP. MZA

 

1.0vI'lIl'l'lﬁl'I'II'

 

 

 

 

 

0.0

o 260 460 600 800 1000 1200 1400 1600 1600 2000
TIME (sec)

Fig. SZMZA Pyrolysis products (after catalytic combustor); EXP. MZA

 

 

CHAR YIELD

(II

Appendix C
243

 

 

NUMBER OF C 8c H ATOMS IN CxHyO

I
0% I v v I I U ‘ t

Hydrogen

    

 

   

Carbon

 

 

300 600 900 'Izroo' 'Is'oor 'Ia'oo'

TIME (sec)
Fig. CHMZA Number of C and H atoms in the products of pyrolysis; EXP. MZA

 

0.8-

0.6-

0.4-

 

l 1— r ' I ' l ' I ' I ' l ' l ‘ I

 

 

.l.,.,.,.,.r.,.,.1_.
200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. YCMZA Char yield (gram char/gram wood); exp. M2A

MASS FLUX (mg/cm’s)

MASS FLUX (mg/amass)

Appendix C

 

 

 

 

 

 

244
0‘12 I I I I Iv I I r I r I
: — H20
‘ --- 002
0.10- - CO
" -— THC
.
0.08-
H o
0.063 2
0.043
1
0.02-
0.00‘ Y ,—-. - w
0 200 400 500 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. SlMlA Pyrolysis products (direct measm'ement); EXP. MIA
0-12 II Iri'l'ﬁ T'I' '
— H20
‘ «'- 002
- 00
0.08-
H20
0.04-
0.00 ‘

 

 

'Irlrf'
02004005008

Fig. 521mg Pyrolysis produc

....r.,.r.r.
OO 10001200140016001800 2000

TIME (sec)
ts (after catalytic combustor); EXP. MIA

 

 

CHAR YIELD

 

0|
—
-
u
d
-I
.4
d
—
J
-
—

A

4- ..

Hydrogen

 

 

Carbon

NUMBER OF C 8: H ATOMS IN CxHyO

O

 

 

I ‘ I n I '* I ‘ I ' I "’1 * Ti ' I ' I '
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig. CHMIA Number of C and H atoms in the products of pyrolysis; EXP. MIA

1.0 -..

 

0.8-

 

 

0.6

0.4

0.2

 

o.o....,.,.,. . . ...
o 200 400 600 800 10'0012'00145016'0018'00 2000
TIME (sec)

Fig. YCMlA Char yield (gram char/gram wood); EXP. MlA

TEMPERATURE (°C)

TEMPERATURE (°C)

Appaldlx'C

 

 

 

 

 

 

 

 

 

 

246
700 I I I I h I I I I I I I I I I I I I I
* _
500 _ FRONT SURFACE # -
-I 2 mm '1
500- ‘
‘ 5.1 mm ‘
400- d
' 10 mm 4
300- ..
200- 22 m -
Ioo-I -
BACK SURFACE ﬂ
0 I I I I I I I I I I I I I I I I I f I
O 200 400 600 800 1000 1200 1400 1600 1800 20C
TIME (sec)
Fig.TD4N Taupaaturemtimeatmiouslocatiousinsidethewood; EXP.D4N
600 I I ' I I I l '— l ' I I I I I r I I
J -I
r -l —I
”00 FRONT SURFACE
400- _
300- ..
200- _
- .I
BACK SURFACE
Ioo-/f/ -
O

 

 

I I I I I I I I I I I I I . I

o 200 400 600 300 1000121001410016l0018100 2000
TIME (sec)

Fig.1'D2N Tanpemtmevs.timeatvuiomlocationsinsidethethewood; EXP.D2N

TEMPERATURE (°C)

TEMPERATURE (°C)

 

 

 

 

 

 

 

247

700 F I ' I ' I ' I ' I I I I T ‘

. FRONT SURFACE .
600- -

-I .
500- 2 m -
400- 6 mm -

1 q
3500- -
200‘ 19.1 mm ' ‘
1 00-/ BACK SURFACE _

-I .

