(Albrecht Kaupp PHD (Auth.) ) Gasification of Rice

Deutsches Zentrum fiir Entwicklungstechnologien - GATE
Deutsches Zentrum fUr Entwicklungstechnologien - GATE - stands for German Appro-

priate Technology Exchange. It was founded in 1978 ·as a special division of the Deutsche
Gesellschaft fiir Technische Zusammenarbeit (GTZ) GmbH. GATE is a centre for the dis-
semination and promotion of appropriate technologies for developing countries. GATE
defines ,Appropriate technologies" as those which are suitable and acceptable in the light of
economic, social and cultural criteria. They should contribute to socio-economic develop-
ment whilst ensuring optimal utilization of resources and minimal detriment to the environ-
ment. Depending on the case at hand a traditional, intermediate or highly-developed can be
the ,appropriate" one. GATE focusses its work on four key areas:
- Technology Exchange: Collecting, processing and disseminating information on technolo-
gies appropriate to the needs of the developing countries; ascertaining the technological
requirements of Third World countries; support in the form of personnel, material and
equipment to promote the development and adaptation of technologies for developing
countries.
- Research and Development~· Conducting and/or promoting research and development
work in appropriate technologies.
- Cooperation in Technological Development: Cooperation in the form of joint projects with
relevant institutions in developing countries and in the Federal Republic of Germany.
- Environmental Protection: The growing importance of ecology and environmental protec-
tion require better coordination and harmonization of projects. In order to tackle these tasks
more effectively, a coordination center was set up within GATE in 1985.
GATE has entered into cooperation agreements with a number of technology centres in
Third World count.ries.
GATE offers a free information service on appropriate technologies for all public and private
development institutions in developing countries, dealing with the development, adaptation,
introduction and application of technologies.
Deutsche Gesellschaft fiir Technische Zusammenarbeit (GTZ) GmbH
The government-owned GTZ operates in the field of Technical Cooperation. 2200 German
experts are working together with partners from about 100 co.untries of Africa, Asia and
Latin America in projects covering practically every sector of agriculture, forestry, economic
development, social services and institutional and material infrastructure. - The GTZ is
commissioned to do this work both by the Government of the Federal Republic of Germany
and by other government or semi-government authorities.
The GTZ activities encompass:
- appraisal, technical planning, control and supervision of technical cooperation projects
commissioned by the Government of the Federal Republic or by other authorities
- providing an advisory service to other agencies also working on development projects
- the recruitment, selection, briefing, assignment, administration of expert personnel and
their welfare and technical backstopping during theirperiod of assignment
- provision of materials and equipment for projects, ·planning work, selection, purchasing
and shipment to the developing countries
- management of all financial obligations to the partner-country.
Deutsches Zentrum fiir Entwicklungstechnologien - GATE

in: Deutsche Gesellschaft fiir Technische Zusammenarbeit (GTZ) GmbH
Postbox 51 80
D-6236 Eschborn I
Federal Republic of Germany
Tel.: (06196) 79-0 Telex: 41523-0 gtz d
Albrecht Kaupp
Gasification of Rice Hulls

Theory and Praxis
A Publication of
Deutsches Zentrum für Entwicklungstechnologien - GA TE
in: Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH
Springer Fachmedien Wiesbaden GmbH

The Author:
Albrecht Kaupp, PhD, staff member of GTZ/GA TE has been working in the fields
of civil engineering, mathematics, and biomass energy conversion systems since
1972. Now project officer for biomass energy conversion systems since 1983. His
field of expertise is gasification of biomass.
CIP-Kurztitelaufnahmc der Deutschen Bibliothek
Kaupp, Albrecht:
Gasification of ricc hulls : theory and praxis ;
a publ. of Dt. Zentrum für Entwicklungstechno-
logien- GATE in: Dt. Ges. für Techn. Zusammen-
arbeit (GTZ) GmbH I Albrecht Kaupp. -
Braunschweig ; Wiesbaden : Vieweg, 1984.
ISBN 978-3-528-02002-6 ISBN 978-3-322-96308-6 (eBook)
DOI 10.1007/978-3-322-96308-6
All rights reserved.

© Springer Fachmedien Wiesbaden 1984
Ursprünglich erschienen bei Friedr. Vieweg & Sohn Velagsgesellschaft mbH, Braunschweg 1984
A word of warning!
This book about gasification of rice hulls is based on

fundamental research and a comparison of various systems.
Because the gasification of rice hulls is very sensitive
to scaling up or down a gasifier, many but not all of the
physical and chemical properties of rice hulls are described
in detail.
To obtain a convincing decision on the most suitable way of
gasification of rice hulls the principle of negative selection
was applied, starting with conventional woodgas producers.
In order to demonstrate the various systems and show their
disadvantages they are discussed in detail at the end of this
book.
However, it must be said that, based on a great deal of

experimental work only two ways of gasification of rice hulls
seem to be technically feasible and promising in the requ~sted
range of 20-130 kWel.
1. An open core downdraft gas producer with no throat and a

slowly rotating grate for ash removal, The limiting factor
for this system is certainly the diameter of the reactor
which should not be below 40 em due to the caking of rice
hulls.
2. Gasification of rice hulls in pelletized form. Although the

densification of rice hulls is difficult and expensive, it
allows their gasification in smaller reactors due to a more
stable fuel-bed.
Option 1 is in the authors opinion the simplest and most

economical way to gasify rice hulls although a great deal of
work remains in order to predict the kinetics of rice hull
gasification andoptimize all the fine details which make the
difference between a good and bad gasifier.
Eschborn, February 16, 1984 Albrecht Kaupp

ACKNOWLEDGEMENT
This study has received much technical, financial and moral support
from many people and institutions. I would like to thank Professor John
R. Goss for his patience. Eldon Beagle, Consultant, for the many hours
he spent in talking with me about his vast experience in the utilization
of rice hulls. Professor Stephan Whitaker, Department of Chemical
Engineering, for his concise, uncompromising teaching and treatment of
the theory of reactor design which was a keystone to most of the theore-
tical parts of this research.
My thanks are extended to the Carl Duisburg Gesellschaft in West
Germany which financed part of this research. The Weyerhaeuser Company
and their two-year scholarship which allowed me to work independently.
Special thanks go to the Briggs and Stratton Corporation, Doug Janish
and Mr. Robert Catterson, whose generous financial support over two
years made this project possible.
The help of Bart Duff, International Rice Research Institute,
Philippines, to secure adequate funding was greatly appreciated.
I thank George Giannini and the workshop people for the many gas
producers and other devices they built for me. Also not forgotten is
Kurt Creamer, Graduate Student, who assisted me in many of the experi-
ments and patiently corrected my English with weekly awards for the
"worst sentence". The illustrations for this work have been done by
Jim Bumgarner. The many hundreds of pages were typed over and over
again by Karin Clawson.
My greatest appreciation goes to Filiz who hates rice hulls, but
certainly does understand the author.
IV
CONTENTS
LIST OF FIGURES IX
LIST OF TABLES XVII
CHAPTER 1. INTRODUCTION 1
References 18
2. OBJECTIVES AND SCOPE 19
3. HISTORY OF GAS PRODUCER ENGINE SYSTEMS 22

Introduction 22
History 22
References 42
4. LITERATURE REVIEW 46
References 49
5. CHEMISTRY OF GASIFICATION OF RICE HULLS 50

Introduction 50
Formation Reactions 51
Reaction Zones 53
Model I 58
Example 62
Computer Program 64
Analysis of the Results 73
Influence of the Moisture Content 73
Higher Heating Values of the Computed Gas Compositions 76
Summary for Model I 77
Modell II 79
Species Concentration Equation 81
The Energy Equation 86
Flame Temperatures of Producer Gas 87
Computer Program 88
Comparison of Theoretical Results and Experimental Data 94
v
List of Symbols 96
References 98
6. PHYSICAL PROPERTIES OF RICE HULLS 99

Densities 99
Introduction 99
Phase Fractions 110
Surface Area of Loose Rice Hulls and Rice Hull Pellets 111
Determination of A~ 0 for rice hulls 112
Apparent Surface ABo of Pellets 115
Weight of a Single Rice Hull 115
Accuracy of the Results 116
Caking and Slagging Behavior of Rice Hulls 118
Slagging of Rice Hull Ash 121
Cause of Slagging 126
Caking of Rice Hulls and Pellets 130
Summary 130
Pressure Drop in a Rice Hull Fuel Bed and Superficial
Velocities 132
Experimental Set Up and Results 133
Discussion of the Results 134
Theoretical Treatment of the Pressure Drop in a Rice
Hull Fuel Bed 138
Size Distribution. of Rice Hulls and Rice Hull Char 141
List of Symbols 144
References 146
7. PHYS!CAL APPEARANCE OF RICE HULLS UNDER THERMAL

DECOMPOSITION 148
Introduction 148
Micrographs of Rice Hulls Before Thermal Decomposition 148
Micrographs of Rice Hulls After Thermal Decomposition 151
Size Reduction of Pelletized Rice Hulls 156
Summary 159
8. LOW TEMPERATURE ENERGY CONVERSION OF RICE HULLS 160

Introduction 160
VI
Products of Pyrolysis 161
Mechanism of Pyrolysis 165
Pyrolysis Experiments in a Pure Nitrogen Atmosphere 169
Heat-up Period for a Single Rice Hull 170
Experimental Set Up and Procedures 172
Discussion of the Experimental Results 174
Composition of the Gas Phase 174
Ultimate Elemental Analysis of Rice Hulls and Char as
Function of Temperature 175
Weight Fractions of Char, Gas, Tar and Water 177
Energy Balance 178
Summary 182
References 183
9. TAR CRACKING IN A RICE HULL AND RICE HULL PELLET FUEL

BED 184
Introduction 184
Tar Conversion in a Downdraft Gas Producer 184
Mechanism of Tar Conversion 186
Design Criteria for Tar Cracking in Past Downdraft
Gas Producers 187
Experimental Set Up 192
Cases Tested 192
Test Material 194
Charred Fuel Bed 195
Filter Train 195
Tar Injection 196
Results and Discussion 197
Water Dissociation in a Hot Rice Hull Char Bed 199
Example 202
Experimental Set up and Procedure 203
Results 205
Summary 210
List of Symbols 211
References 212
VII
10. DESIGN CONSIDERATIONS FOR A RICE HULL GAS PRODUCER 213
Introduction 213
Italian Balestra Type Updraft Rice Hull Gas
Producer (1910 - 1944) 216
Chinese Rice Hull Gas Producer 221
Design Considerations for Ash Removal Systems 224
Ash Removal Designs 226
Summary 237
Design Considerations for the Gas Exit 237
Air Injection Designs 243
Design of a Small (2 - 20 hp) Rice Hull Gas Producer 246
Open Core Gas Producer 256
·Mode of Operation 258
Gas Cleaning Train 278
Sieve Plate Scrubber and Dry Packed Bed Filter 280
Experimental Procedures and Results 284
Summary 295
List of Symbols 296
References 298
VIII
LIST OF FIGURES
FIGURE PAGE
1-1 Energy fractions in gaseous components as a function

of the equivalence ratio ~ 4
1-2 Ignition advancement versus hydrogen content of

producer gas 9
1-3 Soot formation as a function of H/C and 0/C ratio 11
1-4 Soot formation as a function of H/C ratio 12
1-5 Power output as a function of ~ 12
1-6 Ultimate elemental analysis on an ash and moisture

free basis of various biomass fuels 16
1-7 Block diagram of parameters involved in the

gasification process 17
5-l Co-current or downdraft gasification 52
5-2 Accumulative mass loss curve 54
5-3 Differential mass loss curve 54
5-4 Differential thermal analysis 55
5-5 Counter-current or updraft gasification 57
5~6 Equilibrium of the water shift reaction as a function

of temperature in a fluidized bed rice hull gasifier 61
5-7 Kp(T) as a function of T 62
5-8 Range of computed gas compositions as a function of

1\f and ~ 65
5-9 Gas composition, ~ 0, M 0 66
5-10 Gas composition, ~ 0.1, M 0 66
5-ll Gas composition, ~ 0.2, M 0 67
IX
PAGE
5-14 Gas composition, • 0.5, M 0 68
5-22 c2m equilibrium composition at various temperatures,

• = 0.3 72
5-23 Higher heating value of the raw gas for range of T and • 77
5-24 Adiabatic flame temperature as a function of A 91
5-25 Adiabatic flame temperatur~ Tad as a function of

mixture inlet temperature 93
6-1 Fuel bed structure of rice hulls 100
6-2 Apparent volume Va of a single rice hull,

andy, o phase 101
6-3 Ideal surface of waxed rice hulls 109
6-4 Actual surface of waxed rice hulls 109
6-5 Geometry of waxed rice hull for computation of Pa 109
6-6 Schematic of a single rice hull 114
6-7 Modelled cross section of rice hull 114
6-8 Na 2 o - Sio 2 system 123
6-9 K2 o - Sio 2 system 123
6-10 Experimental gas producer for testing the slagging

behavior of rice hulls 127
6-11 Schematic of slag formation in a rice hull fuel bed 128
X
PAGE
6-12 Molten Silica and slag formation 129
6-13 Localized slagging, connecting pellets of rice hulls 129
6-14 Initial stage of rice hull gasification 131
6-15 Final stage of rice hull gasification 131
6-16 Caking of rice hull pellet at low temperatures 131
6-17 Experimental set up for pressure drop testing 133
6-18 Pressure drop through a rice hull and rice hull char
bed as a function of the superficial gas velocity vs
for various bed lengths 135
6-19 Suggested horizontal velocity profile within a

rice hull bed 1111
6-20 Pressure drop across a rice hull bed, experimental

and theoretical results 140
6-21 Size distribution of rice hulls and char removed

from downdraft gasification 142
6-22 Fine particle content of rice hull char from

downdraft gasification 143
7-1 Vertical wall of rice hulls after one hour of

operation 149
7-2 Rice hull-tar conglomerate from updraft gasificatton 149
7-3 Single rice hull before gasification 150
7-4 Outer surface before gasification (x 20) 150
7-5 Outer surface before gasification (x 180) 152
7-6 Inner surface before gasification (x 500) 152
7-7 Inner surface before gasification (x 2000) 153
7-8 Cross section of single rice hull before

gasification (x 550) 153
7-9 Bump on outer surface before gasification 154
7-10 Rice hull after complete combustion at 1200°C (x 20) 154
XI
PAGE
7-11 Rice hull after complete combustion at 1200°C (x 550) 155
7-12 Outer surface after gasification 155
7-13 Single bump after gasification 157
7-14 Inner surface of rice hull after gasification (x 550) 157
7-15 Inner surface Silica skeleton of rice hull after

gasification (x 2000) 158
7-16 Size reduction of pelleted rice hulls under

thermal decomposition 158
8-1 Schematic of producer gas diffusion flame 160
8-2 Orange diffusion flame from raw producer gas 161
8-3 Composition of pyrolysis gas on a nitrogen and

oxygen free basis froma sub-bituminous B coal 167
8-4 DMBA 168
8-5 Rice hull pyrolysis test apparatus 173
8-6 Dry pyrolysis gas composition as a function of

temperature 17 5
8-7 Weight fraction of C, H, ~. 0 consumed as a

function of temperature 176
8-8 Weight fraction of the pyrolysis products as a

function of temperature 179
8-9 Weight fract.lon of the products fr01n Model I at cp 0 179
8-10 Higher heating value of rice hulls and char and

energy lost during the pyrolysis process 181
8-11 Weight fractions of ash, carbon and volatiles

(H, 0, N) in rice hulls and rice hull char 181
9-1 Updraft gas producer 185
9-2 Downdraft gas producer without a throat 185
9-3 Downdraft gas producer with throat 185
9-4 Tar content for beech wood and rice hull producer gas 185
XII
PAGE
9-5 Hot temperature zone of a downdraft gas producer with

wall tuyeres 189
9-6 Approximate position of vertical temperature profile

and throat for optimal tar conversion 189
9-7 Downdraft center nozzle gas producer 191
9-8 Downward creeping fire zone in a downdraft gas

producer 191
9-9 Experimental set up for tar cracking tests 193
9-10 Tar conversion efficiency, experimental results 197
9-11 Longitudinal temperature profiles in tube 200
9-12 Steam conversion in a rice hull char bed Z06
9-13 Dissociation of steam in the presence of glowing carbon Z07
9-14 Steam dissociation with C, Hz, HzO, CO and co2 as

products Z08
9-15 Steam dissociation with c, Hz, H2o and CO as products 209
9-16 Steam dissociation with c, Hz, H2 o, CO, co2 and CH4

as products 209
10-1 Flat slot grate 214
10-2 Shaker grate 214
10-3 Shaker grate 214
10-4 Crossdraft gas producer with no grate 214
10-5 Italian updraft rice hull gas producer of the

Balestra type Z17
10-6 Chinese rice hull gas producer Z22
10-7 Parameters which influence the grate design 2Z6
10-8 Cold test stand for grate performance 2Z8
10-9 Eccentric rotating grate 229
10-10 Gas exit above grate ZJC
XIII
PAGE
10-11 Gas exit below grat~ 230
10-12 Grate with rotating curved wiper 231
10-13 Rice hull char removal as a function of t~e wip~r

speed and distance d 232
10-14 Eccentric wiper grate 233
10-15 Gas producer and engine mounted on a free-swinging

frame 235
10-16 Detail of free-swinging bars and mounted engine 236
10-17 Slightly curved vibrating disk 236
10-18 Producer gas viscosity as a function of temperature 242
10-19 Wall tuyeres without throat, downdraft gas producer 244
10-20 Wall tuyer~s with throat, downdraft gas producer 244
10-21 Center tuyere, without throat, downdraft gas producer 244
10-22 Center tuyer~ with throat, downdraft gas producer 244
10-23 Continuous slot as air inl~t 245
10-24 Open core air Hffusion into the fuel bed 245
10-25 Downdraft gas producer with force- feeding system 248
10-26 Rice hull fuel bed before and after caking 250
10-27 Gas producer with vibration grat~ and gravity flow 251
10-28 Gas producer with wiper grate and gravity flow 252
10-29 Gas producer t.'ith force- feeding system and wiper grate 253
10-30 Gas producer wit~ force-feeding system and water grate 255
10-31 Fixed fire zone in an open core downdraft gas producer

with continuous ash removal system 257
10-32 Batch-fed gas producer ~ith top lighting 258
10-33 13atch-fed gas producer t.7lth bottom lighting 258
XIV
PAGE
10-34 Batch-fed gas producer without ash removal 259
10-35 Air-to-fuel ratio of the gas producer as a function

of the gas flow rat~ 263
10-36 Fire zone velocity (up) and fuel bed velocity (down)
as a function of the gas flow rate 263
10-37 CO and Hz content of the dry gas as a function of gas

production rate 265
10-38 THC content of the dry gas as a function of gas

production rate 265
10-39 co 2 content of the producer gas as a function of the

gas production rate 266
10-40 Nz content of the producer gas as a function of the

gas production rate 2()6
10-41 Higher heating values of producer gas 270
10-42 Superficial gas velocities at 25°C as a function of

gas production rate 270
10-43 Operation time for a 165 em column 272
10-44 Specific gasification rate 272
10-45 Rate of rice hull consumption 274
10-46 Degree of rice hull conversion 274
10-47 Rice hull volume reduction 279
10-48 Efficiencies of the process 279
10-49 Gas cleaning filter train 281
10-50 Steam content of the raw gas ~s a function of gas

flow rate 289
10-51 Approximate temperatnre of upward movlag fire zone 289
10-52 Dust content of gas after gas filter train 293
10-53 Appearance of fiberglass filter papers 294
XV
LIST OF TABLES
TABLE PAGE
1-1 Products of Thermal Decomposition of Biomass 6
5-l Proximate Analysis of Rice Hulls 56
5-2 Relative Intrinsic Gasification Rates at 1 atm

and 800°C 59
5-3 Species Concentration at ~ 0.3 75
5-4 Species Concentration at T 30%,

Wet Basis 75
5-5 Adiabatic Flame Temperature and Measured Maximal

Temperatures of Various Producer Gas Compositions 94
6-1 Ultimate Analysis of Rice Hulls and True Densities

of the Elements 106
6-2 Gas-Solid Phase Distribution in a Gas Producer for

Loose Hulls and Rice Hull Pellets 111
6-3 Maximum Error Data for Rice Hulls and Rice Hull Pellets 118
6-4 Melting Point of Selected Oxidized Minerals 120
6-5 Softening and Melting Temperatures of Biomass Ashes 124
6-6 Rice Hull Ash Composition 125
7-1 Average Size Reduction of Rice Hull Pellets 156
8-1 Fuels and Their C-H Composition 162
8-2 Ultimate Analysis of Pine Sawdust + Bark and Rice

Hulls and Char at 400°C 180
9-1 Chemical Composition of Tar Used in Experiments 194
9-2 Chemical Composition, of Tar 194
9-3 Pellets Ultimate Elemental Analysis, 300°C 195
9-4 Data for 20 Selected Tar Crackiqg Tests
9-5 Producer Gas Composition of a Wood Charcoal

Automotive Gas Producer 201
XVI
PAGE
9-6 Gas Composition of an Automoti~e Crossdraft Gas

Producer. Change of Gas Composition During Use 201
9-7 Energy and 111ass '!Jalance 203
9-8 Elemental Ultimate Chemical Analysis of Rice Hull

Char Used for Dissociation of Steam 204
9-9 Gaseous Products of Dissociation of Steam at 700°C

and 920°C 208
10-1 Producer Gas Composition of the Balestra Unit 220
10-2 Composition of Producer Gas from Downdraft Chinese

Rice Hull Gas Producer 223
10-3 Char Remo~al by Vibration 237
10-4 Terminal Velocity Versus Particle Size Diameter 242
10-5 Ultimate Chemical Analysis of Rice Hulls for Testing 262
10-6 Dry Gas Composition as a Function of the Gas Flow Rate 268
10-7 Residue and Its Chemical Ultimate Analysis 277
10-8 Particulate Content of Air 282
10-9 Water Content of Raw Gas 287
10-10 Moisture Content of Producer Gas after the Gas Cleaning

Train 291
XVII
I. INTRODUCTION
Man discovered fire and with it the combustion process between 0.5
and 1.4 million years ago. It took him much longer, until 1669, to
develop the art of generating a combustible gas from a carbonaceous fuel
(Wyer, 1907). I t may be appropriate to ask why .it took so long. It is
even mor·e appropriate to explore why the history of gasification was so
chaotic and why the heyday of gasification was so short. Over 12,000
large gasifiers were installed in North America within a period of 30
years, providing electricity and street lights to many large cities.
Moreover, in the 1930's and 1940's over 1 million automotive gas produ-
cers were operated worldwide. The period of commercial use of gas pro-
ducers was quite short, considering the almost total disappearance of
the technology after 1950.
The history of gasification is so astonishing and instructive, that
it will be dealt with in a separate chapter.
Insofar as writing a dissertation is a serious matter, the time and
efforts invested in this task should not be wasted by duplicating pre-
vious work or by trying to demonstrate banalities such as a discussion
of· the generation of a co1nbustible gas from a carbonaceous fuel. An
extensive survey in 63 countries has preceded th'is work to provide a
basis for establishing research priorities in this field. ("Kaupp and
Goss 1982).
While conducting this survey it became clear that the level of
disagreement among researchers, engineers and manufacturers is unusually
high. The disagreement begins with the definition of the word
"gasification". Of course, the mistake lies in the attempt to charac-
1
terize a process with an imprecise tool, i.e. language. The litera-
ture is full of catch words such as fast pyrolysis, slow pyrolysis,
pyrolysis without oxygen, pyrolysis with oxygen, starved air combustion,
combustion with excess air, combustion in an oxidizing atmosphere, com-
bustion in a reducing atmosphere and of course gasification itself.
Naturally, finding a meaning for gasification within the mentioned
framework is a hopeless undertaking. It is therefore appropriate to
start at the beginning and make the statement:
"The carbon in the fuel burns above a certain temperature when
in contact with oxygen and this process generates a phenomena
which we sense as heat. In addition, the process generates
certain products, the natures and quantities of which are open
for discussion".
It is a simple calculation to determine the amount of air necessary
to completely burn the solid carbon in the fuel with the assumption that
there are only three products: unburned res !dues other than carbon,
water (HzO) and carbon dioxide (Co 2 ). Whether it is possible to control
the process such that those are the only three components is irrelevant
for this definition. This process is known as stoichiometric combustion
or combustion with 100% theoretical air, and is thus uniquely defined.
The difficulties arise when no restrictions are imposed on the products
and when less than 100% theoretical air is assu~ed. There are ~any
definitions of the meaning of gasification, therefore there is certainly
no danger in postulating still one more which hopefully is more precise.
One can demonstrate a combustion process taking place at less than
2
stoichiometric air. Such a process will be described as occurring at an
equivalence ratio ~ which is defined as:
Mass of Reaction Combustion Air

Mass of Stoichiometric Air for Reaction
By definition ~ is a number lar:ger than 0, although in most combustion
reactors values between 0.1 and 4.0 are achieved. Products from com-
bustion processes are more than just solid residues and gaseous pro-
ducts. An appreciable amount of condensable liquids are also generated
which may or may not be combustible. In converting reactants to pro-
ducts the chemical processes are endothermic or exothermic. However,
energy can be stored in only two modes in the products:
a. Sensible heat which represents translational, vibrational and
rotational energy of the molecules.
b. Chemical energy which upon further oxidation can be converted into
sensible heat.
It is therefore rather intriguing to characterize gasification as
the process which stores the maximum of usable chemical energy in the
gaseous portion of the products. Although at first glance this seems
another weak definition of gasification, it has the advantage of clearly
laying out the objectives. A demonstration is given in Figure 1-1 which
illustrates how the total energy stored in the gaseous phase changes
with ~. One can also deduce another general concept which holds true
for any combustion process: The total energy in the process is shifted
into the gaseous phase with increased ~· In theory one can therefore
utilize the ultimate elemental analysis of a biomass fuel and calculate
3
the equivalence ratio ~ which optimizes the process with respect to the
chemical energy stored in the gaseous phase.
One objective of this work is to determine whether operation at this
optimal ~ can be sustained, and if so to determine the design that will
accomplish this. In order to show the complexity of the problem, one
need only list the major issues involved:
(/)
1-
z
w
z
0
CL
:E
0
u
(/)
:::>
.."'
0
w
(/)
z
z
0
..
1-
u
...a:
>-
"'a:w
z
w
0.2 0.4 0.6 0.8 1.0 1.2
EQUIVALENCE RATIO c/J
Figure 1-1. Energy fractions in gaseous components as a function

of the equivalence ratio ~ (Reed, 1979).
1. A gas producer by itself is of little use. In any installation a
gasifier is either used to:
a. generate a combustible gas which is subsequently burned in a
burner to provide heat, or
b. generate a gas which is used as a fossil fuel substitute in
internal combustion engines. In this case there is a wide
range of applications such as generation of electricity and
mechanical power.
4
2. Almost all combustion processes generate a very complex mixture of
gaseous, liquid and solid products. Some of these products are not
only harmful to engines and burners but also to man. There is the
tendency to minimize this aspect of gasification on a small scale
because it requires some effort to detect and quantify these unde-
sirable products. In order to illustrate this point a partial list
of components of a combustion process is given in Table 1-1.
3. Depending on the usage of the producer gas very different restric-
tions are imposed on its composition. It is not clear whether the
process of thermal decomposition at <P < 1 can in all instances
satisfy these requirements. On the other hand, it is well known as
to what extent the various applications can utilize solid and con-
densable products.
a. A center-fired burner can cope with solid and condensable con-
taminants in the gaseous phase much better than an internal
combustion engine. The degree to which solid carbon particles
and liquid higher hydrocarbons are burned off is however a
strong function of the H/C and C/O ratios. Where the heat is
used for drying purposes, it is important to point out that
direct contact of the hot combustion products of producer gas
with the material to be dried is desirable from an efficiency
standpoint. However it is not known whether combustion of pro-
ducer gas completely destroys undesirable contaminates which
could otherwise affect the product to be dried.
b. Extensive experience with producer gas in internal combustion
5
Table 1-1. Products of Thermal Decomposition of Biomass (Bugge, 1927)
Carbon monoxide Formic acid Methacrylic acid

Carbon dioxide Acetic acid y-Butyrolactone
Hydrogen Propionic acid n-Butyric acid
Crotonic acid iso-Butyric acid
Water !so-Crotonic acid Angelic acid
Tiglic acid ~ 3 -Hexenone-2 2-Methylfurane

~2-Pentenoic acid Methyl n-butyl ketone 3-Methylfurane
y- Valero lac tone 3,6-0ctanedione Dimethylfurane
n-Valeric acid 2-Acetylfurane 2,5-Dimethyltetrahydro-
Methylethylacetic acid Cyclopentanone furane
n-Caproic acid 2-Methyl- ~2 cyclopen- Trimethylfurane
tenone 5-Ethyl-2-methyl-4,5-di-
Isocaprolc acid Methylcyclopentenolone hydrofurane
n-Heptoic acid Cyclohexanone Coumarone
Lignoceric acid Methylcyclohexenone Pyroxanthone
Furoic acid Dimethylcyclohexenone Benzene
Methyl alcohol Toluene

Ethyl alcohol Phenol Isopropyl benzene
Allyl alcohol o-, m, and p-Cresol m-Xylene
Propyl alcohol o-Ethylphenol Cymene
Methylvinylcarbinol 2,4-Dimethylphenol Naphthalene
Isobutyl alcohol 3,5-Dimethylphenol 1,2,4,5-Tetramethyl-
Isoamyl alcohol Catechol benzene
Guaiacol Chrysene
Formaldehyde 2-Methoxy-4-methylphe-
Acetaldehyde nol Ammonia
Propiona1dehyde 2-Methoxy-4-vinylphenol Methylamine
Valeraldehyde 2-Methoxy-4-ethylphenol Dimethylamine
Isovaleraldehlyde 2-Methoxy-4-propylphe- Trimethylamine
Trimethylacetaldehyde nol Pyridine
Furfural 1,2-Dimethoxy-4-methyl- 3-Methylpyridine
5-Methylfurfural benzene Dimethyl pyridine
Hydroxymethylfurfural 2,6-Dimethoxyphenol
2,6-Dimethoxy-4-methyl- Methane
Methylal phenol Heptadecane
Dimethylacetal 2,6-Dimethoxy-4-propyl- Octadecane
phenol
Acetone Propylpyrogallol mono- Eicosane
Methyl ethyl ketone methyl ether Heneicosane
Diacetyl Coerolignol (or - on) Do co sane
Methyl propyl ketone Euppittonic acid (or Tricosane
Methyl isopropyl ketone eupitton) Furane
Diethyl ketone
Ethyl propyl ketone
6
engines has clearly shown that piston engines cannot operate on
producer gas which is contaminated over a specified limit with
solid carbon and condensable products. Restrictions on the
quality of the gas are severe. After cleaning, a practical
limit exists of 5 - 15 mg of solids and tar per kg of gas if
the gas is used as an engine fue 1. A further constraint in
using producer gas as an internal combustion engine fuel is its
combustion characteristics, such as ignition characteristics,
flame speed and soot formation. Thus, the amount of con-
taminants in the producer gas is not the only parameter that
affects its suitability as an engine fuel. The actual com-
position of the gas will also have a pronounced effect on
engine performance.
4. It is of necessity that a gas producer have an effective gas
cleaning train when using the gas for internal combustion engines.
The difficulties in cleaning the gas are rarely documented by scien-
tific data. Some of the issues are:
a. Purification of producer gas is not an impossible task and in
principle not significantly different from the gas purification
that has been performed for a century on a commerical scale.
The technology and knowledge is available.
b. Purification of producer gas results in most cases in bulky
purification systems that often cause high pressure drops
across the system. It is therefore important to find an accep-
table compro~ise between purification efficiency, pressure drop
and installation size.
7
c. In many applications the internal combustion engine is used as
a suction pump. It draws the necessary combustion air into the
gas producer and then draws the products of gasification
through the purification train. A high pressure drop across
the gas producer purification train is therefore highly unde-
sirable for a suction gas producer, because it reduces the
volumetric efficiency of the engine.
5. There are many arguments against using a gas producer to provide
fuel for an internal combustion engine that operates under a
variable load such as an automotive engine. The two most important
are:
a. A gas producer and gas cleaning train replaces only the gaso-
line tank of an ~ngine. For automotive applications, there are
substantial space and weight limitations with regard to the
unit.
b. A gas producer can he optimized for only one gas output rate
and will usually not operate as well at a different rate. In
particular, the formation of large amounts of condensable pro-
ducts are often a problem at non-optimal flow rates when fuels
high in volatiles are used.
6. For any engine application it is desirable to have a gas composition
that does not fluctuate greatly over time for the following·reasons:
a. Some components of the gas, such as hydrogen (Hz), can greatly
influence the flame speed of the producer gas-air mixture,
which in turn would make variable ignition timing as a function
of the Hz content necessary. See Figure 1-Z.
8
b. The ever present moisture in the gas also affects the flame
speed. Moisture also influences the reversible carbon (soot)
reactions.
c. A changing heating value of the gas, which is entirely a func-
tion of its composition, will change the power output of the
engine.
60~------r-------T-------.-------~
%OF HYDROGEN IN PRODUCER
GAS
I /o
0
..,"' 4 0 5 % --=F"'-----+--::::,.....""""=::...t7""'':::::------l
a:
~ 30
0
10
OL-----~~----~~----~-------J
500 1000 1500 2000 2500
ENGINE SPEED rpm
Figure 1-2. Ignition advancement (degrees before top dead center)

versus hydrogen content of producer gas (Anonymous).
7. A gas producer is too often thought of as a simple device that can
generate a combustible gas out of any biomass fuel. This misconcep-
tion has had and will most likely continue to have a considerable
impact on any future developments of the technology.
Having outlined the general issues it is worthwhile to refine the
above points and to discuss the design parameters which are associated
with these points.
l.l Using a gas producer to generate a combustible gas may not have the
advantages one would hope for with regard to fossil fuel savings, econo-
mics and ease of operation.
9
a. Generation of producer gas and subsequent combustion in a
burner is no different from direct combustion of the biomass in
a furnace with respect to the heat released. In both cases the
product of combustion is a gas having a high sensible heat con-
tent. The alleged economic difference lies in the costly puri-
fication of the stack gases from a furnace operation. Whether
gasification has an advantage over combustion in a furnace if
the only objective is to generate heat for drying or space
heating depends on many conditions.
b. Using producer gas as a substitute for fossil fuel in internal
combustion engines is more attractive than direct combustion of
producer gas in that the end product is a higher, more ver-
satile form of energy: mechanical power and electricity. On
the other Hand it also imposes more stringent requirements on
the quality of the gas such as:
Low solid particle content
Low condensable content
High atomic H/C ratio
Steady gas composition
High Hz content (see Figure 1-2)
Producer gas induction temperature near that of the ambientair.
Of special interest is the high H/C ratio. Most producer gas mix-
tures are in an unfavorable range of 0.1-1.4 for their atomic H/C ratio.
The lower number represents a typical updraft producer gas very high in
CO with almost no Hz while the higher ratio of 1.4 is common for
downdraft producer gas.
10
Soot formation can be expected for stoichiometric mixtures of pro-
ducer gas and air as shown in Figures 1-3 and 1-4. It is of course
always possible to supress soot formation by combustion at $ > 1, as is
done in burners. However, combustion at $ >1 in an internal combustion
engine leads to a drastic decrease in the power output of the engine as
sho~~~n in Figure 1-5. One can therefore expect soot formation within the
combustion chamber at an optimal power output. This sensitivity of the
power output with respect to $ is not as severe for an engine operated
on gasoline.
u
x 1
.2
~
E
0
(;j
c No carbon
0
.t:l deposition
ia
u
I
0
ic
Q)
01
0
~,.,
J:
12
Oxygen-to-carbon atom ratio, 0/C
Figure 1-3. Soot formation as a function of H/C and 0/C ratio

(Probstein, 1982)
(-range for producer gas)
2.1 It is often useful to determine the amount of volatile matter given
up when heating a carbonaceous fuel at low temperatures in an oxygen-
free atmosphere. This pyrolytic process is what takes place in the
11
"'ca:
50 CD
UJ
"'
0
:IE
... 7
a:
::;)
"'...a:"'
(/)
"'u
~ 40
6
.. ...>
CD
CL
(/)
"'z ....
..
a.
u
(/)
30
...u
...~ 5~--~f-~1----+--~~--_,
UJ
z
...c:a
z
..
0
CD
a:
u
lL
0
20 ...o41--l+t---
....
-l--+-----+- ---+--_,
c
f-
z ~
0
Z31-++-- --+----t--- -+--_,
UJ
u
a:
UJ
a.
/0
2~--~--~----L---~---4
03 0.2 0.1 40 60 80 100 120 140
H/C (BY WEIGHT) % THEORETICAL AIR
Figure 1-4. Soot formation as a Figure 1-5. Power output as

function of H/C ratio a function of cp
(R. Reed, 1981) (Kennedy, 1940)
upper zones of a gas producer where no oxygen is present and tern-
peratures are around 100°C - 500°C. Of interest are:
a) The fraction of solid, liquid, and gaseous products formed

during thermal decomposit ion at low temperatur es.
b) The amount of carbon evolved and stored in the various phases.
c) The compositio n of the gaseous portion.
3.1 In most gasifiers the proportion of condensabl e products generated
is substantia l. However, this does not mean that these products must
leave the gas producer in a condensabl e form, which will necessitat e
removal by a gas cleaning train for engine use. A properly designed gas
producer includes provisions for the conversion of a major portion of
the condensabl e products into noncondens able gases. However it is not
12
easy to accomplish good tar conversion. The benefits of converting
tarry products inside the gasifier are obvious:
a) Tar is a high energy component of the products of gasification
having a higher heating value of 20-34 MJkg-l depending on its
water content. It is therefore more economical to break it
down into a mixture of combustible gases inside the gas pro-
ducer, instead of removing it mechanically outside the
gasifier.
b) The removal of tar vapors from the gas stream is costly, dif-
ficult and leads to bulky gas cleaning equipment.
c) Tar vapors and liquids have been known to be a health hazard
for decades (Harvey, 1982).
Although the mechanism of tar conversion is difficult to explain, one
may in general postulate three major conversion paths:
d) Thermal cracking at high temperatures which results in cracking
of large hydrocarbon molecules.
e) Thermochemical conversion, which involves not only high tem-
peratures but also a catalyst to promote cracking.
f) Combustion of the tar with oxygen.
Because of cost considerations, the catalytic effect needed for ther-
mochemical conversion should come from the biomass fuel itself, rather
than adding a catalyst just to support tar cracking (see Chapter 9).
4.1 One of the decisions that must be made with regard to gas cleaning
equipment is whether to use wet or dry cleaning devices. This decision
is mainly based on the relative amount of solid and liquid contaminants
in the gas and not so much on their actual make-up. Some of the parame-
ters involved are:

13
a) Solid particle load (mg kg-l of raw gas)
b) Determination of the proportion of solid particles in the raw
gas which can be attributed to physical entrainment as opposed
to those which were generated by reversible carbon reactions
(soot formation).
c) Amount of condensable products in the raw gas, their distilla-
tion point curves and the water content.
d) Dew point of the mixture, which influences greatly the location
of any dry gas cleaning apparatus, since condensation of the
vapors would render it useless.
5.1 Gas production in a gasifier is proportional to the rate of carbon
gasified. It is therefore rather difficult to provide an internal
combustion engine with the necessary gas quality if the engine is fre-
quently changing speeds. Of special interest in this context are:
a) The reactivity of the fuel. It will determine the response
time of the gas producer to a changing gas demand.
b) The ability of the gas producer to maintain the same vertical
temperature profile under varying conditions. This charac-
teristic will determine the amount of unconverted tar leaving a
downdraft gasifier.
c) The ability of a gasifier to operate under a wide range of gas
production rates at similar equivalence ratios.
6.1 Maintaining a steady gas composition and a suitable gas quality is
dependent on two key parameters:
a) Existence of an undisturbed, homogeneous fuel bed, that is free
of channels and caves.
14
b) Steady fuel flow and residue removal.
7.1 It is unrealistic to expect a gasifier to operate satisfactorily on
various types of fuels. One hundred years of gasification research and
commercial applications have clearly shown that the key to successful
gasification is a gasifier specially designed for a particular fuel. It
is of paramount importance that the physical and chemical charac-
teristics of the fuel do not change significantly. The parameters that
influence fixed bed gasification of a biomass fuel are listed with
respect to their importance:
a) Physical shape of fuel particles and fuel bed structure
b) Moisture content
c) Volatile matter
d) Ash content
e) Ash composition
f) Energy content.
A large variety of biomass fuels have very similar chemical compositions
on an ash and moisture-free basis, as shown in Figure 1-6. Therefore,
the differences are attributed to local variations on how the fuel is
processed, where it is grown, and to what climate it is exposed.
Summary
The foregoing discussion on gasification of biomass has emphasized
the complexity of the process which does not only generate a combustible
gas but other products as well. How many factors are involved in the
process is shown in Figure 1-7. In the past the following issues seem
to have dominated gas producer-engiae applications:
a) Contaminations such as tar and dust in producer gas.
15
b) Development of compact, light and efficient gas cleaning
trains.
c) Design of gas producers with a good turn down ratio.
d) Design of gas producers which could gasify inferior fuels such
as bituminous coal, high ash fuels and fuels with low bulk
densities.
e) Fuel preparations such as moisture content, fuel size, shape
and fine particle content.
f) Safety and health hazards of producer gas operations.
100
....... 80
...a:
% 60
"'
Cl
...
%
40
"'
....
31: 20
:.!!.
0
0
31;.
...0
"' "'"" "'
<I
...a:
......
..J ..J 31:
... "'
0
?:
...
..J 31:
"'
..."' ..."' .......... ..."' "' ..."' "' ""
..J ..J <I
a:
... ..."'
Cl a:
...a:31: ...a:
<I 31:
% % ..J "' "'
..J "'31: a: <I
...
O..J
...ii:> z ... ...z "'z

Cl CD
...a: "'
..J C(<l
ii:
...z o"-
Q.
::::o=>
0 z
0 ::::> %
0
::::> c.> ...JI-... ...>- ...... "'z
0
... ... ... a:

% ..J
0 z
0 c.>
;;; .... C(Ul Cl ..J
a: ... ...
0 Cl ::::> ::E a:
..J a: <I
0 uJ ..J ..J 0 Cl 0 ~ 0 ..J % <I <I
iJ: a: c.> 31: a: 31:
c.> c.>
"'
0.. Q. 0 Cl Cl CD CD
Figure 1-6. Ultimate elemental analysis on an ash and moisture

free basis of various biomass fuels (Kaupp, 1982)
16
Collection
Handling and Storage Costs
Fuel Upgrading Ash Content
Chemical Composition Heating Value
Quality Control Moisture

Content
Type
Size
Gas Producer Costs
PRODUCER Environmental Heating Value

Impact
GAS
SYSTEM Reliability Tar Vapor

Content
Producer Gas Gas Mixture Moisture

Content
Particulate
Content
Gas Chemical
Gas Cleaning Composition
Burner Costs
Internal Efficiency
Combustion Engine
Producer Gas
Utilization
External Reliability
Combustion Engine
Environment
Feedstock Safety
Figure 1-7. Block diagram of parameters involved in the gasification

process.
17
I. REFERENCES
1. Bugge, G. 1927. Industrie der Holzdistillationsprodukte.

