Fuel Processing Technology 144 (2016) 145–154

Contents lists available at ScienceDirect

Fuel Processing Technology

journal homepage: www.elsevier.com/locate/fuproc

Kinetic modeling and simulation of throated downdraft gasifier

Ozgun Yucel ⁎, Mehmet Alaittin Hastaoglu
Gebze Technical University, Chemical Engineering Department, 41400, Gebze, Kocaeli, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: An imbert ‘throated’ gasifier which has hourglass heart and varying axial area and nozzles for injection of gasifi-
Received 31 August 2015 cation agents has been studied. The transport of reacting gas-solid two-phase mixture through the gasifier has
Received in revised form 16 December 2015 been mathematically modeled transiently. Downdraft gasification in such a complex system has so many
Accepted 18 December 2015
variables and is quite cumbersome to describe. For the first time throated gasifiers are modeled as realistically
Available online 4 January 2016
as possible. The resulting set of transport, structural, kinetic and auxiliary equations was solved via numerical
Keywords:
methods. The modeling work was experimentally validated using a 10 kW gasifier. The actual geometry of the
gasification experimental setup was used in the model. The model results were in line with the experimental results. Throat-
imbert ed combustion zone causes better distribution of heat and reduces heat loss. With the same core size and solid
downdraft consumption, stratified gasifier output was 24.5% less than throated one. The model exit had 25.85% CO;
modeling 18.25% H2; 7.84% CO2; 2.9% CH4; while the experiment had 35.07–20.77% CO; 18.30–13.66% H2; 13.68–5.95%
simulation CO2; 6.7–1.2% CH4; the rest was N2. The model can be applied to different geometries.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction gasification process involves heterogeneous char reactions that occur

Gasification is a thermochemical process at elevated temperatures
(500–1400 °C) and it uses heat to transform biomass or other carbona-
ceous material into a combustible gas. The product gas mainly consists
of hydrogen, carbon monoxide, carbon dioxide, methane, water vapor,
nitrogen, traces of higher hydrocarbons such as ethane and also various
contaminants such as char particles, ash, tar and oil. The fixed bed gas-
ification systems are classified as updraft, downdraft, cross-draft and
two stage gasifiers according to the method of contacting fuel, direction
of air/gas movement, and types of bed. The fixed bed gasifier, which
may be updraft or downdraft, is simple and generally operates with
high carbon conversion, long residence time, and low gas velocity [1].
The Imbert downdraft gasifiers are suitable to handle biomass fuel
having ash and moisture content less than 5% and 20% respectively
[2]. Downdraft gasifiers usually yield syngas with low tar content
(1–2 g/Nm3) and are suitable for small-scale applications (b 10 MWth).
On the other hand, composition of syngas varies with feedstock and
design parameters.
A chemical process can be modeled with both equilibrium and
kinetic non-equilibrium. Equilibrium models predict the maximum
achievable conversion and the distribution of each species in the prod-
uct streams is subject to thermodynamic constraints independent
from “geometry”. Also equilibrium models do not estimate the differ-
ence between the time needed to reach equilibrium and residence
time. They assume that all reactions reach chemical equilibrium. But
⁎ Corresponding author.
slowly. Kinetic and transport models deal with heat, mass, momentum
and kinetic equations and give the concentrations at any point in time
and space within the system. These models can provide valuable infor-
mation about reaction mechanisms to increase the conversion to
syngas.
Downdraft gasification produces a syngas with very low tar content
and is simple and reliable. For a downdraft gasifier, biomass moves co-
currently with the gas, and passes through drying, pyrolysis, oxidation
and reduction zones. Several downdraft gasifier models are available
in the literature but all of them are defined for stratified (without
throat) gasifiers due to complexity of imbert design, and also they inject
air at the top of the gasifier not at the top of the throated region. On the
other hand, most of them are at steady-state, and assume the existence
of stable zones. Especially in the char reduction zone, usually the
reactions are assumed to run to completion. But the actual conversion
depends on air–carbon ratio, solid residence time and other kinetic fac-
tors, so that equilibrium models overestimate the H2 and CO yields and
underestimate CO2, methane, tar and char, especially for relatively low
gasification temperatures. Sharma [3]. stated that at critical char bed
length and the critical reaction temperature, equilibrium models can
be useful but these phenomena can be modeled depending on initial
composition and initial temperature.
Gasification process can be divided into four steps: biomass drying,
primary and secondary pyrolysis, combustion of products and reduction
of H2, CO2 and H2O with char. Early kinetic models have centered on
(devolatilization) wood or cellulose conversion to gas, char and tar [4].
Later some researchers studied coal gasification and gave valuable
information by combining heterogeneous char reaction kinetics and