O I I I I I I I I I I I I I I I I I I I
0 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fig.1'R4N Tanpaaunemtimeatva'iomlocmionsinsidethewood: EXP.R4N
500 I I I I I I I I I I I I . I I I I I
500- -
FRONT SURFACE

.I
400- —
300- -
2004 -
IOOJ / BACK SURFACE -

O

 

 

 

I I I I I I I I I I I I I I
0 200 400 600 800 10'0012'00140016001800 2000
TIME (sec) .
Fig.1'R2N Temperaturevs.timeatvariouslocationsinsidethewood; EXPRZN

TEMPERATURE CT»

TEMPERATURE (°C)

 

   

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

248
’00 I I r I r I
600- d
500- -
FRONT SURFACE
400~ —
300— _
200— _
100‘ —
20-3 m BACK SURFACE
0 I I I I I I I I I I I I I I I I I I I
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Fhﬂnuﬂv‘nmmaumnmsdmeuwummsbumanhuﬂemnwnuu EXRJWHI

:00 . , . , . 1 . , , . I . I I .
600~
500~ -

.I
400- E

d FRONT SURFACE A
500- __ _
100‘ 10 mm A

_1EE=:"””ﬂﬂﬂﬂf—ﬂﬂﬂﬂﬂﬂjgjr3m
. BACK SURFACE
0 j—

 

 

 

TIME (sec)

' I r I I I ' T ' I t I r I r
0 200 400 600 800 1000 1200 1400 1000

1800 2000

Fig. TMlN Temperature vs. time at vuious lomtions inside the wood; EXP. MIN

TEMPERATURE (°C)

HEAT FLUX (W/Cmas)

AppendixC

 

 

 

 

 

 

 

 

 

 

 

 

 

 

249
foorft..yPIxIrTPmIrI
600— FRONT SURFACE
500‘ 511“]! ..I
400~ -

10.4mm
300‘“ ._I
.I
200‘ _
I 14.8mm 24mm ..
1004 ———
/, BACK SURFACE -
O ' I ' T ' I ‘ I T I T I ' I ‘ I E I P
O 200 400 600 800 1000 1200 1400 1600 1800 2000
TIME (sec)
Phyrnnaﬂ‘nmmaumeWLmmuuwmmukumauhnmemewmﬂ;IQGVTNKN
5'0 ' I T I T I ' T T I T I T I ' I I I
4L0— ‘
Ar””ﬂﬁﬂ——T J
.10- _
I/ I
2.0- I —
1.0— n
(10

 

 

 

ITIIIIITIIIIIIIIII-

O 200 400 600 800 1000 1200 I400 1600 1800 2000
TIME (sec)

Fig.1'HFMN External heatﬂuxes fa experimentson 17% moistwood.

 

 

 

 

 

 

 

250
800 I T I F f I I I F T r I
I FRONT SURFACE __ ..-.-- -‘
7 O— 2.5 mm _/ ... “I
0 . / / ﬁ” q
/ .o"”..‘
A 600- -’ -
U Ignition / 4.5 mm ’_ ,,
gJ ‘ /
.. / ..
LL] 500 / /
0: - / /
E(- 400- / / '4
0: ' / //
LU 300- "
O. . / / -"' I
E] 200 / // ”u,’
i.— .. / // 22.5 M’a"”’ q
100‘ / ”',v/ // q
P; BACK SURFACE ‘
01 V r I r V I V I I F I I f I I T I I I
0 400 800 1 200 1600 2000
TIME (sec)
Fig.1'D4A Twmdmemvu'iouslocaiousimidethemod: EXP.D4A
800 I l I T I I i I j I i I I ﬁr I I
FRONT SURFACE
.I 1.5 mm .4
5 mm
A BOO-I 10.5 nun-I
0L) Ignitio
LL]
0:
E 400- -
M
U I / .
Q 22 mm
2 / I,
E 2004 , ’ -
4 E __
BACK SURFACE
0

 

 

 

I I I - I I I .I I F . I - I -

0 200 400 600 800 1000 1200 1400 I600 1800 2000
TIME (sec)

Fig. TINA Temperature vs. time at various locations inside the wood; EXP. R4A

 

TEMPERATURE (°C)

TEMPERATURE (°C)

Appendix' C
251

 

700 I I I T I I I I"_*Y I I I I I I I I I

FRONT SURFACE
600— ‘

d
d

500- ‘

u

400- T

‘

 

300-/ T

200- T

d
d

BACK SURFACE
100- ‘

 

 

O

 

' ' ' ' I ' ' I ' l ' I '
o 260 460 650 850 IOlOO 12‘00 1400 1600 1800 2000
TIME (sec)

thTan'nmmuummvsﬁmeawummshaﬁanhnﬂemewuﬂ;laﬂhbmk
700 WI

 

I j I I ﬁ r I r I I

500_ FRONT SURFACE

-I

5004 -

400- ﬁ
300~ —
200- 1

. 4

100- BACK SURFACE —

d

O

 

 

 

' I I’I ' I ' I I I ' . ' ' ‘
o 200 400 600 800 1000 12'00 14'00 1600 who 2000
TIME (sec)
Fig. TMZA Temperature vs. time at various locations inside the wood; EXP. MZA

 

LIST OF REFERENCES

 

LIST OF REFERENCES

Atreya. A... "Pyrolysis. Ignition and Fire Spread on Horizontal Surface of Wood." PhD.
Thesis. Harvard University, Cambridge. MA. (1983). Also published as a National Bureau of
Standards report. NBS-GCR—83-449. March 1984.