Theodor Steinkopff Company, Leipzig, East Germany.
2. Harvey, R.G. 1982. Polycyclic Hydrocarbons and Cancer. American

Scientist. July-August.
3. Kaupp, A. and J.R. Goss. 1982. State-of-the-Art Report for Small

Scale (to SO kW) Gas Producer Engine Systems. U.S. Agency for
International Development, S&T/DIV, Room 509, SA-14, Washington,
D.C. 20523.
4. Kennedy, W.B. 1940. Mixtures of Producer Gas and Petrol.

Institution of Engineers 12(9):259-263. Australia.
s. Probstein, R. and R.E. Hicks. 1982. Synthetic Fuels.

McGraw-Hill Book Company.
6. Reed, R.D. 1981. Furnace Operations. Third Edition. Gulf

Publishing Company, Houston.
7. Reed, T.B. 1981. Survey of Biomass Gasification, Volume 2,

Principles of Gasification. Publication IISERI/TR-33-239, Solar
Energy Research Institute, Golden, Colorado.
8. Wyer, S. 1907. A Treatise on Producer Gas and Gas Producers.

Hill Publishing Company, New York.
18
2. OBJECTIVES AND SCOPE
2.1 Objectives:
2.1.1 Determine the physical and chemical properties of rice
hulls affecting the gasification process.
2.1.2 Develop a small rice hull gas producer, taking into con-
sider at lon the unique physical and chemical properties of
rice hulls.
2.1.3 Obtain performance data for the gas producer.
2.2 Scope:
The following physical and chemical properties of the rice hulls used in
this research were determined:
2.2.1 True density of rice hulls.
2.2.2 Apparent density of rice hulls.
2.2.3 Bulk density of rice hulls.
2.2.4 True density of rice hull pellets.
2.2.5 Apparent density of rice hull pellets.
2.2.6 Bulk density of rice hull pellets.
2.2.7 Void space fraction of rice hulls.
2.2.8 Void space fraction of rice hull pellets.
2.2.9 Pore volume of rice hulls.
2.2.10 Pore volume of rice hull pellets.
2.2.11 Apparent surface area of rice hulls.
2.2.12 Apparent surface area of rice hull pellets.
2. 2.13 Ash content of rice hulls and range for data from the
published literature.
2.2.14 Ultimate analysis of rice hulls and range for data from the
published llterdture.
19
2.2.15 Ultimate elemental analysis of rice hull ash and range for
data from the published literature.
2.2.16 Softening and melting points of rice hull ash and reported
phase diagrams.
2.2.17 Average weight of a single rice hull.
2.2.18 Microscale appearance of rice hulls before and after the
gasification process.
2.2.19 Slagging and caking behavior of rice hulls and pellets.
2.2.20 Flame temperatures of the combustion and gasification pro-
cess for rice hulls.
2.2.21 Tar cracking in a rice hull fuel bed.
2.2.22 Tar cracking in a pelleted rice hull fuel bed.
2.2.23 The pyrolysis products of rice hulls in a temperature range
of 250-450 C:
Solid residue and composition
Tar generated
Water generated
Gas generated and composition
Carbon bound in the condensable phase.
2.2.24 Pressure drop across a rice hull and char bed as a function
of the superficial gas velocity.
2.2.25 Size distribution of rice hulls.
2.2.26 Size distribution of the residues of rice hull gasifica-
tion.
The following parameters characterizing the design and performance of a
rice hull gas producer were determined:
20
2.2.27 Ash removal system: design and parameters involved in the
design.
2.2.28 Design of the gas exit and parameters involved such as
superficial gas velocity and solid particle entrainment.
2.2.29 Design of the combustion area of a rice hull gas producer
to optimize temperature profiles and flow fields.
2.2.30 Fuel feeding systems for a rice hull gas producer.
2.2.31 Gas cleaning train for producer gas.
2 .2.32 Gas composition as a function of gas output and furnace
temperature.
2.2.33 Solid particle content of the raw gas stream.
2.2.34 Producer gas output versus size of the rice hull gas pro-
ducer.
2.2.35 Adiabatic flame temperature of producer gas assuming
various losses.
21
3. HISTORY OF GAS PRODUCER ENGINE SYSTEMS
Introduction
The present and future high energy prices which have affected
Developing Countries most have made the possibility of gasification of
waste products such as rice hulls and other crop residues a popular
choice in Third World Countries and elsewhere. The alleged advantages
of gasification seem to be intriguing for several reasons:
a) Decentralized energy conversion system which operates economi-
cally (at present high fuel prices) even for a very small scale
of 5-10 hp output.
b) Gasification has been practiced for over 100 years and was a
very well-established technology.
c) Producer gas was and still is generated in very simple stove-
like combustors on a small scale.
However, the history of gasification provides a realistic picture of
this technology. It is outlined in detail in the following section to
identify the considerable social and economical problems that prevailed
throughout its development and use. This chapter is taken from a publi-
cation in 1982 (Kaupp and Goss) which was based on the analysis of 63
original publications that dealt in length with the social, political
and economical aspects of gasification prior to 1950.
History
The history of gasification can be dated back far earlier than
usually stated. In 1669 Thomas Shirley conducted crude experiments with
carbureted hydrogen and 30 years later Dean Clayton obtained coal gas
22
from pyrolytic experiments. The first patents with regard to gasifica-
tion were issued to Robert Gardner and John Barber in the years 1788 and
1791. Robert Gardner suggested the application of waste heat of fur-
naces to raise steam, by combusting the heated products in a boiler.
However, the first confirmed use of producer gas from coal was reported
in 1792. In this year "1urdock generated gas from coal and used it to
light a room in his house. ~or many years, after Murdock's development,
coal gas was one of the principal fuels used for lighting purposes in
England. It use declined in favor of electricity but the use of pro-
ducer gas still continued and became increasingly important for cooking
and heating. Experiments to gasify wood or at least use the gases
obtained from charring of wood started surpisingly early (in the year
1798) when Lebon tried to gasify wood. In 1801 Lampadius proved the
possibility of using the waste gases escaping from charring of wood.
The process of generating water gas by reaction of water with a hot car-
bon bed was mentioned by Fourcroy in 1804. It took five more years
before it was realized by Aubertot that the stack gases of blast fur-
naces can be combusted and used to roast ore and burn lime. He received
a patent for this process in the year 1812. The first gas producer
built used oil as a fuel and the patent was given to J. Taylor in 1815
who designed and operated the unit. Between the years 1815 and 1839
many patents were issued for utilization of waste heat and stack gas
from blast furnaces. However, the first commercially used gas producer
can be attributed to Bischof who built a large unit at the iron works of
A.ud incourt, France in 1840. During the next 20 years many researchers
and engineers improved the technology. They used low grade fuel and
23
combusted the gases in gas fired furnaces. The real breakthrough came
in 1861 with the Siemens gas producer which is considered to be the
first successful commercial unit. Before the turn of the nineteenth
century there are three more important events to mention. First, the
introduction of the Dowson gas producer in 1878 which was the starting
point of the modern gas producer-engine system. This was the first pro-
ducer that was successfully used for stationary power engines. Second,
the introduction of the Mond by-product process on a large scale in
1889. And third, the introduction of the Bernier suction gas producer
in 1895, which was the beginning of the use of gas producers in small,
compact units. The Mond by-product ·process proved for the first time
that other valuable products such as ammonia could be obtained via gasi-
fication. The residual gas from this process was low in heating value
but still could be used for industrial heating purposes. This process
was also adapted to gasify highly volatile fuels such as peat and brown
coal and several plants were in operation in Japan, the United States
and Europe.
As far back as 1819 a portable gas producing apparatus comprising of
a gas producer and a gas vacuum engine were patented in England. No
record can be found that it was ever fitted on a vehicle. The task to
actually operate a passenger vehicle with producer gas for the first
time ever must therefore be credited to J. w. Parker who covered over
1000 miles with his 2 hp automotive gas producers in Scotland during the
years between 1901 and 1905. It is interesting to note that the inade-
quate protection Bernier got for his patented gas producer-engine
system, permitted other enterprising engineers the opportunity of
24
directly benefiting from his system. Many competing designs were put on
the market in increasing numbers for the next 15 years. One such make
is the Brush Koela plant that was first introduced as a patented device
in 1901 and was actually designed for export to India and other
Developing Countries. The name Koela is the Indian word for charcoal.
The oil engines used during this time period were actually replaced by
producer gas engines. Some companies in England did a brisk business
selling producer gas-engine sets to generate electricity throughout the
country for domestic lighting. The necessity to stay ahead of com-
petitors led some companies to utilization of the waste heat and the co2
generated in the process. However, these early attempts of co-
generation were not very successful, although the general ideas behind
it are no different from today's principles of co-generation. The first
decade of the 20th century was also full of attempts to sp.read the new
concept of suction gas producer-engine systems to other applications.
The Duke of Montrose convinced the British Admirality to introduce
some of the new compact suction plants on ships, because similar experi-
mental units were already in use on barges for channel and river
transport in Germany and France. A small gas producer carried by four
men and used for disinfection purposes was manufactured by J. Pintsch.
The gas, rich in carbon monoxide, was used for killing mice, rats, or
other vermin on farms and ships. The technology of gasification of wood
and charcoal was stepped up, mostly to provide the colonies of the
British and German Empires with gas producers that did not depend on
scarce anthracite coal. H. A. Humphrey had considerable success with.
operating huge pumps on producer gas. Several types of these 1000 hp
25
waterpumps were built in Alexandria (Egypt), Berlin {West Germany) and
Chingford (England). Some enthusiasts considered producer gas the
future fuel for internal combustion engines. On the other hand a talk
given by Ade Clark for the Institution of Mechanical Engineers, London,
in which he discussed industrial applications of the diesel engine
signaled, in 1904, the increasing interest in this new diesel tech-
nology. The manufacture and operation of producer gas plants was in no
way restricted to European countries and their colonies. In fact the
United States Geological Survey had for several years investigated the
economical value of coals and lignites as gas producer fuel. The early
tests done with a pilot plant erected at the Louisiana Purchase
Exposition in 1904 were very encouraging and demonstrated the use of
many coals that could not be combusted in the existing steam-power
plants. The fact that the technology of large updraft gas producers
became more and more reliable encouraged gas engine manufacturers to
build larger and larger units. Before the wide spread use of producer
gas only small gas engines up to 75 hp were found economical to operate
with town gas. However, the cheap producer gas led to the operation of
huge gas engines. The first 600 hp engine was exhibited in Paris in
1900. Larger engines, up to 541)0 hp were put into service in the u.s.
shortly thereafter. The results of a survey of 70 plants out of the 376
existing plants in the u.s. in the year 1909 are published in United
States Geological Survey, Bulletin #416.
With regard to the present situation, the USGS report is important
because i t states for the first tirne the many difficulties caused by
lack of knowledgeable engineers, tack of knowledge and confidence in the
26
technology on the part of the public, inexperienced salesmen not fami-
liar with the details of the engine and the gas producer concept, lack
of types of gas producers that could gasify inferior fuels and the large
number of unsuccessful or only partly successful installations made
during the experimental period of this development. One of the key
problems with gas producer systems that has persistently remained to the
present is quoted from the bulletin:
"It cannot be denied that many of the difficulties charged
to producer-gas power plants are due entirely to incompetent
operators. Some plants have been put out of commission tem-
porarily by the prejudices or the lack of ability and
training of the operators or engineers in charge. A few of
these failures are due to the impossiblity of finding men com-
petent to operate the plants, but many of them have undoubted-
ly been the result of a short-sighted policy on the part of
some manufacturers, who are not willing to give proper and
necessary information about the design, construction, and
operation of the plants made by them. The possibility of a
sale at the time is apparently the only interest they keep in
mind, and the future is allowed to take care of itself".
Sales brochures from many countries and personal contacts indicate
the situation is very much the same today. The demand for better educa-
tion of the designers and builders of gas producer plants and furnaces,
drivers of automotive gas producer vehicles, the existence of special
schools teaching gasification and the demand for higher wages for
27
drivers of automotive gas producer vehicles can be found throughout the
entire literature covering the 100 years of commercial gas producers.
Further development of the automotive gas producer was done by
Porter and Smith in England during the First World War. The impetus for
this work was the possiblity of disruption of gasoline supplies which
had become the dominant fuel for motor transport. Although most of the
early development of automotive gas producers was done in England, wide
spread application during and after the First World War was crippled by
the British taxation system that assigned taxes to cars according to
their weight which included the gas producer. The 1919 special report
of the Inter-Departmental Committee on the employment of gas as a source
of power which dealt at considerable length with the automotive gas
producers and its advantages was not followed by any government action
to put the automotive gas producer in a more favorable tax situation.
A totally different situation prevailed in France. There the use of
wood and charcoal as a fuel had a long history and the French govern-
ment was actively encouraging the development of automotive gas produ-
cers after 1919. Further public awareness of this method to drive an
automobile was greatly increased through ralleys organized each year
since 1926 by the Automobile Club de France. The distances that had to
be covered were between 1600 and 3000 km. One of the greatest names in
the development and manufacture of automotive gas producers was the
Frenchman, Imbert. He filed his first patent for a downdraft gas pro-
ducer in 1923 and many successful designs including the recently built
small automotive gas producers are based on this design. The interest
in the automotive gas producer faded in France during the 1930s and most
28
of the development in this field continued in Germany. In fact the
Imbert Company is again manufacturing gas producer-engine systems in
West Germany. Although the automotive gas producer never played any
role in the development of gasification in the u.s., more than 12,000
stationary gas producers were in operation during the 1920's and 1930's
in the u.s. and Canada. Over 150 companies in Europe manufactured small
and large gas producers for various applications. The gas producer con-
cept was especially appealing for applications in remote areas or
Developing Countries which had brush or timber. For instance, the
British company, Crossley, sold gas producers for remote mines in
Australia and the Tulloch Reading 50 hp truck developed in England was
mostly purchased by the Empire Cotton Growing Cooperation for use in
Nigeria.
The next decade from 1930 to 1940 can clearly be considered as the
development period for small automotive and portable gas producers that
reached its peak during World War II. New concepts and designs such as
crossdraft gas producers were developed or improved. Efforts were
undertaken to build the automotive gas producers lighter and improve the
gas cleaning system which was the vulnerable part of the units. New
units, capable of gasifying more readily available fuels such as bitumi-
nous coal, anthracite and wood, were developed and tested in small num-
bers. The Br.itish gasification efforts were still more directed to
their overseas markets and not so much for domestic use. There were
signs of an increasing critical view toward the automotive gas producer
in France. It was claimed that at least one new gas producer mounted on
a truck was more expensive to run and operate than a comparable gasoline
29
truck despite all government grants and subsidies. It is of interest to
recall the official position of the French and ~ritish governments
during the early 1930s. Authorities in both countries felt at that time
that the automotive charcoal gas producer was more suitable for their
colonies where the supply of gasoline was scarce, and wood that copld be
made into charcoal at very low labor costs was readily available. The
emerging gas producers using wood and low grade coal were not given much of a
chance for general use. History has proven that assessment to be correct.
The first well reported conversion of internal combustion engines,
in this case tractors, to producer g·as drive under economic pressure
occurred during the 1931 to 1934 period in Western Australia. The large
quantities of wood available, the negligible oil resources at this time
and the collapse of the wheat prices during 1930's set the scenario for
a rather hasty, uncoordinated conversion of kerosene tractors to pro-
ducer gas operation. Many farmers, in order to avoid bankruptcy had to
consider all alternatives, including producer gas, although it was well
known that the power loss of the tractors would be considerable. What
happened during these years until the recovery of the wheat prices was
just a small part of what happened later during t"orld War II on a much
broader basis. Many gas producers were failures from the start. Others
deteriorated rapidly owing to faulty construction. Several firms were
interested in the manufacture and sale of such units, but had neither
the money nor time to do the necessary research and development engi-
neering. As a consequence, there were often totally dissatisfied custo-
mers, who after a short trial, resolved they would never again have
anything to do with gas producers.
30
On the other hand, a small number of farmers having ingenuity and
mechanical skill, operated their units very satisfactorily for a number
of years. In this context it should be mentioned that there has never
been an automotive engine especially designed and built for producer
gas, although the technology was wide spread for over 100 years. With
plentiful fossil fuels available during peaceful and stable economical
times, there was no need for the producer gas concept. During emergen-
cies and war times the concept of producer gas engine systems was always
so hastily brought back that there was simply not enough time and money
available to develop a specially designed producer gas, internal com-
bustion engine for automotive use. This explains in part the difficul-
ties some farmers had in converting their kerosene tractors to producer
gas drive. The interest in gas producers faded quickly after the 1930's
depression was over. Only 62 producer gas tractors out of 4548 tractors
in Western Australia were operating at the end of 1937.
In late 1930's the effort of Nazi Germany to accelerate the conver-
sion of vehicles to producer gas operation was the beginning of a world-
wide effort to use the gas producer concept as part of a plan for
national security, independence from imported oil and acceleration of
the agricultural mechanization. A typical example was the Soviet Union.
The build-up of the military as well as rapid expansion of heavy
industry necessitated a major change in the mechanized agricultural
units. The change was directed toward the fuel used. It became
apparent that despite a high priority for the agricultural sector, the
transport of the fuel was becoming a problem. The big agricultural
areas were far from the large oil fields and the distribution of the
31
fuel even when plentiful was one of the biggest problems. The introduc-
tion of gas producer powered tractors and trucks to the Russian farmers
can therefore not be viewed as an emergency measure to reduce the con-
sumption of gasoline and diesel oil. Instead it was viewed as an alter-
native to using fuels available locally and thereby easing the
transportation and distribution problem. Almost all early Russian trac-
tors were powered by gasoline engines which required extensive
rebuilding of the engine to avoid a severe power reduction. (A later
model, the Stalinez C65 tractor, and the Kharkov caterpillar tractor
were equipped with diesel engines). From the design of the gas producer
and its gas cleaning system, it seems most likely that various German
gas producers were used as the basic design for these models. Despite
some criticism about the gas producer concept, its economics and future,
new advanced crossdraft gas producers were built in France. In par-
ticular the Sabatier and Gohin Poulence plants showed an astonishing
performance, equal to most gasoline powered vehicles. However, it
became more and more obvious that good gas producer performance was clo-
sely connected to the quality of the fuel. Plants like the Sabatier, or
later the Swedish Kalle model, were highly reliable and worked well only
with specially manufactured charcoal having carefully controlled
quality. In 1938 most European countries stimulated the use of producer
gas through subsidies for conversion, favorable tax benefits or even
edicts such as in France that required all public transport companies to
change at least 10% of their vehicles to producer gas. The I tal ian
government was even more strict, requiring all buses in public service
to use home produced fuel, wood charcoal, alcohol or home produced
32
petrol and oil. These various measures led to 4500 gas producer
vehicles in France, 2200 in Germany and over 2000 in Italy by the early
part of 1939. England, the country that did most of the pioneer work in
the beginning, however, saw its producer gas program entangled in poli-
tics, resulting in very little conversion to producer gas for vehicles.
Tve learn this from an article written by the Coal Utilization Council
appearing in the Fuel Economist in July 1938. The Director of this
organization complained bitterly about the stubborness of the British
government in this matter. Nevertheless, some British bus companies ran
their city buses on producer gas quite successfully and on schedule.
What happened to the development of the automotive gas producer
after 1939 must be seen in the context of the World War II. From the
number of articles published about gasification in German journals each
year and the work of several national committees on the subject it was
obvious that Ger~any was much better prepared to deal with the logistic
problems associated with the operation of hundreds of thousands of auto-
motive gas producers. However, the most drastic development took place
in Sweden, which experienced a most severe fuel shortage. Other
countries delayed the conversion to producer gas operation, because
there was simply no need for it. For instance, not many automotive gas
producers were seen in Australia in the year 1940, compared to a con-
siderably larger number in New Zealand which was much earlier affected
by the fuel shortage. The United States coped with gasoline shortage by
means of rationing but nevertheless automotive and stationary gas produ-
cers were manufactured in Michigan. They were not available for
domestic use and most of them were sold to China under Lend-Lease terms.
33
"Woman Who Fled Nazis Makes Gas Producers in Michigan Plant for Export
to China" was one of the headlines of several articles that appeared in
the National Petroleum News and Chicago Tribune about this activity.
The development of the European gasification activities was closely
monitored by the Forest Service of the United States Department of
Agriculture and some of the findings have been published. At the end of
1944 it was concluded that wide spread commercial adoption of gas produ-
cers in the United States would not be promoted. Only under special
circumstances in remote areas, gas producer ope rat ion might be accep-
table.
Even after the outbreak of the war, the British government was in no
hurry to regulate or require the use of automotive gas producers. One
of the reasons was the unsuitability of most existing gas producers for
the soft and brown coals of England, which had little anthracite.
Nevertheless, a so-called government emergency crossdraft gas producer
was developed especially for the British coals and low temperature coke
and it was planned to manufacture 10,000 units. The government devel-
oped producer worked reasonably well but ln 1942 i t became increasing-
ly difficult to obtain the necessary low ash coal to run the gas
producer and plans to mass produce the unit were given up. The conver-
sion of vehicles to producer gas drive was therefore mostly restricted
to bus companies and some private companies that installed the station-
ary Cowan Mark 2C gas producer as an emergency power supply to fac-
tories affected by air bombing. Therefore, large-scale conversion of
vehicles took place in Sweden and the countries occupied by Germany
during World War II.
34
In December, 1939, about 250,000 vehicles were registered in Sweden.
At the beginning of 1942 the total number of road vehicles still in ser-
vice was 80,000, about 90% of which were converted to producer gas drive
within 1 1/2 years. In addition, almost all of the 20,000 tractors were
also operated on producer gas. forty percent of the fuel used was wood
and most of the remainder charcoal. Dried peat was used to some extent.
This fast and almost complete conversion was accompanied by the drastic
decline of imported petroleum from 11 million barrels in 1939 to 800,000
in 1942.
It is far more interesting to recall the logistic difficulties
associated with the conversion of gasoline vehicles on a large scale
during World War II, because the technical advances made after 1940 were
not significant and dealt mostly with the improvement of gas cleaning
systems and better alloys for the gas producer shell.
Schlapfer and Tobler, who conducted extensive tests with various gas
producers during the 1930 to 1939 period in Switzerland, pointed out the
human element involved. They argued that most of the converted post
buses running on producer gas in Switzerland did not perform well
because drivers had difficulties getting used to the new driving style
and were reluctant to do the additional work involved. Most troublesome
was the required daily cleaning of the entire cleaning system and the
preparations for the next run, which included the clearing of the fuel
hopper, because overnight storage of the fuel in the fuel hopper caused
considerable starting difficulties due to moisture buildup. It also
became apparent that neither the manufacturers nor the general public
35
really understood the problems associated with a gas-producer operated
bus.
One company in France had a mixed fleet of gasoline and gas producer
driven vehicles. The drivers had to carry out refueling and making
repairs after the day's work at regular rates of pay. The producer
vehicles were constantly having to go to the shop for repairs which the
drivers alledged were beyond what they could do. After the company
decided to pay overtime for the time spent to clean and refuel the gas-
producers, the producer vehicles became at once as trouble free as the
gasoline vehicles.
This situation also existed in Germany prior to World War II. In
the early 20th century, owners of large stationary gas producers talked
about sending their engineers to special schools on gasification and
paid them higher wages to ensure that the gas producer was properly
operated. The German government finally agreed after many years of
rejection, to pay their drivers of automotive gas producers higher
wages, which improved the situation. 'However, the uninformed private
driver remained a persistent problem. ~t the beginning he was faced
with hundreds of makes of gas producers and no manufacturer's guarantee
about the performance. Although one could not prove that some manufac-
turers actually sold equipment they knew would not work, it cannot be
denied that ,nany of them did not know much about the performance of
their units or could only prove reliable performance with high quality
fuel having carefully controlled physical and chemical properties.
Large numbers of unsatisfied customers finally led to government action
in Germany and Sweden as well as in the occupied countries. The number
36
of manufacturers of gas producers was significantly reduced to about 10
with models that had been proven to be successful. However, the fuel
supply and the quality of gas producer fuel was still a problem that
actually was never solved. Until the end of 1941, wood and charcoal
were the fuels most widely used in Germany. The collection and prepara-
tion of gas producer fuel was handled by the Gesellschaft fur
Tankholzgewinnung und Holzabfallverwertung which kept a tight control
on the size, shape and moisture content of the fuel. The fuel could be
purchased at over 1000 official filling stations all over the country.
This service was more or less operated and organized like today' s oil
companies and gasoline stations.
It soon became apparent that at the prevailing wood consumption rate
and the tendency of drivers to use charcoal, there would not be much
forest left within a few years. The construction of charcoal gas-
producers was therefore forbidden in France and Denmark afte.r July 1,
1941 and greatly restricted in Germany and Sweden. The new policy was
to encourage the use of brown-coal, peat coke, anthracite and low tem-
perature coke. It however can be concluded that their use was plagued
by problems with the quality of the fuel, such as high sulfur content,
too much volatile matter, poor physical shape of the various cokes sold,
too expensive production methods and improper handling of the fuel bag~.
Most customers did not understand the differences among the various
fuels they could buy or their influence on the gas producers. The
situation today is about the same and any introduction of small sta-
tionary or portable gas producers on a broad basis would likely lead to
the same difficulties. Some users of automotive gas producers even pro-
37
duced their own fuel out of brush wood collected in the national
forests.
A slightly different situation prevailed in Sweden with its vast
supply of wood. At the beginning the unrestricted use of charcoal led
to various designs of high performance gas producers, which operated
very well as long as they were fired with the specially prepared char-
coal they were designed for. The tar oils from wood carbonization were
also not wasted and used for heavy fuels and as a lubricant. Over 3000
furnaces producing charcoal were in operation in 1944, to provide the
necessary fuel for metallurgical operations and the fleet of gas produ-
cers. Although the officially produced fuel was strictly classified and
quality controlled, not all of the fuel related problems could be
solved. For instance, first grade low volatility fuel of less than 3%
volatiles turned out to be medium volatility fuel with over 8% volatiles
that could not be gasified in most gas producers. Hard, high. grade
charcoal leaving the factories with a low moisture content of 10% and
only a 10% fraction of fines, reached the consumer broken up and
crumbled with a moisture content of over 20% and was therefore rendered
useless. Although the emergency situation was on everybody's mind, the
temptation was high to buy and operate the very convenient, high perfor-
mance gas producers which depended on special fuels.
It is obvious that an automotive gas producer that can be started
within 2 minutes and does not require much cleaning sounded much more
appeaLing for the private customer than one with more flexibility with
regard to the fuel needed to operate the unit. The tendency to modify
the fuel for a gasifier in question instead of investing the time and
38
money to design and construct a gas producer for a fuel in question can
be found throughout the entire history of gasification. This approach
was not changed during the first 100 years of gasification and present
signs indicate that there will be slow progress toward designing gas
producers for specific fuels.
Although the number of accidents related to the use of automotive
gas producers was considerably higher than with gasoline vehicles, most
accidents were due to negligence of the driver. The increasing number
of accidents caused by operators not familiar with their equipment was
of much concern to the Swedish government and the manufacturers. This
was reflected in very detailed operation manuals and the introduction of
a special driver's license for the operation of an automotive gas pro-
ducer. Of concern were simple operational mistakes such as not ven-
tilating the unit after a day's use which resulted in a gas build up in
the gas producer that could explode while the owner was checking the
fuel level next morning. Other operators had the opinion that as long
as the engine was running on the produced gas everything was fine and
switched too early to producer gas drive during the startup period. In
most cases this led to totally tarred up manifold and valves, because
the initially produced gas, although of high heating value was rich in
higher hydrocarbons that condensed in the engine. ~ore serious and not
so easily cQntrolled is the danger of accumulative carbon monoxide
poisoning which occurred frequently according to Swedish reports. The
problems in the past with automotive gas producers, should be viewed in
the light of the enormous task that was undertaken in Europe to convert
hundred of thousands of gasoline vehicles to producer gas drive within
39
three years in a difficult time. An automotive gas producer must also
be viewed as the most advanced gas producer, much more difficult to
design and operate than a stationary unit.
Shortly after World War II, automotive gas producers as well as all
the large stationary units were put out of service because of abundant,
cheap supplies of gasoline, diesel oil, natural gas and electricity.
The change away from producer gas operation was also drastically
reflected in the research done in this field. The number of publica-
tions listed in major engineering indexes dropped sharply from several
hundred a year to less than 10 a year during the 1950 to 1970 period.
It can be said with few exceptions, gasification and in particular small
portable gas producers were a forgotton technology during this time
period. Research done in this field, which can be called a considerable
contribution to the advancement of automotive gas producers, took place
in Sweden during the 1957 to 1963 period. This research was initiated
by the Swedish Defense Department after the Suez Crisis and undertaken
by the National Machinery Testing Institute. The research made con-
siderable contributions to the improvement of the gas cleaning system
and the modifications of diesel engines for gas producer drive.
The 1970s brought an increasing renewed interest in this form of
power generation and a more general look at the complexity of gasifica-
tion. Some o.f the present work concentrates on the revival of the old
ideas and designs and their modification and expansion to fuels dif-
ferent from wood and coal. A worldwide search for small scale gas pro-
ducers in operation and researchers working on the subject as well as
the increasing number of dally inquiries about gasification received,
40
show a considerable interest and demand in small gas producers.
However, the idea still persists that gas producers are a simple stove-
like energy conversion system that is easy to design and operate. The
present demand is therefore also stimulated by the belief that gasifiers
can convert almost any carbonaceous material to useful tnechanical and
electrical energy. This image of a gasification system is far removed
from any reality and in particular the history of gasification has shown
that a fixed bed gasifier providing fuel for an internal combustion
engine is a very selective energy conversion system with little flexi-
bility with regard to the fuel it was designed for. A further handicap
is the little knowledge we have about the behavior of various biomass
fuels under thermal decomposition. This knowledge· is certainly basic
for any further optimization of gas producers and cannot be obtained
easily. On the other hand, amazing performances of gas producer-engine
systems have been reported and verified throughout the history of gasi-
fication. It is not just an assumption but confirmed reality that
trucks have been operating on producer gas for over 300,000 km with no
major repair and less engine wear than that resulting from using only
diesel fuel. Large Italian rice mills have gasified their rice husks
and used the gas to drive the power units used for milling for decades.
The number of quite satisfied owners of small and large gasifiers is
certainly not small and there is lots of evidence that i t can be done.
The history of gasification has also shown that it is not one of the
most convenient technologies.
41
3. REFERENCES
1. Allcut, E.A. 1942. Producer Gas for ~otor Transport, Engineering

Journal 5(4):223-230.
2. Anderson, M.. 1938. Case for the Encouragement of the Producer-

Gas Vehicle in Britain. Fuel Economist 14:245-246, 256-257.
3. Anonymous. 1933. ~ New Gas Producer-Gas Plant for Road

Transport. The Commerical Motor, 20:787-788. January.
4. Anonymous. 1939. Alternative Fuels for Wartime. Gas and Oil

Power. PP• 235-238. October.
5. Anonymous. 1941. Emergency Gas Supplies for Factories. Power

and Works Engineer 36:137-139. June.
6. Anonymous. 1938. Forest Gas for Traction. Engineer 166(4311):

230-231.
7. Anonymous. 1939. Gas as a Substitute for Gasoline Part 1.

Petroleum Times 42(1073:169-170, 189.
8. Foster Wheeler Energy Corporation. 1978. Gas from Coal: A

Volatile Solution. Energy Guidebook. PP• 108-110.
9. Anonymous. 1939.· Gas Producer for Road Vehicles. Engineering.

PP• 631-632. May 26.
10. Anonymous. 1939. Gas Utilization for Automobiles. Gas Age.
December 7.
11. Anonymous. 1943. German Portable Gas Producer Practice.

Engineering 155:423-424. May.
12. Anonymous. 1949. Improvements in the "Brush Koela" Gas Producer.

Engineering 169(4398):395.
13. Anonymous. 1940. Specification for Portable Gas-Producer Fuels.

Engineering. P• 150. February.
14. Anonymous. 1942. Producer Gas Plant Manufacture. Gas and Oil
Power 37(443):147-150.
15. Anonymous. 19 Producer Gas: Present and Future. Gas and Oil
Power 40(473):49-50.
16. Anonymous. 1943. Producer Gas versus Petrol Operation in

Germany. Petroleum Times 47(1193):190.
42
17. Anonymous. 1942. The P.s.v. Gas Producer. Bus and Coach. PP•
228-230. November.
18. Anonymous. 1945. The Soviet Producer-Gas Tractors. Gas and Oil
Power. pp. 89-95. Ma~ch.
19. Anonymous. 1929. The Tulloch-Reading Gas Producer for Motor

Vehicles. Engineering 127:641-644.
20. Bailey. M.L. 1979. Gas Producers for Motor Vehicles: A

Historical Review. Department of Scientific and Industrial
Research, Chemistry Division, Report CD 2279. New Zealand.
21. Branders. H.A. 1941. Producer Gas is the Motor Fuel of Finland.
Automotive Industries. pp. 482-485, 522-523. May.
22. Breag, G.R. and A.E. Chittenden. 1979. Producer Gas: Its
Potential and Application in Developing Countries. Tropical
Products Institute, Report G130, London, England.
23. Brownlie, D. 1940. Producer-Gas Driven Vehicles. The Iron and

Coal Trades Review. PP• 121-123.
24. Campbell, J.L. 1942. Gas Producers: An Outline of the

Compulsory Government Tests in Australia. Automobile Engineer
32(422):156-158.
25. Clarke, J.S. 1940. The Use of Gas as a Fuel for Motor Vehicles.
Institute of Fuel Journal 13(70):102-117.
26. Dimitryev, A.P. 1940. Automotive Gas Generators Used in USSR.

Automotive Industries 83(10):534-535, 551.
27. Dunstan, W.N. 1945. Gas Engine and Gas Producer Practice in
Australia. Engineer 180(4688):400-401.
28. Egloff, G. and M. Alexander. 1944. Combustible Gases as

Substitute Motor Fuels. Petroleum Refiner 23(6):123-128.
29. Egloff, G. 1947. Fuels Used in Sweden. Petroleum Engineer

18(5):86-88.
30. Forbes, W. 1939. Experiments with Gas Producer Vehicles in

Cardiff. Passenger Transport Journal. PP• 201-205. November.
31. Fowke, W.H. 1938. Operating Results with Producer Gas. Bus and
Coach 10(2):84-86.
32. Freeth, E.E. 1939. Producer Gas for Agricultural Purposes.

Journal of the Department of Agriculture of Western Australia
16(4):371-414.
43
33. Gall, R.L. and J.D. Spencer. 1966. Caking Coal Behavior in
Gas-Producer Tests. Coal Ag~ 71(2):128-130.
34. Goldman, B. 1938. Fuels Alternative to Oil for Road Transport

Vehicles. Fuel Economist 14(248-252. July.
35. Goldman, 13. and N.C. Jones. 1939. The Modern Portable Gas
Producer. Institute of Fuel 12(63):103-140.
36. Goldman, B. and N.C. Jones. 1938. The Modern Portable Gas
Producer. The Engineer 166:248-252. December.
37. Goldman, B. and N.C. Jones. 1939. The Modern Portable Gas
Producer. The Petroleum World 36(460):3-5.
38. Greaves-Walker, A.F. 1916. The Design and Construction of a

Producer-Gas House for Clay Plants. Transactions of American
Ceramic Society 18:862-866.
39. Hurley, T.F. and A. Fitton. 1949. Producer Gas for Road
Transport. Proceedings of the Institution of Mechanical Engineers
161:81-97.
40. Kralik, F. 1943. Rail Car with Charcoal Gas Producer. The
Engineers' Digest. PP• 24-25. December.
41. Lang, W.A. 1943. Alternative Fuels for Motor Vehicles.

Engineering Journal 26(8):449-454.
42. Langley, F.D. 1942. The Revival of Suction-Gas Producer. Gas

and Oil Power 37(446):236-240.
43. Lindmark, G. 1944. Swedish Gas Producer Buses. Bus and Coach.
PP• 266-269. April.
44. Littlewood, K. 1977. Gasification: Theory and Application,

Progress in Energy and Combustion Science 3(1):35-71.
45. Lustig, L. 1947. New Gas Producer for Dual Fuel Engines. Diesel
Progress 13(5):42-43.
46. Mellgren, s. and E. Andersson. 1943. Driving with Producer Gas.

National Research Council of Canada RP 15/43. Ottawa, Canada.
47. Miller, R.H.P. 1944. Gasogens. u.s. Department of Agriculture,

Forest Service, Forest Products Laboratory. Madison, Wisconsin.
48. Overend, R. 1977. Wood Gasification: An Old Technology with a

Future? Sixth Annual Meeting, Biomass Energy Institute Symposium.
Winnipeg, Manitoba, Canada. October 12.
44
49. Pavia, ~.E. 1942. Woodgas Producers for Motor Vehicles.
Institution of Engineers Journal 14(12):279-292.
50. Rambush, ~.E. 1923. Modern Gas Producers. Van Nostrand Company.
~ew York.
51. Renton, c. 1940. Producer Gas Tests in the Queensland Railway

Department. Institution of Engineers Journal. PP• 274-278.
Australia. October.
52. Ridley, c. 1944. Temporary Fuels, A Consideration of the

Prospect of Their Permanency. Automobile Engineer 34(446):63-67.
53. Roberts, R.P. 1933. Producer Gas Equipment on Tractors in

Western Australia. Journal of the Department' of Agriculture <?f
Western Australia 15(4):391-402.
54. Ruedy, R. 1944. Wood and Charcoal as Fuel for Vehicles.