http://dx.doi.org/10.1016/j.fuproc.2015.12.023
0378-3820/© 2015 Elsevier B.V. All rights reserved.
146 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154
transport phenomena [5]. For the degradation of biomass three-step
and one-step mechanisms are proposed in order to predict particle
mass loss [6–9]. Also catalytic effect of char on the heterogeneous reac-
tions in the reduction zone has been studied [10,11]. First dynamic
model for stratified gasification was presented by Di Blasi [12]. without
taking into account the pressure drop along the bed (although they
have modeled the pressure drop it has not been used in simulations).
Tinaut et al. [13]. built a steady state model also including hydrocarbon
reforming (tar and methane reforming). Di Blasi and Branca [14]. pro-
posed a model for the open-core downdraft gasification of wood pellets,
which permits a dual air entry: from the top section and at a certain
height of the packed bed. But they studied top-stabilized front (second-
ary air up to 50%) and double-front stabilization (secondary air up to
60–70%). In reality, air is just introduced at an intermediate level for
imbert gasifiers. A very small amount of air could enter from the top
when solid fuel is introduced into the gasifier at the top which is
negligible for downdraft gasification. Furthermore, these models do
not describe transversal variation of throated gasifier.
Fixed bed gasifiers have surrounding jacket in order to minimize
heat loss. There is no work in the literature that evaluates the energies
of the inner wall and the gas exiting from the reactor passing thorough
the jacket which are important in reality. Dynamic behavior of the
downdraft gasifier cannot be examined realistically without considering
these energies. In this work the wall and exit gas are also considered in
the set of energy equations of the downdraft reactors. In the present
study the mathematical model is presented for gasification of the
wood pellets in an imbert gasifier which has hourglass heart shape
and permits the air feed at intermediate levels of the reactor. Also exper-
iments have been performed to validate the model.
Modeling work on stratified gasifier was started similar to a previous
study [14] where gas flow was calculated based on conservation of total
gas mass. However, in this work, we have changed it completely by
using momentum transfer approach, and since we had used a circula-
tion pump in the experimental system, we have approached the reality
more closely. We have changed all the parameters by using the actual
system values and we have tried to develop as a realistic model as pos-
sible. The gas velocity is evaluated from continuity equation in general,
but in our system it is evaluated from the pressure to simulate vacuum-
driven system. The energy equation for the inner wall was also solved.
Additionally the exit gas was modeled as it leaves the packed bed
zone while traveling between the inner and outer reactor walls. All
the transport equations including water–gas shift, tar reforming and
methane reforming reactions were solved for the exit gas as well. This
meant that the relevant PDEs were solved simultaneously for all the
gaseous species. Thus the cold gas concentrations at the exhaust of the
gasifier were determined.
2. Experiment
Several experiments were conducted on a 10 kW downdraft
fixed-bed gasification system which is developed by All Power Labs
Inc. The scheme for the gasification system is illustrated in Fig. 1. The
reactor is cylindrical with an internal diameter of 28 cm and a height
of 55 cm. At the inside, however, the reduction cone tapers to 20 cm
and the heart cone tapers to 7 cm in order to increase solid residence
time in the pyrolysis and combustion zones. Gasifier has a throat
where the unit narrows below the flame zone so constriction restricts
the tar content in the produced gas by forcing the volatiles to pass
through the combustion zone. In the reduction zone the diameter
expands to 19 cm then 23 cm and 38 cm respectively. Flue gas exits
through a 21 cm equivalent diameter.
Temperatures at 6 different heights inside the gasifier were
measured with a group of 6 in-line K-type thermocouples with 8 mm
diameter. Two groups of thermocouples were located at the center
and 5 cm off center. A perforated iron grate is installed at the bottom
of the gasifier and the grate was shaken at adjustable intervals in
order to dispose the ashes continuously from the gasifier and to prevent
bridging and dead zones inside the gasifier. We have designed a pneu-
matic eccentric shaking system for this purpose. The gasifier is filled
with batches of charcoal from the grate up to air injection nozzles and
then followed by the wood pellets. Air is used as an oxidizing agent
for biomass gasification. The top of the gasifier is sealed to prevent
(primary) air inlet, pressure drop also prevents the primary air from
entering the filled gasifier. The gasifier has 5 (secondary) air injection
nozzles just above the hearth of the gasifier.
The experimental set-up has a limitation of working with only few
kPa of vacuum (around 1–2 kPa in this work). This is applied to the
gasifier in order to provide gas flow through the gasifier by an air
compressor connected with the flue gas. More vacuum would cause
low quality gas since the reduction reactions will not have adequate
time for completion, less vacuum wouldn't let enough oxygen to get
into gasifier so gasifier temperature would drop dramatically: either
way detrimental to the operation of the gasifier.
The gasifier was ignited by a butane torch from the ignition port
until white flue gas is seen at the outlet. The gas obtained from the gas-
ifier enters the cyclone and packed bed filter after traveling between the
inner and outer walls of the gasifier. Particulates are removed while
passing through the cyclone and the tar is collected by packed bed filter.
Afterwards, a portable infrared syngas analyzer, Wuhan Cubic Syngas
Analyzer Gas board 3100P, was used for the measurement of the con-
centrations of CO, CO2, CH4, H2 and O2 in the sample gases simulta-
neously while heating value (calorific value) is calculated
automatically by an NDIR and Thermal Conductivity Detector. Gas is
also collected and analyzed using a GC-7890A. Finally, the flue gas is
burnt out with a swirl burner. Otherwise this gaseous mixture can be
used to generate energy at gas engine or separated in order to obtain
H2, CH4, CO etc. via membrane. Flue gas also can be used to manufacture
of ammonia, methanol, and other valuable chemicals.
Wood pellets used in the experimental runs had according to the
elemental analyses 50.67% C; 6.18% H; 2% N; 0.18% S and the rest was
O. The cylindrical pellets were roughly between 1 and 1.5 cm in size.
Other parameters can be seen in Table 1. While the feeding rate of the
fuel was determined from the loading and operating time; the flow
rates of the air injected and product gases were calculated from the or-
ifice plate measurements. During the experiments flow rates of the
gases and the pressures at the gasifier exit, air injection point and the
pressure at the filter exit were recorded. Vacuum is applied at the latter
point. Additionally temperatures were recorded at the selected points.
3. Mathematical model
Gasification operation is not a steady state process. For example,
fixed bed gasifiers are filled with fixed amount of charcoal to startup
and ignited with blowtorch for a while. Also, there is a fluctuation in
fuel and air consumption. On the other hand, air input is a function of
pressure in the gasifier. Therefore the gasifier must be modeled as tran-
sient. The following partial differential equations have been derived
typically comprising of conservation equations in the gas and solid
phases for a one-dimensional unsteady state system. Within the bed
1-D modeling is used considering variable cross-sectional area of the re-
actor. At all grid points, contributions of the reactor walls were
accounted for in the energy balance equation and at the air injection
grid air is considered both in mass and energy balance equations.
The variation of the cross-sectional area in the throated section leads
to varying reactor diameter which is embedded in the cross-sectional
area Ac. Note that the reactor diameter is directly used in the calculation
of various transfer coefficients such as overall wall heat transfer coeffi-
cient as well as particle heat and mass transfer coefficients via Reynolds
number.
The model describes the processes of drying, pyrolysis, combustion,
and heterogeneous char reactions. For simplification, the particles fed
are assumed to have uniform size and shape. As the reaction continues,
 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154 147