Bamford. C.H. Crank. 1., and Malans D.H.. "The Combustion of Wood Part I." Proc.
Cambridge Phil. Soci., 42. pp. 166-182 (1945).

Bradbury. A.G.W.. and Shaﬁzadeh. F., "Role of Oxygen Chemisorption in Low-Temperature
Ignition of Cellulose," FI Combustion and Flame. 37. pp. 85-89, (1980).

Broido. A.. and Nelson. M.A.. "Char Yield on Pyrolysis of Cellulose." Combustion and Flame.
24, pp.263-268. (1975).

Camahan.B.. Luther. H.A.. Wilkes. 1.0.. "Applied Numerical Methods." Wiley Publishing Co.
(1969).

Carslaw. HS. and Jaeger. J.C.. "Conduction of Heat in Solids." 2nd edition. Oxford
University Press, 1959.

Chan. W.C.R.. "Analysis of Chemical and Physical Processes During The Pyrolysis of Large
Biomass Pellets." PhD. Thesis. U of Washington. (1983).

Chan. W.C.R.. Kelbon, M.. and Krieger. B.B.. "Products Formation in the Pyrolysis of Lrage
Wood Particles." Preprint. To be published in Fundamentals of Thermochemical Biomass
Conversion. (1984).

Chan. W.C.R.. and Kreiger, B.B.. "Modeling of Physical Processes of a Large Biomass Pellets
with Experimental Veriﬁcation. ASC Fuel Div.. Preprint, (1984).

Croce. P.A.. Combustion Science and Technology. 14. pp 221-228, (1975).

Cullis. C.F.. Hirsciller, R.P.. Townsend. RP, and Visanuvimol, v., "The Pyrolysis of Cellulose
under Conditions of Rapid Heating." Combustion and Flame. 49. pp. 235-248, (1933).

Delichatsios. MA... and de Ris. J. (1983). "An Analytical Model for the Pyrolysis of Charting
Materials.” Fall Technical Meeting, Eastern Section, The Combustion Institute. p.68. (1983).

252

253

Ehrlich, L.W., "A numerical Method of Solving a Heat Flow Problem With Moving
Boundries." JACM, 5(2). pp. 161-176. (1958).

Eckert, EKG. and Drake, RM. "Analysis of Heat and Mass Transfer,” McGraw-Hill
Publishers. New-York. (1972).

6008. A.W.. "The Thermal Decomposition of Wood." Wood Chemistry. 11. ACS Nomograph
Series. No. 97. 2nd Ed.. Reinhold Pub. Co.. New York (1952).

Havens. 1A.. Welker, JR. and Sliepcevich. C.M.. "Pyrolysis of Wood - A Thennoanalysis," J.
ofFire and Flammability. 2. p 32. (1971)

Havens. J.A.. Hashemi, H.T.. Brown, LE, and Welker. J.R.. ”A Mathematical Model of the
Thermal Decomposition of Wood," Combustion Science and Technology, 5. pp. 91-98 (1972).

Hileman. F.D.. Wojcik. L.H., Futrel. JM.. and Einhom. I.N., "Thermal Uses and Properties of
Carbohydrates and Lignin." Academic Press. New York. p. 49 (1976).

Kansa. E.J., Perlee. HE... and Chaiken. R.F.. "Mathematical Model of Wood Pyrolysis
Including Internal Forced Convection." Combustion and Flame. 29. pp. 311-324 (1977).

Kanury. M.A., "Burning of Wood. A Pure Transient Conduction Model," J. of Fire and
Flammability. 2. p. 191 (1971).

Kanury. M.A., "Rate of Burning of Wood. (A Simple Thermal Model)," Combustion Science
and Technology. 5, pp. 135-146. (1972).

Kanury. M.A., "Rate of Charring Combustion in a Fire." 14th Symposium (lnt’l) on
Combustion. The Combustion Institute, Pittsburgh. PA, p. 1131 (1973).