National Research Council of Canada (1157).
55. Skov, N.A. and M.L. Papworth. 1975. The Pegasus Unit. Pegasus
Publishers Inc. Olympia, l~ashington.
56. Telford, W.M. 1949. Some Notes on the Design of Mobile Producer
Gas Units. Institute of ~ngineers Journal 12(11):299-304.
Australia.
57. Telford, W.M. 1941. Some Notes on the Design of Mobile Producer
Gas Units. Gas and Oil Power 36:179-181.
58. Tookey, W.A. 1952.. Suction Gas Plant Development Fifty Years
Ago. ~ngineer 193(5028):754.
59. Twelvetrees, ~. 1944. Paving the l-lay for Producer Gas Operation.
Bus and Coach. PP• 104-107. February.
60. Walton, J. 1940. Alternative Fuels. Automobile Engineering

30:91-92. March.
61. Woods, M.W. 1940. An Investigation of the High-Speed Producer

Gas Englne. Engineer 169(4401):448-450.
62. Woods, M.W. 1938. Producer Gas Vehicles. Instltution of

Engineers Journal 10(3). Australia.
63. ~~yer, s.s. 1906. A Treatise on Producer Gas and Gas Producers.
Hill Publishing Company.
45
4. LITERATURE REVIEW
Gasification research and commercial production came to an almost
complete standstill after World War II. The current revitalization of
the technology is due to the growing belief that fossil fuel supplies
will be exhausted within 50-100 years and that transitional energy tech-
nologies will be necessary until direct utilization of the sun's energy
can be instituted on a large scale.
The present work done in the field of biomass gasification is
insignificant when compared to the work done between 1840 and 1950. An
extensive literature review was therefore conducted which identifed
about 1200 publications of worthwhile content. More than 640 of these
publications were obtained, analyzed and organized by the following
keywords:
a) Internal Combustion Engine
b) Developing Countries
c) Gas Cleaning
d) Chemistry of Gasification
e) Portable Gas Producer Engine Systems
f) Stationary Gas Producer Engine Systems
g) Biomass Fuels
h) Fossil Fuels
i) Theory of Ga~ification
j) Economics of Gasification
The results have been published by Kaupp and Goss ( 1982). Also incor-
porated in the report is a survey conducted in 63 countries concerning
present research efforts in this field. Of over 500 persons contacted,
46
148 in 34 countries either had experience with biomass gasification or
are presently working on it.
Rice hulls are one of the most abundant crop residues in the world
(an estimated 70 million tons in 1982). This byproduct of rice milling
often causes disposal problems. Many studies have been undertaken with
regard to the utilization of rice hulls for a wide range of applica-
tiona. An excellent survey of the utilization of rice hulls as an
energy source has been done by Beagle (1978). Books on the various
applications have been published by The American Cereal Chemist
Association (Houston, Chapter 12, 1972) and Angladette (1966). Ultimate
analyses of rice hulls as well as ash content measurements are given by
Beagle (1978), Houston (1972), Jones (1953) and Borasio (1928) for
various locations. A number of 40 - 80 hp rice hull gas producers were
in operation for decades in Italy. Descriptions and sketchy performance
data are given by Garbarino (1911) and Francesco et al (1917) about the
80-100 hp Italian rice hull gas producers. The journal IL Giornale di
Risicoltura has regularly published information on rice hull gasifica-
tion as well as on the properties and utilization of rice hulls, most of
which appeared between 1908 and 1950.
The German engine maker, Humboldt-Deutz Motoren AG, did extensive
testing on rice hull gas producers during the 1930s. Possenti (1944)
describes a successful Italian rice hull gas producer, named the
Balestra unit after the engineer who built and operated it for over 30
years in the gas plant of Vercelli.
Tucker (1944) investigated some physical properties of rice hulls
such as thermal conductivity and true density. However, his investiga-
47
tive methods leave ·some doubt about the validity of the published data.
The most recent reports of existing rice hull gas producers are
listed in the FAO publication RAPA No. 61 (FAO Bangkok, June, 1982).
The publication discusses rice hull gasification efforts in Malaysia,
Thailand, The Peoples Republic of China and the Philippines.
The Egyptian government shut down their last remaining rice hull gas
producer because one worker died of carbon monoxide poisoning while two
others developed respiratory diseases (Beagle, 1982). A well-preserved
(1925) rice hull gas producer and engine (Crossley-Premier) was reported
by Beagle (1983) in Burma. It may therefore be concluded that rice hull
gasification on a larger scale (40 - 150 hp) is well documented and was
successful at least in Egypt, Italy and Burma. Recent work and reports
about development units as well as laboratory units at the university
level must be viewed with great caution. Well-documented tests on rice
hull gasification in a fluidized bed are given by Van den Aarsen (1982).
48
4. REFERENCES
1. Angladetta, A. 1966. LeRiz. G.P. Maisonneuve and Larose, Paris.
2. Beagle, E. 1978. Rice husk conversion to energy. FAO

Agricultural Services Bulletin #31.
3. Beagle, E. 1982. Personal communication. October, 1982.
4. Beagle, E. 1983. Personal communication. January, 1983.
5. Borasio, L. 1928. Sulla composizione chimica della lolla di

riso. IL Giornale di Risicoltura, Anno XVII, No. 3.
6. Francesco, et al. 1917. Processo industriale per la utiliz-

zazione della lolla. Premiata Tipo litografia Gallard! and Ugo,
Vercelli 1917.
7. Garbarino, G. 1911. Utililzzazione integrale della lolla di

riso come combustibile. November 1911.
8. Jones, J.D. 1953. New refractory from vegetable source.

Canadian Metals. January.
9. Humboldt-Deutzmotoren AG. 1933. Gas producer for using vege-

tables wastes, e.g., rice husks, nut shells of cacao, coffee or
palm kernels. Chem. Abstracts Vol. 27:5520.
10. Humboldt-Deutzmotoren AG. 1934. Gas producer for using husks and
similar vegetable wastes. Chem. Abstracts, Vol. 28:3565.
11. Kaupp, A. and J.R. Goss. 1982. State-of-the-Art for small scale
(to 50 kW) gas producer engine systems. u.s. Agency for
International Development, Washington, D.C. 20523.
12. Possenti, A. 1944. Gasification plant for rice hulls in the gas
plant of Vercelli. Risicolture Vol. 30:243-251 and Chem.
Abstracts Vol. 37:3247.
13. Tucker, P.W. 1944. A study of the physical and chemical proper-
ties of rice hulls. M.S. Thesis, Louisiana State University,
Department of Chemistry.
14. Van den Aarsen, F.B. et al. 1982. Performance of a rice husk
fueled fluidized bed pilot plant gasifier. First International
Producer Gas Conference, Colombo, Sri Lanka. November.
49
S. CHEMISTRY OF GASIFICATION OF RICE HULLS
Introduction
Gasification of coal and biomass had its heyday at a time when very
little was known about the process except that incomplete combustion
generated a gas which in turn could be further oxidized (burned).
The simple definition of gasification given earlier as a process
that stores a maximum of chemical energy in the gas phase does not
include any suggestions about how this is achieved, nor does i t disclose
the principles behind this conversion process.
Any air-blown combustion process yields a range of gases such as:
CO, COz, CH4 , CzHz, CzH4 , CzH6 , c 3 ~, c 4 ~, Nz,
Hz, HzO(g), NOz, NO, 0, Oz, N and H
as well as condensable vapors consisting of higher hydrocarbons.
Advanced computer programs as well as experimental results have
shown quite clearly that at the prevailing gas producer temperatures of
800-1300°C the products which contribute more than 1% to the total gas
volume are restricted to CO, COz, CH 4 , Nz, Hz, HzO(g) and in some
instances Oz •
One elemental chemical ultimate analysis of dry rice hulls shows
that 100 kg of them can be represented by:
C3.45 H4.94 °Z.33 N0.04 (Si0z)O.Z6 (see Table 6-1)
This representation accounts for 99% of the constituents of rice hulls.
It is therefore up to the individual to decide how the gasification pro-
cess will be characterized based on the given input and output species.
One should realize that there are many options available and that those
so
models which most closely resemble a real situation are in most cases
not solvable. The non-solvability of models is not primarily caused by
insufficient numerical methods. The major handicap and missing link are
the physical parameters of the fuel and the design parameters of the gas
producer that must be incorporated into the governing equations, either
as dependent variables or as boundary conditions.
The complexity is shown by a list of properties involved in
modelling the gasifier performance:
a) Temperature Field
b) Flow Field
c) Species Concentration Profiles
d) Intrinsic Reaction Rates of the Homogeneous Reactions
e) Boundary Conditions
f) Mass Transport across Phase Boundaries and Heterogeneous
Surface Reactions.
Formation Reactions
It is customary to list the formation reactions of the major pro-
ducts as well as their interaction. One should not conclude that for-
mation of these species can be adequately characterized by this simple
set of equations (at 25°C, 1 atm):
Combustion c + oz = co2 - 394 MJ/kg-mole (5-1)
Water Gas c + HzO CO + Hz + 131 MJ/kg-mole (5-2)
Water Shift Reaction CO + HzO = COz + Hz - 41 MJ/kg-mole (5-3)
Boudouard Reaction c+ COz =2 co + 173 MJ/kg-mole (5-4)
Methane Reaction c + 2Hz CH4 - 75 MJ/kg-mole (5-5)
51
It is difficult to determine whether C02 is formed directly according to
Equation 5-4 or whether i t is the final result of a sequence of reac-
tions such as:
c + 1/2 02 t co (5-6)
CO + 1/2 02 t C02 (5-7)
In addition, methane can be formed in many ways in homogeneous reactions
such as:
CO + 3 H2 t CH4 + H20 (5-8)
C02 + 4 H2 t CH4 + 2 H20 (5-9)
The basic set of equations listed above does allow one to identify
various zones in a gas producer as shown in Figure 5-l for a downdraft
gas producer where fuel and air are flowing in the same direction.
PYROLYSIS
E
u COMBUSTION
w
(.)
z REDUCTION
<X
SHIFT
~ 20 REACTION
Cl
10
ASH LAYER
----Grote
0 400 600 800 1000
TEMPERATURE, °C
Figure 5-l. Co-current or downdraft gasification.
52
Reaction Zones
Splitting up a gas producer into strictly separate zones is cer-
tainly not realistic but is never~heless conceptually useful. Most com-
monly we have:
a) Pyrolysis Zone
b) Combustion Zone
c) Reduction and Shift Zone
It is necessary to distinguish between reversible reactions which may be
in equilibrium and irreversible reactions which can hardly be modelled.
It is therefore difficult to find an adequate mathematical description
of the processes taking place in the pyrolysis zone. The pyrolysis pro-
cess, or low temperature decomposition of biomass, is highly irrever-
sible. It involves the release of oxygen, hydrogen and nitrogen in the
fuel along with some carbon. The products evolved are called the vola-
tile matter of the fuel. It is not difficult to determine the tem-
perature at which the release of volatiles ceases i f the experiment is
done in a pure nitrogen atmosphere. In such a case the accumulative and
differential mass loss curves exhibit the typical behavior shown in

Figures 5-2 and 5-3 for wood.
a) Almost no release of vapors and gases below 250°C.
b) Rapid devolatilization between 250 and 400°C, the rate reaching
its peak at 375°C.
c) Very sharp drop in the release of products after 375°C with
termination of the process beyond 450°C.
53
100
1-
J:
(!)
IU 50
~
0
100 400 500
1 O c
Figure 5-2. Accumulative mass loss curve
(modified from Shafizadeh, 1976).
It is noted that in a technical sense pyrolysis and devolatilization
are defined as taking place in an oxygen-free atmosphere. Therefore,
after the release of all volatiles, carbon can no longer co~bine in an
oxygen-free atmosphere and the process stops completely.
z
0
i=
cr
:E
a::
0
u...
....
u
::I
0
0
a::
CL.
u...
0
....
1&.1
cr
a::
200 250 300 350 400 450

TEMPERATURE, •c
Figure 5-3. Differential mass loss curve. (Shafizadeh, 1976).
54
The products of pyrolysis of biomass are a mixture of water and con-
densable higher hydrocarbons, gases, and solid residue. A long list of
trace species often found in the condensate of biomass was included
earlier (Table 1-1). The proportions of the various products of the
pyrolysis of rice hulls at different temperatures will be given later
(Chapter 9).
In practice, thermal devolatilization occurs with at least some oxy-
gen present (unless, of course, special steps are taken to exclude it).
One of the best known practical applications of devolatilization of
biomass is the ancient skill of charcoal production. By burning some of
the carbon, the temperatures that are necessary to drive off the vola-
tiles are achieved (around 300°C). It is noted that the devolatiliza-
tion of biomass is not self-sustaining. The overall heat of reaction
oscillates about zero; thus the reaction switches from endothermic to
exothermic as shown in Figure S-4.
'·· '\
\
/ "-
_,.,............
/.
.r··,
'\ '\. .............
. I . '
'·
~·
__.........-.....,·.........\j_____ ,.,... '
0 '··,WOOD
"
"' ..... __....,.--·- ·-·-·
,_. \
\
<l
'·'LIGNIP-1
0
0
..,
z CELLULOSE
j
..........
.......... ,... -............ ,,,
\, :' "---- .....................
\,,,, ' I '.,
............. XYLAN
450
TEMPERATURE (°CI
Figure S-4. Differential thermal analysis. (Shafizadeh, 1976).
A common laboratory method for characterizing biomass fuels is the
proximate analysis. In this test the amount of volatile matter, ash and
55
residual carbon in the ash is determined by heating a sample of the
biomass to a specific temperature. However, the product distribution is
strongly dependent on the temperature, rate of heating and the type of
furnace. In addition, the presence of even minute amounts of air can
greatly influence the results of these tests. It is therefore not
surprising that a wide range of proximate analysis data exist for a
given biomass fuel. In general the proximate analysis is not a reliable
tool for judging the usefulness of a biomass fuel as a smokeless
feedstock for burners and gasifiers because the tests are done in an
unspecified atmosphere. Beagle (1978) lists results from previous work.
His range for rice hulls are given in Table 5-l. Differences can also
occur because of varietal characteristics and processing variables, etc.
Table 5-l. Proximate Analysis of Rice Hulls (% weight, dry basis)
Ash 15.8-23.0
Volatile Matter 56.4-69.3
Fixed Carbon 12.7-17.4
Two widely used methods for pyrolysis analysis of coal are the
Fischer and Gray-King Assays. Probstein and Hicks (1982) have carried
out a linear least square fit of a large number of standardized 500°C
assays for coal. Their work proves that is is possible to predict with
high accuracy the fraction of tar + light oils and water released. One
interesting part of the evaluation is the linear dependence of the above
components on the H/C and 0/C ratios. Unfortunately research on the
thermal decomposition of biomass has not yet reached the state where a
large number of results have been obtained under identical conditions.
56
The equations for coal are given below and will be compared later to a
500°C assay of rice hulls.
Tar + Light Oils (mass%) 29.1 c-

H
12.1 (5-10)
19.7 c0 + 1.64 (5-11)
One widely used gas producer, the downdraft system, has already been
introduced (see Figure 5-1). Another commonly used system, the updraft
gas producer, is shown in Figure 5-5. It is noted that no matter how a
moving (fixed) fixed bed gas producer is built, there will always be a
low temperature zone where pyrolysis takes place, thus a considerable
amount of condensable higher hydrocarbons will be generated. In deve-
loping a useful model of the gasification process one must exclude this
zone and its products because it is very difficult to postulate the pro-
ducts formed, the mode of formation of these products and also the
manner in which they are again decomposed.
Gas
Figure 5-S. Counter-current or updraft gasification

(Skov, Papworth, 1975).
57
It will be shown in Chapter 8 that the mixture of condensable hydro-
carbons and H2 0 formed is not a small fraction of the pyrolysis products
but amounts to over 30% (by weight) of the rice hulls. On the other
hand, many gas producers are either operated with devoiatilized fuels
(charcoal) or at high enough temperatures to induce tar cracking. In
these cases, there is no need to account for the products of pyrolysis.
Model I
One of the simplest models which will be called Model I, is based on
three assumptions.
Assumption 1:
a) The two input streams, fuel and air, are given as (see 5.1):
Fuel: c3 • 45 H4 • 94 o 2 • 33 (Sio 2 >0 • 26 (100 kg of dry rice hulls)
Air: N1 • 58 o0 • 42 (1 kg-mole air)
For an equivalence ratio of 4J .. 1 the chemical composition of the two
feed streams is:
It is noted that Silicon (Si) is already in its oxidized form in the
rice hulls and will not be oxidized by the combustion air. Silicon is
most often found in plants within a Silicate System. The representation
as Sio2 orily serves to indicate that there will be no oxygen consumed by
this mineral in rice hulls. With biomass fuels having an ash content
smaller than 1% such as corn cobs and most woods, the ash may be
58
neglected. However as shown in Figure 8-11, the ash content of rice
hulls is too high to neglect in computations.
At first, one may find it strange to represent a mixture of solid
carbonaceous fuel and air in the form of atomic ratios which does not
reveal the actual nature (gas, liquid, solid) of the inputs. However,
the product concentrations are functions only of temperature and
pressure and not the original arrangements of the atoms. Thus, for a
computer program "combusting" a given fuel, i t does not matter whether
the original atoms are in a gaseous, solid or liquid form.
Assumption 2:
All possible reversible reactions go to equilibrium. This assump-
tion is not completely valid for two reasons:
a) The pyrolysis process is highly irreversible and there is a wide
range of intrinsic gasification rates for the Equations 5-l through 5-5.
How the intrinsic rates compare is given by Walker et al. (1959) and
listed in Table 5-2.
Table 5-2. Relative Intrinsic Gasification Rates at 1 atm and 800°C
Reaction Relative Rste
C + 02 t C02 105
C + H20 t: CO+ H2 3
C + C02 t 2 CO 1
C + 2 H2 t Clft. lo-3
It is noted that the water shift reaction (Equation 5-3) is the only
significant homogeneous reaction taking place. The assumption of
equilibrium is therefo!'e justified bqt nevertheless difficult to prove
for several reasons:
59
a) In an updraft gas producer the off gas is always mixed with the
products (gases, tar and water) of the pyrolysis zone. Any reliable
verification of the water shift reaction reaching equilibrium must be
done below the pyrolysis zone. In addition a "true" temperature must be
measured.
b) The same difficulties arise with the off gas from a downdraft
gas producer. In a well-functioning unit, no uncracked pyrolysis pro-
ducts, except water, will be present in the raw gas. However, one still
faces the difficulties involved in determining a "true" temperature for
the gas sample.
To verify whether the equilibrium state of the water shift reaction
is reached, it is best to use a fluidized bed gas producer, because a
more homogeneous temperature and species concentration field is present.
In fact, by definition the species concentrations and temperature field
are independent of the spatial coordinates in a fluidized bed gas pro.:.
ducer.
The degree to which the water shift reaction approaches equilibrium
in such a unit is shown in Figure S-6. The conversion efficiency, n, is
represented by:
(C02)max - (C02)meas
n
(C02)max- (C02)equil
The graph indicates that the product distribution is dominated by
pyrolysis products in the low temperature range. In the high tem-
perature range, the graph shows that the water shift reaction is near
60
completion. Based on these results it is of interest to check the tern-
perature sensitivity of the equilibrium constant, ~ (T).
0.9
0.8L-~--~----~--~----~----
700 800 900
REACTOR TEMPERATURE,°C
Figure S-6. Equilibrium of the water shift reaction as a function of

temperature in a fluidized bed rice hull gasifier
(Van den Aarsen et al., 1982).
Kp (T) is given by:
[C0 2 ] [H 2 ]
e- li G(T)
0
(5-12)
RT [CO] [~OJ
Kp (T) as a function of T is shown in Figure 5-7. The standard free
energy change for the reaction when going to completion at 1 atm is
given by li 0 G(T). Its value can be computed in KJ/g-mol by:
61
where ! represents the species CO, co2 , H2 o and H2 • Then ~ 0 G(T) is given
by:
n
~OG(T) "' I ~o
i
G(T) (5-14)
!=1
The coefficients in Equation (5-13) are tabulated by Probstein and Hicks
(1982).
-
1-
Q.
:::.&::
5
CO+ H 2 0 =CO2+ H 2
1- 4
z
<(
1- ~
(/) 1-
z <(
3
0 II
(..)
Q.
~ 1-
~ <(
~ 2
CD
-
...J
~
0
w
500 600 700 800 900 1000 II 00

TEMPERATURE T, °C
Figure 5-7. Kp (T) as a function ofT.
Example
The following example shows that it is not always possible to deter-
mine the validity of Assumption 2.
A typical raw gas compos! tion as measured in one of the small rice
hull downdraft gas producers is the following':
62
Species % Volume
COz 12.6
co 15.3
Hz 10.8
H20 10.1
The associated equilibrium temperature is 840°C. The measured tern-
perature was 742°C. At 742°C, the equilibrium composition should be:
Species % Volume
C02 13.4
co 14.5
H2 11.6
HzO 9.3
From the example it is clear that the water shift reaction due to its
small energy change with respect to large temperature changes is very
stable. The vastly different response times of the gas analyser com-
bined with the inaccuracy of high temperature measurements in a convec-
tive gas stream prohibits any firm conclusions from pilot plants.
Assumption 3
No condensable products except for water, minerals and solid carbon
are present in the product stream.
Basically this means a "tar-free" gas, thus no carbon or hydrogen is
bound in higher hydrocarbons (> C5~). This assumption cannot be valid
for most gasifiers. It may be justified for well-designed downdraft gas
producers.
63
Computer Program
Based on these three assumptions, the raw gas compositions of a rice
hull "reactor" operating at various equivalence ratios have been com-
puted. Reactors operating without an external heat source are sometimes
labeled gas producers whenever the gas contains enough chemical energy
to allow subsequent combustion in a furnace or internal combustion
engine.
The computations have been done with the STANJAN program developed
by R. Reynolds (1982). This program was selected over the other
available programs for the following reasons:
a) The program uses one of the largest thermodynamic data bases
presently available (JANAF Tables, 1971).
b) It allows the user to specify the products. For the purposes of
this work, the following product species were chosen:
Solids: c
Liquids: H2 0 (.l)
Gases: CH30H, CH4, C2H2, C2H4, C2H6, H, H2, H20(g), HO, N, N2 ,
25, N02, O, 02, C(g).
c) The program accounts for condensable phases and for the products
of disassociation reactions which become important at higher tempera-
tures.
The results of the computations are given in Figures 5-9 through
5-22. Species having mole fractions less than 0.01 are not shown in the
graphs. It is noted that the gas composition has been computed over a
temperature range which will not be attainable without external heating
for small values of 4>. For instance, sustaining the gasification pro-
64
cess at cjl = 0.1 and T "' 1000°C is hardly possible, because the heat
generated from the small amounts of carbon burned is not enough to match
the heat losses from the reducing reactions and from the system (see
Section 9.4).
Another limitation lies in Assumption 3. It will be shown in
Chapter 9 that insignificant tar conversion occurs at temperatures below
800°Co The graphs are therefore only realistic within a narrow tem-
perature band with a lower limit of 800°C. The upper limit will be the
adiabatic combustion temperature at a given cp.
Let c be the raw gas composition, then
c = c(cjl, M, T)
where: equivalence ratio
M = moisture content in % weight, wet basis

T temperature, °C.
The range of cases shown in Figures 5-9 through 5-21 for the temperature
interval 500°C to 1400°C is shown in the grid in Figure 5-8.
0
....<t 0.5
a:: 0.4
41
(.)
z
0.3
lJJ
...J 0.2
<t
>
:::::> 0.1
0
lJJ 0
0 10 20 30 40
MOISTURE, M
Figure 5-8. Range of computed gas compositions as a function of

M and cp.
65
""c:r
(I)
cp: 0
::t:
Q.. M=O
(/)
c:r
~
z
z
0
....u
c:r
a:
"-
_. 0.1
0
:E
1000 1200 1400
TEMPERATURE,°C
Figure 5-9. Gas composition, ~ = 0, M = 0
~
(I)
c:r
t
::t:
Q..
cp =0.1
(/) 0.6 M=O
c:r
~
z 0.5
co
~ 0.4
~---------------------- H2
~ 0.3
c:r
~0.2
~ 0.1
·:::::E
600 800 1000 1200 1400

TEMPERATURE,°C
Figure 5-10~ Gas composition, ~ = 0.1, M = 0
66
w
CJ')
<(
:I:
0..
CJ')
<(
<.!> cp = 0.2
z M=O
z
0 0.4
t-
co
0
<( 0.3 Nz
0:: Hz
~
0.2
....J
0 0.1
::e
800 1000 1200 1400
TEMPERATURE I 0 c
Figure 5-ll. Gas composit ion, ¢ = 0.2, M= 0
cp = 0. 3
M=O
Nz
-----------co
600 800 1000 1200 1400

TEMPERATURE I 0 c
Figure 5-12. Gas composit ion, ¢ = 0.3, M= 0
67
w
en
c:s:
:I:
a.. 4> = 0. 4
en M=O
c:s:
<.!)
N2
z
0 0.4
I-
u
c:s: co
0::
\L 0.2
Hz
..J
0 0.1 H20
::E C02
600 800 1000 1200 1400
TEMPERATURE,°C
Figure 5-13. Gas composition, ¢ = 0.4, M = 0
LLJ
en
<X
:r:
0.
en
0.6
<X
(!) 0.5
z
0 0.4
1- 4> =0.5
(.) 0.3 M=O
<X
a:: 0.2
LL..
...J
0 0.1
:E
600 800 1000 1200 1400
TEMPERAT URE, °C
Figure 5-14. Gas composition, ¢ = 0.5, M= 0
68
lLJ
cp = 0. 3
CJ)
c:r 10% MOISTURE, WET BASIS
:I:
a..
CJ)
c:r
C)
z 0.5
z 0.4 N2
co
0
t-
(.)
c:r 0.2
a: H2
LL.
.....J 0.1 H20
0 C02
~
600 800 1000 1200 1400
TEMPERATURE, °C
~
en
<(
X
Q.
cp = 0. 3
CJ)
<t
(!)
20% MOISTURE ,WET BASIS
z 0.5
~ 0.4
- - - - - - - - - - - - - N2
t- 0.3
(.)
<t
a: 0.2 -:=~~~=====:::co
H2
LL.
H20
_J 0.1
0 C02
:E
600 800 1000 1200 1400
TEMPERATURE I 0 c
69
LaJ cp = 0. 3
(I)
<( 30% MOISTURE 1 WET BASIS
:t:
a.
(f)
~ 0.6
~ 0.5
t; 0.3
<(
a: 0.2
IJ...
...J 0.1
0
:E
600 800 1000 1200 1400
TEMPERATURE, °C
Figure 5-17. Gas composition, ~ = 0.3, M = 30
LaJ
(f)
<(
:I: cp = 0. 3
a.
(f)
40% MOISTURE 1 WET BASIS
<(
(!)
z
z 0.4
0
~ 0.3
u
<(
a: 0.2
IJ...
..J 0.1
0
:E
600 800 1000 1200 1400
TEMPERATURE, °C
70
LIJ
(/) cp = 0. 4
<X 30% MOISTURE, WET BASIS
:I:
a.
(/)
<X 0.6
(!)
z
z
0
1-
u 0.3
<X
a::
LL.
.....J 0.1
0
~
600 800 1000 1200 1400

TEMPERATURE, °C
LIJ c/>=0.5
(/)
<X 30% MOISTURE,WET BASIS
:I:
a.
(/)
<X 0.6
(!)
z
r---~-----------------------N2
z 0.4
0
1- 0.3
u
<X
a:: 0. 2
LL.
.....J 0.1
0
~
800 1000 1200 1400
TEMPERATURE, °C
71
w cp = 0.6
en
<t 30% MOISTURE, WET BASIS
:t:
Q..
en
<t
(!)
600 800 1000 1200 1400

TEMPERATURE I °C
z -6
t
0
....u -8
<(
a::
LL.
....J
0
-12
:E
0
-14
C) -16 cp = 0. 3
0
....J
-18
600 800 1000 1200 1400

TEMPERATURE , °K
Figure 5-22. c2Hm equilibrium composition at
various temperatures, ¢ = 0.3
72
Analysis of the Results
Although no mass transport, kinetics and local energy balances are
incorporated in Model I some important conclusions can be drawn.
a) Hydrocarbons CH4 , c 2 ~, c3 ~ are not formed from molecular or
atomic hydrogen present in the gas phase. At temperatures above 800°C
none of these hydrocarbons are formed in significant amounts. (See
Figure 5-22). Consequently, large quantities of gaseous total hydrocar-
bons in the sample gas are an indicator that the unit is generating a
highly contaminated raw gas and operating at a low temperature.
b) Operating a reactor at a low equivalence ratio results in high
CO and H2 fractions at high temperatures. However in practice such con-
di tions are not possible without external heating. This last statement
has important implications with regard to the design of a gas producer
and with regard to the condition (moisture) of the rice hulls.
In order to achieve the favorable combination of low cp and high tem-
peratures it is necessary to:
(i) Heavily insulate the gas producer.
(ii) Preheat the combustion air with the sensible heat of
the raw gas, using a heat exchanger.
(iii) Use rice hulls that are as dry as possible.
Influence of the Moisture Content
The moisture content of rice hulls depends on handling and milling
procedures (commercially 8-12%) as well as the local storage and cli-
mate. Moisture in the feed material has a marked effect on the raw gas
composition and efficiency of the process. There are significant dif-
73
ferences between fluidized bed gasifiers, updraft gasifiers and
downdraft gasifiers with respect to their ability to cope with high
moisture contents of the rice hulls when on an experimental basis.
In an updraft gas producer the moisture in the fuel does not come
into contact with a hot carbon zone (> 800°C). Therefore the only
detrimental effect of a high-moisture content is a reduction in effi-
ciency, because more carbon must be combusted to provide the necessary
heat for evaporation of the water. In addition, the gas stream loses a
fraction of its sensible heat, in heating up the wet feed material.
In downdraft and fluidized bed gas producers with bot tom feed, the
water evaporated from the fuel must pass through high temperature zones
rich in carbon, where it is converted to a certain extent to water gas.
It will be shown in Section 9.4 that the water gas reaction (5-2) is
slow and not very effective in a rice hull fuel bed.
The effect of the moisture content on the gasification reactions and
product composition is shown in Figures 5-13 through 5-21. The first
set of graphs (5-15 through 5-18) shows an operation at <j> 0.3 with
rice hulls having a 10%-40% moisture content. It is noted that the
water vapor content of the gas increases significantly at the expense of
CO with insignificant change in the hydrogen content of the raw gas.
The trend is shown for one specific temperature (T == 1000°C) in Table
5-3.
74
Table 5-3. Species Concentration at T = 1000°C and cp 0.3
Species % volume at various M
10% 20% 30% 40%
co 26.1 22.3 18.6 15.0
Co 2 6.6 8.2 9.5 10.3
Hz 20.6 21.2 21.5 21.2
HzO 8.6 12.8 18.0 24.0
c~ 0 0 0 0
In practice, an increasing moisture content necessitates an
increasing <P in order to provide the necessary heat of reaction to reach
reasonable gasification temperatures (1000°C and higher). The heating
value of the gas is greatly decreased. (See Figures 5-19 through 5-21).
In this set of figures the equivalence ratio <P is varied using a fixed
moisture content of 30%. As can be seen from Table 5-4 the raw gas com-
position shows a sharp decrease in the combustible fractions of CO and
Hz with increasing cp, a corresponding increase in the sensible heat will
take place.
Table 5-4. Species Concentration at T = 1000°C, M = 30%, wet basis
Species % volume at various p
0.3 0.4 0.5 0.6
co 18.6 14.7 11.4 8.6
C0 2 9.5 10.6 11.7 12.6
Hz 21.5 16.3 12.1 8.7
H2 0 18.0 19.4 20.4 21.1
c~ 0 0 0 0
75
This behavior of the gasification reactions is one of the most often
overlooked disadvantages of gasifiers that are capable of cracking tar
and decomposing moisture. The high temperatures and significant energy
needed for the conversion of both tar and water can only be achieved by
drastically increasing the equivalence ratio for biomass having a high
moisture content. This in turn leads to internal burning of the com-
bustible fractions, CO and H2 , and a shift from the chemical energy to
sensible heat.
Higher Beating Values of the Computed Gas Compositions
The higher heating values of the raw gas composition are given in
Figure 5-23. The heating value decreases significantly with increasing
cp, because of the shift from chemical energy to sensible heat as shown
in Figure 1-1. However, a more interesting feature of the raw gas com-
position is its nearly constant heating value at temperatures above
900°C, for all .p. The reason for this phenomenon is that the water
shift reaction dominates at higher temperatures causing only a shift in
the gas composition towards a higher CO/H2 ratio. The water shift reac-
tion has little effect on the total chemical energy since H2 and CO have
about the same higher heating value (286 MJ/kg-mol for H2 and 283
MJ/kg-mol for CO, at 25°C and 1 atm). I t is noted that operation at
higher temperatures than what is required to crack tar will therefore
not result in a gas with a higher heating value.
76
13
¢:0.1
12
I I
0'
10
~
........
-:>
9 ¢:0.2
:!'!
(/)
8
<t
<.!) 7
!: 6 ¢: 0.3
<t
a::
5
lJ...
0
¢: 0.4
> 4
:I:
:I: 3 </> =0.5
2 ¢: 0.6
¢ : 0. 7
4>: 0.8
600 800 1000 1200 1400

TEMPER A TURE , o C
Figure 5-23. Higher heating value of the raw gas for range of T and $•
Summary for Hodel I

a) The raw gas composition was computed as functions
ci = c1 ($, M, T), i = 1,2,3,N.

b) The results apply only to gasifier designs that allow the fuel
moisture and pyrolysis products to be converted into gaseous
products.
77
c) The computed gas composition must be viewed as the best that
can be achieved at any given temperature, irrespective of
whether this temperature is obtainable with the given ¢>• One
exception is the methane content of the gas. It is usually
higner in experimental data, since total tar conversion rarely
takes place.
d) It will be shown later (Section 9.4) that the steam conversion
is far from completion at any temperature in a rice hull fuel
bed with its high void space fraction. The actual water con-
tent of the gas is therefore higher than the computed one.
e) The gasification of rice hulls with a very high moisture con-
tent () 30%) seems to be impractical in downdraft units and
fluidized beds. The issue is not the decreasing efficiency of
the process but the high cp needed to sustain temperatures
favorable for tar cracking and moisture conversion.
f) Hydrocarbon gases in the product stream are unconverted pyroly-
sis products and indicate low temperature gasification.
It is noted that this simple concept of gasification, viewing the
reactor as a black box with a known input stream, does not account for
mass transport phenomena, kinetic rates or local energy balances.
However the advantage of Model I is that it is easily solvable and is
very helpful in pointing out general trends. Taking into account
intrinsic kinetic rates of the various reactions and mass transport
limitations, Groeneveld (1981) in his dissertation, compared more
sophisticated models to his experimental data from a small charcoal and
wood downdraft gas producer.
78
One could also follow a completely different path, as outlined by
Whitaker (1981) and argue that since the process must occur according to
the basic laws of mechanics, i t is always possible to write down the
"exact" governing differential equations for the process.
Model II
The difficulties with Model II are the following:
a) The "exact" equations express all variables and physical para-
meters as a function of the spatial coordinates and time.
b) Boundary conditions, which depend heavily on the design of the
gas producers, must be established.
c) Any measurements taken constitute an average bulk phase result;
it is therefore necessary to volume average the point equations
in order to compare theory and experimental data.
d) The "exact" differential equations must be modified to a point
where they can be expressed in terms of physical parameters
that are readily measurable such as temperature, mass velocity
fields, thermal conductivity, void space fractions and dif-
fusion coefficients.
e) An order of magnitude analysis must be performed to exclude all
terms that do not contribute significantly to the process.
f) All processes take place within a three-phase system - 1. void
space in the bed; 2. solid phase in rice hulls; 3. gas phase in
rice hulls.
It is noted that the limitations of this method are not due to a lack of
computer programs to perform the necesary computations, but are related
79
to the lack of knowledge concerning the magnitudes of the physical para-
meters and their relative importance.
Although it is beyond the scope of this work to outline in detail
the complications involved in establishing a set of solvable equations a
short outline is given. The species continuity equation is given as
(5-15)
The energy equation will be represented by
DT !!?.
pc - = - V•q - V•I Ji Hi + Tt5
p Dt Dt
n n
+ v v t + t - I Ht Ri + I Hi V•Ji (5-16)
i=l i=1
It is noted that:
a) The index, i, stands for the various major gas species, CO,
b) Both equations are coupled by an unknown velocity field which
can be determined by the Navier Stokes equations.
c) The displacement of a gas molecule by diffusion (caused by con-
centration gradients) is ignored because it is insignificant
when compared to convective displacement.
d) Based on typical designs of rice hull gas producers no
variations of ci in the radial direction is assumed and there-
fore only concentration gradients in the axial direction must
be taken into consideration.
80
Species Concentration Equation
As already mentioned it is necessary to convert the unworkable point
equation into a volume averaged differential equation. This is done by
first defining the size of the control volume V in which physical and
chemical properties are considered constant. The general structure of a
rice hull control volume is shown in Figures 6-1 and 6-2. The various
phases and numerical values of their fractions have been experimentally
determined and are reported in Chapter 6. We define for the B-phase:
(ci> =l.v fVB

ci dV (5-17)
(f)
1
v If
va
dV (5-18)
for any parameter f we consider constant within the control volume v.
For purely technical reasons one may wish to define
dV
The relationship between (f)S and (f) is given by (f)
represents the void space fraction given as
(5-19)
The process of volume averaging then leads to
V <vi ci> +.!.v f

ABo
c 1 (vi - w) • n
ao
dA (5-20)
81
It is noted that <Ri> is the "production" rate of species i in the gas
phase. This term is non-zero only for the homogeneous water-shift reac-
tion. All other reactions are assumed to be heterogeneous. They are
accounted for by the integral term
f
A~o
dA
which describes the niass flux (mole m- 2 s- 1 ) of species i across the
apparent surface Aa 0 as shown in Figure 6-2. The velocity w is the
velocity in the interphase.
It is, of course, up to the individual to decide which model will
best fit this interfacial mass transport. It is noted from Figure 6-2
and the micrographs 7-14 and 7-15 that the interfacial area A~ 0 is the
sum of the areas associated with the solid carbon phase and the pore
entrances. The rice hull velocity in the reactor is assumed to be iden-
tical with the average of w.
Comparing micrograph Figures 7-4 and 7-12 one can see that rice
hulls exhibit a rather unusual behavior during thermal decomposition.
Their outer surface appears to be impermeable both before and after
gasification. In fact, no visible pores could be detected at a magnifi-
cation of 6000x. On the other hand, the inner surface becomes extremely
porous. If one assumes that no appreciable surface reactions occur on
the internal pore surfaces, then a manageable representation of the
interfacial mass flux can be developed. Based on Equations (5-l)
through (S-5) we can assume the following sequence of events:
a) A gaseous species i (for instance co2 ) is diffusing through a
82
boun4ary layer towards the carbon surface.
b) co2 is adsorbed by the surface at a specific rate kad • Cc 02 •
c) co2 will be present at the carbon surface at a concentration of """

ca.
d) co2 will react according to C + co2 t 2 co.
e) The product CO will be desorbed and will diffuse back into the
gas phase.
The missing link is the constituent equations for c) and d). For
instance, if one assumes first order irreversible surface kinetics and a
surface concentration proportional to the bulk phase concentration ci we
have
{5-21)
at the surface and the integral term is now modelled by
1
v {5-22)
with {5-23)
and a 60 , the interfacial specific area (m2 /m3) that has been measured
and is given in Chapter 6. Equation {5-21) now reduces to
{5-24)
Since <vici) is not measurable it is necessary to express it in terms of
known parameters.
Let v
i
= <v
i
>6 + V'i {5-25)
83
(5-26)
Thus, the point concentration and velocity of species. i are split up
deviation.
fact one
It is not unreasonable to assume that <vi>
simply chooses a very small control

-
into the sum of the measurable average bulk phase values and a spatial
volume
<~>
so
0.
that
In
the
deviations integrated over the volume cancel one another. Applying the
volume averaging rule to equations (5-24) through (5-26) we have
(5-27)
Based on theoretical considerations (Carbonell, Whitaker, 1983) one may
let (5-28)
where D is the hydrodynamic dispersion tensor. We therefore have:
Gasification of rice hulls can be described as a radial symmetric pro-
cess. It was previously assumed that Drr is zero and that no convective
transport in the radial direction takes place. This leads to
where vi = vz is assumed.
It is noted that Dzz refers to the axial dispersion coefficient for
species i in a mixture. From equat lon (5-30) it can be seen that the
solution of this fairly simple differential equation contains several
physical parameters such as
84
a) Void space fraction, £13
b) Dispersion coefficient, Dzz
c) Specific surface area, al3o
d) Linear adsorption isotherm, K
e) Surface reaction rate, "k
These parameters must be determined and are, of course, functions of z
although some of them, such as £ 13 , can be assumed to be constant (see
Chapter 6). In addition, the velocity field must be established. The
bulk phase velocity <vz> is certainly a function of z because the volu-
metric gas output is larger than the air input and the density of the
gas mixture is proportional to the temperature ratio. The horizontal
velocity field in small rice hull gas producers is unusual, as will be
explained later (see Section 6.6). In the case where there are homoge-
neous bulk phase reactions (water shift reaction) a constitutive
equation for (Ri>l3 must be found. Because of the highly endothermic
nature of the water gas and C02-reduction reactions the term (Ri)l3 will
be a function of z.
In summation: At this point the solution to the "exact" set of dif-
ferential equations can not be found due to the lack of data concerning
physical parameters. It is noted that this type of approach may also be
used to derive a set of equations that take into account the influence
of the internal pores of rice hulls. In addition, the above mentioned
axial dispersion coefficient Dzz does not include turbulent effects,
which may be overwhelming.
85
The Energy Equation
The complete equation (5-16) cannot be solved in this form. A check
of the dimensions shows that each term represents the rate of change of
i nterna 1 energy per un it Control Volume (J m-3 s-1). However one
can show that the following terms contribute insignificantly and will be
neglected:
n n
a) L Hi 'V • J i and L J i Hi which describe t'Qe energy transport·
i=l i=l
by diffusion of the individual species and are small.
b) The reversible component of the work term T~ v • 'Vp which
accounts for the compression and expansion of the gases is
neglected.
c) The irreversible component of the work term v .v : t which
accounts for the heat dissipation due to friction will also be
neglected.
In summation, the overall energy equation can be greatly simplified.
The result is the following:
n
pc
p
v•VT -'V•q + ~ - L iii Ri (5-31)
i=l
The term ~ represents an internal heat source, which describes
radiation. Radiation can not be neglected at the temperatures that pre-
vail in gas producers. The energy change due to the highly endothermic
and exothermic reactions taking place in a gas producer are represented
by:
86
n
L Hi Ri (5-32)
i=l
where Ri is the molar rate of "production" of the i-th species and ~ is
its partial enthalpy (J /mole mixture). It is noted that the latter term
includes the heat of format lon and not just the sensible heat change.
It is therefore given by Equation (5-33).
At this point, further simplification is only possible if the
effects of radiation within a packed bed can be assessed. It has been
concluded that an exact description of the rice hull gasification pro-
cess by the previously outlined method must be preceded by a more
detailed analysis of the physical and kinetic parameters involved.
Flame Temperatures of Producer Gas
Introduction
Equilibrium compositions for a wide range of equivalence ratios,
rice hull moisture contents and temperatures have been computed in
Section 5.2. Once the equilibrium composition is known, one can then
examine the possible adiabatic flame temperatures and the parameters
which influence these temperatures.
Using producer gas for external combustion (Stirling Engine.) or
direct combustion (furnace, boiler applications) is less complex than
using producer gas in an internal combustion engine, because of the
extensive cleaning of the gas which must take place before the gas can
be used in an I.e. engine. Some applications such as crop drying can
require a heat exchanger because the crop may otherwise be contaminated
(odor, particles) by the products of combustion. It is therefore impor-
87
tant to anticipate the various maximum temperatures that may occur. The
following cases have been studied:
a) Flame temperature as a function of the equivalence ratio, A,
for combustion of producer gas.
b) Flame temperature as a function of the products of combustion.
c) Flame temperature as a function of the mixture inlet tem-
perature.
d) Flame temperature as a function of the solid carbon content of
the raw gas.
e) Comparison with reported experimental data.
Computer Program
Let g be the adiabatic flame temperature; g then represents a func-
tion g(~, T, M, A) inc where:
T Temperature of equilibrium gas composition, °C
M Moisture content of the fuel, % weight, wet basis
~ Equivalence ratio for the gasification process as defined in
Chapter I.
A = Equivalence ratio for the combustion of producer gas.
The JANAF program was partially utilized to execute the computations.
It is noted that JANAF allows for the following options:
a) T, p specified
b) H, p specified
c) s, p specified
d) p specified, s same as last run
e) p specified, s and composition same as last run
88
f) T specified, S same as last run
g) H and p same as last run
h) T and p specified, compos! tion same as last run
i) u and V specified
j) s and V specified
k) v spec! fied, s same as last run
1) v specified, s and composition same as last run
m) U and V same as last run.
The same set of products as that used for the equilibrium calculations
was allowed to form. The computation was executed along the following
path.
a) Input of 100 kg of rice hulls at a given moisture content, M.
b) Determination of the amount of air needed at a given ~·
c) Determination of the atomic ratio of the rice hull-air mixture
d) Computation of the equilibrium composition with JANAF option a)
at p = 1 atm and T = 500°C - 1400°C.
e) Constant pressure expansion of the raw gas to a predetermined
temperature Tf and computation of the enthalpy of the raw gas:
n n
H
g L
i=l
Hi 298
,
+ L
i=l
cpi (T) dT (5-33)
Tf represents the t_emperature of the mixture of producer gas
and air as fed into the burner. This step was done with JANAF
option h).
89
f) Determination of the mass of raw gas generated over an input/
output mass balance.
g) Determination of the stoichiometric amount of air needed to
combust the raw gas.
h) Computation of the atomic fractions of the air-producer gas
mixture for various A's. (0.5 ( A ( 1.5).
i) Computation of the total enthalpy HR of the air-producer gas
mixture at Tf • -
j) Determination of the adiabatic flame temperature Tad of the
products when ~(Tf) = Hp(Tad>. This last step was done with
JANAF option b) with a two-dimensional iteration, guessing Tad
and then adjusting the product composition.
Flame Temperature as a Function of the Equivalence Ratio A
A typical raw gas composition:
co = 17.1%
C02 = 8.2%
H2 = 13.8%
H20 = 21.8%
N2 = 39.1%
was assumed. For this composition the adiabatic flame temperature Tad as
a function of A is shown in Figure 5-24 at a mixture inlet temperature
of 200°Co It is noted that in theory the peak temperature is reached at
A = 1. In practice a higher A is needed to burn off most of the CO,
hydrocarbons and solid carbon.
90
wu
::E 0
<X
..JW
u..ct: 1400
::>
<->t-
j:<X
1300 co = 17. I
c:xct: C02 = 8.2
IDW
c:xa.. 1200
-::E
ow H20=21.8
"'o
<XI- 1100
N2 =39. I
1000
0.5 0.7 0.9 1.1 1.3 1.5