Fig. 1. Gasifier geometry.

although their size decreases, they keep their shapes. The internal diffu- mechanism sudden collapse of the bed is prevented and continuity is
sional resistances of the particles are neglected. Porosity of the bed is as- provided in the structure of the bed and the operation.
sumed constant along the reactor. Note that although particle size is Conservation of wood which flows is a quasi-one-dimensional con-
decreasing during the process, the number of particles per unit volume tinuity equation
increases along the reactor keeping the porosity fairly uniform and
constant. ∂ðρs Ac Þ ∂ðρs Ac us Þ m
Another important point is in actual reactors bridging and dead ¼ þ Ac ∑ ϑ sj r j ð1Þ
∂t ∂z j¼1
zones occur, which collapse in an erratically oscillating cyclic manner.
However due to the introduction of pneumatic eccentric shaking
The first term is the rate of change of solid mass with respect to time.
The second term represents the downward convection of particles and
the last term is the rate of consumption of the wood due to pyrolysis.
Table 1
For the species in the solid and gas phases ρ is defined as a partial
Operating conditions.
density. Here us is superficial velocity which is the volumetric flow
Property Value rate divided by the cross-sectional flow area. In multiphase flow, it is
Bed diameter 0.1 m not true velocity (physically real) but it is a fictitious quantity which is
Porosity 0.5 convenient for parameter estimation such as transfer coefficients.
Initial particle diameter 0.01 m Conservation of moisture which is considered to be a part of the solid
Initial solid density 416 kg/m3
phase is
Solid mass 3.5–4.0 kg/h
Air-fuel ratio 1.2 kg/kg
Moisture 7.28 wt.%
Ash content 1 wt.% ∂ðρm Ac Þ ∂ðρm Ac us Þ m
Char ratio 0.32 wt.% ¼ þ Ac ∑ ϑmj r j ð2Þ
∂t ∂z j¼1
148 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154