Kanury. A.M.. and Blackshear. P.L.. "On the Combustion of Wood. 11: The Inﬂuence of
lntemal Convection on Transient Pyrolysis of Cellulose." Combustion Science and Technology.
2. pp. 5-9 (1970).

Kanury. A.M.. and Blackshear. P.L.. "Pyrolysis Effect in the Transfer of Heat and Mass in
Thermally Decomposing Organic Solids." Eleventh Symposium (Int’l) on Combustion. p 517
(1967).

Kanury. A.M.. and Backshear, P.L.. "Some Considerations Pertaining to the Problem of
Wood-Burning," Combustion Sciences and Technology, 1. pp. 339-355 (1970).

Kanury. M.A., and Holve. D.J.. "Transient Conduction With Pyrolysis (Approxim ate Solutions
For Charﬁng of Wood Slabs," J. of Heat Transfer, Transactions of ASME. 104. p.338. (1982).

Kashiwagi. T., "Radiative Ignition Mechanism of Solid Fuels," Fire Safety Journal, 3. 185-
200, (1981).

Kashiwagi. T., Ohlemiller. T.J.. Wemer. K.. "Effects of External Radiant Flux and Ambient
Oxygen Concentration on Nonﬂaming Gasiﬁcation Rates and Evolved Products of White Pine,"
Combustion and Flame 69. pp. 331-345. (1987).

254

Kashiwagi. T., T]. Ohlemiller, K. Werner, "Wood Gasiﬁcation at Fire Level Heat Fluxes."
Combustion and Flame, 69. pp. 155-170. (1987).

Kilzer, FJ.. and Broido, A.. Pyrodynamlcs. 2. pp. 151-163 (1965).

Krause. Jr.. RF, and Cam. RG.. "Rate of Heat Release Measunnents Using Oxygen
Consumption." J. of Fire and Flammability. 12 pp. 117 (1980).

Kung. H.C.. "A Mathematical Model of Wood Pyrolysis." Combustion and Flame, 20. pp.
185-195. (1972).

Kung, EC, ”The Burning of Vertical Wooden Slabs." 15th Symposium (Int’l) on Combustion,
The Combustion Institute. Pittsburgh. PA. p. 243. (1975).

Kung. H.C.. and Kalelkar, A.S.. "On the Heat of Reaction in Wood Pyrolysis." Combustion
and Flame. 20. pp. 91-103, (1973).

Landau, H.G.. ”Heat Conduction in a Melting Solid," Quarterly of Appl. Math. FBVIlI. No. 1,
(1949).

Lee. C.. and RF. Chaiken, and JM. Singer, "Charting Pyrolysis of Wood in Fires by Laser
Simulation," 16th Symposium (Int’ l) on Combustion, (1976).

Lee. C.K.. and LR. Diehl. "Combustion of Irradiated Dry and Wet Oak." Combustion and
Flame. 42. pp. 123-138. (1981).

Lewellen. P.C.. W.A. Peters, and LB. Howard. "Cellulose Pyrolysis Kinetics and Char
Formation Mechanism.” 16th Symposium (Int’ 1) on Combustion. p. 1471. (1976).

Loehrke, KL. and Nagib, H.M.. "Control of Free-Stream Turbulence by Means of
Honeycombs: A Balance Between Suppression and Generation." Journal of Fluids
Engineering. (September 1976).

Madorslty, S.L., "Thermal Degradation of Organic Polymers." RE. Kreiger Publishers, New
York. (1975).

Madorslty. S.L., V.E. Hart. and SJ. Staus. Res. Nat’l Bur. Stad.. 56 (6). pp. 343-354 (1956).
Martin. S.. Tenth Symposium (Int’l) on Combustion, pp. 877-896 (1965).

Matsumoto, T., T. Fujiwara. and J. Kondo "Nonsteady 'lhermal Decomposition of Plastics,"
12th Symposium (int’l) on Combustion, p. 515 (1969).

Min. K.. and H.W. Emmons "Drying of Porous Media." Heat Transfer Fluid Mechanics
Institute. pp. 1-18 (1972).

Mitchel. U., and Frobel, E. "Lower Limit for the Velocity Fluctuation Level in Wind Tunnels."
Experiments in Fluids, 6. pp. 49-54, (1988).

Myers. G.E.. "Analytical Methods in Conduction Heat Transfer," McGraw-Hill Publishers,
New York. (1971).