EQUIVALENCE RATIO A
Figure 5-24. Adiabatic flame temperature as a function of A
Flame Temperature as a Function of the Products of Combustion
From a computational point of view, it is easiest to perform an
energy balance if one assumes that the only products of combustion are
In reality, however, many radical and dis so-
elation products are formed as listed in section 5.1.
The data for Figure 5-24 take into account the major dissociation
reactions that occur in combustion processes. It is therefore of
interest to determine whether the use of a restricted set of products
such as C02, H20, 02 and N2 will significantly alter the results.
91
Computation of the adiabatic flame temperature under the latter assump-
tion leads to Tad = 1542 °C at A z 1 as compared to Tad = 1530°C when
dissociations are taken into account. The error from assuming a simple
set of products is therefore smaller than 1% and is not significant.
Flame Temperature as a Function of the Mixture Inlet Temperature
In many applications it is desirable to feed the producer gas-air
mixture at elevated temperatures into a furnace. There are several
reasons for this practice:
a) The wide range of boiling points of the tarry products makes it
necessary to avoid cooling down the gas too much before i t is
combusted.
b) Often the tar vapors in the raw gas are in a saturated state or
slightly above the dew point. Mixing of producer gas with cold
combustion air will therefore lead to the formation of tar
droplets, which are more difficult to combust than tar vapor.
These tar droplets may ·condense out along with the water formed
during combustion and accumulate within the furnace.
c) In some applications it is desirable to reach furnace wall tem-
peratures as high as possible (Stirling Engine, Carnot cycle).
The effect of the inlet temperature Tf of the gas-air mixture into
the furnace is shown in Figure 5-25. It is noted that raising the inlet
temperature of the mixture by 500°C will result in an increase in Tad of
only 325°C. This modest increase is not due to higher dissociation
activity, as already shown. It is instead the result of the increase of
Cp(T) with temperature for all gaseous species.
92
A=
wu
~ 0
<f
...J w
l4. 0:::
::::::>
u I-
I- <f
<f 0:::
CD w
<f a..
~
•171}
co = 8. 2
C02
0
<f I-
w H2 = 13.8
H20= 21.8 "·
N2 = 39. I
100 200 300 400 500 600

MIXTURE INLET TEMPERATURE,°C
Figure 5-25. Adiabatic flame temperature Tad as a function

of mixture inlet temperature.
Flame Temperature as a Function of the Char Load in the Raw Gas
Direct firing of producer gas can not be assessed on the basis of a
dry gas analysis, since the dust-laden, tar-laden gas is burned directly
1n a furnace. One obvious concern is whether large amounts of solid
carbon in the raw gas stream will significantly alter the adiabatic
flame temperature. The raw gas composition is the same as in the pre-
vious cases. Three cases: 0 grams, 6 grams and 20 grams of solid car-
bon per cubic meter of raw gas at an inlet temperature of 200°C were
considered. The adiabatic flame temperatures for the three cases at
A = 1, were:
93
Tad,O
Tad,6
Tad, 2 0 = 1640°C
A significantly higher temperature may therefore be expected when high
char loads occur in the raw gas.
Comparison of Theoretical Results and Experimental Data
One of the most extensive studies involving bulk phase temperatures
of producer gas flames as well as wall and flue gas temperatures was
conducted by Shoffstall, Waibel (1978). The results given in Table 5-5
were obtained by preheating the combustion air to 180°C and mixing it
with producer gas at ambient temperatures.
Table 5-5. Adiabatic Flame Temperature and Measured Maximal

Temperatures of Various Producer Gas Compositions
Gas co
Lurgi oxygen 18.5 40.2 29.4 9.4 0.6 1.9 1735 1541
Winkler oxygen 32.9 41.2 20.0 3.0 1.0 1.9 1830 1634
K-T-oxygen 52.1 34.5 9.2 0.5 1.0 1.9 1970 1782
Wellman-air 26.3 14.3 7.4 2.6 46.9 1.9 1620 1535
Winkler air 21.1 13.0 6.9 0.6 56.5 1.9 1415 1402
From the data one can conclude that the maximum temperatures obtained
were very close for air blown gasifiers to the computed adiabatic flame
temperatures. For example, the calculated Tad for the Winkler Air case
is 14l5°C as compared to a measured maximum temperature of 1402°Co In
addition, furnace wall temperatures of 1090°C and flue gas temperatures
of 1285 °C were reported in this case. It is noted that although pro-
94
ducer gas heating values are only a fraction of the heating values of
natural gas (15%-20%), adiabatic flame temperatures and measured tem-
peratures in a furnace are only 200-300°C lower.
95
5. LIST OF SYMBOLS
A equivalence ratio for combustion of producer gas
A80 total surface area of rice hulls within the control volume V (m2)
a 80 specific surface area of rice hulls (m2;m3)
cp specific heat of mixture at constant pressure (J kg-1 K-1)
ci molar concentration of species i
ci surface concentration of species i (g-mole/m2)
cpi(T) specific heat of species i at constant pressure (J kg-1 K-1)
Dzz effective diffusion coefficient
Hg total enthalpy of producer gas mixture (J kg-1)
Hi, 298 heat of formation of species i at 298 K (J kg-1)
Hi specific molar enthalpy of species i
total enthalpy of the products of gasification or combustion

(J kg-1)
total enthalpy of the products of gasification or combustion

(J kg-1)
molar diffussive flux of specie i (J m-2 s-1
equilibrium constant at pressure p and temperature T
surface reaction rate parameter
M moisture content of rice hulls (% weight, wet basis)
orientation of the interphase area
p pressure of gas mixture (N m-2
q heat flux vector due to conductive energy transport (J m-2 s- 1 )
R universal gas constant (J/g-mole K)
rate of formation or consumption of species i due to bulk phase

reactions (g-mole/s)
temperature of producer gas-air mixture (°C)
96
Tad adiabatic flame temperature (°C)
t stress tensor
vi specie velocity vector
v velocity vector of the bulk phase gas mixture (m s-1)
vi spatial deviation of specie velocity vector
V control volume (m3)
v8 void space volume (m3)
w velocity vector of the interphase area
8 coefficient of thermal expansion at constant pressure (K-1)
p density of gas mixture (kg m-3)
£ 8 void space fraction of fuel bed
<f> volume averaged parameter or physical constant f
~ energy source due to radiation (J m-3 s-1)
~ equivalence ratio for gasification of rice hulls
K linear adsorption isotherm
97
5. REFERENCES

Agricultural Services Bulletin #31. Rome.
2. Groeneveld, M.J. 1980. The co-current moving bed gasifier. PhD

Thesis, Technische Hogeschool Twente, Netherlands.
3. Carbonell, R.G. and S. Whitaker. 1982. Heat and mass transport

in porous media. Proceedings of NATO Advanced Study Institute of
the Mechanics of Fluids in Porous Media, University of Delaware.
July 18-27.
4. JANAF. 1971. Thermochemical Tables. 2nd Edition, U.S.

Department of Commerce, National Bureau of Standards.
5. Probstein, R.F. and R.E. Hicks. 1982. Synthetic Fuels. McGraw-

Hill. New York.
6. STANJAN Program. 1981. R. Reynolds Thermosciences Division,

Department of Mechanical Engineering, Stanford University,
Stanford, california 94305.
7. Shafizadeh, F. et al. 1976. Thermal Uses and Properties of

Carbohydrate and Lignines. Academic Press. New York.
8. Shaffstall, D.R. and R.T. Waibel. 1978. Lo-Btu gas combustion

research. Institute of Gas Technology, Chicago, Illinois.
9. Van den Aarsen, F.G. et al. 1982. Performance of a rice husk

fueled fluidized bed pilot plant gasifier. First International
Producer Gas Conference, Colombo, Sri Lanka. November.
10. Whitaker, s. 1982. Lecture notes: Reactor design. Department

of Chemical Engineering, University of California, Davis.
98
6. PHYSICAL PROPERTIES OF RICE BULLS
Densities
Introduction
As outlined previously, gasification is a complicated interaction of
a gas-solid phase system. Based on a literature study and a survey of
existing units, it is concluded that the physical properties of a fuel
are usually underestimated while the influence of the heating values and
ash content are overestimated with regard to the performance of a gas
producer.
In the author's opinion, fuel properties should be ranked as follows
with respect to their importance on the producer gas quality:
a) Physical shape of particle and fuel bed structure
b) Moisture content
c) Volatile matter
d) Ash content
e) Ash composition (natural and tramp ash)
f) Energy content.
The reasons for the above ranking will be outlined in subsequent
chapters. This chapter will examine the fuel bed structure of a rice
hull gas producer in detail.
Figure 6-1 shows a control volume V which may be viewed as the
entire fuel column of a gas producer or only a very small section of it.
An enlarged schematic sketch of a single rice hull and its various pha-
ses is given in Figure 6-2. Four phases are identified:
a) The S-phase having a volume v8 represents the void space within
99
Figure 6-1. Fue 1 bed structure o f rice hulls •
100
Figure 6-2. Apparent volume V of a single rice hull, and y, a phase.
a
101
the fuel bed if the rice hulls are treated as a nonporous
material.
b) The a-phase having a volume Va is the region the rice hulls
occupy within the control volume v.

c) The y-phase having a volume Vy is the total pore volume of the
rice hulls within V. The volume Vy includes both the closed
and open pores of the rice hulls.
d) The a-phase represents the true solid matter of the rice hulls.
The various fractions of each phase are defined as:
where i ~a, e. y, o.
It is noted that:
One can experimentally determine the following densities:
a) Bulk density:
b) Apparent density:
c) True density:
The above densities were measured for rice hulls and pelletized rice
hulls. The pellets were made by a ring-die pellet mill producing a 9.5
mm diameter pellet of length 20-22 mm. A series of over 130 measure-
menta over a period of 1 year were taken. All data given are on a dry
basis.
102
Bulk density Pb
The bulk density Pb· was determined in a plastic cylinder of known
volume V = 6914 cm3 and a height to diameter ratio of 3:1. This ratio
corresponds roughly to the height - diameter ratios of small gas produ-
cers. The following average bulk densities for pellets Pbp and rice
hulls Pbh were measured:

-3 range 0.55-0.60
0.58 g em,
Pbp
Pph 0.10 g cm-3, range 0.09-0.11
True density p t
The true density, as well as the apparent density, were measured
with a Beckman Air Comparison Pycnometer Model 930. True density
measurements have been made by Tucker (1944). Beagle (1978) and Houston
(1976) are assumed to have used Tucker's numbers in their publications.
However, the fact that a single rice hull when immersed in water even-
tually sinks is indisputable proof that the reported true density
Pt = O. 735 g cm- 3 is in error. Attempts were made to repeat Tucker's
tests with water and with toluene, which is a better wetting agent. The
underlying principle of the test is to fill the pores with water and
measure the volume change in the water bath. This method did not work,
however, because too much air was trapped in the pores. Applying long-
term vacuum to release the trapped air did not help. Thus, a more
sophisticated method, air pycnometry, was employed.
In an air pycnometer, a weighed sample of rice hulls is exposed to a
gauge pressure of 1 atm in a container of known volume. The air is
allowed to diffuse into the open pores and the volume displacement by
the true solid phase plus the closed pores is measured with a precision
103
piston displacement within an accuracy of 0.05 cm3 and better. The
accuracy of the instrument is frequently tested with various precision
steel balls of known volume. It is noted that the "real" true density
of rice hulls can not be measured due to the following complications:
a) The fuel possesses closed pores inaccessible to air.
b) Micropores open to the air require a long time (sometimes many
hours) to reach pressure equilibrium. Because the measurements
are based on the air volume comparison at 1 atm gauge pressure,
the errors introduced through prolonged pressurizing of the
instrument can be significant and result in erratic data.
c) There is always the possihility of surface reactions, depending
on the gas used for the measurement.
d) Any size reduction by milling destroys only an insignificant
number of pores because the finest available sieves separate
particles larger than 35 1Jm• This is a large fraction. of the
average thickness of a rice hull, which is 50 llm•
The measured true density, Pt• on a dry basis is therefore defined as:
( V _ x 1 x2 )
(6-1)
100 p
w
x1 sample weight (wet basis, gram)
x2 = moisture content in percent (wet basis, gram)
V measured volume (cm3)
Pw = density of water, 1 g cm-3.
The air was allowed to diffuse into the pellets and rice hulls for at
most 30 minutes and for at least 5 minutes before a measurement was
104
taken. Eighty-three measurements of the sample volume V, with the
Beckman Pycnometer using 6 different lots of pellets and hulls
resulted in:
Pt = 1.41 g cm- 3 , average dry basis; range 1.37 to 1.45 for rice hulls
Pt 1.61 g cm- 3 , average dry basis; range 1.58 to 1.63 for pellets.
The considerably higher true density for pellets is caused by com-
paction and by the destruction of a large number of closed and open
pores during the pelletizing process.
For comparison, 14 tests using 3 different samples of rice hulls
ground so that they passed through a 60 mesh screen (250 ll particles)
resulted in:
Pt = 1.60 g cm-3, average dry basis; range 1.59 to 1.61
A lower bound for the true density may be computed from an elemental
ultimate analysis of rice hulls. The underlying assumption is the con-
servation of mass and complete combustion of the fuel to co2 and H2 o.
Table 6-1 lists the mass fractions for· rice hulls on a dry basis of C,
0, H, N, Si, Fe, Mg, Ca, Zn, K, Na, S, and Po These constituents of
rice hulls account for 99.9% of the mass of rice hulls. Based on the
ultimate analysis and on the conservation of mass, the computed true
density of dry rice hulls is Pt = 1.56 g cm- 3 • Because the densit.ies of
H, 0, N within the rice hulls are not known it was assumed that H and 0
are represented by water and we have therefore:
105
where xi are the mass fractions from Table 6-1 and Pi the species den-
sities. The small amount of nitrogen was neglected in the eomputation.
It should be pointed out that true densities reported on a wet basis
are always lower and differ drastically due to the moisture in capillary
openings, which will give a distorted value of Pt•
Table 6-1. Ultimate Analysis of Rice Hulls and

True Densities of the Elements
Element Mass Fraction % True density, Pt*

Dry Basis g cm-3
c 41.44 1.8-2.26
0 37.32
Si 14.66 2.0-2.5
H 4.94
K 0.59 0.86
N 0.57
s 0.30 2.0-2.9
p 0.07 1.82-2.69
Ca 0.06 1.55
Na 0.035 0.97
Fe 0.006 7.86
Mg 0.003 1.74
Zn 0.006 7.14
*Data from Handbook of Chemistry and Physics, 62nd Edition, CRC Press.
106
Apparent Density Pa
The apparent volume, Va• is defined as shown in Figure 6-2. It
includes the solid matter as well as all open and closed pores. In
order to measure the apparent volume of a sample of rice hulls i t is
necessary to close all pores that terminate at the surface. This is
done by filling the pores with ~ax that is highly viscous at 125°C. Due
to the tedious procedure only two samples of 1000 rice hulls each were
tested. It is relatively easy to remove excess paraffin from a large
object such as pellets. One can therefore compute the apparent density
by:
(6-2)
where m mass of unwaxed sample
volume measured of waxed sample.
However, equation (6-2) may not be used with small objects such as rice
hulls. Electron microscope pictures showed that samples differed signif-
icantly with regard to the amount of visible eJicess paraffin on the
surface and in the pores. If one assumes that no excess paraffin
exists, the apparent density will be underestimated because a real
sample is more accurately characterized by Figure 6-4. The ideal case
is shown in Figure 6-3.
A more accurate determination of the apparent density may be
obtained from Pt and the paraffin density Pp as follows:
107
The measured volume is: V (6-3)
m
according to Figure 6-5.
The apparent volume is: Va = Z1 + Z2
Thus, the true volume is
apparent density on a dry basis is:
X
x 1 (1 - ~)
p 100
= _ _,_--:;=-
(6-4)
t vt
Consequently, since z2 + z3 (6-5)
we get:
(6-6)
where Vm = measured volume of waxed rice hulls, (cm3)

x 2p sample weight, waxed rice hulls, (gram)
x1 = sample weight, unwaxed wet basis, (gram)
x2 = moisture content in percent.
Pp density of paraffin 0.888 g cm-3.
It is noted that Pt is known with high accuracy. One can therefore com-
put a new Vm and x2p that agrees with the measured pt. The apparent
density Pa is then given as:
where Va Vm• A series of 34 measurements resulted in:
108
Rice hu II sur face
Figure 6-3. Ideal surface of waxed rice hulls.
Paraffin Rice hull surface
··············~ 1
Open pore
Figure 6-4. Actual surface of waxed rice hulls.
Paraffin
Figure 6-5. Geometry of waxed rice hull for computation of Pa·
109
-3 dry basis, range 1.14-1.22 for pellets
1.18 g em,
Pa
-3 dry basis, range 0.60-0.72 for rice hulls.
0.65 g em,
Pa
Phase fractions Ea• £B• Ey, £a
The phase fractions £ 1 can be computed as follows:
VB
v
a
v- VB and £
a
1 --=
v
1 - £B (6-7)
v pb
VB = v- va and £B = 1 -v-=
a
1
Pa
(6-8)
V = V - V - V and £ (6-9)
y B 0 y
v0 v- vy (6-10)
In practice, three porosities are important:
a) The total void space fraction £t = £B + £y (6-11)
b) The porosity of the fuel £f = 1 (6-12)
c) The void space fraction between the rice hulls
£ = 1 (6-13)
B
These porosities are given in Table 6-2 for rice hulls and rice hull
pellets.
llO
Table 6-2. Gas-Solid Phase Distribution in a Gas Producer
for Loose Hulls and Rice Hull Pellets
Pellets Hulls
True Density, Pt• g cm-3 1.60 1.42
Apparent Density, Pa• g em-3 1.18 0.65
Bulk Density, Pb• g cm-3 0.58 0.10
£f 26% 54%
£13 51% 85%
£t 64% 93%
Surface Area of Loose Rice Hulls and Rice Bull Pellets
Introduction
Since most gasification reactions are heterogeneous, the specific
surface area Aso (m 2 tm3 ) plays an important role in the process because
of its influence on the reactivity of the fuel bed. It is noted that
the apparent surface ABo refers to the surface area one would find if
the rice hulls were nonporous, as indicated by Va in Figure 6-2. This
surface area is measurable with an image analyzer and electron micro-
scope. The real solid-gas interface in a porous media is usually deter-
mined by the BET method (Brunauer-Emmett-Taylor). This surface area
includes all internal surfaces resulting from the porous structure of
rice hulls. It can be startlingly high, on the order of 106 m2 kg-1 for
activated charcoal from coconut shell, wood, and bituminous coal
(Mantell, 1968).
Determination of the BET surface is based on vapor adsorption and is
beset with difficulties. The theory is essentially an extension of the
Langmuir theory, which is based on two very tenuous postulates:
111
a) The adsorbed vapor covers the entire surface with a uniform
monolayer.
b) The amount of adsorbed vapor can be detected from the adsorp-
tion isotherm.
The assumption of a monolayer can hardly be justified and even extending
the theory to multilayer adsorption (BET surface) leaves some doubt as
to the accuracy of the method.
Any useful determination of the solid-gas interphase area at
various stages in the gasification process of rice hulls must certainly
be preceded by the development of an appropriate method of obtaining
reliable samples and of evaluating the maximum error for surface
measurements. There does not appear to be a convincing solution to
these problems at present.
Determination of Aso for rice bulls
The apparent surface area A~ 0 of a single rice hull has been experi-
mentally determined under the following assumptions and procedures:
a) Electron microscope pictures of cross sections were used to fit
a curve that will describe the regular appearance of the
geometrical bumps on the outer surface of the rice hulls. The
shape of the bumps, as shown in the electron microscope photo-
graphs (Figure 7-4 is an example) was measured at 6 points.
The shape was approximated by a sixth-degree polynomial,
P(x) (6-14)
112
The assumed cross section is shown in Figure 6-7 with the mean
measured dimensions.
b) The bumps were assumed to be rotationally symmetric around
their vertical Y axes. The surface area of each bump was com-
puted by
36
2
s f
0
2nx(l + [P'(x)] ) dx (6-15)
c) A composite picture, consisting of 6 sections at a magnifica-
tion of 25 was then used to establish the total enclosed plane
area of three single rice hulls. This circumference was then
traced by a computer-guided image analyzer that automatically
computed the enclosed area.
d) The number of bumps present on the outer surface was
established by counting them and checked by superimposing a
72 ~ grid over the composite picture.
It is noted that a "single" rice hull looks as shown in Figure 6-6.
The rice hull consists of two parts, each of which can be naturally
separated into two additional parts. The average computed apparent sur-
face area was Aeo = 3855 m2 /m 3 and is based on the computed cross sec-
tion shown in Figure 6-7 and Pbh previously given. It is noted that the
roughness of the outer surface (bumps) increases the total outer surface
area by about 12%.
113
Figure 6-6. Schematic of a single rice hull.
OUTER SURFACE
INNER SURFACE
Figure 6-7. Modelled cross section of rice hull.
114
Apparent Surface Ae 0 of Pellets
The surface area Aso of pellets has been experimentally determined
using the following assumptions and procedures:
a) The pellets were made with a ring die a pelletizer that
is normally used to manufacture alfalfa hay pellets for
livestock.
b) The particular die ring used generates pellets with a diameter
of 9.6 mm and an average length of 22 mm. However, during sub-
sequent handling a considerable number of pellets were broken
into pieces of various lengths. It was found that approxima-
tely 30% of the pellets remained unbroken and that 70% had
been broken.
c) The additional surface area generated by breaking a single
pellet into two pieces has been taken into account. The
average apparent surface area was A130 = 302 m2 /m3.
Weight of a Single Rice Bull
In order to obtain a complete description of the geometry of a rice
hull bed in a gas producer, deterillination of the average weight of a
single rice hull is necessary. The measurements were based on a selec-
tion of 50 single rice hulls that had both hull parts, as shown in
Figure 6-6. Weight measurements were performed with a Kahn-Electro
Balance with an accuracy of + 10-6 gram. The range for these rice
hulls was ms 2.944 to 3.564 mg (dry basis) with an average ms of
3.087 mg.
One of the difficulties in measuring the weight of a highly porous
small object such as a rice hull is its moisture content. The measuring
115
chamber of the electro-balance is lighted; thus, there is a heat source
which evaporates small amounts of equilibrium moisture within seconds.
The average decline in the rice hull weight during the measurement pro-
cess was 30 vg.
Accuracy of the Results
Because a wide range of measurement techniques (some of them novel)
have been used to characterize a rice hull fuel bed, it is appropriate
to perform an error analysis for the data reported in Chapter 6.
All data has been reported on a dry basis to allow comparison of
various samples and to eliminate the "moisture content" variable which
can considerably distort experimental data.
The error analysis is based on the following:
a) Random errors are not included in the analysis.
b) Instrument errors and data transformation errors are based on
manufacturer's data and on the type of instrument.

c) An upper bound for the maximal error has been computed in the
following manner.
For instance, the dry bulk density, Pb• is given by:
(6-16)
where x1 sample moisture content in % wet basis
x2 mass of wet sample, gram
x3 volume of sample conta1ner,-cm3.
Furthermore,
116
(6-17)
and consequently,
(6-18)
where is the largest possible error.
In this case,
< + + (6-19)
Assumptions:
~x 1 0.5% (error in the moisture content measurement)
~x 2 5 gram (error in wet sample)
Furthermore, assumed measurement errors of + 1 mm for the diameter and
height of the sample container with a total volume x3 = 6914.26 cm3
results in
= 0.016
Finally,
< 0.0055 + 0.0021 + 0.016 0.024 ~ 2.4%
Based on this technique maximum errors in the data have been established
and are listed in Table 6-3.
117
Table 6-3. Maximum Error Data for Rice Hulls and Rice Hull Pellets
Maximum Error %
ProEertl Pellets Hulls
Bulk Density Pb 2.4 2.8
Apparent Density Pa 4.2 5.6
True Density Pt 1.3 3.4
Porosity Ef 15.5 8.3
ea 6.4 1.4
Et 2.1 0.5
Weight of Single Rice Hull or Pellet 0.9 1.1
Surface Area * 3.8

*insufficient data
Although electron microscope evaluations and image analyses have an
error of less than 3%, only a few samples (3 in our case) can be eva-
luated because of the tedious procedures involved. This can hardly be
viewed as a representative sample from a population of 33 ,000/liter.
Some improvement in the accuracy may be achieved by correlating the sur-
face area to the weight of a single rice hull based on the model given
in Figure 6-7. From a measurement standpoint it is much easier to
obtain a reliable weight distribution from a large sample population of
rice hulls.
Caking and Slagging Behavior of Rice Bulls
Introduction
The present interest in utilizing a wide variety of biomass fuels
for gasification has resulted in an increased awareness of the phase
changes of solid fuels. However, inadequately presented results and
118
inappropriate descriptions of the slagging process have resulted in some
confusion regarding the process. It is therefore essential to clarify
the differences between caking and slagging (including the causes of
each) before the experimental results are presented.
Any solid material will change from a solid to a liquid or gas when
heated up to high enough temperatures. Coal ,and biomass fuels are no
exception and will melt down at high temperatures. For instance, one
can envisage what would happen if a sample of rice hulls was placed in
an inert atmosphere and heated. It is noted that rice hulls consist of
c, H, 0, N, SiOz, K2 o, NzO and some other trace minerals most likely in
their oxidized form. The first elements that will "melt off" and change
from the solid phase to the liquid and gas phases are H, 0, N and some
C in form of compounds. This process is called devolatllization or
pyrolysis and will be discussed in Chapter 8 in detail. Pyrolysis
results in the formation of free carbon radicals which recombine with
the freed atomic hydrogen to form a liquid (tar, oils and water) that is
easily evaporated at higher temperatures. In addition, a gas phase con-
sisting mainly of CO and COz is formed directly. At higher temperatures
some of the low temperature mineral oxides such as KzO and NazO will
reach a transition state. In this state the minerals soften before
becoming completely liqui fied. It is noted that despite the high tem-
peratures, no carbon will be burned off since the atmosphere is inert.
The final result is therefore a molten mass of minerals which entomb
solid carbon and high temperature mineral oxides such as SiOz, MgO and
CaO. The melting points of selected minerals and mineral oxides as well
as carbon are given in Tahle 6-4.
119
Table 6-4. ~elting Point of Selected Oxidized Minerals
Mineral Oxide Melting Point OC*
CaO 2614
MgO 2852
K20 350
Na 2 o 1275
Fe 2o 3 1560
p2o5 580
SiOz: Quartz (amorphous) 1460-1610
Trydimite 1703
Cristobalite 1723
C: amorph 3652
*Data taken from the Handbook of Chemistry and Physics, 62nd Edition,
CRC Press, 1982.
It is noted that the softening and melting points of biomass must be
determined experimentally because they are strongly dependent on the
mass fractions of K20 and Na 2 o.
One could also envision placing a sample of rice hulls into a
highly oxidizing atmosphere and adding heat. In this case all the car-
bon in the fuel will be quickly burned off. The residues from this pro-
cess (minerals) will melt down if the temperature reaches the softening
and melting temperatures of the particular sample.
A third phenomena is closely associated with the low temperature
pyrolysis process. It has already been mentioned that unstable free
radicals are formed in the 250-450°C range. Those which react with
120
hydrogen from the fuel form tar, whereas the rest tend to form bonds
with one another. In the case of rice hulls this latter process leads
to formation of large fluffy conglomerates which are easily breakable
yet which nevertheless obstruct gravity flow. This phenomena is called
caking of the fuels. It is widely observed in the gasification and com-
bustion of certain bituminous coals. Caking of coal was prevented by
stirring the bed or by preoxidizing the coal, which results in a product
known as coke. The notion of caking has so far not found a place in the
biomass gasification literature.
In summation: Three forms of biomass conglomerates are sometimes
found within the gas producer:
a) Conglomerates that have been generated in an inert atmosphere
within the gasifier. These rock-like structures are made up of
melted minerals and entrained carbon (slagging).
b) Conglomerates that have been generated in a highly oxidizing
atmosphere within the gasifier, resulting in the meltdown of
minerals (slagging).
c) Conglomerates that have been formed by free radicals which
react with one another at low temperatures (caking).
Depending on the type of biomass and the conditions under which it is
gasified, all three processes may occur. Also, some interaction does
occur between these processes.
Slagging of Rice Hull Ash
According to some authors, rice hull ash is highly prone to slagging
(Jenkins 1981, Goss 1980, Groeneveld 1982) due to its high silica con-
tent (16-22%) and alleged low mineral melting point (800-1000°C).
121
Based on this information, the author has tried unsuccessfully to build
a rice hull gas producer which induces slagglng of rice hulls. In the
course of these attempts several tests were conducted that showed that
rice hull ash has no tendency to slag below 1400°C if localized com-
bustion with excess oxygen does not occur. Osman (1982) has measured
the softening and melting point temperatures of various biomass ashes.
His results are listed in Table 6-5. The experiments were conducted in
a nitrogen atmosphere. The biomass ash was made into small cones and
the softening temperature was defined as the temperature at which the
peak of the cone toppled over and began to deform. It is noted that the
softening temperature of the biomass fuel will be the lowest temperature
at which the possibility of fuel flow obstruction exists.
From the table it is seen that the rice hull ash has the highest
softening and melting point of all tested biomass fuel residues. It is
noted that when present in large amounts low melting point temperature
mineral oxides such as NazO and KzO will significantly lower the melting
point of the ashes. This is shown in the phase diagrams taken from
Toropov (1972) for the NazO - Sio 2 and K2 o - Sio2 system (Figures 6-8
and 6-9). It is noted that SiOz is found in three forms in rice hull
ash: quartz, cristobalite and trydimite. Jones (1953) has already
pointed out that rice hull ash contains a negligible proportion of
quartz and a high proportion of cristoballte and trydimite. This was
later confirmed by Bartha (1974). Jones (1953) also reported that rice
hull ash, when mixed with 3-15% CaO (lime), showed no tendency to soften
nor was there a sign! ficant change in the volume of the ash at tem-
peratures of 1667°C.
122
•
18]
CRISTOBL IT E
+LIQUID
TRIDYMITE
+LIQUID
LIQUID
QUARTZ
~ +LIQUID
40 60 80 100
No2 0 WEIGHT % Sj02
Figure 6-8. Na 2o- Si0 2 system (Levin, 1964).
LIQUID
1700
u
0
w
1500 a:::
•
~
SOLID
CRISTOBAL ITE
SOLID
CRISTOBALI TE
TRIDY MITE
a
::::> [] SOLID
TRIDYMITE
1-
<f
a:::
1300 w
Cl..
~
w
1100 1-
900
100
K2 0 t MOLE % t Si 0 2
Figure 6-9. K20 - Si0 2 system (Levin, 1964).
123
Table 6-5. Softening and Melting Temperatures of Biomass Ashes*
Biomass Ash Softening oc Melting oc

Bean straw 900 1150
Safflower 770 1430
Rice hulls 1439 )1650
Cotton gin trash 1010 1380
Barley straw 925 1100
Corn stalks 820 1091
Rice straw 823 1190
Wood chips 1050 1190
Corn fodder 1010 1180
Almond shell 790 1440
Corn cobs 900 1020
Tree prunings 770 1550
Walnut shell 820 1225
Olive pits 850 1480
Almond shell 860 1350
Good rice straw 1060 1250
Partially decomposed rice straw 850 1280

*sample values are highly dependent on ash composition (fluxing effect).
Because low melting point minerals affect softening point temperatures,
it is of interest to look at the available data on rice hull ash com-
position collected by Beagle (1978), Houston (1976) and Kaupp, Goss
(1982) shown in Table 6-6. This data represent values determined as
long ago as 1910 and on four continents (North America, Europe, Asia,
Australia). Comparing the listed values for K20 and Na 2 o with the phase
diagrams in Figures 6-8 and 6-9 one can conclude that local variations
124
Table 6-6. Rice Hull Ash Composition
Specie % Weight Fraction
Si0 2 90-97
CaO 0.2-1.5
MgO 0.1-2
KzO 0.6-1.6
Na 2 o Traces-1.75
P2o5 0.3
803 0.1-1.13
Cl 0.15-0.4
Fe 2 o3 Traces-0.4
Mno 2 Traces
in the ash composition will not result in a softening temperature below
1400°C. A review of the experimental data and operational reports for
the large (60-100 hp) Italian rice hull gas producer (see Section 10-1)
did not reveal a single reference to slagging problems due to the
rotating grate and a low superficial velocity. Considering the
available evidence, it is concluded that rice hulls will not slag under
normal operation of a gas producer (temperatures during normal operation
will always be under 1400°C). Nevertheless, severe slagging problems
have been reported and there is a real possibility that slagging will
occur in a rice hull gas producer as will be outlined later. However,
the false conclusions put forth concerning the slagging potential of
rice hulls result from a misinterpretation of the true cause of
slagging, which has to do with the rice hull fuel bed structure and not
directly with the high ash content of rice hulls or its 'silica content.
125
Cause of Slagging
In Chapter 6 the fuel bed structure of rice hulls was examined in
detail. It was concluded that due to the high void space fraction
(Es ~ 0.85) and low bulk density (100 kg m-3), the homogeneous fuel bed
structure is easily disturbed. From coal gasification it is known that
channel formation within the coal bed, which can be caused by caking or
by a large proportion of fine particles, is very detrimental to the
gasification process. Channel formation usually leads to an erratic gas
composition and also leads to the formation of hot spots. Because the
size distribution of rice hulls is within a very narrow range (see
Figure 6-21, Section 6.7) channel formation due to the presence of fine
particles will not occur. However, caves and channels are readily
formed in a rice hull fuel bed when settling creates shear zones
(cracks) and the local superficial gas velocity exceeds the average
value. Experiments to determine the cause of slagging have been per-
formed in an updraft gas producer with a 30 em firebox diameter and with
a fixed flat grate (Figure 6-10). It is noted that gasification can not
be sustained in such a gas producer because there is no continuous ash
removal. The fire zone will therefore move upwards from the ignition
point at the grate, consuming layer after layer of unburned rice hulls
until the entire rice hull column has been charred. The combustion air
flow rate was set at 50 Nm 3 h- 1 , which represents a superficial velocity
of at least 0.19 m s-1 within the column. The actual superficial velo-
city within the column is higher since the gas output is greater than
the original air input by the amount of solid fuel converted to gaseous
products. l.fhen the fire zone reached the gas exit port, the combustion
126
GAS EXIT
4
(I)
w
...J
Q..
~ 3 RICE HULLS
0
u
0
~
a: 2
w
X
1-
FLAT GRATE
AIR-
Figure 6-10. Experimental gas producer for testing the slagging

behavior of rice hulls.
air was shut off and the fuel bed was examined layer by layer. A sche-
matic sketch of the fuel bed structure is shown in Figure 6-11. The
fuel bed developed characteristic caves that were glazed with a layer of
molten snow-white rice hull ash. A sample of this form of slagging (as
described in the introduction) is shown in Figure 6-12. The cause of
this slag formation is clearly localized complete combustion of the rice
hulls due to high localized airflow. The formation of a cave or channel
allows air to reach the carbon surface at an equivalence ratio of ~ > 1.