Conservation of char is Conservation of mass for the exiting gas is

∂ðρc Ac Þ ∂ðρc Ac us Þ m
∂ρei ∂ðρei ue Þ m
¼ þ Ac ∑ ϑ cj r j ð3Þ ¼ þ M i ∑ ϑij r j ð10Þ
∂t ∂z j¼1 ∂t ∂z j¼1

The following differential equation describes conservation of gas
species flowing through void region with contribution of intermediate
level injection at the z0 position for nitrogen, oxygen and water content
of air. The Dirac Delta function is a function that is zero everywhere
except one point ∫L0δ(z - zi)dz = 1 where i represents O2, CO, CO2, H2O,
CH4, H2, T1, T2, N2 and k = O2, H2O, N2.
The energy equation for the outer wall of the gasifier is not solved
but the natural convection at the reactor outside is taken into account
to estimate heat loss when solving energy of the product gas. Heat
transfer rates from solid to gas, solid to wall, gas to wall, wall to product
gas and product gas to ambient air are respectively:


q000 sg ¼ hsg Ap vp T s  T g ð11Þ

q000 sw ¼ 4hsw ðT s  T w Þ=D ð12Þ

∂ðρi Ac Þ ∂ ρi Ac ug k m
ε ¼ þ ∑ ρi0 ui0 Ai δðz  z0 Þ þ Ac M i ∑ ϑ ij r j ð4Þ
∂t ∂z i¼k j¼1

q000 gw ¼ 4hgw T g  T w =D ð13Þ

The net rate of change in the gas mass is shown below and it is used q00 we ¼ hew ðT e  T w Þ ð14Þ
to estimate total gas density where W _ gi is mass flow per unit length for
specific gas specie which is ρgiAcug. q00 ea ¼ hea ðT e  T a Þ ð15Þ
Conservation of total gas mass is
qij ¼ qji ð16Þ
   
∂ ρg Ac ∂ ρg Ac ug n m
ε ¼ _ gi δðz  z0 Þ þ Ac M i ∑ ϑij r j
þ ∑W ð5Þ
∂t ∂z i¼1 j¼1 3.1. Auxiliary equations

The solid phase in the gasifier is composed of different species such We apply vacuum to the gasifier in order to provide one-directional
as biomass, moisture, char, ash and the energy equation of the total solid flow. This vacuum condition is not the same at every point along the
phase is derived from the enthalpy form of the energy conservation of gasifier. In addition to contributions of changes in the number of
all solid species. Solid temperature is determined from its enthalpy moles and temperature, pressure drop in the bed is caused by friction,
cpi(Ts − T0). The conservation of energy for the solid phase is form drag, expansion, contraction, etc. This pressure drop is calculated
from Ergun's equation. Thus we can calculate the vacuum and primary
n   n _ sii H si flow rate at the top of the gasifier. Void fraction of the feed at the top
∂ ∑i¼1 W i H si ∂ ∂T s ∂ ∑i¼1 W m
¼ Ac k s þ þ ∑ ΔH j r j is estimated and the voidage is assumed to be constant within the reac-
∂t ∂z ∂z ∂z j¼1
 Ac q000 st ð6Þ tor.