 

 

255

Panton, KL. and Rittmann. J.G.. 13th Symposium (Int’l) on Combustion. The Combustion
Institute. Pittsburgh. PA. p. 881. (1971).

Parker. WJ.. "Determination of the Kinetic Parameters and the Heat of Combustion of the
Volatiles for Wood." Preprint. Center for Fire Research. NBS. (1985).

Quintiere, 1., Fire and Materials," 5. p. 51 (1981).

Roberts. A.F.. and Clough. 6.. "Thermal Decomposition of Wood in an Inert Atmosphere."
9th Symposium (lnt’l) on Combustion. the Combustion Institute. Pittsburgh. PA. p. 158. (1963).

Roberts. A.F.. "A Review of Kinetics for the Pyrolysis of Wood and Related Substances."
Combustion and Flame. 14, pp. 261-272 (1970).

Roberts. A.F. "Problems Associated with the Theoretical Analysis of Burning of Wood." 13th
Symposium (lnt’l) on Combustion, p. 893, (1971).

Roberts. A.f. "Heat of Reaction During Pyrolysis of Wood.” Combustion and Flame. 17, p.79
(1971).

Rogers. FE. and Ohlemiller. T.J., "Cellulosic Insulation Materials. 1. Overal Degradation
Kinetics and Reaction Heats," Combustion Sciences and Technology, 24, 129. (1980).

Rohsenow. WM. and Hartnett. J.P.. "Handbook of Heat Transfer," McGraw-Hill Book Co.
(1973).

Shaﬂzadeh. F., Sarkanen, K.V., and Tillman. D.A.. "Thermal Uses and Properties of
Carbohydrates and Lignins." Academic Press. New York. (1976).

Shaﬁzadeh, F., "Chemistry of Pyrolysis and Combustion of Wood." Int’l Conference on
Residential Solid Fuels. Oregan Graduate Center, Portland. Oregan (1981).

Shivadev, U.K., and Emmons. H.W.. "Thermal degradation and Spontaneous Ignition of paper
Sheets in Air by Irradiation," Combustion and F lame. 22, p. 223 (1974).

Siau. F.J.. "Flow in Wood." Syracuse University Press. Syracuse, New York (1971).

Simms, D.L.. and Law, M., "The Ignition of Wet and Dry Wood by Radiation." Combustion
and Flame. 11. p. 377 (1967).

Susott. R.A.. ”Characterization of the Thermal Properties of Forest Fuels by Combustible Gas
Analysis,” Forest Science. 28, 2, pp. 404-420 (1982).

Tang, W.K.. "Effects of Inorganic Salts on Pyrolysis of Wood. Alpha-Celluolse, and Lignin.
Determined by Dynamic Thermogravimetry." US. Forest Services Research Paper, FPL71,
Forest Products Laboratory. Madison, Wisconsin (1967).

Tinney. E.R., ""Ihe Combustion of Wooden Dowels in Heated Air," 10th Symposium (Int’ 1) on
Combustion. pp. 925-930 (1965).

Volvlle. C.. Delfau. "Mass Loss Rate Measurment on Solid Materials Under Radiative

 

256

Heating," Combustion Science and Technology, 36. pp. 1-18, (1984).

Volvelle. C.. Akrich. R. and DelfauJ.L.. "Thermal Degradation of Solid Materials Under A
Variable Radiant Heat ﬂux." 20th Symposium (Int’l) on Combustion. The Combustion Institute.
Pittsburgh. PA. p. 1674. (1984).

Volvelle. C., Delfau. J.L.. Reuillon, M., Bransicr, 1., Laraqui, N.. "Experimental and Numerical
Study of the Thermal Degradation of PMMA." Combustion Science and Technology. 53,
pp.187-201. (1987).

Welker. JR. "The Pyrolysis and Ignition of Cellulosic Materials: A Literature Review." J. of
Fire and Fammability. 1. p.12 (1970).

Wesson. H.R., Welker. JR. and Sliepcevich. C.M., "The Piloted Ignition of Wood by Thermal
Radiation." Combustion and Flame. 16 p.303 (1970.

Wichman. W.S., and Atreya, A., "A Simpliﬁed Model for the Pyrolysis of Charting Materials."
Combustion and Flame. 68. pp.231-247. (1987).

Wilson, D.G.. Solomon, AD. and Dogs. P.T.. "Moving Boundry Problems." Academic Press,
New York. (1978).

Zardy. C.A., and Pyle, D.L. Proceedings of the Indians Academy of Sciences, J. of
Engineering Sciences. 5. part 4. p. 264 (1982).