Because of the poor conductivity of the fuel bed and radiation within
the caves, localized temperatures remain high, which results in melting
of the rice hull ash. The installed thermocouples showed only normal
gasification temperature of 900-1200°C. Thus, one might be tempted to
127
RICE
[} HULL
COLUMN
MOLTEN
SILICA
COATING CAVE OR
CHANNEL
Figure 6-11. Schematic of slag formation in a rice hull fuel bed

(caves not to scale).
wrongly conclude that slagging took place at low temperatures since the
thermocouples were not placed where the slagging was actually occurring.
The same experiment was repeated with rice hull pellets. Although
rice hull pellets do not easl ty alLow channel formation, they do tend
to form hot spots, which in isolated cases leads to melting of the ash
which in turn leads to bonding between pellets as shown in Figure 6-13.
It has been concluded that slag formation is caused by high superficial
air velocities in the fuel bed which lead to localized meltdown
(glazing) of the rice hull ash. In order to determine the approximate
superficial velocity at which disturbances in the fuel bed can be
expected, experiments in a cold rice hull and rice hull char fuel bed
were conducted. The results are given in Section 6.6.
128
Figure 6-12. Molten silica and slag formation.
Figure 6-13. Localized s1agging, connecting pellets of rice hulls.
129
Caking of Rice Bulls and Pellets
As previously argued, caking of biomass is strictly a low tem-
perature phenomena caused by pyrolysis or thermal destruction of the
fuel. It is also a phenomena that is closely related to the size of the
gas producer, because caking can go unnoticed or can be prevented
(through the use of mechanical stirrers) in larger units. The caking
behavior of rice hulls is most easily demonstrated in an open core gas
producer which operates without air nozzles as shown in Figure 10-27,
Section 10-4. To induce caking, the gas producer is first fueled with a
good gasification fuel, such as corncobs, until a sufficiently deep hot
carbon layer develops. The gasification process is continued by the
addition of rice hulls through the open top of the gas producer. A
characteristic slight shrinkage of the rice hulls will take place
throughout the entire rice hull column. In addition, a very distinct
gap between the gas producer wall and the rice hull bed will develop.
This effect is shown in Figures 6-14 and 6-15 which shows a top view of
the rice hull fuel bed during thermal decomposition.
In order to make caking more visible, pellets can be used. A typi-
cal bond is shown in Figure 6-16 (black line). This figure should be
compared to Figure 6-13 which shows a bond caused by local over-
oxidation and meltdown of the ash.
Summary
Rice hull ash does have a very high softening temperature ()
1400°C), but nevertheless tends to slag if the fuel bed structure is
disturbed by high superficial gas velocities. Rice hulls do cake during
the pyrolysis process which may obstruct gravity flow. This property
130
along with the insignificant size reduction of rice hulls during gasifi-
cation are some principal disadvantages in rice hull gasification.
Figure 6-14. Initial stage Figure 6-15. Final stage

of rice hull of rice hull
gasification. gasification.
Figure 6-16. Caking of rice hull pellet at low

temperatures (300 - 500°C).
l3l
Pressure Drop in a Rice Bull Fuel Bed and Superficial Velocities
It was outlined in the previous section that disturbance of the fuel
bed will lead to slag formation and to erratic gas compositions. It is
therefore of interest to determine. the minimum gas velocity which will
cause disturbances in a rice hull fuel bed and rice hull char bed. Also
of importance is the pressure drop across such a system, because this
pressure drop must·be overcome by the intake manifold vacuum of a given
engine, if the gas is to be used in the engine without special pumping.
It is noted that the Italian rice hull gas producers were 1.5 m in
diameter to power 60-100 hp engines (see Section 10.1). It is assumed
that 3 Nm3 of producer gas are necessary to generate one brake-horse
power hour and that the gas will be at approximately 200°C when leaving
the upper layers of the rice hull bed. Under these assumptions, the
superficial gas velocity will be 7 .s em s -1 • If the tComputation is
based on the air entering the lower charred layers in an updraft gas
producer, then the resulting superficial gas velocity will be 10 em s -1 •
The latter computation is based on the following: (1) 2 kg of rice
hulls are needed in order to generate one brake-horse power hour; (2)
the equivalence ratio is cj> = 0.4; (3) air enters the .fuel bed at a tem-
perature of 600°Co It is concluded that ln the past, updraft rice hull
gas producers operated at very low superficial velocities, on the order
of 0.1 m s- 1 (throughout the fuel bed). Downdraft gas producers
operated on wood blocks have superficial gas velocities of 0.5-3 m s- 1
in the fuel bed, which are consider.::lbly higher (Nordstroem 1957, Kaupp
and Goss 1982).
132
Experimental Set Up and Results
The pressure drop across a rice hull and rice hull char fuel bed was
measured in a vertical tube as shown in Figure 6-17. The rice hulls
were supported by a fine (3 mm) wire mesh. The volumetric flow rate of
the incoming air was measured with a calibrated sharp edge orifice plate
( 4% accuracy). The pressure drop across the bed was measured with an
inclined manometer (+ 2.5 mm H20 reading accuracy). The wire mesh
induced a pressure drop of less than 2.5 mm HzO at all tested flow
rates. At the maximum flow rate the corresponding Reynolds number was
2200, which indicates the presence of a laminar flow field.
oE
f
I()U
-___.I_ PRESSURE TAP NO.2
WIRE MESH
SCREEN
MANOMETER
~
... ~
oe
-+ PRESSURE TAP NO. I
.,
rt')U
A I R -==:=~11:=::=~_j _ _._+_
ORFICE PLATE
Figure 6-17. Experimental set up for pressure drop testing.
133
Tests were conducted using 25 em, 50 em, 75 em and 100 em deep
unburned rice hull and rice hull char beds. The objective of the
experiment was three fold:
a) To determine the pressure drop across beds of various lengths;
b) To determine the superficial velocity which results in channel
formation, using visual inspection;
c) To compare the results for unburned rice hulls and rice hull
char.
The results of the experiments are given in Figure 6-18.
Discussion of the Results
a) Channel Formation
For the purposes of this experiment, one must distinguish bet-
ween the fluidizing velocity, vs, and the· velocity at which
channel formation takes place. The velocity at which single
unburned rice hulls or charred rice hulls were lifted up from
the bed is defined as the fluidizing velocity. Channel for-
mation does not necessarily start at this velocity. Based on
the observations, the fluidization of single rice hulls (not
broken rice hulls) was always accompanied by channel formation.
However, in the case of rice hull char, channel formation
started at a lower superficial gas velocity than the fluidiza-
tion of individual charred rice hulls. The following conclu-
sions were drawn:
i) Severe visible channel formation occurred at super-
ficial air velocities of 8.5-9 em s- 1 for charred rice
hulls. Channel formation was accompanied by a reduction
134
Fluidization
75 em, Char
a.
<3 7
0..
0
a:
0
I ~·
........-- __._ I I • I ... ~ ·-L~"'"
8
0 I 1 I I : J I j I I I I .,
0 I 2 3 4 5 6 7 8 9 I0 II 12 I 3 14 15 16 17 18 19 20 21
SUPERFICIAL GAS VELOCITY, em/sec
Figure 6-18. Pressure drop through an unburned rice hull bed (A)
and rice hull char bed as a function of the super-
ficial gas velocity vs, for various bed lengths .
......
w
V1
in the pressure drop despite an increased flow rate.
ii) The convex shape of the curves for rice hull char indi-
cates,that slight disturbances in the fuel bed must have
taken place at lower superficial gas velocities.
iii) The unburned rice hull bed was considerably more stable
than the char bed. Rice hull fluidization and channel
formation began at a vs of 20-23 em s-1 •
iv) Initial fluidization and channel formation never oc-
curred at the center or at the wall of the tube. In-
stead these phenomena always began at a short distance
away from the wall. This observation may suggest a
horizontal velocity profile within the bed as shown in
Figure 6-19.
Based on these results it is understandable why classic Italian
I
VI
E
>-
.......
u
0
_J
w
>
7.5 0 7.5
DISTANCE, em
Figure 6-19. Probable horizontal gas velocity profile within a

rice hull bed.
136
rice hull updraft gas producers were operated at such low
superficial gas velocities and why they were made so large as
compared to today's compact downdraft gas producers.
b) Pressure drop
The pressure drop (reported as a function of the superficial
velocity vs) was measured up to the fluidizing velocity. The
results are shown in Figure 6-18. The pressure drop ~p can be
viewed as the following function:
where
e6 void space fraction
p gas density, kg m-3
vs superficial gas velocity, m s-1
L length of rice hull column, m
y = surface roughness
DP particle diameter, m.
aL
Intutitively we should have ~ = constant, in the absence of
entrance and exit disturbances. The rice hull char bed
follows this law closely while the unburned rice hull bed data
indicates that large entrance and exit disturbances must be
present. Although the results were double checked, no obvious
reason could be found to explain why the difference in pressure
drops between a fuel bed with a height of 50 em and a fuel bed
with a height of 75 em was considerably smaller than the dif-
ference between a 25 em and 50 em fuel bed. The following was
concluded:
137
i) The pressure drop (em H2 0/m) is about 5.75 times higher
in a rice hull char bed than in an unburned rice hull
bed. This ratio is independent of the height of the
fuel bed.
ii) Pressure drops are automatically kept low because super-
ficial velocities above 10 em s- 1 will disturb the fuel
bed and interfere with the gasification process.
iii) The maximum pressure drop will be about 12 em H20/m.
Theoretical Treatment of the Pressure Drop in a Rice Bull Fuel Bed
The pressure drop through a fixed bed can be predicted accurately if
the input parameters are known with a certain precision. In general we
Ergun ( 19 54) has proposed an
equation to predict this pressure drop based on 4p (£ 8 , p, vs• L, DP).
He did not include the surface roughness.
150 (1 - EB)
Ergun equation: ----:::----=-
Re
+ 1. 7 5 (6-20)
D
where A = ---2-.. ~
p v 2
L
s
6 v
D =~
p A
p
D v
Re =-l!.E..~
ll
V volume of particle
p
A surface area of particle

p
138
A review by McDonald et al (1978) of correlations for pressure drops
through a fixed bed confirms the accuracy of the Ergun equation. The
authors suggested that the additional parameter of surface roughness of
the particles be incorporated into the equation. Their conclusions are
given as follows:
3
A EB 180 (1 - E )
B + 1.8 (smooth particles) (6-21)
1 - EB Re
3
A EB 180 (1 - EB)
1 - EB Re
+ 4.0 (roughest particles) (6-22)
Solving for the parameter 6p yields:
6p
-- [150 (1 - EB)
Re
l
+ 1.75J (1 - EB)
p Vs 2 L
(Ergun)
2
180 (1 - EB) ] p Vs L
6P = [ Re + 1. 8 ( 1 - E ) - - - (smooth particle)
B D 3
p EB
180 (1 - E ) ]
6p =[ Re B + 4.0 (1 - EB) (rough particle)
Because it is difficult to draw a line between smooth and rough par-
ticles, the Ergun equation will be applied to the experimental data for
a rice hull bed of 1 m height.
The Ergun equation indicates that 6p will be very sensitive to
changes in Dp and Es, the physical parameters that can be measured with
the least accuracy. We assume
(6)(4.75 mm 3 )
D 0.246 mm
p 2
(115.66 mm )
139
where Ap and Vp were computed from the data given in Sections 6.2 and
6.3. The other physical constants are given as:
p 1.204 kg m-3, 1.1
The best fit for the experimental data was obtained by choosing Dp =
2.28 x 10-4 instead of Dp = 2.46 x 10-4 with a fixed £ 8 = 0.85. The
result is given in Figure 6-20.
It is noted that the Ergun equation overestimates the pressure drop by
at most 15%. However this equation can not be used to predict pressure
drops at superficial velocities higher than 21 em s-1 because of distur-
bances in the fuel bed.
E
0'
N
:z:
E
u
a_ Ergun Equal ion
<]
a_
0
a:
0
w
a:
::>
(J)
(J)
w
a:
a_
0L--L--L--L--L--L--~-L--~-L--~~--~~--~~--~~--~~---L--L-~~
0 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21
SUPERFICIAL GAS VELOCITY, em /sec
Figur~ 6-20. Pressure drop across a rice hull fuel bed,

experimental and theoretical results.
140
Size Distribution of Rice Bulls and Rice Hull Char
Introduction
The high void space fraction in a rice hull bed, the large pore
volume generated during gasification, and the minimal size reduction of
rice hulls undergoing gasification make the residue of gasification
(the char) a potentially efficient filter material for the raw gas. On
the other hand, the friable consistency of rice hull char makes it dif-
ficult to handle without pulverizing a sizeable fraction. Pulverization
may even occur in the lower layers of the gas producer, caused by
interaction with the grate and by higher superficial velocities through
the grate and gas exit in downdraft gas producers. The carryover of
rice hull char has been a problem in the units tested. It is therefore
of interest to establish a size distribution curve and to compute the
extent to which a disengagement zone would separate fine solids (< 1 mm)
from the gas stream. While these theoretical considerations will be
discussed in Chapter 10, the size distribution of rice hulls and rice
hull char is given here.
Size Distribution
The size distribution was determined using a set of Tyler wire mesh
screens with square openings ranging from 0.05 mm to 2.5 mm. The
screens were mounted in a standard screen shaker for the size distribu-
tion tests. It is noted that a particle falling through the wire mesh
screen of size X during the shaking process was considered to be smaller
than X. However, the particle is not necessarily smaller than X in all
dimensions. Any rice hull remaining on the screen was considered to be
larger than X. Thus it is possible that a particular particle could
141
have fallen through given a different orientation of particle with the
screen.
The results for rice hulls and char (obtained from a downdraft
gasifier) are shown in Figures 6-21 and 6-22. Two types of char were
tested. In one case the char was tested within hours after gasification
while the second batch was tested after smoldering had ceased, 48 hours
later. It is noted that when char from a rice hull gas producer is
stored in a steel drum, char will smolder for extensive periods (days).
Smoldering of the gasifier residue completes burnout (carbon removal).
The silica skeleton formed is increasingly friable as the carbon content
is reduced and any test performed will show considerable size reduction.
With respect to the experiments, it is noted that:
a) Uncontrolled smoldering significantly increases the fine par-
ticle content of the char. For instance, the fraction of par-
><
z A Rice hulls as received from milling operation
<t
:X: B Hot char from downdraft gasification
1-
C Cold char after smoldering has ceased (48
hours later l
z lll!ID: Expected range of size

~ 02 distribution for rice hull char
1- as filter material
~0
a::
i.L
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3
WIRE MESH SCREEN OPENING ,X ( mm l
Figure 6-21. Size distribution of rice hulls and char removed from
downdraft gasification.
142
><
z
<(
::r: C = Cold char after smoldertng
~
has ceased (48 hours later).
a:: Continuation of curve C for
w
<.!>. the,;Small particle range.
a::
<(
...J
en
w
...J
u
...,:.
a::
<(
0..
LL.
0
z
0
~ .91
u
<(
a:: 10 30 50 70 90 110 130 150 170
LL.
WIRE MESH SCREEN OPENING, X (MICRON)
Figure 6-22. Fine particle content of rice hull char from downdraft
gasification.
ticles smaller than 500 ~m increases from 15% to 30%.
b) Particles smaller- than 600 ~m are mostly generated during the
gasification process since only 2% were present before gasifi-
cation as compared to 23%-36% afterwards.
c) The largest fractional increase is observed for the category of
particles smaller than 1.5 mm. Initially, 25% of the particles
fell into this category. This fraction increased to 82% after
gasification.
d) The fraction of particles smaller than 50 ~m is around 2%
(particles smaller than 50 ~m will be more difficult to remove
from the gas stream in conventional cyclones and wet
scrubbers).
143
6. LIST OF SYMBOLS
Aeo specific surface area of rice hulls (m2/m3)
Ap surface area of single rice hull (m2)
Dp rice hull diameter (m)
L length of rice hull column (m)
m mass of rice hulls (kg)
ms mass of single rice hull (mg)
v control volume (cm3)
apparent volume of rice hulls (cm3)
ve void space within the fuel bed (cm3)
pore volume of rice hulls (cm3)
volume of the solid matter of rice hulls (cm3)
measured volume (cm3)
volume of single rice hull (cm3)
Re Reynolds number
fluidizing velocity (m s-1)
X wire mesh screen opening (mm)
weight fractions of elements
y surface roughness
p gas density (kg m-3)
apparent density of rice hulls (g cm-3)
bulk density of rice hulls (g cm-3)
density of paraffin (g cm-3)
true density of rice hulls (g cm-3)

density of water (g cm-3)
phase fractions, where i =a, e. y, 0
144
£f porosity of rice hulls
Et total void space fraction of a rice hull fuel bed
~ equivalence ratio
6p pressure drop across a rice hull bed (em HzO/m)
~ gas viscosity (kg m-1 s-1)
145
6. REFERENCES
1. Bartha, P. and A. Huppertz. 1974. The structure of silica in

rice husks and their crystallization. Proceedings, Rice
By-Product Utilization International Conference, Valencia, Spain.
2. Beagle, E. 1978. Rice husk conversion to energy. FAO Agricul-

tural Services Bulletin #31.
3. Beagle, E. 1981. Personal Communication, Davis, California, USA.
4. Ergun, s. 1952. Fluid flow through packed columns. Chern. Eng.

Prog. 48(89).
5. Goss, J.R. 1980. Personal Communication, Davis, California, USA.
6. Groeneveld, M. 1982. Personal Communication, Berlin, West

Germany.
7. Houston, D.F. 1972. Rice chemistry and technology. Monograph

Series IV, Chapter 12. American Cereal Chemist Association, St.
Paul, MN.
8. Jenkins, B.M. 1980. Downdraft gasification characteristics of

major California residue-derived fuels. PhD Thesis in Engineer-
ing, University of California, ~avis.
9. Jones, J.D. 1953. New refractory from vegetable source.

Canadian Metals. January.
10. Kaupp, A. and J.R. Goss. 1982. State-of-the-Art for small scale
(to 50 kW) gas producer engine systems. U.S. Agency for
International Development, Washington, D.C. 20523.
11. Levin, Ernest "!. and Carl R. Robbins. 1964. Phase diagrams for
ceramists. The American Ceramist Society, Columbus, Ohio.
12. Mantell, C.L. 1968. Carbon and Graphite Handbook. Inter-

Science Publishers, New York.
13. McDonald, I.F. 1979. Flow through porous media - the Ergun
equation revisited. Ind. Eng. Chern. Fundam. 8(3).
14. Nordstroem, o. 1964. Redogerelse for riksnamndens for ekonomisk

forsvarsberedskap forsknings-och forsoksverksamhet pa gengasomra-
det vid statens maskinprovningar 1957-1963. Overstyrelsen for
Ekonomiskt Forsvar, Box 7830, S-10398, Stockholm, Sweden.
15. Osman, E.A. 1982. A study of the effects of ash chemical com-
position and additives on fusion temperature in relation to slag
146
formation during gasification of biomass. PhD Thesis in Engineer-
ing, University of California, Davis.
16. Toropov, N.A. et al. 1972. Handbook of phase diagrams of sili-

cate systems. Translated from Russian, Israel Program for
Scientific Translations, Jerusalem.
17. Tucker, P.W. 1944. A study of the physical and chemical proper-
ties of rice hulls. M.S. Thesis, Louisiana State University,
Department of Chemistry.
147
7. PHYSICAL APPEARANCE OF RICE BULLS UNDER THERMAL DECOMPOSITION
Introduction
During my first trial runs using rice hulls in standard wood gasi-
fiers of the updraft and downdraft type, it became apparent that the
quickly changing flow properties and the unique thermal decomposition
characteristics of rice hulls cause severe operational problems in small
gasifiers of 15 em to 30 em hearth diameter. For instance, common slot
grates used in coal and wood gasification can neither remove the rice
hull ash continuously nor prevent the rice hulls from falling through.
Another extreme example of the variable nature of rice hull flow prop-
erties is shown in Figure 7-1. An experimental hopper with 70° and 45°
walls was used to feed the rice hulls into a downdraft gas producer.
The rice hulls were loosely poured into the hopper, but because of
gravity feeding the rice hulls became compacted and formed a vertical
wall during the trial run. Additional difficulties were encountered in
updraft gasification. Tar vapors condensed in the cooler upper layers
and formed a sticky coating of charred rice hulls and tarry liquids, as
shown in Figure 7-2. The frequent flow distortlon and channel genera-
tion within the fuel bed could not he attributed to one particular
cause. The next step was to examine closely the physical appearance of
rice hulls and rice hull pellets before and after gasification to see if
any insights into fuel bed disturbance could be found.
Micrographs of Rice Bulls Before Thermal Decomposition
A schematic diagram of a single rice hull has already been shown in
Figure 6-6. A micrograph picture is shown in Figure 7-3. The geometri-
cally unique outer surface of rice hulls with its thousands (about
148
T Rice
t
70cm hulls
35
em
Figure 7-1. Vertical wall of rice hulls after one hour of operation.
Figure 7-2. Rice hull-tar conglomerate from updraft gasification.
149
Figure 7-3. Single rice hull before gasification (X 20).
Figure 7-4. Outer surface before gasification (X 180).
150
4000-5000) of bumps is shown in Figures 7-3 through 7-5. The inner sur-
face of a single rice hull, on the other hand, is almost completely flat
as shown in Figures 7-6 and 7-7. It is noted that even extensive
scanning at a magnification of 6000X did not reveal any visible pores on
the inner surface. Therefore pores on this surface must be smaller than
0.17 ~m given that the human eye can detect 1 mrn objects on the surface
of the rice hull. One of the cross-sections used to determine the sur-
face area of rice hulls (see Figure 6-7) is shown in Figure 7-8. A
single bump on the outer surface is shown in Figure 7-9. It is again
noted that rice hulls contain 15%-22% ash, most of which is Si02, as
outlined in Chapter 6. Silicate systems have a marked effect on the
stability of a rice hull undergoing thermal decomposition, as shown in
the following micrographs.
Micrographs of Rice Bulls After Thermal Decomposition
The remarkable stability of a rice hull that has undergone thermal
decomposition can be seen in Figure 7-10, which shows a single rice hull
after complete combustion at l200°C. All volatiles and carbon have been
released leaving a snow-white silica skeleton that has the same shape
and surface structure as the original rice hull. The slim appearance is
due to the intense heat, which tends to curl the si 1 ica skeleton. The
high temperatures have virtually no effect on the bumpy outer surface,
as can be seen in Figure 7-11. Although temperatures in a gas producer
are in the 900-1200°C range, snow-white residues shown in the figure are
not found because the process operates at ~ < 0.5.

A typical outer surface of a gasified rice hull is shown in Figure
7-12. The appearance of a single bump after gasification is shown in
151
Figure 7-5. Outer surface before gasification (X 550).
Figure 7-6. Inner surface be fore gasification (X 500).
152
Figure 7-7. Inner surface before gasification (X 2000).
Figure 7-8. Cross-section of single rice hull before gasification

(X 550).
153
Figure 7-9. Bump on outer surface before gasification (X 2000).
Figure 7-10. Rice hull after complete combustion at 1200°C (X 20).
154
Figure 7-11. Rice hull after complete combustion at 1200°C (X 550).
Figure 7-12. Outer surface after gasification (X 550).
155
Figure 7-13. It is noted that the initially rounded bumps on the outer
surface became sharp peaks after gasification. No pore formation was
observed at this magnification. A completely different situation pre-
v~ils along the inner surface of a rice hull, which is shown in Figure
7-14. From Figure 7-14 it can be concluded that the volatiles and car-
bon must be tightly woven into the silica skeleton of a rice hull. An
indication as ·to the density of this network of pores can be seen from
Figure 7-15.
Size Reduction of Pelletized Rice Bulls
Because of their stable silica skeleton, the size reduction of rice
hulls under thermal decomposition is minimal. This is shown for
pelleted rice hull~ in Figure 7-16. In this case the remaining carbon
in the pellets was completely burned off by long-term exposure (6 hours)
to medium temperatures (900°C) in an atmosphere of air. The average
measured spatial reductions are given in Table 7-1.
Table 7-1. Average Size Reduction of Rice Hull Pellets
Before Gasification After Combustion
Length (em) 1.85 1.54
Diameter (em) 0.95 0.77
Computed Volume (cm 3 ) 1.31 0.71
From Table 7-1 one may conclude that a 45% reduction in volume will
occur for a single dimension reduction of 17%-19%. It is noted that
depending on the gas producer and the ash removal design the char resi-
due is brittle and disintegrates. For instance, in a downdraft or
updraft gas producer the rice hull char will remain intact in most
156
Figure 7-13. Single bump after gasification (X 2000).
Figure 7-14. Inner surface of rice hull after gasification (X 550).
157
Figure 7-15. Inner surface silica skeleton of rice hull after
gasification (X 2000).
Figure 7-16. Size reduction of pelleted rice hulls under thermal de-
composition. (Left: before, Right: after gasification).
158
cases. On the other hand, the QSe of sand as a heat transfer media (as
in a fluidized bed gas producer) results in a fine dust-like residue.
Summary
It has been shown that rice hulls, because of their high ash content
(in the form of a stable silica skeleton), exhibit unusual behavior
during thermal decomposition. This behavior is not found in most other
biomass fuels and is characterized below:
a) Minimal size reduction.
b) Stable silica skeleton is friable.
c) Large differences in the number of pores t~at open up along the
inner and outer surfaces.
d) Highly porous char residue (see Figures 7-14 and 7-15).
159
8. LOW TEMPERATURE ENERGY CONVERSION OF RICE IIDLLS
Introduction
Combustion of a carbonaceous fuel includes a pyrolytic component.
Pyrolysis takes place when the fuel is heated to a temperature between
2.50-500°C. At these temperatures, the carbonaceous fuel will begin
releasing its so-called volatile components. The pyrolysis process can
be. characterized as the first stage of combustion. This stage takes
place at relatively low temperatures of 250-500°C. Pyrolysis is not
limited to solid fuels. A diffusion f.lame of producer gas from rice
hull gasification and a schematic sketch of the flame is shown in
Figures 8-1 and 8-2.
~-
GAS =====:::J.;:~£_ -----BURNED
MIXTURE GASES
COLORED VISIBLE FLAME

ENVELOP <
lmm
PREHEAT ZONE
Figure 8-1. Schematic of producer gas diffusion flame.
In this case, oxygen is diffusing towards the very thin flame envelope
from the outside. It will be consumed at the outer surface of the flame
envelope. Consequently the inner preheat zone is the pyrolysis zone.
In this zone low temperatures of 300-600°C are sufficient to induce the
generation of pyrolysis products such as solid carbon (soot) which is in
turn generated from the cracking of hydrocarbons in the gas. This pro-
cess takes place within a fraction of a second in an oxygen-free
atmosphere.
160
Figure 8-2. Orange diffusion flame from raw producer gas
(solid particles, tar).
Rice hulls fed into a fluidized bed gas producer will be heated up
within a fraction of a second to 500°C in an oxidizing atmosphere as
shown in Section 8.3. Thus, although this process is completely dif-
ferent from that mentioned above, it also forms a similar range of pro-
ducts.
In updraft rice hull gas producers, the rice hulls move through a
spatially fixed temperature gradient (see Figure 5-1). In both cases
they will undergo a slow pyrolysis (on the order of minutes) in an
oxygen-free atmosphere (< 2.5% oxygen). Therefore the following param-
eters characterize a pyrolysis process:
a) Temperature
b) Amount of time used to heat the reactants
c) Amount of oxygen present, which is used in a combustion mode to
provide the heat necessary to sustain pyrolysis.
Products of Pyrolysis
The process can generate liquid, solid and gaseous products most of
which are highly undesirable i f the gas is used as an engine fuel. For
161
instance, the separation of carbon and hydrogen in the pyrolysis zone of
a diffusion flame results in solid carbon (soot). The presence of this
solid carbon gives the flame a characteristic orange color.
The pyrolysis zone in a rice hull gas producer generates a con-
siderable amount of liquid products. The mixture is called tar which is
mainly higher hydrocarbons and water. A list of trace species that are
also found in tar is given in Table 1-1. An important issue in the
pyrolysis of coal or a biomass such as rice hulls is how to maximize
the desired product(s). In general, converting a biomass solid fuel to
a liquid or gaseous fuel is matter of adding hydrogen to the fuel. This
can be seen from Table 8-1 which lists the C-H composition of some
fuels.
Table 8-1. Fuels and their C-H Composition
llir.e hulls CHo.7 (solid)

Coal CHo.8 (solid)
Water-free tar CH1.5 (liquid)
Light oil CH1.8 (liquid)
Gasoline CH2.25 (volatile liquid)
Natural gas CH3.9 (gas)
Two major pyrolysis processes have been classified according to their
products. The following section discusses each of these often used
processes:
a) Charcoal production
This ancient method was mainly used to generate a smokeless
fuel with a high energy content. Thus the product of interest
is the solid phase. Although charcoal production is mostly
associated with wood, it can be done with most other biomass
162
fuels as well, including rice hulls. Although large commercial
charcoal production installations often use the gaseous phase
to generate the necessary heat to devolatilize the biomass, it
is difficult to find a use for the condensable phase. Heavy
and light oils can only be marketed as fuels or engine oils if
their viscosities and boiling point curves can be guaranteed
within a narrow range. In addition a variety of corrosive
acids are generated during pyrolysis and may harm burners or
engines. Nevertheless the Italian rice hull gas producers
operated prior to t.forld War II generated a sizeable amount of
heavy tar (asphalt) that was a very valuable byproduct at this
time (see Section 10.2).
Almost all charcoal production equipment burns part of the fuel
to generate the heat necessary for the pyrolysis process. As
shown in Figure S-4, the pyrolysis process is endothermic or at
best thermally neutral. It is noted that optimization of the
desired product, charred rice hulls, and minimization of the
smokiness of the fuel cannot both be achieved. This problem
becomes critical with a high ash fuel such as rice hulls. It
will be shown later that the removal (using heat) of all nitro-
gen, hydrogen and oxygen in the fuel (the smoke generators)
cannot be achieved without the removal of a considerable amount
of carbon. A smokeless rice hull charcoal is therefore charac-
terized by a high ash content. A large fraction of the initial
energy stored in the fuel is lost to the gaseous and liquid
phases, which are generally not utilized.
163
b) Tar production
There are a few situations in which the objective is to opt!-
mize tar production. In these cases the carbon ash residue
(char) and the gases are only byproduc ts. Based on a very
large amount of research done on the pyrolysis of coal, i t is
generally agreed that rapid heating increases the tar yield
significantly and that at high heating rates and particles
smaller than 50 ).lm, the particle size does not influence the
tar yield (J .B. Howard 1981). Based on their results, the
various rates of temperature increase during pyrolysis can be
defined as follows:
Rapid (flash) pyrolysis 103 - 105oc s-1
Intermediate pyrolysis 10 - 100°C s-1
Slow pyrolysis
Rapid pyrolysis which optimizes the tar yield is therefore well achieved
in a fluidized bed gas producer although tar production is certainly not
the objective of such a unit (see Section 8.3). In order to market the
liquid products, fractional distillation and removal of the acids in the
tar is necessary. Thus utilization of the liquid phase is only feasible
on a large scale.
In the gasification of rice hulls for engine fuel, the objective is
to have as little tar in the gas as possible. However, for all types of
gas producers one uses, tar will always be generated because pyrolysis
will always be occurring. To reduce the amount of the tar in the raw
gas three independent paths can be followed:
a) Very slow volatilization which may not decrease tar formation
164
as much as desired.
b) Tar cracking within the gas producer (Chapter 9).
c) Pyrolysis of the rice hulls prior to their use as a gasifier
fuel (Section 8.3).
Mechanism of Pyrolysis
The formation of volatile matter depends primarily on the hydrogen-
to-carbon ratio of the fuel. Equations S-10 and S-11 are used to pre-
dict the anhydrous tar and water fractions of the pyrolytic products of
coal for a 500°C assay. It is noted that the yields of the various pro-
ducts are very dependent on the type of laboratory assay used. The most
commonly used tests for coal in the u.s. are the Fischer and Gray King
assays. Both are carried out at up to S00-600°C or until the evolution
of volatiles ceases. The devolatilization of coal has been described by
Probstein (1982) as follows:
"During active decompostion and temperatures around
S00°C, i t is the weak aliphatic links of the coal that are
ruptured with subsequent release of the large molecular ring
free radicals. The tar and light oils are formed from the
stabilization of these free radical fragments by the hydrogen
evolved".
This quote may serve as a reminder that the description of the pyrolysis
process in terms of reversible kinetics or even in terms of the reac-
tions described in Equations S-1 through S-9 has resulted in false
conclusions concerning the overall gasification process.
No governing theory exists for the pryolysis process. In fact, the
uncertainty about the process .is so great that there is disagreement as
165
to whether the product yields change with heating rate (Anthony, Howard
1976 and Howard 1981). The main obstacle to a unified theory is the
difficulty in distinguishing between primary and secondary reactions
during the process.
The primary reactions are those involving the breakdown of the
hydrogen, oxygen, hydroxyl and nitrogen bonds in the fuel. Once this
stage is reached, gaseous and liquid products are generated that will
undergo homogeneous and heterogeneous reactions, which are secondary
reactions. These secondary reactions give a false impression as to the
importance of the highly irreversible primary reactions. It will be
shown in Chapter 9 that high temperatures in the gasifier alone can
catalyze the conversion of all tarry liquids into gases. Thus the pri-
mary pyrolysis products have therefore been converted into an
equilibrium mixture of gases and liquids whose composition is dependent
on temperature and pressure.
In order to accurately assess the gasification process, a clear
distinction must be made between the products of pyrolytic reactions and
the products of gasification reactions. To demonstrate this, the
methane generation is used as an example. It has been shown in Chapter
5 that methane as well as the gaseous higher hydrocarbons (Cz~· c3 ~
and c 4 ~) are present only in small amounts (< 1%) in the equilibrium
compositions. On the other hand, methane is found in large quantities
(3% - 12%) in the reported dry gas compositions from downdraft, updraft
and fluidized bed gas producers. The fact that these high methane con-
tents are primarily caused in the pyrolysis zone and that the methane
content remains relatively stable at high temperature (~ 1000°C) is
166
shown in Figure 8-3, reported by Goodman et al (1958). The interesting
feature of this figure is that the methane content increases with
increasing temperature up to 600°C and that it remains relatively stable
up to a temperature of 1000°Co These findings do not contradict the
well-known observation that the methane fraction generated through
reversible reactions such as 5-5 and 5-8 will decrease with temperature
and becomes negligible above 800°C (Chapter 5). Instead, Goodman' s

results indicate the difficulty in finding a model for the pyrolysis
process.
70~--------------------------------.350
........
........
"'
, "'Net yield
/------
'EGl /
...
(,)
Gl
/
/
Q. /
/
Gl
/
E
:I /
0
>
c
~
"iii
0
Q.
E
0
0
700 800 900 1000

Pyrolysis temperature, 'C
Figure 8-3. Composition of pyrolysis gas on a nitrogen and oxygen

free basis from a sub-bituminous B coal (Probstein, 1982).
167
In particular, tar formation and tar yield in a gas producer must be
explored further not just for operational reasons, but also because of
the significant health problems associated with the compounds that make
up tar. It is well known that polycylic hydrocarbons (PAH's) which are
products of pyrolysis have an exceptionally tumorigenic potential
(Harvey 1982). In particular, DMBA, shown in Figure 8-4 is one of the
most potent carcinogens known. Harvey reports that the problem was
identified as far back as 200 years ago by a London surgeon, Percival
Pott, who related the soot in chimneys to the high incidence of epithe-
lioma of the scrotum among chimney sweeps.
CH3
7,12- DIMETHYL BENZ [ol ANTHRACENE

(OMBA)
Figure 8-4. DMBA.
In order to obtain some data on the pyrolysis of rice hulls an,l on
the products of rice hull pyrolysis, laboratory experiments were per-
formed in a nitrogen atmosphere. Charring of rice hulls was also per-
formed using slight oxidation.
168
Pyrolysis Experiments in a Pure Nitrogen Atmosphere
The objective was to observe the formation of the solid, gaseous and
liquid products as a function of temperature. Therefore a constant heat

rate and an oxygen-free atmosphere were used for all tests. Because the
literature is full of many definitions especially concerning distinc-
tions between the phases, a set of definitions are presented:
Defintions
a) Volatiles: All gaseous and liquid products evolved up to tem-
perature T.
b) Tar: The anhydrous fraction of the liquid products.
c) Tarry liquids: The entire liquid phase.
d) Fuel carbon: The amount of carbon that was initially present
in the fuel as measured by the elemental chemical ultimate ana-
lysis (% weight, dry basis).
e) Char: The solid phase (residue) after pyrolysis.
f) Base carbon: The carbon left (fixed) in the char.
g) Volatile carbon: The carbon that has been evolved during
pyrolysis.
h) Ash: The mineral fraction of the fuel or char.
i) Weight fraction consumed: The fraction of each of C, H, 0, and
N in the liquid and gas phases after pyrolysis.
The heat-up period for rice hulls is an important parameter with
respect to tar yield and is computed in the following section.
169
Heat-up Period for a Single Rice Bull
The heat-up period for a single rice hull can be closely approxi-
mated by a nontransient solution of the energy equation. This assump-

tion is justified if one can show that the resistance to heat transfer
within the hull is negligible as compared to the resistance within the
fil m. Thi s means NBi = .b2.

2k << 1 • Under these assumptions the dimension-
<h>A
---- t
p c v
less normed temperature function e(t) e p
describes the heat-up process.
(T) average rice hull temperature
T1 temperature of the environment the rice hull is exposed to
T0 initial temperature of the rice hull
(h) surface averaged film heat transfer coefficient
p V = weight of single rice hull
A surface area of a rice hull
cp specific heat of a rice hull
b thickness of rice hull
k thermal conductivity.
One of the difficulties involved in this computation is in obtaining a
close estimate of the film heat transfer coefficient, (h), which appears
in the Blot number NBi and the temperature function. A review of
experimental heat transfer data which has been compiled by Whitaker
(1972) has been used to estimate (h). The particle size diameter is
given by:
170
6 V. ( 6 )( 4. 7 5) mm 3
D =___!?.= ............~~~~ 0.246 mm
p
Ap (115.66) mm 2
where the volume and surface area have been measured as outlined in
Chapter 6. The characteristic length L* is then given as
L* (0.246) ( 0 • 85 ) 1.394 mm
1-0.85
Depending on the required gas quality a 15 em diameter rice hull gas
producer can generate between 2 and 10 Nm3/h of producer gas. Therefore
-1
the real gas velocity u* is given by u* 0.185 m s
assuming Q 10 Nm3 /h
0.85
A 0.01767 m2
we have
· pu*L*
N =--=2.13
Re \.1
assuming p 0.33 kg m- 3
Correlations given by Whitaker (1972) do not cover this unusual combina-
tion of high void space fraction and low Reynolds number. One may
conclude from Figures 6 and 7 of Reference 7 that NNu <2 is justified,
and therefore
171
2k -1 -1 -1
h = L* = 249 J m s K
-1 -1 -1
k 0.1737 J s m K
<h>A
- pc
-- v. t - 4.4St
0(t) e P = e
Cp = 2094 J kg-1 oc-l (average for wood, not available for rice
hulls).
p V. = 3.087 X 10-6 kg.
(249)(85 X 10-6 )
The Biot Number is then given by NBi 0.061.
(2)(0.1737)
This means the assumption of a nontransient solution of the energy
equation is justified.
Example: A single rice hull exposed to a hot glowing carbon bed at
1000°C will heat up to 990° from an initial temperature of 20°C within
0.5 seconds. It is noted that the uncertainty lies with (h). Varying
the Nusselt number from 1 - 10 changes the calculated heat-up period
from 1 to 0.1 seconds.
It is therefore concluded that fluidized bed gasification induces a
rapid pyrolysis process while the following experiments simulate a slow
pyrolysis process at heating rates of 0.3 - 0.4°C s- 1 • The later heat
rates are common in large-scale updraft and downdraft gas producers.
Experimental Set Up and Procedures
The experiments were conducted in a tube furnace connected to a tar
and water sampling train as shown in Figure 8-5. Tar and water were
172
#
Thermocouple I
Plastic Filler Cartridge

#
Steel Piston Thermocouple 2
Nitrogen
Figure 8-5. Rice hull pyrolysis apparatus.
filtered and separated by using a temperature resistant glass wool
arrangement. Condensation and separation of tarry liquids usually
results in inacurracies because separation is not complete.
Thermocouple #1 indicated the final temperature of the experiment while
thermocouple #2 (in conjunction with a heating collar) was used to keep
the cooler end of the tar trap at 105°C to avoid condensation of water
before it reached the ice bath. The pyrolysis experimental procedure
was:
a) Loading of the tube with 4-5 gram of oven-dry rice hulls.
b) Purging of the entire system with nitrogen.
c) Generation of a vacuum (25 em H20) to draw vapors and gases
from the sample into the filter train and gas bags.
d) Heat-up rate of 0.35°C s- 1 until the sample reached the desired
temperature.
e) Rapid cooling of the tube with water after the desired tem-
perature was reached.
f) Immediate sealing of all liquid and solid products in zip bags
to avoid evaporation.
173
In some of the low temperature experiments (250-300°C) small amounts
of light oils were carried over into the water collection system.
Discussion of the Experimental Results
The following data were collected:
a) Composition of the gas phase at 250, 300, 350, 400 and 450°C.
b) Ultimate elemental analysis of the rice hulls and of the solid
residue of pyrolysis at 250, 300, 350, 400 and 4)0°C.
c) Weight fractions of the products of pyrolysis (char, gas, tar
and water).
d) Weight fractions consumed of the elements C, H, N, 0 during
pyrolysis.
e) Higher heating value of rice hulls and the char at 250, 350,
400 anii 450°C.
f) Cumulative mass loss during pyrolysis.
Composition of the Gas Phase
The composition of the gas phase as a function of temperature is
shown in Figure 8-6. Gas samples were tested for o2 , CO, co 2 , CH4 ,
c2 H4 , c2 H6 and H2 • Hydrogen and c2 ~ were only found in very small
amounts (< 1%) in the gas. The low hydrogen fraction indicates that
there is' not enough hydrogen present during low temperature pyrolysis to
stabilize the radicals and prevent caking of the rice hulls.
As expected, the methane content was between 1% and 4% and did not
increase with decreasing temperature. The predominant gaseous products
were co 2 and CO. No significant change in the CO content could be
detected while a slight decrease in C02 occurred with increasing tem-
perature.
174
_.
0~
~ 60 • • • ••
z
0
• C02
I-
u
• ••
<X
0:: I •• • co
LL.. 30 H2 ,C2H4,C2H6 <I%
w
_J 20
0
~ 10
CH 4
250 300 350 400 450
TEMPERATURE, °C
Figure 8-6. 'Dry gas composition as a function of temperature.
Ultimate Elemental Analysis of Rice Hulls and Char as a Function of
Temperature
Charring of rice hulls is a wasteful process with regard to the fuel
energy lost. Its only purpose is to generate a more reactive and a less
smoky fuel. As outlined in Chapter 6, in the special case of the gasi-
fication of rice hulls in small gas producers, it may also prevent the
caking of rice hulls during thermal decomposition at low temperatures.
The object is therefore to remove as many of the tar forming elements
(H, 0, and N) as possible while preserving most of the carbon in the
fuel. Release of volatiles and the retention of carbon cannot be
simultaneously optimized. This is shown in Figure 8-7. Because the
175
0
w 0
::E
:::> H
CJ)
z
0
u
z 0.6 N
._
0
u
<t
a:: c
1.1.. 0.4
._
:I: 0.3
(.!)
25 250 300 350 400 450

TEMPERATURE,°C
Figure 8-7. Weight fraction of C, H, N, 0 consumed as a function of
temperature.
tests were conducted in a nitrogen atmosphere the release of volatiles
becomes insignificant above 450°C. However, at this temperature only
87% of the oxygen, 82% of the hydrogen, and 64% of the nitrogen in the
fuel were released along with 45% of the carbon. It is noted that the
release of carbon between 300 and 450°C is insignificant, whereas large
fractions of 0, H and N are released in this temperature interval. If a
smokeless fuel is desired i t is advisable to char the rice hulls at a
higher temperature, especially since the carbon loss at 450°C is not
176
much greater than the carbon loss at 300°C. These results may not be
used to predict the results of large-scale rice hull charring, because
partial burning of a layer of rice hulls is allowed.
The outcome of any pyrolysis process in which some oxygen is present
is heavily dependent on the length of time it takes to complete the pro-
cess. An example is that of gasified black rice hull char at 200 -
300°C which, when exposed to ambient air, turns gray and completely
changes its size distribution (Section 6.7) within 24 hours.
Weight Fractions of Char, Gas, Tar and Water
The results are shown in Figure 8-8. It is noted that no signifi-
cant change is observed in the temperature interval of 350°C - 450°C,
thus the bulk of the volatile matter is released below 350°C. Because
of the heating rate of 0.3 - 0.4 °C s- 1 , the process is classified as
slow pyrolysis which yielded 45% char and 37% tarry liquids which were
separated i'l}tO 10% tar and 27% water. The balance was gases mainly CO
and co 2 • It is noted that in an ideal downdraft gas producer, all of
the tar would be converted into gases while a large fraction of the
water would not be decomposed, because the water-gas reaction is far
from equilibrium.
177
In an updraft gas producer, which usually generates a gas with a 0%
- 2% oxygen content (unreacted o2 in the air), devolatilization takes
place in an almost inert atmosphere. Figure 8-8 indicates why updraft
gas producers generate excessive amounts of tarry liquids. These
liquids necessitate extensive cleaning of the gas.
Model I introduced in Chapter 5 cannot be used to predict the yields
of pyrolytic processes as outlined. The product distribution based on
Model I is shown in Figure 8-9. It is noted the gas phase fraction is
overestimated while the liquid phase fraction is underestimated. The
model therefore fails to take into account the generation of tar which
is not readily converted into gases.
Energy Balance
It is recalled that pyrolysis or charring of a biomass fuel has
three main objectives:
a) Production of a less smoky, clean burning fuel that does not
produce tar.
b) Production of a fuel with a higher heating value than the or-
iginal feedstock it was made from.
c) Production of a more reactive fuel.
On the other hand the generation of charred biomass has the
following drawbacks:
a) A considerable fraction of the energy in the initial feedstock
is lost in the gas and liquid products.
b) The charred product is more delicate to handle and easily
breaks up, such as soft wood charcoal or rice hull char.
178
en
1-
u
::>
~ i- 60
a:: :X:
Q.. ~50
cnw Char
c:n :;: 4o
>-~
..Jo H20+Tar
0
a:: H20
>-
Q..
Gas
Tar
250 300 350 400 450

TEMPERATURE, °C
Figure 8-8. Weight fraction of the pyrolysis products as a function of

temperature.
z
0
1-
v
cr
«
IL
en 0.3
en
cr 0.2
2
0.1
300 500 700

TEMPERATURE, °C
Figure 8-9. Weight fractions of the products from Model I at ~ 0.
179
To show the striking difference between wood charcoal production and
rice hull char production, the experimental results are compared to
results from pyrolysis of pine sawdust and bark at 400°C in Table 8-2.
From Table 8-2 it is obvious that rice hull pyrolysis does not generate
the desired products. The charring process does not result in a product
with a much higher energy content.
Table 8-2. Ultimate Analysis of Pine Sawdust + Bark,

Charcoal and Rice Hulls and Char at 400°C
Specie Wood* Charcoal Rice Hulls Char

% weight, dry basis
Carbon 52.3 75.3 39.0 43
Hydrogen 5.8 3.8 4.9 2.2

Oxygen 38.8 15.2 33.9 ll.2
Nitrogen 0.2 0.8 0.5 0.3
Sulfur o.o o.o < 0.01 < 0.01

Minerals 1.4 3.4 21.6 43.3
HHV MJ kg- 1 19.7 27.1 15.3 16.3
*Data from Tillman (1978)
It is noted that the reason for the very unfavorable results is the ini-
tial high ash content of the feedstock. This yields a product of
very little economical value. In this context it is of interest to look
at the higher heating value of the ch'lrred rice hulls as a function of
pyrolysis temperature as shown in Figure 8-10. From Figure 8-10 it can
be seen that only charring at very low temperature of 250° will result
in a reasonable energy loss of 15%. However, low temperature charring
(< 250°C) does drive off an insignificant amount of volatiles and makes
180
t-
(/)
0
------- _,., " 0.5 z

0
...J
-
16.5 0.4 t-
HHV
..
(.)
I
<t
.a::
Ot
16 0.3 a::
....,
~ 15.5 0.2 >
(!)
> a::
:I: 15 0.1 UJ
:I: z
UJ
300 350 400
TEMPERATURE, oc
Figure 8-10. Higher heating value of rice hulls (A) and char and
energy lost (B) during the pyrolysis process.
0
H
25 300 350 400 450
TEMPERATURE, °C
Figure 8-11. Weight fractions of ash, carbon and volatiles (H, 0, N) in
unburned rice hulls and rice hull char, as a function of
temperature.
181
the process impractical. This is shown in Figure 8-11. For instance,
slight charring at 250°C results in a product of 40% carbon and 60%
volatiles and ash, whlle charring at 450°C will yield a product of 45%
carbon, 45% ash and 10% volatiles.
Summary
The slow pyrolysis of rice hulls has been investigated at tem-
peratures of 350°C to 450°C. The primary results are the following:
a) An average yield of 10% dry tar, 27% water, 18% gas and 45%
char.
b) The charred rice hulls have a slightly higher heating value of
16 MJ kg-l compared to 15.3 MJ kg-l for rice hulls.
c) The energy fraction lost due to charring amounts to 45-55% on a
dry basis.
d) The rice hull char consists of approximately 45% ash, 45% car-
bon and 10% remaining volatiles.
182
8. REFERENCES
1. Anthony, D.B. and J.B. Howard. 1976. Coal devolatilization and

hydrogasification. American Institute of Chemical Engineers
Journal 22.
2. Goodman, J.B. et al. 1953. Low temperature carbonization assay

of coal in a precision laboratory apparatus. U.S. Bureau of Mines
Bulletin 530, Government Printing Office, Washington, D.C.
3. Harvey, R.G. 1982. Polycyclic hydrocarbons and cancer. American

Scientist. July-August.
4. Howard, J.B. 1981. Fundamentals of Coal Pyrolysis and

Hydropyrolysis in Chemistry of Coal Utilization. Second
Supplementary Volume. Wiley and Sons. New York.
5. Probstein, R.F. 1982. Synthetic Fuels. McGraw-Hill Book

Company. New York.
6. Tillman, D.A. 1978. Wood as an Energy Resource. Academic Press.

New York.
7. Whitaker, s. 1972. Forced convection heat transfer correlations

for flow in pipes, past flat plates, single cylinders, single
spheres, and for flow in packed beds and tube bundles. AICHE
Journal 18(2).
183
9. TAR CRACKING IN A RICE BULL AND RICE BULL PELLET FUEL BED
Introduction
The renewed interest in small gas producer-engine systems (5-100 hp)
has mostly involved downdraft (co-current) gas producers because of
their ability to convert condensable pyrolytic products into nonconden-
sable combustible gases. The evolution of gas producers for use with
internal combustion engines has taken place as shown in Figures 9-1
through 9-3. The trend toward downdraft designs is due to the dif-
ficulties that occur in cleaning the highly contaminated (tar and water)
gas generated in an updraft gas producer. However, the myth that
downdraft gas producers generate a "tar-free gas" has never been proven
and the fact is that all known operational downdraft and updraft gas
producer-engine systems are equipped with extensive gas cleaning trains
for tarry liquids and solid contaminants. The notion of a tar-free gas
has been persistently associated with downdraft gas producers in scien-
tific literature. It is therefore of interest to:
a) Analyse the reasons why downdraft gas producers could poten-
tially generate a tar-free gas.
b) Determine why this process is difficult to control.
Tar Conversion in a Downdraft Gas Producer
Measurement Techniques
It was stated in Chapter 8 that the tarry vapors (which may be a
mixture of higher hydrocarbons and water in mist form) have a wide range
of boiling points, ranging from 100°C - 400°C. Large quantities of fine
char particles in the exit stream (700 mg/Nm3. - 6000 mg/Nm3) act as
184
R AIR
R ' f AIR
-F~~i
FUEL AIR ' {lAIR
FUEL
c=i c::::a c::::l c=l
-GAS -GAS
Figure 9-1 Figure 9-2 Figure 9-3

Evolution of gas producers for use with internal combustion engines.
I
8000 I
I
I
\
\
,., \
E \
\
'e
Cl \
\
..... \
z \
w \
..... \
z \
0
(.) '\ \
a::
<X
.....
' ' ' .....
..... ......................
,_
-- ....... _
Rice husk gasification
Beech wood gasification
100 700 800 900

REACTOR TEMPERATURE, °C
Figure 9-4. Tar content for beech wood and rice hull producer gas
(Van den Aarsen, 1982).
185
nuclesation sites for tarry vapors. The first stage in processing the
raw gas was usually to cool it at the gas producer exit in an ice bath.
Thus, the resulting condensate is a tarry liquid that has already lost
part of its high temperature fraction and contains solid carbon. This
product is called the tar content of the raw gas and is usually measured
in mg Nm- 3 of dry gas. A typical result is given by Van den Aarsen
(1982) in Figure 9-4 for beech wood and rice hulls. These experiments
have been performed in a 30 em diameter fluidized bed gas producer. The
reported data indicate that the tar content of the raw gas is strongly
dependent on the temperature.
Groeneveld (1980), in a study of a downdraft gas producer with a
center tuyere, points out that "tar-free" means a tar content of less
than 200 mg Nm- 3 of dry gas. Based on the stringent requirements con-
cerning maximum contamination in producer gas that is to be used in
engines (less than 10 mg of total contaminants per Nm3 of producer gas),
it seems inappropriate to speak of "tar-free" gas in relation to a
downdraft gas producer. However, use of the term "tar-free" is
justified for producer gas obtained from charred fuels. In an ideally
carbonized fuel, all the volatiles have been driven off. Only carbon
and minerals remain as residue. As shown in Chapter 8 a highly car-
bonized fuel cannot be economically produced from rice hulls.
Mechanism of Tar Conversion
In general, there are three ways in tihich large hydrocarbon mol-
ecules can be broken into smaller ones in a downdraft gas producer.
a) Thermal decomposition, i.e. applying heat to the tar products
in an oxygen-free atmosphere.
186
b) Thermochemical decomposition, i.e. applying heat to the tar in
the presence of a catalytic surface. In practice, this
involves utilization of the hot carbon-mineral surfaces
available in a gas producer.
c) Combustion of tar with an oxygen carrier such as air.
All three mechanisms will contribute to tar conversion as follows:
a) It will be shown later that high temperatures provide enough
energy to split higher hydrocarbons in the absence of any
carbon-mineral surface.
b) In the presence of a carbon-mineral surface the tar cracking
process consumes carbon. This reaction may be represented as:
C+ C H ~ Products (9-1)
n m
c) Tar, like any other carbonaceous substance, can be burned in
air.
Design Criteria for Tar Cracking in Past Downdraft Gas Producers
It is noted that tar and water vapors are generated in a temperature
zone of 100°C - 450°C above the air injection level in a downdraft gas
producer. It is therefore extremely important (with regard to tar
cracking) to generate a "hot" cross section below the pyrolysis zone so
.that tar and water vapors formed in the pyrolysis zone will pass through
this hot layer and be partially converted. The first step towards more
efficient tar cracking was therefore to incorporate a throat into the
design of a downdraft gas producer. This design change is shown in
Figures 9-2 and 9-3. Gumz (1950) presented a schematic of the approxi-
mate "hot" zone distribution in a downdraft gas producer (Figure 9-5)
187
with wall tuyeres. In this context, hot temperature zone means that
temperatures are at least 1000°C. These are zones in which one can
expect sufficient tar cracking in a rice hull fuel bed. Figure 9-5
points out the drawbacks of this commonly used design.
a) Cold regions are generated between the tuyeres (air nozzles).
b) .With this design a homogeneous hot cross section is not
generated below the pyrolysis zone.
It was therefore necessary to introduce the throat as a design feature
in downdraft gas producers (see Figure 9-3). Under steady-state con-
ditions, one can expect an ideal vertical temperature distribution as
shown in Figure 9-6. It is noted that the only purpose of the throat is
to facilitate tar cracking. This design feature improves neither the
energy content of the gas nor the kinetics of the process. On the
contrary, the throat is a physical hinderance to the downward flow of
the fuel. The fuel bed must be reduced in size so that they can pass
through the narrow passage. The flow velocity, however, remains
constant. This is the very reason why downdraft gas producers with
throats cannot be used for rice hull gasification. The minimal size
reduction and stable configuration of a rice hull fuel bed during ther-
mal decomposition leads to jamming and blockage at the throat.
While writing the State-of-the-Art for small-scale gas producer
engine systems for USAID the author analyzed many designs (Kaupp, Goss
1982) and came to the following conclusions:
a) It was recognized that oxygen penetration into the fuel bed is
very limited. It is of the order of em or 5 particle diameters
from the point of ignition (Groeneveld 1980, Hedden 1982).
188
Figure 9-5. Hot temperature zone of a downdraft gas producer with
wall tuyeres (Gumz, 1950).
400 800 1200

TEMPERATURE, oc
Figure 9-6. Approximate position of vertical temperature profile

and throat for optimal tar conversion.
189
b) Groeneveld, while testing a center tuyere model shown in Figure
9-7, found a drastic ~ncrease in tar generation when the
distance between the air nozzle and throat was not "right".
c) Many wall and center nozzle designs have been tried (including
combinations) in order to combat the very limited oxygen
penetration without much success.
d) Oxygen penetration and dispersion cannot be arbitrarily
increased by increasing the air jet velocity (Groeneveld 1980).
e) The hot glowing carbon zone is mainly generated by radiation
and not by combustion of carbon. It is therefore nearly
impossible to scale up downdraft gas producers to large hearth
diameter without losing the tar cracking .potential of the unit
because of cold spots.
f) The wall and middle tuyere downdraft gas producers with throats
and uninsulated metal walls were a typical product of the auto-
motive gas producer industry during the 1930's to 1950's, and
were operated on charcoal. These units are not necessarily

efficient in tar cracking.
g) Reliable tar conversion is dependent on the maintenance of the
optimal vertical temperature profile shown in Figure 9-6.
h) It is difficult to maintain the optimum location of the "hot"
fire zone as shown in Figure 9-8a (Beagle, 1980). Pownward
creeping fire zones, as shown in Figures 9-8b,c, are not uncom-
mon. It is noted that the fire will always move up, indepen-
dent of the design (updraft, downdraft). Downward creeping
fire zones are therefore caused by uncontrollable ash removal
systems.
190
Figure 9-7. Downdraft center nozzle gas producer
(Groeneveld, 1980).
AIR_l l_AIR
a) b) c)
Figure 9-8. Downward creeping fire zone.
i) The danger of a fluctuating vertical temperature profile has
led to designs that call for a very short distance between
grate and throat. This will improve tar cracking at the
expense of the reduction and shift reactions and will lead to
high grate and raw gas temperatures. No reliable data is
available as to the optimal distance between throat and grate.
In order to test the hypothesis that only high temperatures and a flat
191
horizontal temperature profile across an entire cross section are needed
to ensure sufficient tar cracking, special experiments were performed.
Experimental Set Up
In order to obtain an objective picture of tar cracking in a rice
hull fuel bed, 38 tests were performed in the experimental set up shown
in Figure 9-9. Twenty of these tests were used to draw conclusions. In
order to simulate the processes taking place in a downdraft gas pro-
ducer, a temperature profile similar to that observed in a downdraft gas
producer was generated inside the tube.
The tube diameter was 27 mm and must be viewed as a compromise bet-
ween generating a constant cross-sectional temperature profile and being
!lble to use rice hull char and charred pellets as the fixed bed. In
order to test the influence of the fuel bed structure, carbon surface
and temperature on tar cracking, it was necessary to exclude the com-
bustion of tar. A constant nitrogen flow of 0.3-0.7 liters/min was used
to carry the tar-cracking products into a filter train and to exclude
air.
Cases Tested
Let f be the conversion efficiency of a fuel bed where
weight of recovered tar

f
we i g h t o f injected tar then f = f(Ea,
v
Aao•
v
T). (See Table 9-4).
The following cases have been tested:
a) Ee ~ 75% - 85%, A130 ~ 3000 - 4000 m2 m- 3 (charred rice hulls as
a fuel bed).
b) Ee ~ 40% - 50%, A130 ~ 300 - 400 m2/m-3 (charred rice hull
pellets as fuel bed).
192
NITROGEN
1----FIL TER TRAIN~ fsA;P~SINGI

I· FURNACE WITH ·I
COMBUSTION TUBE
Figure 9-9. Experimental set up for tar cracking tests.
1-'
1.0
w
c) c8 = 100%, A80 = 0 (empty tube)
d) Maximum temperatures in test tube:
cases a) through c).
Test Material
The same type of tar was used for all tests. This tar was origi-
nally generated during the updraft gasification of rice hulls and was
kept frozen until the tests were performed. Its chemical composition is
given in Table 9-1.
Table 9-1. Chemical Composition of Tar Used in Experiments
Species % weight, wet basis
Carbon C 67.86
Hydrogen H 9.31
Nitrogen N 0.85
Ash 0.20
This wet sample contained 6.18% water. It is noted that this tar is not
representative of tar generated in updraft gasification. The com-
position of the tar, particularly the water content, is strongly depen-
dent on the sampling location and on the temperature in the upper
layers. For instance, another sampling technique was used during a dif-
ferent run which resulted in a tar composition as shown icy Table 9-2.
Table 9-2 Chemical Composition of Tar
Carbon 17.16
Hydrogen 9.40
Nitrogen 0.44
Ash 0.70
194
The water content of this sample was 81.3%. The first sample was pre-
ferred for these tests because of its low water content and high visco-
sity.
Charred Fuel Bed
The rice hulls and rice hull pellets used as the fuel bed were
precharred at 300°C and 400°C respectively in a furnace for a period of
30 minutes. An ultimate elemental analysis for the pellets charred at
300°C) is given in Table 9-3.
Table 9-3. Pellets Ultimate Elemental Analysis, 400°Co
Carbon 45.63
Hydrogen 2.51
Nitrogen 0.62
Oxygen ll.24
Residue 4ci.oo
The pellets had a moisture content of 2.32%.
It has already been shown in Chapter 8 that the charring of rice
hulls does not result in a tar-free fuel because some hydrogen and oxy-
gen will remain in the fuel. This is again demonstrated in Table 9-3.
To avoid the collection of this remaining fuel tar, special procedures
were followed.
Filter Train
The wide range of tar boiling points and the remaining fuel tar make
it necessary to perform the tests with a filter train that can be
attached during the experiment and that can be exposed to a temperature
of 400°C without losing its tar collection ability.
195
Based on previous experiments, a lightweight aluminum tube filled
0
with a combination of fiberglass and 30 A molecular sieve pellets, as
shown in Figure 9-9, was both convenient and accurate.
Tar Injection
The tar was injected with an 18 gauge 3-inch spinal needle and
disposable syringe. Because of the highly variable bulk density of
loose rice hulls and pellets a wide range of tar to fuel ratios of 1:1.5
to 1:10 were used. It is noted that 1 kg of dry rice hulls forms about
380 grams of liquids or 100 grams of dry tar. The tar to fuel ratio in
a gasifier is therefore within the tested range (see Figure 8-8). The
tar was injected at a rate of 0.6 - 1 g/min.
Test Steps
a) The combustion tube was filled with about 9 gram of rice hulls
or 45 gram pellets (see Table 9-4).
b) A constant nitrogen flow was ~stablished and the sample was
heated.
c) When the smoke from the righthand side of the combustion tube
ceased (usually at 500°C) the prepared filter train was
attached. Special precautions were taken to collect tar vapors
d) Tar injection was initiated.
e) The nitrogen carrier gas and all gaseous products were cooled
down to 7 - 12°C in an ice bath which surrounded the molecular
sieve and fiberglass wool.
f) A material balance was performed. The balance used had an
accuracy of one tenth milligram.
196
g) Samples of the gas were taken (downstream from the filter tube)
during injection and analysed for COz, CO, H2 , CH4, c2 H2 , c2 H4
Results and Discussion
Data for 20 selected tar cracking tests are given in Table 9-4. The
conversion efficiency is shown graphically in Figure 9-10 as a function
of temperature T with £a as parameter.
~ • Pellets
• Hulls
~
(.)
• Empty Tube
z
1&.1
(.)
'"-
'"-
1&.1
z
0
en
a::
1&.1
>
z
0
(.)
680 800 920
MAXIMUM TEMPERATURE IN TUBE, •c
Figure 9-10. Tar conversion efficiency, experimental results.
The longitudinal temperature profiles in the tube at the three maxi-
mum temperatures of 680°C, 800°C and 920°C are shown in Figure 9-11.
The odd numbers in Figure 9-11 refer to the temperature profiles in the
packed tube, while even numbers refer to the profiles in the empty tube.
Figure 9-10 confirms the stated hypothesis along with an additional
twist.
a) Efficient tar cracking cannot •be expected to occur at temper-
atures below 900°Co It should be noted that high temperatures
197
1-'
\0 Table 9-4. Data for 20 Selected Tar Cracking Tests
(X)
Tar Conversion
Test Temp. Fuel Welght1 Spent 2 Injected Recovered Efficiency
# CC) Type (Gram) (Gram)
* (Gram) (Gram) (%)
1 920 Pellets 44.98 6.91 T 8.14 1.22 86.2

2 920 Pellets 27.82 4.62 T 6.85 0.97 85.9
3 920 Pellets 34.12 8.33 w 8.04 4.80 40.2
4 no Pellets 30.03 6.68 w 8.12 6.36 21.7
5 920 Hulls 9.73 2.10 T 10.08 2.03 79.9
6 920 Hulls 8.96 2.86 T 6.42 1.12 82.5
7 920 Accuracy test with water: 5.1% Loss
8 920 None 0 0 T 7.94 2.65 67-7
9 920 None 0 0 T 4.85 1.65 66.0
10 800 None 0 0 T 7.15 3.93 45.0
11 800 None 0 0 T 6.85 3.84 43.9
12 800 Pellets 31.95 6.20 T 7.96 2.57 67.7
13 800 Pellets 23.51 7.43 T 9.00 3.14 65.1
14 800 Hulls 8.01 1~51 T 5.92 2.01 66.0
15 800 Hulls 8.95 1.03 T 7.06 2.27 67.8
16 680 None 0 0 T 7.02 6.20 11.7
17 680 None 0 0 T 5.26 4.37 16.9
18 680 Pellets 29.61 6.63 T 3.04 1.50 50.1
19 680 Hulls 29.49 N.D. T 6.07 3.13 48.4
20 680 Corncobs 9.99 1.43 T 6.27 2.68 57.2
*T - Tar W = ii'ater
1 rnltial weight of the charred rice hull or rice hull pellet bed.
2Amount of carbon and remaining volatiles driven off during heat up period and tar cracking.
must be maintained over the whole cross section of the gas pro-
ducer.
b) The availability of a carbon surface for tar cracking becomes
crucial at lower temperatures but does not play a significant
role at temperatures above 900°C.
c) The conversion efficiency seems to be independent of £e•

However, it is noted that rice hulls have a much higher surface
are~ per unit volume than pellets.
d) In the temperature range tested (680°C - 920°C) the conversion
efficiency is a linear function of temperature for both the
packed bed and empty tube cases.
Water Dissociation in a Bot Rice Hull Char Bed
Introduction
The moisture content of rice hulls has a marked effect on the gas
quality as shown in theoretical Model I outlined in Chapter v. However,
the effect of moisture content on gas quality is somewhat controversial.
Jenkins (1980) in his PhD thesis argues that moisture contents of less
than 20% (by weight) are necessary for fuels which are to be used in
downdraft gas producers. Various other sources (Ekman 1972; Hose 1982)
report that rellable operation and good gas quality were achieved in a
downdraft gas producer using fuels with moisture contents of 30% - 40%.
It is noted that the term fuel moisture refers to the moisture
trapped in the pores and on the surface of the fuel. This term does not
include the chemically bound water which forms during the pyrolysis
state. It has been shown in Chapter 8 that on the average 23-27 grams
of water are released from 100 grams of rice hulls with a O% moisture
199
0:
1000
u.
.2
:E'
....0
.
u
....0:
....:::>
<l 600
....a..
0:
2
....
....
400
300
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
DISTANCE , em
Figure 9-11. Longitudinal temperature profiles in tube.
content during thermal decomposition at temperatures of 350°C - 450°C.
A qualitative assessment of the conversion of the moisture in the fuel

during the gasification process results in the following conclusions:
a) The bulk of the hydrogen in the producer gas mixture comes from
the conversion of water to CO and Hz with carbon acting as a

catalyst. This process may be represented by
C + HzO ~ CO + Hz + 131 MJ/kg-mole

It will be shown in the following section that considerable amounts of
COz and small amounts of methane are also formed. Very little hydrogen
(< 1%) is formed from atomic hydrogen in the fue 1, as discussed in
Chapter 8. It is therefore beneficial to have a high fuel moisture con-
tent in order to increase hydrogen formation. In particular, pre-
charred biomass fuels yield a gas which is low in hydrogen when
200
gasified. Another characteristic· of the gasification of dry fuels is
the decreasing hydrogen content of the gas when a batch fed gas producer
is used. These phenomena are shown in Tables 9-5 and 9-6. In Table 9-5
the gas compos! t ion from an automotive charcoal gas producer is given,
while in Table 9-6 the declining hydrogen content of the gas with
operating time is illustrated.
Table 9-5. Producer Gas Composition of a Wood Charcoal Automotive

Gas Producer (Anonymous, 1942)
SEecies % volume
co 31.9
COz 1.4
CHi. 0.6
Hz 4.1
Nz 62.0
Table 9-6. Gas Composition of an Automotive Crossdraft Gas Producer.

Change of Gas Composition during use (Anonymous, 1937).
% by volume
After a journey of co
10 kilometers 24.9 2.1
54 kilometers 2.8 26.4 11.8 2.3 56.5
198 kilometers 2.7 27.8 7.4 1.8 60.2
218 kilometers 2.4 29.1 7.1 0.8 60.8
246 kilometers 3.1 30.5 4.1 o.s 61.8
250 kilometers 2.7 29.9 3.2 0.3 63.8
b) Although the fuel moisture is the main source of hydrogen pro-
duction, it has a cooling effect on the gasification process
and will also lower the efficiency of the gas producer. It is
201
noted that the water gas reaction is highly endothermic. In
addition, all fuel moisture must be heated to temperatures
above 700°C before significant conversion takes place, as will
be shown in the following section. In a downdraft gas pro-
ducer, where all moisture must pass through the hot oxidized
zone (see Figure 5-l), large amounts of water may still remain
unconverted. If, however, a high proportion is converted,
gasification cannot be sustained due to cooling down of the
fuel bed (see Table 9-6).
Example
The case of an airblown, downdraft gas producer is used in which
only the following three reactions are assumed to go to completion:
C + o2 + 3.76 Nz t co2 + 3.76 N2 - 394 MJ/kg-mole (9-2)
co2 + C ~ 2 CO + 173 MJ/kg-mole (9-3)
C + H2 o 1: CO + H2 + 131 MJ/kg-mole (9-4)
Based on the elemental ultimate analysis of rice hulls given in Table
8-2, we can represent 30.7 kg of dry rice hulls as CH 1 • 5 o0 • 65 NO.Ol
(SiOz)o.tz• According to the experimental results in Chapter 8, about
27%.of the initial weight of a sample of dry rice hulls is released as
water during gasification. It is assumed that the rice hulls have a
moisture content of 20% on a wet basis. Thus 38.375 kg of rice hulls
generate 15.965 kg of water (which includes the chemically bound
water). Two cases are considered: (i) 50% conversion of steam and
(ii) 100% conversion according to reaction 9-4. The results are given
in Table 9-7.
202
Table 9-7. Energy and Mass Ralance
Steam conversion efficiency
50% 100%
Carbon consumed for dissociation

of steam (% mole) 44.3 88.7
Energy consumed MJ 58.0 116.0
Additional carbon consumed in a

thermally neutral reactor (% mole) 26.2 52.5
Carbon balance (% mole) +29.4 -41.1
The two cases show that while 50% steam aissociation is theoretically
possible without combusting all the carbon, 100% steam dissociation is
not possible. In addition, the example shows that the real issue in the
gasification of wet biomass fuels is not their high moisture contents
but is the degree to which steam conversion in the hot carbon fuel bed
takes place. It is noted that it takes 4.3 MJ of energy to heat and
evaporate 1 kg of water at 20°C to a final temperature of 1000°C. An
additional 7.3 MJ is required to dissociate the steam to CO and Hz by
react ion 9-4. The question of steam dissociation in a rice hull char
bed at various temperatures is treated in the following paragraphs.
Experimental Set up and Procedure
A slightly modified version of the experimental set up shown in
Figure 9-9 was used. The modification involved the placement of
heating bands at both ends and a thermocouple in the center of the tube
ln order to record the maximum temperature during steam injection. The
test steps were:
203
a) The combustion tube was filled with char which had undergone
low temperature combustion at about 600°C (see Table 9-7).
b) A constant nitrogen flow was maintained while the char was
being heated to the appropr.i.ate temperature.
c) When the desired maximum temperature in the tube was achieved
and the temperature profile (as shown in Figure 9-11) was sta-
bilized, a known amount of water (~ 5 grams) was injected.
d) The unconverted water was collected in a fiberglass wool-
molecular sieve filter which was situated in an ice bath. The
average filter temperature was 6°C - 8°C.
e) Gas samples were taken during the 700°C and 900°C test runs.
f) Steam dissociation at four temperatures (700, 800, 900 and
1000°C) was recorded.
To assure that carbon depletion· did not take place during a test run, 15
grams of rice hull char were used as the fuel bed for each 5 grams of
water injected. The elemental ultimate chemical analysis of the char is
given in Table 9-8.
Table 9-8. Elemental Ultimate Chemical Analysis of Rice Hull Char

Used for Dissociation of Steam.
seecies % wei~ht 1 drx: basis
c 24.56
0 3.37
H 0.21
N 0.26
Minerals 71.60
204
It is noted that the dissociation of 1 kg of water requires 0.67 kg of
carbon according to equation 9-4. The 3.7 gram of carbon available in
15 grams of rice hull char as given in Table 9-8 are· therefore suf-
ficient and did not lead to carbon depletion during testing.
Results
While the dissociation 9f tar is mainly a function of temperature,
as shown in the previous section, the dissociation of steam cannot be
achieved by high temperature alone and requires a carbon surface. It
can therefore be expect~d that steam dissociation will be a strong func-
tion of the fuel bed structure and carbon content of the char. No
attempts have been made to examine the steam dissociation as a function
of carbon availability in the fuel bed. Instead, typical char from rice
hull gasification (Table 8-2) was used as a fuel bed in the test tube
and the same procedures for filling the tube were applied to all
reported tests. Temperature was the only parameter which was varied
Elemental ultimate chemical analysis of the
char was not performed after the dissociation took place. Since steam
dissociation occurred along a vertical temperature profile, typical of
that found in a downdraft gas producer, it does not make sense to per-
form a carbon balance because of locally different carbon consumption
rates. The re~ults of the tests are shown in Figure 9-12. The squares
indicate tests in which 0.5 grams of water were injected every minute,
while the round dots refer to tests in which 0.2 grams of water were
injected every 20 seconds. The resulting differences in steam disso-
elation between the two modes of injection were insignificant.
It is concluded that at low temperatures (700°C) almost no disso-
205
ciation takes place (< 6%) and that large amounts of steam will be part
of the raw gas leaving the gas producer. At temperatures of 1000°C, a
considerable fraction (roughly 19%) of the water still remains uncon-
verted. These experimental results are of course in complete
disagreement with the results obtained from computer simulated
equilibrium computations such as the one shown in Figure 9-13, where
only H2 and H2 0 are specified as products.
Q
w
1-
<[
u
0
(f)
(f)
0.7
Q
~
0.6
0
N
I 0.5
u..
0 0.4
z
0 0.3
1-
u 0.2
<[
a::
u.. 0. I
700 800 900 I000

TEMPERATURE, oc
Figure 9-12. Steam conversion in a rice hull char bed.
Squares: water injection at 0.5 g every minute.

Dots: water injection at 0.2 g every 20 seconds.
206
~
10 0
/
v-
.,w 10
/
~
su ~0
0
10
v
.,.,0 40 /
0
...
0
2 / H2
0
400 eoo aoo 1000
TEMPERATURE (°C)"
Figure 9-13. Dissociation of steam in the presence of glowing carbon

( Skov, 19 74) •
The dependence of the equilibrium calculations on the set of products
used in the computations is shown in Figures 9-14 to 9-16. The com-
putations have been executed with the JANAF program (see Chapter 5).
Three sets of products were assumed:
a) C, Hz, HzO, co (Figure 9-14)
b) c, Hz, HzO, co, COz (Figure 9-15)
c) c, Hz' HzO, co, COz, CH4 (Figure 9-16)
It is noted that the steam dissociation process does generate a con-
siderable amount of COz and CH4 as shown in Table 9-9.
207
z 0.9
0
......
u 0.8
<t
a::
LL 0.7
(/)
(/)
<t
0.6
~
a 0.5
LL.J
......
a:: 0.4
LL.J
> 0.3
z
0
u
z 0.2
~
0.1
400 600 800 1000

TEMPERATURE , oc
Figure 9-14. Steam dissociation with C, Hz, HzO, CO, COz as products.
Table 9-9. Gaseous Products of Dissociation of Steam at

700°C and 920°C as \nalysed by Gas Chromatograph
% volume
Species 700°C 920°C
co 27.1 34.5
COz 39.3 4.6
Hz 28.1 59.7
Cllf. 4.1 1.2
CnHm 2.7 Traces
From Table 9-9 it can be seen that only option c) is a realistic set of
products for the steam dissociation rt!actlons. However, since the reac-
tion is far from completion, the model does overestimate the converted
208
0.6
0.5
0.4
0.3
0.2
0.1
400 600 800 1000

TEMPERATURE, o C
Figure 9-15. Steam dissociation with C, Hz, HzO, CO as products.
c
oz
wo
1-t-
o:::u
wc::r
>O:::
z~
0(/)
U(f)
Zc::r
:::> ~ 0. 2
0.1
400 600 800 I 000

TEMPERATURE, °C
Figure 9-16. Steam dissociation with C, Hz, HzO, CO, COz,
CH 4 as products.
Z09
water fraction (see Figure 9-12 and 9-16). Based on these results, the
water dissociation efficiency will probably be around 50% at tem-
peratures between 900 and 950°C. The computed positive carbon balance
of +29.4 (mole %) from Table 9-4 is therefore a higher bound and is not
really available if one takes the following into consideration:
a) The carbon which is tied up in the condensable portion of the
pyrolysis product is only accounted for if complete tar
cracking takes place.
b) The gasification process does not burn off all the carbon. A
fraction of the carbon remains in the char residue.
c) A fraction of the carbon is needed to make up for the energy
lost by convection and radiation, the energy used to heat the
fuel and air, and the energy consumed in tar cracking
and pyrolysis reactions which are endothermic.
Summary
It has been argued that standard downdraft gas producers with
throats will not generate a "tar-free" gas. These designs are not
suitable for rice hull gasification. In addition, efficient tar conver-
sion can only be expected at temperatures above 1000°C in a· rice hull
fuel bed. Tar conversion is a strong function of temperature and the
high conversion efficiencies (> 95%) will not be achieved through sur-
face reactions or catalytic effects involving the minerals in rice
hulls. In addition, steam conversion does not contribute significantly
to the Hz content of the gas below temperatures of 900°C. It is there-
fore advisable to design a rice hull gas producer in such a way that
reactor temperatures of 1000°C are achieved.
210
9. LIST OF SYMBOLS
A Angstroem, 10-9 m
Ae 0 apparent surface area of rice hulls or rice hull pellets
T maximum temperature of horizontal temperature profile
e6 void space fraction of fuel bed
f conversion efficlency of tar cracking
211
9. REFERENCES
1. Anonymous. 1942. Gas Producers, Automobile Engineer, ~ovember ~.

1942. PP• 433-464.
2. Anonymous. 1937. Producer gas for road vehicles.

163(4248):682-884.
3. Ekman, E. 1972. Peat as a fuel in Finland. United ~ations,

Economic and Social Council, Seminar on Technologies for the
Utilization of Low-calorie Fuels. Varna Bulgaria. ~prll.
4. Groeneveld, M.J. 1980. The co-current moving bed gasifier. PhD

Thesis, Technische Hogeschool Twente, Netherlands.
5. Gumz, w. 1950. Gas Producer and Blast Furnaces. John Wiley and
Sons. New York.
6. Hedden, K. 1982. Personal communication. Colombo, Sri Lanka.

November.
7. Hose, F. 1982. Personal communication. Davis, California.

August.
8. Kaupp, A. and J.R. Goss. 1981. State-of-the-Art for small-scale

(to 50 kW) gas producer-engine systems. u.s. Agency for
International Development, Washington, D.C.
Publisher, Inc. Olympia, Washington.
10. Whitaker, s. 1972. Forced convection heat transfer correlations

for flow in pipes, past flat plates, single cylinders, single
spheres and for flow in packed beds and tube bundles. AICHE
Journal 18(2).
212
10. DESIGN CONSIDERATIONS FOR A RICE HULL GAS PRODUCER
Introduction
It was mentioned in Chapter 4 that small rice hull gas producers
have never been used commercially. In this context small refers to a
unit with a furnace diameter smaller than 50 em. Commercial installa-
tions like the Italian Balestra unit or the Chinese unit (which is
attached to the rice mill in Suzhou in Jiangsu Province) have furnace
diameters of 1.5 - 2.5 m and mechanical power outputs of 60 - 186 hpo
It is not possible to apply the large amount of knowledge acquired
during the t-lorld War II experience with automotive gas producers to the
design of a rice hull gas producer. The differences between the feed
stocks and the types of applications are striking and are pointed out
here again. A typical World War II downdraft gas producer had the
following operational characteristics:
a) Good fuels: charcoal, hardwood, coke, and anthracite were
used.
b) A gas output of 200 Nm3 h-l was not uncommon in 40 em diameter
furnaces. This means that up to a 100 hp engine could be

driven with this unit.
c) World tolar II units operatlng on wood, charcoal, coke, or
anthracite did not have char removal problems. The residue
generated by the gasification process amounted to o~ly 1% - 5%
(by weight) of the feed input. This small amount of residue
could either be shaken down through a grate (as shown in
Figures 10-1 through 10-3) or simply left at the bottom of the
gas producer and removed periodically (2-5 hours). The latter
213
Two piece frame
Lever arm
Figure 10-1. Flat slot grate Figure 10-2. Shaker grate

(Rambush, 1923). (Skov, 1975).
Partial combustion zone
Distillation zone
·Reduction zone
Gas
Figure 10-3. Shaker grate Figure 10-4. Crossdraft gas

(Schlapfer, 1937). producer with no
grate
(Harter-Seberick,
1937).
option was the most common procedure for residue removal in
crossdraft gas producers, as shown in Figure 10-4.
d) The high bulk density of the fuel (~ 250 - 300 kg m- 3 for wood,
500 - 800 kg m-3 for coal) enabled one to design units with
small fuel hoppers that could operate for 2-4 hours without
refueling.
214
e) A small firebox and an appropriately located throat in a
downdraft gas producer permit the development of a homogeneous
fire zone which aids the partial cracking of tar and disso-
elation of water (Chapter 9).
On the other hand a typical rice hull gas producer has the following
operational characteristics:
a) A fuel is used which has a high natural ash content (15% - 23%
by weight) and a low bulk density (100 kg m-3).

b) A gas output of 200 Nm3 h-l is achieved in large diameter fur-
naces (> 1.5 m).
c) The residue after gasification amounts to 30% - 40% of the ini- .
tial volume of the feedstock and 25% - 35% of its initial
weight. An effective continuous ash removal system is there-
fore required.
d) The low bulk density of rice hulls necessitates either a very
large hopper or a mechanism for continuous feeding.
From previous chaptoi!rs it is concluded that the three parameters which
influence rice hull gasification the most are the following:
a) Continuous ash removal.
b) Superficial gas velocity.
c) Diameter of the gas producer.
215
This work has concentrated on very small-scale units, having furnace
diameters of 15 - 30 em. Nevertheless it is useful to describe in
detail two large rice hull gas producers that have been operating com-
mercially for quite some time and for which some performance data
exists.
Italian Balestra Type Updraft Rice Hull Gas Producer (1910 - 1944)
Although this unit has been discussed in several papers, two papers,
one by Garbarino (1911) and the other by Borasio (1944), have been
translated to extract relevant information. A schematic of this rice
hull gas producer-engine system is shown in Figure 10-5. The feeding
mechanism consisted of an elevator (b) which transported the rice hulls
into the small hopper on top of the gas producer. Large amounts of air
leaking through the feeding arrangement could easily cause an explosion
in the gas producer. Thus it was essential to have the feeding column
always filled with rice hulls. The large cylinder reaching into the gas
producer was necessary in order to keep the rice hull bed always at the
same level. Raising the r.ice hull fuel bed would restrict the disen-
gagement zone and would result in a large carryover of fine particles
(see Section 10.4). Lowering the level could lead to the development of
a channel through the fire zone which would in turn cause poor co2
reduction. The knife gate (c) was kept open during operation. The
diameter of the furnace was 1.5 m at the bottom, 1.9 m at the top. It
had a refractory lining as shown in Figure 10-5. The ash removal system
consisted of two spiral-shaped wipers (g and h) made out of cast iron.
The wipers rotated in a continuous step-like motion generated by a belt
216
F ig u r e 1
0 -S - Italian
updraft
r ic e h u ll
(G a r b a r in gaa prod
o , 1 9 1 1 )· ucer of
th e B a le
a tr a ty p e
217
drive using transmission i. This motion caused partial crushing of the
charred rice hulls which could then fall through. The rice hull ash
then slid down the cone-shaped bottom of the gas producer and was
flushed out with water in the ring channel (k). It is noted that this
gas producer did not have a water seal at the bottom, since it was an
updraft suction gas producer. One interesting feature of this unit is
the way in which the air was introduced at the bottom. The unit had a
cylindrical steel sleeve (e) at the bottom which could be raised and
lowered by the drive (d). Full opening of the sleeve exposed the entire
lower part of the gas producer to the air, while closing of the sleeve
prevented air from entering. Therefore, the necessary combustion air
had to move through the two ash layers at the bottom and was introduced
into the center through the step-like arrangement of the grate.
After leaving the gasifier, the producer gas entered a cooling and
~~pansion tower (n). It is noted that this tower's main function was to
cool the gas from an inlet temperature of 130°C to 40°C. The arrange-
ment consisted of a series of baffles with two wat~r spray nozzles at
the top (1). The spray caused condensation of most of the tar which was
then collected in the centrifugal tar separator (o). Two types of gas
cleaning trains (p) were used: a dry packed bed (coke) and a wet packed
bed (coke, with spray nozzle on top). Because of the sizeable pressure
drop across the system, a suction fan (q) was sometimes installed at the
end of the gas cleaning train. This procecture increases the volumetric
efficiency of the engine.
It is not clear from the literature what the actual horsepower out-
put of this gas producer was. Horsepower outputs of 25 - 60 hp are men-
218
tioned. An average gas output of 160 m3 h-l is reported. However, no
reference to the temperature of the gas is given. In Chapter 1 i t was
noted that there are many ways to operate a gas producer and that a
reliable way of describing the syste•n is by the equivalence ratio lj>.
The combination of an updraft gas producer and a gas engine requires
extensive gas cleaning because of the large amounts of tarry liquids
generated which must be removed from the gas stream. However, one
should not conclude that the reason for building an updraft unit was
because downdraft gas producers did not exist. The decision to build an
updraft unit jo7as also based on the fact that a highly valuable by-
product, tar: was generated. It is noted that the gasification of 100
kg of rice hulls yielded 122 - 158 kg of producer gas, 2 - 4 kg of
"Catrame" and 20 - 25 kg of char. It is worthwhile to point out that
Catrame is the Italian word for the material which is commonly used for
road surfaces (asphalt). During the 1910 - 1950's Catrame was a very
valuable product. It was used to pave barn floors and small backyards,
and to do repair work on roads. Thus, the operators of Italian rice
hull gas producers did not wish to suppress tar generation. In fact the
units were operated at very low furnace temperatures (700 - 900°C) which
yielded a fair amount of tar. The compos! tion of the producer gas
generated as well as its heating value is given in Table 10-1.
219
Table 10-1. Producer Gas Composition
Species % Volume
co 25.0 - 26.1
co 2 4.2 - 4.5
H2 4.8 - 8.6
CH4 7.1 - 7.2
Cn an Traces
N2 51.2 - 56.1
02 2.1 - 2.5
HHV 6.25 MJ kg-1 (6.72 MJ m-3)
It is noted that the gas has a very high heating value and contains a
considerable amount of methane. Based on the given mass balance, it is
concluded that the unit operated at a q, of about 0.3. Considering all
thls one can conclude the following:
a) The unit was operated such that the superficial gas velocity
through the bed was very low at roughly 5 em s- 1 - 10 em s- 1
(see Chapter 6).
b) Furnace combustion temperatures were low (in the 700 - 900°C
range) which resulted in a producer gas with a high heating
value of 5.4 MJ kg-l to 6.25 MJ kg- 1 •
c) Tar was considered a valuable by-product and not a pollutant or
health hazard.
d) The gas cleaning train and centrifugal tar separator were
capable of cleaning the gas t~ the extent necessary for use in
an internal combustion engl11e.
e) The gas engine powered all equip~ent in the rice mill. Not all
220
the gas which was generated could be used and some was simply
flared off through pipe (m) or used for heating service (Figure
10-5).
Rice Bull Gasification Plant in Suzhou, Jiangsu Province, China
'!'his plant is described in detail in the FAO report RAPA No. 61,
Bangkok, Thailand, June 1982. A schematic diagram of its design is
given in Figure 10-6. It is noted that this unit operates in the
downdraft :node with the air inlet (1) on top. Because of the large
quantities of ash which must be removed from the unit, it was necessary
to set the entire plant in a water jacket, which creates a hydraulic
seal. The char gradually settles to the bottom of the water jacket and
is then pumped up as a slurry (8) with a common mortar pump.
The grate design is different from the Balestra unit. Instead of
having one or two wipers rotating, the sieve plate in the Chinese unit
itself rotates. This grate also features an annulus through which the
gas is drawn. The annulus also provides a disengagement zone for larger
particles in the gas stream. This 'l!ethod of drawing off the gas has
been widely used in automotive downdraft gas producers. Despite the
water jacket, the gas exit temperature is still high at 300 - 400°Co
The recorded gas composition is given in Table 10-2.
221
0
0
0
.....
I
0
0
lU
"' 2
11
•.
1. Fuel & Air inlet 7. Ash settling pond

2. Cooling water jacket 8. Ash removing tube
3. Furnace erate reduction box 9. Cooling water intake
4. Gas outlet 10. Cooling water effluent
5. Rotary grate II. Gas
6. Grate support 12. Ash
Figure 10-6 . Chinese rice hull gas producer (Rapa No. 61).
222
Table 10-2. Composition of Producer Gas From Downdraft
Chinese Rice Hull Gas Producers
Species % Volume
co 15.4
co2 7.25
N2 65
LHV (lower heating value) 2.6 MJ kg-1 (3.0 MJ m-3)
It can be seen that the gas is of very poor quality with respect to its
heating value and has only about one-half the heating value of the gas
from the Balestra unit. This rice hull gas producer can be viewed as an
updraft gas producer turned upsicie down. The raw gas from this unit
should contain less tar than the gas from updraft units. It is claimed
that the extensive filtering system will reduce the impurity content of
the raw gas to less than 30 mg Nm-3. 'Due to the low heating value of
the gas, the gas-engine generator set consumes 5.1 '113 of producer gas
per Kilowatt-hour. However, it is noted that the efficiency of the
overall plant is good at 13% (61% gas produc<!r efficiency, 22% engine
efficiency computed from availahle data). The underlying reason for
installtng this plant was two-fold:
a) A rice mill with a capacity on the order of .50 ton/day paddy
powered by a rice hull gas producer like the one described
above saves 3300 - 4180 barrels of diesel oil per year.
b) The same rice milt produces 3571 tons of hulls a year. There-
223
fore, 0.5 - 0.7 hectares of land would be needed to store these
rice hulls, which cannot be burned in the open. The gasifica-
tion process reduces the rice hull volume by 60%.
Most rice mills in the Developing Countries are much smaller than the
ones mentioned above (1 - 2 ton/day). It is therefore useful to deter-
mine the design criteria for a small rice hull gas producer that could
power a 5 - 15 hp engine. In order to optimize the gas producer-engine
system, it is necessary to treat the various parts of a gasifier as
separate components which do not interact with one another.
The author has been testing various gas producers, gas cleaning
trains and ash removal systems and has explored the following:
a) Continuous ash removal systems.
b) Ash removal by vibration.
c) l"orce feeding systems.
d) Downdraft gas producers without throat.
e) Various air injection systems.
f) Various gas cleaning trains.
g) Gas exit designs for rice hull gas producers.
Design Considerations for Ash Removal Systems
Introduction
Because of the large amounts of rice hull char that must be removed
during the gasification process, a continuous ash removal system is
essential. Experiments with an open core 15 em diameter gas producer
showed that the residue retained 67% of the initial volume of the feed
input, if the caked rice hulls were not disturbed. If the fuel bed was
stirred, the residues retained 37% of their original volume, i.e. before
224
gasi ftc at ton. After the char was discharged into the ash bin, it
occupied 32% of it;; original volume and had a hulk densLtj of 117 kg
m- 3 (dry basis). The volume fraction occupied by the char is denoted
here by E. It is now of interest to compute the velocity vr of the
charred residue through the rice hull gas producer as a function of the
rice hull consumption rate R (kg h-l as received), the gas producer
diameter d, and the volume fraction E. We have therefore:
where Pb bulk density of the rice hulls as received.
EKamples
A 15 em diameter gas producer operating an engine with an output of
5 hp requires about 10 kg of rice hulls per hour (p = 120 kg m- 3 ). This
would result in:
= .il:.:l7)(0.37)(10 -1
v 1.74 m h
r
(0.15) 2 (120)
On the other hand, the large Italian rice hull gas producer with a· fire-
box diameter of 2 m and a power output of 180 hp would require 540 kg
h-l of rice hulls (Beagle, 1978). This results in:
(1.27)(0.37)(340) -1
v 0.53 m h
r (4)(120)
From the two examples it is obvious that the ash removal velocity is
significantly higher in smaller units. It is noted that the two ash
removal systems discussed in Sections 10.1 and 10.2 have limitations
with respect to the rate of char removal.
225
Eight different types of rice hull char removal systems have been
tested and the three most successful are discussed here.
Ash Removal Designs
Parameters which influence the grate design are shown in the diagram
in Figure 10-7. In general, the followfng ash removal options are
available:
a) No grate at all.
b) Fixed grate allowing for periodic removal of the ash.
c) Mechanical, rotating grate allowing for periodic removal of the
ash.
d) Mechanical, rotating grate allowing for continuous ash removal.
Rate of
ash removol
Superficia I
gas velocity
l
and flow field Pressure drop
/across the grote
Size
distribution--~..
of the char
~
8 .,."4t---Construction
material
/
Bulk density
of the char r
Construction
Type of
drive
cost and
maintance
Figure 10-7. Parameters which influence the grate design.
Option a), which has been widely used for fuels with very low ash con-
tents, is impractical for continuous rice hull gasification because of
226
the large volumes of char generated during gasification. Options b) and
c) also result in unsatisfactory gas producer performance because of the
high char velocities of 0.4 - 1.8 m h- 1 , as previously outlined. It is
noted that based on the above calculations, periodic ash removal at 30
minute intervals would cause a fire zone rise of 0.20 and 0.9 m for the
range of char velocities and lead to a breakthrough of the fire zone at
the top which in turn would result in deteriorating gas quality. The
gasification process will therefore be sustained only for a short period
of time in rice hull gas producers with fixed grates or with periodic
ash removal systems.
Based on these observations, it was concluded that a controlled con-
tinuous ash removal system is necessary for reliable continuous gasifi-
cation of rice hulls.
When dealing with a difficult fuel such as rice hulls, it is very
important to eliminate factors that may influence the performance of a
grate, but are not caused by the design of the grate. For instance,
even a well-designed grate will not help if the rice hull fuel bed above
the grate cakes. In this case, gravity flow stops altogether. In order
to avoid misinterpretation of grate performance, the grates were first
tested on a cold test stand that simulated a 15 em diameter downdraft
gas producer (Figure 10-8). A good ash removal system should meet ·the
following criteria:
a) It must· remove the char in any form (loose, lumped, mixed with
small pieces of slag.
b) The rate of ash removal must be controllable.
c) Because force feeding of rice hulls into the firebox column is
227
Figure 10-8. Cold t es t s t a nd for grat e pe rformanc e.
one option with regard to overcoming gravity flow problems and
channel formation, the ash removal system should be able to
remove densified char.
d) The design should be simple.
Design I
The design shown in Figure 10-9 is a copy of the large grates used
in past coal gasifiers. It was used in a 30 em diameter downdraft gas
producer of the type showed in Figure .5 -l by Williams and Horsefield
(1982) to test the gasification characteristics of California crop resi-
dues. The version shown in Figure 10-9 fits a 15 em diameter gas pro-
ducer and can be installed as shown in Figures 10- 10 and 10- 11.
228
GAS AND CHAR
GAS AND CHAR
Figure 10-9. Eccentric rotating grate.
1. Performance Characteristics and Properties.
The eccentricity of the rotating grate causes the char to be pressed
through the horizontal slats. The construction of this fairly complex
grate takes about 20% of the total time needed to build a small 15 em
diameter gas producer. Although the grate is very effect.i.ve for wood
chips and corncob char, it handles rlce hull char in a very erratic
229
><
w
VI
0
<.:>
a
a
Figure 10-10. Gas exit above grate. Figure 10-11. Gas exit below grate.
manner. One of the major difficulties is that it cannot remove con-
solidated matter. In such a case the grate generates free caves or
simply plugs up. 13ecause the superficial velocities between the plates
are high in a small gas producer, it is not advisable to pass the gas
through the grate as shown in Figure! 10-11. One big advantage of the
eccentric grate is that it slowly and continuously stirs the lower part
of the gas producer. It is therefore suited for gas producers that
operate at low temperatures (700 - 800°C) and lower ··s. The author con-
siders this design a non-optimal solution to the problem.
Design I I
A completely different approach to the problem is represented in
Figure 10-12. This design consists of a solid disk which is fixed at a
distance d below the bottom of the firebox. A curved wiper slowly rota-
tes over the top of this disk. It is noted that the ash is not removed
through holes in the disk but is pushed over the edge of the disk and
into the ash bin.
230
-)(
I.&J
c
en
0
(,!)
Ash
Figure 10-12. Grate with rotating curved wiper.
1. Performance Characteristics.
This grate was tested extensively on the cold test stand since it
holds the most promise with regard to meeting the criteria mentioned
above. The results are given in Flgure 10-13. Rice hull char was used
for these tests. The wiper height was 7 mm and the disk diameter was 30
em. The firebox column had a diameter of 15 em. This grate design
removes the char .in a very steady, continuous manner. It has the
following advantages:
a) The amount of char removed can be controlled by four
parameters:
231
I
..c
...,0
c
UJ
>
0
2
UJ
a:
:J:
VI
<
....J
....J
:::1
:J:
UJ
u
a:
WIPER SPEED, rpm
Figure 10-13. Rice hull char removal as a function of the wiper speed
and distance [d]. Wiper height 7 mm, disk diameter 30 em
and firebox diameter 15 em.
i) The overlap between firebox diameter and disk diameter.
ii) The distance d between the firebox and the disk.
iii) The height of the wiper.
iv) The wiper speed.
b) This system is capable of removing consolidated matter. The
wiper simply slices it off piece by piece. It can also handle
slag as long as the slag is not larger than the distance d.
c) It makes possible removal of char which has been compacted by
232
force feeding of the rice hulls into the firebox.
d) The system is simple to build and easy to replace.
The main disadvantage of this type of grate is the ease with which the
fuel bed can be disturbed near the grate. Because of the high pressure
drop caused by passing ·the gas through the space between the disk and
the bottom of the firebox, it is recommended that the gas be drawn off
well above the grate, as shown in Figure 10-14.
Design III
This design is a slightly modified version of Design II. It is
shown in Figure 10-14. This design allows better utilization of the ash
bin. It slices the lowest layer of the char off the plate and is there-
fore also suitable for periodic ash removal in low ash fuel gasifica-
-
tion.
)(
w
1/)
0
<.!)
·. Ash
Figure 10-14. Eccentric wiper grate.
233
Design IV
Small gas producer-engine systems (5 - 15 hp) cannot be justified by
their fuel savings alone, they must be very simple and very cheap. It
is therefore important to explore the possibilities for removing the ash
with a grate that does not need an external drive. In this small range,
the engine and gas producer can be mounted together on a free-swinging
frame. Here, the engine vibration facilitates ash removal. It is noted
that past automotive gas producers had very few difficulties with regard
to fuel flow and bridging of the fuel in the fuel hopper. This was due
to engine vibration and perhaps to rough roads. In order to test the
effect of engine vibration on the gasification process and on ash remo-
val, the engine and gas producer were mounted together on a free-
swinging frame as shown in Figures 10-15 and 10-16. It is important to
leave the bars unbolted at the ends because this would inhibit vibration
in the gas producer. The small Briggs and Stratton 10 hp engine caused
a very satisfactory vibration in the corn-cob fuel bed and also in the
grate disk. The ash remova~ over the disk can be fine tuned by varying
the distances d and a. It is noted that the gas should be drawn off
above the grate because even slight adjustments of the distance d
resulted in large pressure drop changes across the grate.
Another version of the vibrating grate is shown in Figure 10-17.
This version is a slightly bent disk that facilitates the char removal
by vibration better than the flat disk in the case where the distance d
(Figure 10-12) is very small.
The results of two tests with rice hull char and corncob char are
given in Table 10-3.
234
Bin Hopper
Fire zone of
gas producer
Circular fixed plate
Threaded shaft to
support plate
. Internal
Handle to screw~ combustion
engine
Metal
up or down to adjust d frame
\ Two swinging channel bars to transfer

Rubber \ en.gine vibration to fixed plate and gas Rubber
stopper producer stopper
Figure 10-15. Gas producer and engine mounted on a free-swinging frame.
235
Figure 10-16. Detail of free-swinging bars and mounted engine.
~
c
0
N
~
·-
~
lL
~
~
c==:=
Figure 10-17. Slightly curved vibrating disk.
236
Table 10-3. Char Removal by Vibration
Char Velocity Mass Flow Rate

Fuel m h-1 kg h-l
Corncob char 0.71 2.0
Rice hull char 1.88 7.4
Distance, d, was set at 38 mm. Distance, a, was 63 mm for this disk.
The tests showed the potential for simply using the engine vibration to
induce horizontal char removal by means of an excited disk underneath
the firebox. This method of char removal will be specific to the par-
ticular gas producer, to the engine, and also to the way in which the
two are mounted on a free-swinging frame.
Summary
Of the eight grates tested, four have proven reliable. It is noted
that Design I stirs the lower part of the char bed whereas the other
three designs have no mixing effect on the bed. However, Design I is
also the most unreliable with respect to a constant ash removal rate. In
small units the gas should be drawn off above the grate with these
designs. Passing the gas through the grate will result in high super-
ficial gas velocities that cause entrainment of char particles. These
char particles are difficult to separate from the gas stream.
Design Considerations for the Gas Exit
One of the most important obstacles to the use of producer gas in an
internal combustion engine is the impurities in the gas. In general, it
is not difficult to generate a combustible gas. However the large
237
amounts of water vapor, tar, char particles and solid carbon entrained
in the gas stream cause serious operational difficulties in the gas
cleaning apparatus and in the engine. This section addresses exclusive-
ly the topic of separation of fine particles from a gas stream by
settling. It is noted that past gas producers of the updraft design had
a large disengagement zone as shown in Figure 10-5 for the Italian
Balestra unit.
Downdraft units such as the Chinese rice hull gas producer and the
automotive Imbert gasifier use an annulus to disengage particles and to
act as a heat exchanger for the sensible heat of the gas (Figure 10-6).
On the other hand, a common complaint about past crossdraft gas pro-
ducers had to do with the excessive amounts of solid particles entrained
in the gas stream.
The question as to whether a low superficial exit gas velocity and a
long vertical disengagement zone facilitate separation of char and gas
can be answered quite accurately on a theoretical basis. The usual
simplifying assumptions that allow an exact representation of the ter-
minal velocity of a particle are:
a) The particle is spherical and rigid and no slip exists between
the particle and the gas.
b) The gas is homogeneous within the length scale of the par-
ticles.
c) The settling velocity is so low that all inertia effects are
negllgible.
d) The flow field is laminar.
Of course, it is obvious that one of the assumptions, a) is not at
238
all satisfied for rice hull char. But as pointed out by Heiss (1952)
even for nonspherical particles where the terminal velocity depends on
the orientation of the particle, the highest ratio between maximum and
minimum velocity is less than 2.
Example
For a typical gas exit temperature of 600°C the average viscosity of
the producer gas will be n = 356° 10- 7 kg m- 1 s- 1 , as shown in Figure
10-13. Assuming that the gas behaves according to the ideal gas law,
the calculated gas density Pg is 0.33 kg m-3. The terminal velocity of
a particle may be computed usiag Stokes theorem if Re < 0.2. We there-
fore need to look at particles with diameters
d <
The density of rice hull char, Ps• has been determined to be between 2.0
- 2.1 g cm-3 depending on the amount of carbon left in the char (2.3 g
cm- 3 is the true density of the rice hull ash). Consequently d < 85 JJID
for laminar flow settling. The terminal velocity of a 85 JJID particle is
given as
(p - p ) g d2
s g 0.25 m s-l (10-1)
18n
If we assume that the particle accelerates from u 0 to ut within the
gas stream we have:
239
u(t)
_<_P...;;s_~P::.~g.._)_g_p...;.s_dz_ ( 1 -
(10-2)
18n
this equation yields:
t* (10-3)
where t* is the time needed for the particle to reach 99% of its ter-
minal velocity. In this case:
2
(4.6)(85) 2 (10-6 ) (2300)
t* 0.1 s
(18)(356)(10- 7 )
The distance traveled when the particle reaches 99% of its terminal
velocity is:
h 0.785 u • t* (10-4)
t
(0.785)(0.25)(J.l) 1.9 em
One may conclude the following from the example.
a) For all practical purposes the terminal velocity is reached
instantaneously by particles with diameters d < 0.1 mm.
b) The distance traveled in reaching 99% of ut is minimal.
c) Particles with diameters d > 0.1 mm may be removed efficiently·
by gravitational settling in large updraft gas producers.
It is noted that an updraft rice hull gas producer with a diameter
of 2 m at the top and with a gas output of 400 m3 h-1 at 200°C will
have a superficial gas velocity us 0.035 m s-1. Because of the
240
instantaneous terminal velocity of the particle (0.25 m s-1), it is
likely to remain in the gas producer. On the other hand, a small rice
hull downdraft gas producer with a gas output of 12 m3 h-1 at 600°C,
will have a superficial gas velocity of 0.42 m s-1 if the gas is drawn
off through a 10 em diameter pipe or 0.15 m s- 1 if the gas is drawn off
through an annulus with an inner diameter of 25 em and an outer diameter
of 30 em.
The above examples illustrate that the retention of small particles
of d < 0.1 mm in the gasifier is also a function of the gas exit design.
From the size distri.bution curve (Figure 6-21) we also can conclude that
rice hull char contains about 48% - 60% of particles smaller than 0.1
mm. For larger particles with diameters d > 0.1 mm, Stokes theorem can-
not be applied and the general set of equations relating the size of a
sphere to its settling velocity must be written:
(10-5)
(10-6)
p u d
g t
Re = n
(10-7)
The drag coefficient c0 is given by Equation (10-6) for the transitional
region, 0.2 ( Re ( 500, which is from Schillar and Nauman (1933). For
the turbulent region Re > 500 Cn is roughly constant at c0 = 0.44
(Allen, 1975).
This system of equations must be solved by iteration. The visco-

sity, n, of the gas is strongly dependent on the temperature and is
given in Figure 10-18 for a typical producer gas composition.
241
500
T
.,
1 400
.....
E
.,.
I 3QQ
Q
19'Yo co
E=- 9'Yo C02
>- 200 15 "lo H2
.,
t- 4"/o
15"/o
CH 4
H2 0 ( voporl
0
u 3 8"/o
., 100 N2
>
TEMPERATURE T, •c
Figure 10-18. Producer gas viscosity as a function of temperature.
The terminal velocity of char particles smaller than 100 11m has been
computed using equation (10-1) and is shown in Table 10-4. The assump-
tions made are that the raw gas exit temperature is 600°C and that the
density, Ps = 2.0 g cm-3.
Table 10-4. Terminal Velocity Versus Particle Size Diameter
Particle size, um 100 75 so 25

Terminal velocity m s-1 0.31 0.17 o.os 0.02
From Table 10-4 i t can be concluded that removal of particles smaller
than 100 11m is difficult to achieve by settling and requires a very low
superficial gas velocity at the gas exit.
242
Air Injection
In the previous sections evidence has been represented to show that
the method of drawing the air into the fuel bed is crucial for reliable
rice hull gasification (Figures 6-11, 9-5, 9-7, 9-8). The majority of
past downdraft gas producers had either wall tuyeres, a center tuyere,
or a combination of the two. The following geometrical configurations
have been tested with corncobs and rice hulls.
a) lolall tuyeres with and without throat (Figures 10-19 and 10-20).
b) Center tuyere with and without throat (Figures 10-21 and
10-22).
In order to avoid misinterpretation of the results, a series of tests
were also conducted with a high quality gasification fuel (broken
corncobs) of similar bulk density ('" 200 kg m-3). All tests were con-
ducted in a 1 'i em diameter refractory-lined gas producer equipped with
either the vibrating or rotating grate shown in Figures 10-12 and 10-17.
The purpose of these tests was not to determine the arrangement
which optimizes tar cracking. Instead the purpose was to determine the
influence of the method of air entry on the gas composition. The
results of the experiments are given by Kaupp (1982) in unpublished
monthly reports to the Briggs and Stratton Corporation. The following
was concluded:
a) None of the designs shown in Figures 10-19 through 10-22 are
suitable for rice hull gas producers because the fuel bed is
disturbed by the high superficial air velocity at the end of
the nozzle.
b) The above designs work well for corncobs, a fact that has been
known for decades.

243
1 "',(C
~
AIR
Figure 10-19. Wall tuyeres Figure 10-20. Wall tuyeres with

without throat, throat, downdraft
downdraft gas gas producer.
producer.
Figure 10-21. Center tuyere Figure 10-22. Center tuyere

without throat, with throat,
downdraft gas downdraft gas
producer. producer.
244
c) A small rice hull downdraft gas producer cannot have a throat,
since the insignificant size reduction of rice hulls leads to
blockage at the throat.
In order to avoid disturbance of the rice hull fuel bed, two different
methods for drawing the air into the gas producer have been tested.
a) A continuous slot around the circumference of the gas producer
which replaced the eight tuyeres (Figure 10-23).
b) Diffusion of the air through the fuel bed from above in an open
top gas producer (Figure 10-24).
Based on experimental results with corncobs and rice hulls, it was
concluded that diffusion of the air through the top is by far the most
convenient way of introducing air without disturbing the fuel bed.
1 _1 ~
Ill!
AIR
AIR
Figure 10-23. Continuous slot Figure 10-24. Open core air diffu-
as air inlet. sion into the fuel
bed.
245
Design of a Small (2 - 20 hp Mechanical Power Output) Rice Hull Gas
Producer
Introduction
It was outlined in Chapter 3 that automotive gas producers which
were fueled primarily with good quality charcoal, wood and coal. In
addition large rice hull gas producers (60-200 hp) have been in opera-
tion for several decades prior to 1950 (Beagle, 1983). However,
designing a small simple rice hull gas producer to power an internal
combustion engine is a very challenging task for the following reasons:
a) The fossil fuel savings in small engines are relatively small.
It is therefore necessary to have a design that can be inexpen-
sively manufactured and that can be replaced after a few years
of operation.
b) Contaminants in the raw gas (tar, solid particles) must be
removed to a large degree within the gas producer itself
because sophisticated gas cleaning equipment is either una-
vailable on a small scale or prohibitively expensive at this
scale.
c) Any continuously moving part on a small rice hull gas producer
must be driven by either an electric gear motor or a compli-
cated drive arrangement involving the internal combustion
engine. It is therefore highly desirable to avoid the incor-
poration of moving parts into the design.
Using these criteria, the approach of negative selection was used to
show that the presently available designs for small, fixed-bed gas pro-
246
ducers (crossdraft, downdraft, updraft) are not suitable for rice hull
gasification. It is emphasized that it is possible to generate a com-
bustible gas from rice hulls in some of these units at least in the
laboratory. The objective of this work, however, has been to explore
the behavior of rice hulls during thermal decomposition in order to
establish parameters which would be helpful in improving designs.
In the previous chapters it has been outlined that the caking of
rice hulls, the method of char removal and the method of injecting the
necessary combust ion a! r are cruc ia 1 with regard to reliable operation
of a gas producer. Based on these experimental results the following
system·s have been excluded and are considered unworkable for the gasifi-
cation of rice hulls on a small scale.
a) Downdraft gas producers with a throat.
b) Downdraft gas producers with wall tuyeres or center nozzles or
a combination of both.
c) Crossdraft gas producers.
d) Gas producers without a continuous ash removal system unless
the units are batch fed.
e) Gas producers without a force feeding system unless the units
are batch fed.
For instance, many versions of a downdraft unit were tested. These
early models included the following:
a) Force feeding of rice hulls using an auger system with throat
and center nozzle air inlet (see Figure 10-25).
b) Force feeding of rice hulls using an auger system without
throat and center nozzle air inlet (not shown).
247
BIN HOPPER
E
u
0
0
N
-GAS EXIT
r-T""r-----,
ASH
Figure 10-25. Downdraft gas producer with force-feeding system.
248
In both case the ash was removed by an auger as shown.
Although both systems could be operated for short periods of time
(30 180 minutes), they have many design deficiencies as outlined
below:
a) The center air nozzle generated caves in the fuel bed and
induced slagging.
b) A force feeding system cannot move the rice hulL char past the
throat, which jams the auger (insufficient size reduction of
rice hulls).
c) The system requires at least one external drive if the center
pipe is water-cooled and two drives if the feeding and ash
removal systems are separately driven.
It is noted that the raw gas generated had a sufficient higher heating
value (3. 7 - 4.4 MJ m- 3 ) and was fairly clean (the flame was blue to
slightly orange).
Four other types of downdraft gas producers have been designed and
tested. The results are presented in Chapters 6 through 9. All four
prototypes have several common features:
a) The air is allowed to diffuse through the top.
b) All are equipped with a continuous ash removal system as
described in Section 10.3.
c) The gas is drawn off above the grate.
d) High fire zone temperatures of ;;. 1000°C should be maintained to
ensure proper tar cracking and steam conversion.
The design shown in Figure 10-27 is the simplest version. It is
equipped with a vibrating grate and is mounted on a free-swinging frame
249
with the engine (see Figures 10-15 and 10-16). The design was tested
extensively with corncobs and performed very well. The fire zone could
be maintained anywhere desired within the fuel column (Kaupp, 1982).
The rate of ash removal was controlled by adjustment of the distance
between disk and firebox. Attempts to achieve continuous operation
using rice hulls failed because of the caking of rice hulls (see Section
6.5). A schematic of the structure of the rice hull fuel bed before and
after gasification is shown in Figure 10-26. The vibration induced by
the engine did not break the carbon bonds established during the pyroly-
sis process because of the air gap between the fuel bed and the wall of
the gasifier (see Figure 6-15). A slightly modified version, which was
equipped with an effective and controllable wiper grate, did not perform
any better because the wiper cannot remove rice hull char that does not
move down onto the grate (Figure 10-28). A more complicated version is
shown in Figure 10-29. With an auger mounted on top of the unit, the
caked rice hulls could be pushed through the firebox column at a steady
pace.
Figure 10-26. Rice hull fuel bed before (A) and after caking (B).
250
FUEL
AIR
~ jj
.........
REFRACTORY
:·.·'.-':.
" ·:-:
.': .·: "·..~
... ·...···'· ...
...· ...··: ,
;'
.,
. :.:. ·... ...
...... ' 1 ~ •
···..·.-:
:
.....:·
VI BRAT ION
GRATE
Figure 10-27. Gas producer with vibration grate and gravity flow.
251
FUEL
....:··
.......
...... t
900 ...... :
.-~'......
_~
l',l.. :
··,...
700 ... ·.:
-----
:' l
•::·
:·.:·:
..:...:~· -~·
~ 600 \!) ~GAS
~WIPER
FIXED PLATE
DRIVE
Figure 10-28. Gas producer with wiper grate and gravity flow.
252
DRIVE
REFRACTORY
900
700
\V
--GAS
~600
WIPER
FIXED PLATE
DRIVE
Figure 10-29. Gas producer with force-feeding system and wiper grate.
253
They are then removed from the grate by the wiper and discharged into
the ash bin. Although this system does overcome the problems asso-
elated with caking, it introduces other complications. A careful co-
ordination of the ash removal rate and the feeding rate becomes
necessary. In addition, the feeding and ash removal rates must be coor-
dinated with the gas output required by the engine. In order to avoid
these difficulties and save one external drive, the grate was replaced
with a hydraulic water seal (Figure 10-30). The water surface acts as a
grate. It is noted that a rice hull column sitting on a water surface
does not sink nor does it absorb a significant amount of water ( .. 1
inch/24 hours). Although the rice hull char is more porous than rice
hulls and is therefore more likely to take up water and sink, the time
required for this process is long when compared to the char velocity in
the column. The water seal has the additional advantage of allowing
continuous discharge of the ash into the water. The ash can then be
removed from the water by hand. The water seal can also be used as an
internal wet scrubber and cooler condenser. Despite the advantages of
this system, it cannot be considered a practical solution to the
problem of rice hull gasification on a small scale.
After testing all four prototypes over a period of 18 months, the
concept of a "continuous" rice hull gas producer with a moving fuel
column and a locally fixed fire zone was abandoned. Instead, a batch-
fed unit with a moving fire zone and no ash removal during operation was
designed and tested. The designs in Figures 10-27 through 10-30 show
controlled char removal with the fire zone fixed in space while the air
is diffusing in through the top as shown in Figure 10-31.
254
I
DRIVE
OPEN FUEL
HOPPER
60
.·..
.....
::-.:·
I:·.
,'
50
>::
....
..:.•
......-..•.
.;;~·
~,·;
··..
ASH BIN
40
l~~
""'1•------- 60 em - - - - - -......~~
Figure 10-30. Gas producer with force-feeding system and water grate.
255
When gasifying rice hul ts in a batch-fed unit, one cannot fix the
position of the fire zone because the volumetric reduction o.f the fuel
bed of 60% - 68% is insufficient. Two choices are available with regard
to the operation of this type of a unit:
a) The fire is initiated at the top. The fire will then move down
through the rice hull column leaving charred rice hulls behind.
Below the flre zone will be uncharred rice hulls. This method
results in a raw gas which is highly contaminated with tar
because the gas does not pass through a hot column of char
(Figure 10-32). (Similar to an updraft gas producer).
b) The fire is initiated at the bottom of the column. The fire
zone wilt then move upwards against the air flow leaving a hot
charred rice hull column behind.
This method (b) meets the necessary prerequisites for proper tar
cracking and steam conversion. It also results in the generation of a
carbon surface for co 2 reduction to CO. Nevertheless, it, is not obvious
whether such a system can generate enough gas for continuous operation
over a four-hour period. A detailed evaluation is given in the next
section. It ls noted that several researchers have simultaneously
explored this new concept using wood fuel. Published literature con-
cerning their work did not exist at the time of writing and information
exchange was by personal communication (Bob Reines 1982, Tom Reed 1983,
"Buck" Rogers Company 1983).
Open Core Gas Producer
The gas producer used in a set of six tests is shown in Figure
10-34. The gas exit of the test unit is located at the top instead of
256
AIR
RICE
HULLS
FIRE
GRATE
GAS
Figure 10-31. Fixed fire zone in an open core downdraft gas producer
with continuous ash removal system.
in the middle. The middle position of the exit would be a better loca-
tion because the dry gas cleaner would not be higher than the top of the
producer itself. The design is based on the results reported in the
previous chapters. The design features include the following:
a) A large annulus which serves as a heat exchanger and as a
separation zone for coarse particles.
b) A steel restriction at the bottom which eliminates to some
degree the wall flow shown in Figure 6-15.
c) A hydraulic seal, which allows one to empty the gas producer
easily by lifting it out of the seal. The hydraulic seal can
also be used as an internal wet scrubber and cooler for the hot
producer gas.
257
AIR AIR
~~~ ~~~
----- RICE
FIRE HULLS
-----
~
RICE I
HULLS
GRATE
GRATE
=== ===
~
GAS
+
GAS
Figure 10-32. Batch-fed gas Figure 10-33. Batch-fed gas

producer with top producer with
lighting. bottom lighting.
d) A wet, sieve-plate scrubber followed by a dry bed scrubber of
rice hull char.
Mode of Operation
The unit was filled with 20 - 40 grams of charcoal (wood or rice
hulls) and was lighted by dropping a burning piece of paper through the
top onto the charcoal. The necessary suction was provided by a suction
pump, as a substitute for actual engine suction. When the entir~ cross
section was glowing red, the column was filled with rice hulls. The
purpose of the tests was to obtain all relevant operational data as a
function of the gas flow, using a batch-fed operat.ional mode (Figure
10-33).
258
Bolts
steel
f To engine
Dry filter
e Flexible
,
u
CD
connection
4
/Stopper
3
Sieve plate
_...,__ scrubber
"-Drain with valve
Water seal and Rubber pod

wet scrubber
Figure 10-34. Batch fed gas producer without ash removal

1, 2, 3, 4 location of thermocouples.
259
From a theoretical and practical point of view, the following data
are important with regard to rice hull gasification:
a) Air to fuel ratio.
b) Equivalence ratio.
c) Velocity of the upward moving fire zone.
d) Velocity of the downward moving rice hulls.
e) Gas composition CO, C02, THC, H2, H20, N2, 02•
f) Higher heating value of the dry gas.
g) Superficial gas velocity within the fuel bed.
h) Time of operation for a 165 em column.
i) Specific gasification rate.
j) Rate of rice hull consumption.
k) Degree of rice hull conversion.
1) Volume reduction of rice hulls.
m) · Carbon conversion.
n) Efficiency of the process.
In addition, the efficiency of and the pressure drop across the gas
cleaning train has been tested. It is noted that all results are given
as a function of the gas flow rate which was varied between 2.76 and
9.9 m3 h-l (25°C, 1 atm). The gas flow rate was measured with a sharp-
edged orifice plate. A very clean gas is necessary in order to measure

the flow rate with this method. The fire zone temperature has a direct
relationship with the gas flow rate. This is shown in Figure 10-51.
260
Air to Fuel Ratio
The air to fuel ratio is the amount of air used (by weight in kg) to
gasify 1 kg of dry rice hulls. It can be computed once the total gas
flow and the fraction of nitrogen in the dry producer gas are known. It
was assumed that air consists of 21% oxygen and 79% nitrogen. The
results are given in Figure 10-35. It is noted that the air to fuel
ratio will vary according to the amount of producer gas demanded by the
engine. It is not possible to determine an optimal air to fuel ratio
without specifying the parameter to be optimized in a gas producer-
engine system. The observed range of air to fuel ratios was between
1.28 and 2.1.
Equivalence Ratio +
The meaning of the equivalence ratio ¢ as well as its importance has
been discussed in Chapter 1 and shown in Figure 1-1.
From the test results (shown in Figure 10-35) a linear relationship
between the air-to-fuel ratio and the equivalence ratio is apparent. It
is noted that the equivalence ratio is computed by first determining the
stoichiometric amount of air, i.e. the amount needed to combust the rice
hulls completely, to C02 and H2 0 with only the minerals as residue. The
ultimate chemical analysis and the chemical formula of the rice hulls
used for all tests is given in Table 10-5. One kg of these rice hulls
requires 4.70 kg of air for stoichiometric combustion.
261
Table 10-5. Ultimate Chemical Analysis of Rice Hulls for Testing
Species % weight, dry basis
c 39.00
H 4.90
0 33.97 by difference
N 0.53
Char residue 21.60
The measured moisture content was between 10.1% and 12% on a wet basis.
The equivalence ratio at which the higher heating value of the gas
was maximized was 0.32. This is shown in Figure 10-41.
Velocity of the Fire Zone and Velocity of the Rice Bull Fuel Bed
Because no ash removal takes place, the high ash content of the rice
hulls results in an upward moving fire zone. On the other hand, the
volume reduction of rice hulls following combustion and their caking
behavior result in a downward movement of the rice hull fuel bed
(Figure 10-26). The test was terminated when the entire 165 em long
column was filled with char and the fire zone had reached the very top
of the unit. Both fuel and fire zone velocities are important when
considering long-term operation and when considering the use of a forced
feeding system as shown in Figure 10-30.
The fire zone velocity is defined as the ratio of the length of the
fuel column to the total operating time. The downward fuel velocity was
measured two to three times during each run. The results are given in
Figure 10-36.
262
• '4>
Ul o 'AIR TO FUEL RATIO
-9-
,..
a: 0
0 1-
Q1.8 0.45 <t
a:
1-
<t
...
u
a:
~ 1.6 0.40
...-'z
;;;;) <t
"- >
...
;;;;)
0
,
1- 0
a: 14 0.35
<t
1.2 0.30
3 4 6 8 9 10
PRODUCER GAS FLOW RATE, m3 h- 1 (25°C, I otm)
Figure 10-35. Air to fuel ratio of the gas producer as a function

of the gas flow rate.
1.5
I'
/
'c /
/
'c
E /
/
/
e
E ,.,.,. ........
/ E
___ ........ ::.. ........

u
w 0
z w
0 1.1 ........ <D
...
N
-r
........ ...
_J
~
.... ....
;;;;)
....
.... 0.9 0
,..
0 ,..
1-
1-
u0 u
0
_J
w ...>
_J
>
0.5
3 4
PRODUCER GAS FLOW RATE,m3h- 1 (25°C,Iatm}
Figure 10-36. Fire zone velocity (up) and fuel bed velocity (down)
as a function of the gas flow rate.
263
Gas Composition
From the theoretical considerations discussed in Chapter 5, the
expected behavior of the various components of the gas is as follows:
a) As a function of the gas output, the volume fraction of CO and
Hz will reach a maximum within the measured range. The frac-
tion of CO and Hz will decrease with decreasing flow rate
(lower <P) because the rate of steam conversion and CO genera-
tion by Equations (S-2 and 5-4) are a strong function of tem-
perature (Figure 9-12).
A similar decrease in CO and Hz will occur at higher flow rates
(higher <P) due to internal oxidation and to the water-shift
reaction. This typical behavior of CO and Hz is shown in
Figure 10-37. The combined CO and Hz content was maximized at
a flow rate of 7.5 m3 h-l or a <P of 0.44. However, this does
not mean that the chemical energy content of the gas is maxi-
mized under these conditions (see Figure 10-41).
b) The methane content of the gas decreases with increasing gas
flow rate. As outlined in Chapter 5, the methane content will
decrease drastically with increasing temperatures. This is
shown in Figure 10-38. Due to the unique mode of operation,
which results in a long rice hull char column downstream from
the fire zone, lower flow rates and an increased methane con-
tent did not result in a measurably higher tar content in the
raw gas. It is noted that the amount of total gaseous hydro-
carbons was measured (THC). However, it is assumed that only
traces of Cz~ and c3 ~ are present (see Figure 5-22).
264
/
,.,..- ----------
.....
, _..___
"
/0
I "
I
I "
I
I
I
0 I
u I
I"
I
"'2:::> I
I
I
...J I
0 I
> I
I
~
PRODUCER GAS FLOW RATE, m' h-1 (25"C,I aim)
Figure 10-37. CO and H2 content of the dry gas as a function of gas

production rate.
4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE,m5h-1(25"C,Iotm)
Figure 10-38. THC content of the dry gas as a function of gas

production rate.
265
12.0~----~----~----~------~----~----~-----4----~
2 4 5 6 9 10
PRODUCI;:R GAS FLOW RATE, m3 h- 1 (25"C 1 1 otm)
Figure 10-39. COz content of the producer gas as a function of the

gas flow rate.
53,~----~----~----~L_ ____L __ _ _ __ L_ _ _ __ L_ _ _ _ _ _L___ ~
2 4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE, m3 h-1 (25"C,I atm)
Figure 10-40. Nz content of the producer gas as a function of the

gas flow rate.
266
c) The amount of co2 in the gas will be minimized within the range
of operation. An increase in COz must be expected at low and
high gas flow rates due to insufficient reduction of COz and
internal oxidation respectively. This is shown in Figure
10-39.
d) The volume fraction of nitrogen in the gas is measured by dif-
ference and therefore must follow the co2 pattern. This is
shown in Figure 10-40.
All results are based on gas samples that were oxygen free. It
is emphasized that during the start-up period (5 - 20 minutes)
oxygen is present in the gas stream. Th:i.s period can be con-
siderably shortened by allowing the fire zone to develop over
the entire cross section before adding the rice hulls. A rice
hull fire zone tends to develop very slowly in a horizontal
direction if there are unburned rice hulls in the vertical
direction. Once the fire zone was fully developed, no oxygen
could be detected in the gas stream.
A summary of the range of gas compositions as a function of the gas
flow rat~ is given in Table 10-6.
267
Table 10-6. Dry Gas Composition as a Function of the Gas Flow Rate
Flow rate co COz THC Hz

m3 h-1 % volume
Z.75 15.9 14.3 6.1 6.3 57.4 Z7.9
4.97 17.6 13.1 5.1 9.6 54.6 Z7.0
5.74 18.6 13.0 3.0 11.1 54.3 Z6.8
7.46 19.0 1Z.7 Z.6 11.3 54.4 Z6.8
8.68 18.4 1Z.9 z.z 11.1 55.4 Z6.9
9.9 17.5 13.4 z.z 10.3 56.6 Z7.Z
The same mode of operation has been extensively tested using corncobs in
the unit shown in Figure 10-Z7. For instance, a continuous 7-hour test
with corncobs resulted in ZZ% CO and 15% Hz (Kaupp, 198Z). It is to be
noted that the relatively low CO and H2 content of the gas is not due to
the mode of operation or to the design but is due to the structure of
rice hulls and the make up of a rice hull fuel bed.
Higher Heating Value of Producer Gas
The higher heating value of the gas was computed based on the
following heating values of the individual gas species at Z5°C:
CO 11.57 ~J m-3
Hz 11.69 MJ m-3
36.41 MJ m-3
The results are given in Figure 10-41. It is emphasized that the be-
havior of the higher heating value as a function of the gas output is
very typical of that for downdraft producer gas operation in general.
The following characteristics of this behavior are noted:
268
a) Maximization of the CO or H2 contents does not result in maxi-
mization of the chemical energy content of the gas. For
instance, the CO content can be as high as 32% (by volume)
under conditions favoring the water shift reaction in such a
unit (Kaupp, 1982). However the heating value will drop
sharply because almost no CH4 and very little H2 (< 5%) will
exist under such conditions.
b) It is noted that even small increases in CH4 at lower flow
rates can offset moderate losses in H2 and CO because CH4 has
three times the heating value of CO or H2•
c) Operation at very low gas flow rates also means low fire
zone temperatures. Thus, the chemical energy content of the
gas will decrease again in this range because the large
decreases in CO and H2 cannot be offset any longer by increases
in CH4 • The chemical energy content of the gas was maximized
at a flow rate of 4 m3 h-l or a ~ of 0.32 whereas the combined
H2 and CO content was maximized at a flow rate of 7.5 m3 h-1 or
an equivalence ratio of 0.44.
The higher heating value of the gas lies between the 3.0 MJ m-3
reported for the Chinese downdraft gas producer (Section 10.2)
and the 6.72 MJ m- 3 reported for the Balestra unit (Section
10.1).
Superficial Gas Velocity, v 8
The superficial gas velocity, vs, within the rice hull column is
defined as the gas velocity that would result if the cylinder was empty.
269
E
;;
.'
E
4 5 6 9 10
PRODUCER GAS FLOW RATE, m3 h- 1 (25°C, I otml
Figure 10-41. Higher heating values of the producer gas.
'
en
E
u
,:
....
g II
...J
"">
...J
!!
u
"- 7
a:
""...
~
en
5
3 4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE, m3 h-1 (25°C, I olm)
Figure 10-42. Superficial gas velocities (25°C) as a function of

gas flow rates.
270
Its values are given in Figure 10-42 for the range of producer gas out-
puts. It is noted that the numbers are computed using a temperature of
25°C. From the graph it can be seen that high gas outputs of more than
6 m3 h-l lead to superficial gas velocities that can slightly disturb
the fuel bed, as shown in Figure 6-18.
The actual gas velocity, va, at any point in the gas producer can
only be computed if the temperature and void space fraction, £8 , is
known at this point. The relation between vs and va is then given as:
v (10-8)
a
where T1 is the fuel bed temperature.
lim
This equation assumes that V + 0 £ 8 = £a where £a is the area void
space and V is a control volume of cross-sectional area A and height ~z
with ~z + 0.
Operating Time for a 165 em Column
Due to the high ash content of rice hulls a short to moderate
operating time can be expected at various flow rates, as shown in Figure
10-43. During the first 5 - 15 minutes of operation, the gas is not
suitable for engine use. However, the gas composition then remains
stable up· to the point where the fire zone reaches the very top of the
unit. This system can be rnade continuous at additional expenses by
adding a force feeding system and a continuous ash removal system.
271
z"
::1
::>
..J
0
u
E
u
on.,
~ ...
.....
0:::>
oz
...
IL-
:::E
:!
.....
C>
zz
::>
a::
100
3 4 5 6 9 10
PRODUCER GAS FLOW RATE, m3 h- 1 (25•C,I otm)
Figure 10-43. Operation time for a 165 em column.
'
.<=
•
"''
.E
.,.,"'
"'<t.....
a::
z
2
.....
<t
~
IL
(/)
<t
(!)
u
~
...u
~
(/)
100
3 4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE,m3h- 1 (25"C,Iatml
Figure 10-44. Specific gasification rate.
272
Specific Gasification Rates
The specific gasification rate is an indicator of the size of the
gas producer needed for a given output. This parameter is strongly
dependent on the type of gas producer and on the fuel. For instance,
values of 100 - 300 kg m2 h -l are considered normal for large coal
updraft gas producers (Gumz, 1952), while values of 225 kg m-2 h-1 to
5000 kg m- 2 h-l are not uncommon in downdraft gas producers operated
with wood, corncobs and charcoal (Jenkins, 1980; Groeneveld, 1982). It
is pointed out that the assessment of the specific gasification rate in
a downdraft gas producer with a throat depends on which cross-sectional
area is used as a reference. Because this is not always speci fled,
there is a high uncertainty of 200 - 300% in reported values. From
Figure 10-44 it can be concluded that the tested open core downdraft gas
producer has a low specific gasification rate for rice hulls of 110 -
210 kg m- 2 h-l similar to that found in coal gasifiers.
Rate of Rice Bull Consumption
The rate of rice hull consumption, on a dry basis, is shown in
Figure 10-45. A range of 2.11 to 4.05 kg h-l for the tested gas flow
rates was observed. It is noted that the volume of producer gas gener-
ated per kg of rice hulls gasified increased from 1.78 m3 h-1 to

2.44 m3 h- 1 with increasing flow rate but with a reduction in energy
content.
It is assumed that the producer gas can be converted into mechanical
shaft power with an overall efficiency of 15% in a small gasoline
engine. The resulting range of horsepower outputs (1.0 to 2.2) is also
shown in Figure 10-45. Because of the lower volumetric efficiency when
273
• ~Rice Hull Consumption
• ' Horsepower Output
>-
a:
c
'
.c 1-
..."' ::l
a.
1-
z ::l
0 0
...,a:
1-tJ)
a..-
:::E<Jl
::>Cl ;!::
tJ)al 0
z
0
2 ...,a.
u tJ)
a:
_J 0
_J X
;;;;)
X
...,
u
a:
4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE, m3 h- 1 (25"C, I atm)
Figure 10-45. Rate of rice hull consumption, and computed

horsepower output at 15% engine efficiency.
z
0~
;:;;(/)
a::<t
"'"'
~~ 70
So
-'
-'
::l
:z::
"'u
«
4 6 9 10
PRODUCER GAS FLOW RATE , ml h- 1 ( 25"C, I aim l
Figure 10-46. Degree of rice hull conversion.
274
operating on producer gas, and the lower energy dens! ty of the gas, a
gasoline engine must have a gasoline rating at least twice the desired
power output when operated on producer gas. Based on evaluations of
past systems, 1.75 - 3.00 kg of rice hulls (as received) were necessary
to generate one horsepower-hour. The results given in Figure 10-45 show
a dry rice hull consumption of 1.76 - 2.0 kg per horsepower hour and are
therefore in agreement with past data.
Degree of Rice Bull Conversion
Due to the high ash content of rice hulls, one has to cope with
large amounts of residue following gasification. The extent to which
rice hulls are reduced in weight following gasification is shown in
Figure 10-46. The theoretical limit of a 78.4% conversion or a 21.6%
residue retention by weight is based on the mineral content. Not all
the minerals stay in the gas producer. The bulk of the mineral matter
has a very high melting point and therefore an even higher boiling
point. However, some fractions are indeed boiled off at lower temper-
atures. The characteristic white spots on the inside walls of a rice
hull gas producer and piping system suggests that at least some silicon
is given up as SiO and is later oxidized to Si02 , which is in the form
of a very fine powder. Based on the assumption that no minerals are
leaving the gas producer, a weight conversion of 66.3% to 76.8% takes
place. It is pointed out that this represents a conversion of 84.6% to
97.9% of the combustible matter C, H, 0, N. The degree to which carbon
was converted is shown in Table 10-7.
275
Volume Reduction of Rice Bulls
Because of the mode of operation, the volume of residue following
gasification is identical to the volume of the inner steel cylinder. In
fact, the residue is a loosely connected cake of rice hull char. Based
on the experiments, a volume reduction of only 42% to 57% can be
expected (Figure 10-47). Gasification of rice hulls, therefore, is cer-
tainly not a solution to rice hull disposal because 58% to 43% of the
original volume (undisturbed) is present after gasification in the form
of char. It has been previously pointed out that when the fuel bed is
stirred or when the rice hulls are discharged through a rotating grate,
32% - 37% of the initial volume will be occupied by the residue.
Carbon Conversion
The term carbon conversion refers to the degree to which the carbon
in the fuel has been converted into gaseous and liquid products. The
amount of solid carbon carried out in the gas stream is negligible. The
ultimate chemical analysis of the rice hulls was given in Table 10-5.
The ultimate chemical analysis of the residue as a function of the flow
rate is given in Table 10-7. If xf is the weight fraction of carbon in
the rice hulls and xc is the weight fraction of carbon in the residue
then the carbon conversion T is given in percent on a dry basis by:
T = 100 (1 - me xc) where mf and m is the dry mass of the rice hulls
mfxf c
and their residue respectively.
27.6
Table 10-7. Residue and Its Chemical Ultimate Analysis
Carbon
Conversion Species % weight
% c 0 H N Mineral
3.76 67.4 37.78 3.78 0.65 0.29 57.5
4.97 81.1 26.78 3.97 0.16 0.19 68.9
5.74 85.0 22.43 3.31 0.29 0.17 73.8
7.46 87.9 19.34 2.15 0.27 0.14 78.1
8.68 90.3 16.26 2.51 0.22 0.11 80.9

9.9 90.6 15.75 1.55 0.21 0.09 82.4
From Table 10-7 one can conclude that only 67.4% to 90.6% of all the
carbon in the fuel is utilized. It is noted that even at the higher
flow rates (and higher temperatures), moderate amounts of volatiles
are still present in the residue. The residue obtained from the test
with the lowest flow rate has a chemical composition similar to the
residue obtained from low temperature pyrolysis (see Figure 8-7 and
8-8).
Efficiency of the Process
The efficiency of a gas producer is defined here as:
chemical energy output, cool dry gas

energy input from rice hulls
Because the total volumetric flow was measured as well as the total con-
sumption of rice hulls, n can be computed. The following assumptions
were used:
a) The higher heating value of rice hulls is 15.3 MJ kg-l on a dry
basis (measured).
277
b) The higher heating value of the gas is used for the chemical
energy output.
The efficiencies, which vary between 55.8 and 66.5%, are shown in Figure
10-48. This efficiency is most commonly called the cold gas efficiency
of a gas producer. It is emphasized that the efficiency curve must
reach a maximum over a range of g~s outputs. At low gas flow rates too
much carbon remains in the rice hulls. At high flow rates, the carbon
utilization is quite good (see Table 10-7), but the chemical energy
content of the gas deteriorates (see Figures 5-23 and 10-41), with a
higher sensible heat fraction.
Gas Cleaning Train
Introduction
The difficulties encountered in the cleaning of producer gas have
hampered the widespread use of producer gas as an engine fuel. It is
reported that the generation of producer gas from biomass is not a dif-
ficult task. However, much effort is required to generate producer gas
in a form which does not require complicated and expensive gas cleaning
equipment and operates with a low pressure drop.
The present lack of reliable data with respect to the performance of
gas cleaning equipment is mainly due to measurement difficulties and to
the need for expensive monitoring equipment.
As part of this research, several very different methods of gas
cleaning have been tested and are reported by Kaupp (1982). A gas
cleaning train in general must perform several tasks:
a) Cooling of the gas to near ambient temperature.
278
II)
u;
"'>-
CD 60
0::
c
It
z
2
1-
u
c 50
::;)
w
0::
w
2i
::;)
...J
0
>
40
3 4 5 6 7 8 9 10
PRODUCER GAS FLOW RATE, m3 h-1 (25°C,I atm)
Figure 10-47. Rice hull volume reduction.
,..
u
...!::!z
...
......
z
0
u;
"'
~ 60
z
0
u
,..
"'"'...
...z
50L-----~----~-----L-----L----~----~~----~--
3 4 6 7 8 9 10
PRODUCER GAS FLOW RATE, m3 h-1 ( 25"C ,I atm l
Figure 10-48. Efficiencies of the process.
279
b) Removal of solid particles from the gas stream.
c) Removal of tar vapors and steam from the gas stream.
It is therefore somewhat unreasonable to expect a single device to
accomplish these different tasks.
During the heyday of automotive gas (charcoal) producers, a solid
particle load of 5 - 15 mg m-3 of gas at ambient conditions was an
accepted maximum standard. It was claimed that higher solid particle
loads will result in decreased engine life. It is noted that the raw
gas carries solid particles, tar vapors and steam. No regulations
existed with regard to the allowable amounts of tar vapors and steam in
the gas stream.
The following systems were tested (Kaupp, 1982):
a) Dry packed bed filter (rice hull char).
b) Dry cyclone.
c) Wet cyclone.
d) Sieve plate scrubber in combination with a dry packed bed
filter of rice hull char.
Only option d), which has proven to be a highly efficient combination,
will be analyzed in detail.
Sieve Plate Scrubber and Dry Packed Bed Filter
The system and its dimensions are shown in Figure 10-49. Cooling
alone of the gas by passing it over or through a water film is not ade-
quate for efficient gas cleaning. One of the foremost problems in gas
cleaning is the removal of very fine particles and the separation of tar
mist from the gas stream. The difficulty lies in the condensation of
280
f
VOID
SPACE
DEMISTER
50
em
RICE HULL
CHAR
~---- 20 - - - . !
em WIRE MESH
SCREEN
RAW PRODUCER
GAS
-
f
STOPPER
SIEVE PLATE, 2 mm
HOLES
DRAIN
Figure 10-49. Gas cleaning filter train.
281
the higher hydrocarbons (tar) at temperatures between 100°C and 400°C.
This tar mist will still remain in the gas stream and will partially
adhere to the wall surfaces and to dust in the gas. These tar vapors
and solid particles give the gas a foggy appearance.
There Is considerable disagreement over how clean producer gas must
be in order to give an internal combustion engine long and trouble-free
operation. However, it is a simple fact that "clean" producer gas, when
observed through a glass viewing port, is completely invisible to the
human eye. This rule of thumb immediately separates the good gas produ-
cers and gas cleaning trains from the poorly functioning systems. It is
noted that "clean" air still contains a sizeable amount of particles as
shown in Table 10-8. Comparing these figures with the requirement that
producer gas has a maximum of 5- 15 mg m-3 (at 25°C), it is clear that
reducing the particulate content of the gas to such low limits is a for-
midable task.
Table 10-8. Particulate Content of Air (SERI)
Air In mg m-3
Rural areas and suburbs 0.5 - 1
Cities 2
Industrial centers 4
Streets with heavy traffic 20
Dusty highways, farm work
with tractors over 200
The purpose of this work was to design a simple, inexpensive, and
reliable gas cleaning apparatus that does not generate an excessive
282
pressure drop. The excessive pressure drop caused by most gas cleaning
trains is a serious problem. ~ost gas cleaning devices, such as fabric
filters and packed bed filters, are able to clean the gas to a certain
degree, but the excessive· pressure drop caused by these filters makes it
necessary to install a pump between the gas producer and the engine. In
most installations, the engine itself is used to overcome the entire
pressure drop across the gas producer and gas cleaning train, which can
cause a severe drop in the volumetric efficiency of the engine and
therefore a substantial power drop (Kaupp, 1982). Taking all this into
consideration, a sieve plate scrubber followed by a dry bed filter of
rice hull char was chosen.
It is emphasized that efficient tar and fine particle removal
requires a highly atomized water spray and therefore a large surface
area and turbulent motion to impinge the contaminants. In order to
avoid the need for a high pressure water line and a pump, the gas itself
was used to induce turbulent mixing of gas and water.
Designing. sieve plate scrubbers for a given gas flow rate and
pressure drop is an old technology and treated in detail, by Treybal
(1980) and Calvert (1972). The tested scrubber, however, is not a stan-
dard design and has no water flowing across the plate. The hole
diameter, d0 , is 2 mm. The holes are located at the corners of equi-
lateral triangles at a distance d = 4 mm between centers. The ratio of
total are, A, to hole are, A0 , is then given by
A
0
-=
A
283
(Treybal, 1980). Therefore the area available for the gas flow is 22.6%
of the total cross-sectional area. In order to avoid excessive wall
flow and large bubbles, the sieve plate should be pressed against a
steel ring to get a tight fit at the wall. The entire gas cleaning
train was made from plexiglass in order to observe the motion of water
and gas. A wet sieve plate scrubber can be operated without a con-
tinuous water flow for small units although continuous water flow is
preferred because it enables cooling of the gas to near ambient tem-
perature. The water temperature of the wet scrubber should never reach
the boiling point because the repulsive forces of boiling water will
actually repel the particles instead of collecting them.
The tests reported here were conducted without a continuous water
flow. This is one reason why the space below the sieve plate is rela-
tively large even though this part does not contribute to gas cleaning.
The apparatus was also tested with a continuous water flow during the
gasification of corncobs in the unit shown in Figure 10-28 (Kaupp,
1982).
Experimental Procedures and Results
The principle of operation was as follows:
~efore starting the gas producer the sieve plate scrubber is filled
with water to a level 1 - 3 em above the sieve plate depending on the
desired pressure drop and on the gas flow rate. After the gas producer
was started, the incoming raw gas is finely divided by the sieve plate
into many small bubbles. Above the sieve plate a highly turbulent mix-
ture of water and gas (froth) is formed which provides an excellent
284
environment for impinging tar droplets and solid particles. At the same
time the gas is cooled. T,o/'hen the gas leaves the foam bath it contains
entrained water mist which will be partially eliminated in the packed
bed filter. The packed bed filter is purposely located above the wet
scrubber in order to allow drainage of the collected water back into the
wet scrubber. As the saturated gas cools, the moisture in the gas con-
denses and creates a film covering the rice hull char, which provides
additional particle removal in the dry bed. Gas sampling was done
before and after the gas cleaning train. The continuous gas sampling
before the gas cleaning train was used to determine the moisture content
of the raw gas at the gas exit and the gas composition. The continuous
gas sampling after the gas cleaning train was used to determine the
moisture content and the solid particle content of the gas.
Consequently, the following measurements were taken:
a) Moisture content of the raw gas at the gas exit of the gas pro-
ducer.
b) Moisture content of the gas after leaving the gas cleaning
train.
c) Particle content and liquid higher hydrocarbon content con-
densing out above 105°C.
d) Pressure drop across the wet sieve plate scrubber.
The tar and solid sampling was done according to EPA Method 115,
which is used to sample the contaminants in stack gases.
285
Moisture Content of the Raw Gas Before the Gas Cleaning Train
Two sets of tests were performed:
a) The gas was sampled with a stainless steel tube reaching into
the middle of the gas producer (shown in Figure 10-28). The
stainless steel tube was located 5 ern above the wiper grate.
These measurements can therefore be considered a true measure
of the water content of the raw gas as a function of tern-
perature and producer gas output. The results are given in
Figure 10-50.
It is noted that the moisture content data are reported as
grams per m3 of gas at 25°C, and as percent volume. The latter
representation was obtained by assuming that the ideal gas law
holds. If we let m be the mass of water condensed from each
m3 (measured at 25°C) of producer gas in an ice bath, then this
water would occupy a volume of
vw
where p 1 atm and T > 100°C.
3 T • 1
Each m of producer gas occupies a volume Vg =~at temperature T.
vw
Consequently, the water occupies a volume fraction £
vg + v • The
w
experimental results confirm the general trend outlined in Section 9.4.
The only sources of steam in the gas producer arise from the chemically
bound water in the fuel, the natural moisture present in the rice hulls,
286
and the water generated by the water shift reaction. The moisture con-
tent of the air has been ignored. Because almost no steam conversion
takes place at lower temperatures (< 700°C), a significantly higher
water content in the raw gas must be expected as shown.
Moisture content of the raw gas obtained from the design shown in
Figure 10-34.
In this test series, the water content of the gas was measured as it
left the gas exit, before entering the wet scrubber. It is noted that
this data is heavily influenced by the following:
a) The temperature at the gas exit.
b) The temperature at the bottom of the gas producer column.
c) The height of the hydraulic water seal, and water evaporated
from it.
The results are given in Table 10-9.
Table 10-9. \-later Content of Raw Gas
Gas Flow Water Content
m3 h-1 -3
s m % volume
3.76 152.6 17.1
4.97 134.6 15.4
5.74 207.5 21.7
7.46 166.2 18.2
8.68 133.0 15.3
9.9 142.2 16.2
In general, one can expect steam contents of 15% - 20% by volume in the
hot raw gas. With respect to the "true" steam content of the raw gas as
287
shown in Figure 10-50, it is of interest to note that the steam content
decreases with increasing fire zone temperature caused by higher air
input rates.
The approximate temperature of the upward moving fire zone as
observed in the refractory insulated design (Figure 10-28) is shown in
Figure 10-51. The temperatures were measured 1 em away from the wall
and should be viewed as approximate values because thermocouples
reaching into a rice hull bed disturb the fuel bed structure. Due to
these observed temperatures, the design shown in Figure 10·-34 (which
does not have a refractory walt) uses a stainless steel tube as inner
cylinder.
Moisture Content of the Gas After the Gas Cleaning Train
Producer gas leaving a downdraft gas producer will always be
saturated with moisture at temperatures below 100°C. It is therefore of
interest to compute the amount of water released when cooling producer
gas. The specific humidity, w, is rtefined as the ratio of the mass of
water vapor to the mass of noncondensihle gas. Assuming the ideal gas
law holds, we have:
where Mv and Mg are the molecular weights of water and producer gas
respectively. The molecular weights of the measured producer gas are
within the range of 26.8 to 27.9 as shown ln Table 10-6. An average of
288
..1: ''
''
w
liE
:::>
.,. '' ..J
...z
0
' ' .....---- >
f!
...wz ....................... ...z
0
u
.................. .! ...wz
w
a: ........................... 0
u
...
:::>
w
a:
...
Vl
:::>
0
liE Vl
0
liE
90 3~----..J4------~5------6L-----~-----8~----..J 10
9
PRODUCER GAS RATE,m 3 h- 1 ,(25"C,Iatm)
Figure 10-50. Steam content of the raw gas as a function of gas

flow rate.
w
z
0
N
w
a:
;;:
....
0
w
a:
...:::>
"'a:w
a..
liE
...w
4 5 6 8 9
PRODUCER GAS RATE,m3h- 1 ,(25"C,Ialm)
Figure 10-51. Approximate temperature of upward moving fire zone.
289
p
v
27.1 is assumed and therefore w 0.664 p
g
The relative humidity, ¢, is defined as the ratio of the actual mass of
vapor to the mass of vapor required to produce a saturated mixture.
pv
Therefore ¢ = --- where p is the saturation pressure of the water vapor
Ps s
w p
g
at the mixture temperature. Consequently ¢
0.664 • ps
Because no vapor pressure curves could be found for producer gas-water
mixtures, the curves for air-water mixtures are used. The computational
error is probably not large since both air and producer gas consist
mainly of nitrogen.
The mass of water vapor carried at ¢ 1 and T 50°C in a producer
gas mixture is therefore:
mass of vapor
mass of mixture
p
-2
Since ¢ = ~
p
it follows that P v p 12350 N m and therefore
s
s
p. ·= (1.01322)(10 5 ) - 12350 88972 N m- 2

g
(0.664)(12350)
Then w 0.092
88972
and mass of vapor 0.0842

mass of mixture
When cooling down this mixture to 30°C we have mass of vapor 0.029
mass of mixture
290
Assuming that the mass of gas is preserved during the cooling process,
it is concluded that 1 kg of saturated producer gas at 50°C will release
56.9 grams of water when cooled to 30°C. This calculation demonstrates
that wet scrubbers should operate at near ambient temperatures. It is
noted that the gas leaving a wet scrubber is not only saturated but also
contains water mist.
The moisture content of the gas leaving the sieve plate scrubber and
dry filter arrangement is given in Table 10-10.
Table 10-10. Moisture Content of Producer Gas

after the Gas Cleaning Train
Flow Rate Moisture Content
m3 h-1 m-3 % volume (> l00°C)

~
2.76 71.59 8.9%
4.97 108.13 12.8%
7.46 213.3 22.2
8.1)8 170.3 18.6
9.9 180.0 19.4
It is noted that the mass of moisture in the gas in Table 10-10 refers
to the amount that was condensed out when cooling the gas from the dry
gas filter exit temperature down to l3°C - l7°C. The results indicate
that at low flow rates the highly absorbent rice hull char can eliminate
a significant amount of moisture. For instance, the dry char filter
material, which weighed 1037 gram before the test, picked up 577 grams
of water at the 4.97 m3 h-l flow rate.
291
Particle Content of Producer Gas
The particle content of the gas was measured downstream from the
various gas cleaning arrangements, which are listed as follows:
a) Sieve plate scrubber only with no dry bed filter, operated at a
gas flow rate of 7.46 m3 h-l (data point 1).
b) Sieve plate scrubber followed by a dry bed filter filled with
rice hull char (30 em height). A stainless steel wire mesh
demister was placed on top of the rice hull char. Although the
steel wire mesh demister does not aid in gas cleaning, it does
prevent rice hull char from being entrained in the gas stream.
The gas flow was 8.68 m3 h-l and 9.9 m3 h-l (data points 2 and
3).
c) Sieve plate scrubber followed by a dry bed filter filled with
compacted rice hull char (30 em height) and the demister on
top. The gas flow rate was 4.97 and 3.76 m3 h-1 (data points 4
and 5 respectively).
All data for particle content are given in Figure 10-52.
292
"' I
•
I
...EE
...z
...z
3
loJ
•
0
...
u
~ 5 5
0 •
2
4
• •
GAS FLOW RATE, m3 h-1 ( 25°C ,I otm I
Figure 10-52. Dust content of gas after gas filter train.
The following conclusions can be drawn:
a) A packed bed of rice hull char will further reduce the solid
particle content of the gas leaving the wet sieve plate
scrubber. This is not obvious if one considers the fact that
rice hull char contains many fine particles that could be
entrained into the gas stream.
b) The particle load in the gas stream was very low for all tested
gas flow rates and was well under the accepted maximum for
engine applications.
The appearance of the fiberglass filter papers used for sampling is
shown in Figure 10-53. All measurements were done with a precision
balance with an accuracy of 10- 5 grams. Due to humidity in the environ-
ment, the mass measured contains an uncertainty factor of ~ 3 mg. This
293
1 2
3 4
Figure 10-53. Appearance of fiberglass filter papers for
total collection of 1) 2.5 rng; 2) 4.3 rng;
3) 17.4 rng; 4) 33.5 mg .
294
mass represents the amount of moisture a filter paper carries when
exposed to ambient air.
Pressure Drop Across the Wet Sieve Plate Scrubber
The pressure drop across the sieve plate scrubber is strongly depen-
dent on the following:
a) The dry plate pressure drop.
b) The height of the water above the sieve plate.
c) The elimination of producer gas flow between the rim of the
plate and the wall of the sieve plate scrubber.
The pressure drop across the wet scrubber was kept at 2.5 - 5 em H2 0 at
all flow rates. The pressure drop can be easily regulated by adjustment
of the water level above the sieve plate ( 1 - 3 em). The generated
foam height (2 - 8 em) was sufficient to clean the gas as shown.
Summary
The combination of a wet sieve plate scrubber followed by a dry
packed bed filter has proven to be a very successful gas cleaning train.
However to avoid large amounts of water being carried out either in mist
or vapor form it is suggested the sieve plate scrubber be operated at
low temperatures. This may either require very low flow rates for a
given cross section or a modest continuous water replacement.
295
10. LIST OF SYMBOLS
A total area of sieve plate (m 2 )
A0 hole area of sieve plate (m2)
Cn drag coefficient
d diameter of gas producer, particle diameter, pitch of sieve plate

holes.
d0 diameter of sieve plat~ holes (mm)
E volume fraction occupied by the char
g standard gravitational acceleration, 9.81 m s-2
h distance traveled during time t* (em)
Mg molecular weight of producer gas
Mv molecular weight of water
m mass of water condensed out of the gas stream (kg)
me mass of residue (kg)
mf mass of rice hulls (kg)
Pg partial pressure of the gas (N m-2)
Ps saturation pressure of water vapor in alr (N m-2)
Pv partial pressure of water vapor (N m-2)
R rice hull consumption rate (kg h-1, wet basis)
Re Reynolds number
T temperature (°C)
T1 fuel bed temperature CC)
t* time n~eded to reach 99% of the terminal velocity ut (s)
u(t) particle velocity as a function of time (m s- 1 )
ut terminal particle velocity (m s-1)
v8 superficial gas velocity (m s-1)
296
va actual gas velocity (m s-1)
volume at T 100°C of collected water vapor (m3)
volume at T 100°C of producer gas (m3)
xc mass fraction of carbon in rice hull residue (kg)
Xf mass fraction of carbon in rice hulls (kg)
£ volume fraction occupied by water in a gas-water vapor mixture
£s void space fraction
f:,z height of control volume V (em)
w specific humidity
~ equivalence ratio and relative humidity
n pro d ucer gas v i scos i ty (kg m-1 s-1)
T carbon conversion
297
10. REFERENCES
1. Allen, T. 1974. Particle Size Measurements. Halsted Press, New

York.

Agriculture Services Bulletin #31. Rome.
3. Beagle, E. 1983. Personal communication.
4. Groeneveld, M.J. 1980. The co-current moving bed gasifier.

Thesis, Technische Hogeschool Twente, The Netherlands.
5. Calvert, s., et. al. 1972. Wet Scrubber System Stl.!dy. Volume I:
Scrubber Handbook. NTIS, Springfield, VA 22161.
6. Garbarino, G. 1911. Utilizzazione integrale della lolla di riso

come combustibile.
7. Harter-Seberick, R. 1937. Motor gas-producers and their fuels,

their state of development with special reference to the use of
fossil fuels. Fuel 16(1). London.
8. Heiss, F. and J. Coull. 1952. Chemical Engineering Progress

48(3) :133-140.
9. Jenkins, B.M. 1980. Downdraft gasification characteristics of

major California residue-derived fuels. PhD Thesis in
Engineering, University of California, Davis.
10. Kaupp, A. 1982. Technical reports Ill - 14 to the Briggs and

Stratton Corporation.
11. Peiyi, Jia, et. al. 1982. Utilization of agricultural wastes for
energy conversion and product processing. FAO Bangkok Report
RAPA No. 61.
12. Rambush, N.E. 1923. Modern Gas Producers. Van Nostrand Company,
New York.
13. Reed, T. 1983. Personal communication. Solar Energy Research

Institute, Golden, Colorado.
14. "Buck" Rogers Company. 1983. Personal communication.
15. Reines, R. 1982. Personal communication. Open University at

Milton Keynes, Great Britain.
16. Schillar, L. and A.Z. Nauman. 1933. Verein Deutsche Ingenieure

Journal 77:318.
298
17. Schlapfer, P. and J. Tobler. 1937. Theoretische und praktische
Untersuchungen uber den Betrieb von Motor Fahrzeugen Mit Holzgas.
Schweizerische Gesellschaft fur das Studium der ~otorbrennstoffe,
Bern.
18. SERI. 1981. Gengas, the Swedish classic on wood-fueled vehicles.

Tipi t.ror'kshop Boo'ks, Allenspark, Colorado 80510.
Publisher Inc., Olympia, Washington.
20. Treybal, R. 1981. Mass Transfer Operations. McGraw-Hill

Publishers, New York.
21. IHlliams, R.O. and B. Horsfield. 1977. Generation of low-BTU

fuel gas from agricultural residue experiments with a laboratory-
scale gas producer.
299
INDEX
PAGE
Air injection 243-245

Air-to-fuel ratio 260
Ash analysis 125
Ash removal syst~m 217,221,224-236
Balestra updraft gas producer 216-220

Biomass ultimate chemical analysis 16
Blot 1\lumber 172
Carbon conversion efficiency 275

Channel formation 134
C-H composition of fuels 162
Characteristic length 171
Charcoal production 162-163
Char load in raw gas 93
Chemical energy 3
Chemlcally bound water 179
Chemistry of gasification 50-98
Chinese rice hull gas producer 221-224
Combustion 10
Computer programs 64-73
Condensable products 63
Control volume 100
Cyclone 279
Degree of rice hull conversion 273

Design considerations 213-215
Oevolat llizat ion 55
Diffusion flame 160
Dispersion tensor 84
Dry bed filter 280
Efficiency of a gas producer 277

Energy equation 86-87
Equilibrium constant 51-62
Equivalence ratio A 90
Equivalence ratio ~ 3,261
Ergun equation 138
Flame temperatures
coruputed 87-94
measured 94
Formation reactions 51
Free radicals 165
Gas cleaning train 7,277-297

Gas coruposition 262,267
Gas exit design 237-242
300
PAGE
H/C ratio of producer gas 10
Health hazards 167
Higher heating value of producer gas 76,268,270
Higher hydrocarbons 166
History 22-49
Ignition Advancement 9
Internal combustion engines 8-10
Intrinsic gasification rate 59
Light oils 165

Linear adsorption isotherm 85
Measurement accuracy 116
Mechanis~ of pyrolysis 165
Melting point
minerals 120
rice hulls 124
biomass fuels 124
~ethane content 264-265
Moisture content
r'iw gas 285-287
rice hulls
Molten Silica 129
Open core gas producer 251-260
Parameters involved in gasification 17

Particle cont~nt, producer gas 2<n-292
Pellets 129,131,158
Phase diagrams
Power output 12
J:>resure drop
system 132-135
sieve plate 293
Products of combustion 91
~roximate '!nalysis 56
Pyrolysis 2,160-133
char composition 175
char heating value 181
experimental set up 169
governing theory 165-166
gas composition 173
products of, 161
product distribution 177
process efficiency 181
rapid (flash) 64
test apparatus 173
301
PAGE
Range of gases so
Rapid Pyrolysis 164
Reaction zones 53
Reversible reactions 59
Rice hulls
apparent density 107-110
bulk density 103,111
ash content 50,56,68,106
caking of, 118-121 '130-131
char 135,140-142,1 49,175-182
cross section 114
fixed carbon 56
flow properties 148-149,246-2 56
heat-up period 170-172
higher heating value 181
inner surface 152
micrographs 148-159
moisture content 73,199
outer surface 150-154
particle diameter 139
pore structure 100,110,158
proximate analysis 56
weight of, 115
rate of consumption 273
shape 114
size distribution 141-143
size reduction 156,215,275
slagging of, 121-131
surface area 112-114
surface roughness 113,138
tar formation 162,168,179
true density 103-106
ultimate chemical analysis 106,180
volatile matter 175-177,181,2 76
Sieve plate scrubber 280

Softening temperatures, biomass 124
Soot formation 11 '160
Species concentration equation 80-85
Specific gasification rate 271
Steam dissociation 198-209
degree of conversion 206
gas composition 208
models 207-209
Superficial gas velocity 132-137,269
302
PAGE
Tar
conversion efficiency 197-199
conversion mechanis~ 186
.cracking in downdraft gas producer 184,187-199
energy content 13
experimental set up for cracking 192-196
measured in producer gas 185
production 164
r~lease during pyrolysis 164-166
re111oval 13
thermal cracking 13
Terminal velocity of particles 242
Updraft gas producer 57
Velocity
char
fire zone 262
fuel bed 262
profile 136
Volume averaging 81
Volume reduction of rice hulls 275
303

(Albrecht Kaupp PHD (Auth.) ) Gasification of Rice

Uploaded by

Copyright:

Available Formats

(Albrecht Kaupp PHD (Auth.) ) Gasification of Rice

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Albrecht Kaupp PHD (Auth.) ) Gasification of Rice

Uploaded by

Copyright:

Available Formats

Deutsches Zentrum fiir Entwicklungstechnologien - GATE

Deutsches Zentrum fUr Entwicklungstechnologien - GATE - stands for German Appro-

Deutsche Gesellschaft fiir Technische Zusammenarbeit (GTZ) GmbH

Deutsches Zentrum fiir Entwicklungstechnologien - GATE

Gasification of Rice Hulls

Springer Fachmedien Wiesbaden GmbH

CIP-Kurztitelaufnahmc der Deutschen Bibliothek

All rights reserved.

This book about gasification of rice hulls is based on

However, it must be said that, based on a great deal of

1. An open core downdraft gas producer with no throat and a

2. Gasification of rice hulls in pelletized form. Although the

Option 1 is in the authors opinion the simplest and most

Eschborn, February 16, 1984 Albrecht Kaupp

he spent in talking with me about his vast experience in the utilization

of rice hulls. Professor Stephan Whitaker, Department of Chemical

Engineering, for his concise, uncompromising teaching and treatment of

tical parts of this research.

My thanks are extended to the Carl Duisburg Gesellschaft in West

Germany which financed part of this research. The Weyerhaeuser Company

and their two-year scholarship which allowed me to work independently.

Special thanks go to the Briggs and Stratton Corporation, Doug Janish

years made this project possible.

The help of Bart Duff, International Rice Research Institute,

Philippines, to secure adequate funding was greatly appreciated.

Kurt Creamer, Graduate Student, who assisted me in many of the experi-

again by Karin Clawson.

My greatest appreciation goes to Filiz who hates rice hulls, but

certainly does understand the author.

2. OBJECTIVES AND SCOPE 19

3. HISTORY OF GAS PRODUCER ENGINE SYSTEMS 22

5. CHEMISTRY OF GASIFICATION OF RICE HULLS 50

6. PHYSICAL PROPERTIES OF RICE HULLS 99

7. PHYS!CAL APPEARANCE OF RICE HULLS UNDER THERMAL

8. LOW TEMPERATURE ENERGY CONVERSION OF RICE HULLS 160

9. TAR CRACKING IN A RICE HULL AND RICE HULL PELLET FUEL

1-1 Energy fractions in gaseous components as a function

1-2 Ignition advancement versus hydrogen content of

1-3 Soot formation as a function of H/C and 0/C ratio 11

1-4 Soot formation as a function of H/C ratio 12

1-5 Power output as a function of ~ 12

1-6 Ultimate elemental analysis on an ash and moisture

1-7 Block diagram of parameters involved in the

5-l Co-current or downdraft gasification 52

5-2 Accumulative mass loss curve 54

5-3 Differential mass loss curve 54

5-4 Differential thermal analysis 55

5-5 Counter-current or updraft gasification 57

5~6 Equilibrium of the water shift reaction as a function

5-7 Kp(T) as a function of T 62

5-8 Range of computed gas compositions as a function of

5-9 Gas composition, ~ 0, M 0 66

5-10 Gas composition, ~ 0.1, M 0 66

5-ll Gas composition, ~ 0.2, M 0 67

5-12 Gas composition, ~ 0.3, M 0 67

5-13 Gas composition, ~ 0.4, M 0 68

5-14 Gas composition, • 0.5, M 0 68

5-15 Gas composition, • 0.3, M 10 69

5-16 Gas composition, • 0.3, M 20 69