∂P 150μ g ð1  εb Þ2 1:75ð1  εb Þρg 2

Conservation of energy for the gas phase is ¼ u u ð17Þ
∂z 2 3
dp ε dp ε
3
b b
n   n _ gi H gi
∂ ∑i¼1 W i Hgi ∂ ∂T g ∂ ∑i¼1 W m
ε ¼ Ac kg þ þ ∑ ΔH j r j Temperature, density and pressure are coupled through the
∂t ∂z ∂z ∂z j¼1 equation of state
n
_ gi H gi δðz  z0 Þ  Ac q000
þ ∑W ð7Þ
gt
i¼1 ρg RTg n n
P¼ with M g ¼ ∑ ni Mi = ∑ ni ð18Þ
Mg i i
where q'''gt is total heat transfer rate from gas phase, gas to solid and gas
to wall including radiative heat transfer (q'''gw + q'''gs). W is the mass The solid velocity depends on the consumption of the fuel. As the
flow rate per meter which is equal to ρgiAc. solid biomass pellets are consumed they descend gradually under the
Conservation of energy for the inner wall is influence of gravity and pass through the drying, pyrolysis, oxidation
and reduction zones in the gasifier.
  00
Variation of the solid velocity is
∂ρw cpw T w ∂ ∂T w q Ac
¼ kw  we  q000 wt ð8Þ
∂t ∂z ∂z Vw
∂Q s m
¼ Ac Mc ∑ ϑcj r j =ρc0 ð19Þ
∂z j¼1
In order to complete the heat transfer source terms, heat transfer by
convection between the gas and solid phases and between each of these
Particle number density is obtained by mass balance using unreacted
phases and the gasifier wall must be known. Also it is necessary to solve
shrinking particle model. For particle with low ash content, it can be
energy of product gas passing between the inner and outer walls of the
assumed that the external radius of the particle will shrink according
gasifier. q'''wt is the total heat transfer rate from wall to gaseous and solid
to the mass loss and ash layer can be neglected. The particle radius
phases (q'''wg + q'''ws).
decreases in proportion to the particle conversion to the 1/3rd power
Conservation of energy for the exit gas is
and also equals to reaction-front. Minimum particle radius depends on
n   the initial ash content of wood particle which is experimentally
∂ ∑i¼1 W ei Hei ∂ ∂T e
¼ Ac kg determined (%1.82 in this work).
∂t ∂z ∂z
∂ ∑i¼1 W_ ˙ ei Hei  1=3
n m
þ þ ∑ ΔH j r j þ Sðq00 we  q00 ea Þ ð9Þ 6ð1  εÞ dp Qs
∂z Ap vp ¼ with ¼ ð20Þ
j¼1 dp dp0 Q s0
 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154 149

Gas and solid properties such as the density, heat capacity, thermal vaporization is seen to continue even after water runs out. In our ap-
conductivity, and viscosity are naturally functions of temperature as proach the amount of water in the wood controls the rate. Since the
well as of composition, and they change along the gasifier and vary temperature never falls below condensation temperature in the process
with time. The specific heat capacities are calculated using the polyno- it allows a one directional rate for modeling. However, a more detailed
mial coefficients in cp = a + bT + cT2 + dT3. The polynomial coefficients model should consider water partial pressure, vapor pressure, free
are given in Table 2 for various components in the system in kJ/kmol; water and structural water, which requires intra pellet mass transfer,
except (*) in kJ/kg [15]. and thermodynamic equilibrium. This makes the model quite
The heat and mass transfer coefficients are calculated from Reynolds, cumbersome.
Prandtl and Schmidt numbers using correlations summarized below The devolatilization of biomass is a complex phenomenon involving
[16–18]. a large number of chemical reactions. But the level of complexity of the
model depends on the specific case of interest and it is usually
2:06cpg G 0:575 2=3 considered as a first order-one-step global reaction to produce gases
hsg ¼ Re Pr ð21Þ
ε in fixed bed gasifiers. For a single particle, pyrolysis is a surface reaction
directly related to the reaction-front, but this form of pyrolysis is also
2:06G 0:575 2=3
km ¼ Re Sc ð22Þ widely accepted due to its simplicity. In the case of wood gasification
ερg
fast and flash pyrolysis is more dominant than pyrolysis, so intra-
particle resistance and physical properties of the remaining fuel is
kg ¼ 4:8∙104 T 0:716
g ð23Þ assumed uniform. Tars undergo secondary pyrolysis. Combustion of
volatile products follows a second order reaction. Even though the

2=3
μ g ¼ 1:98∙105 T g =300 ð24Þ combustion reactions of hydrogen and light hydrocarbons are infinitely
fast in comparison to other reactions, we have not changed or modified
 1 these reaction rates in spite of the mathematical and programming
t wo 1
hea ¼ þ ð25Þ difficulties.
kwo ha
Heterogeneous reactions of the char with water vapor, CO2, H2 and
O2 are modeled using the ash segregation model (exposed-core
kw ¼ 14:6 þ 1:27∙102 T w ð26Þ
model). There can be small amounts of O2 and H2 in the char besides
carbon, thus the elemental composition is taken as C = 95 wt.%, H =
ks0 ¼ 1:3∙103 þ 0:5∙104 T s þ 0:63∙106 T s 2 ð27Þ 2 wt.%, O = 3 wt.%. Mass transfer to gas phase from the char also de-
pends on diffusion through the surrounding gas film, so effective reac-
ksr ¼ 0:1952∙0:85∙T s 3 ð28Þ tion rate is introduced taking into account this aspect. Effective
reaction rate depends on the chemical reaction rate, the mass transfer
ð1  εÞ coefficient, and the concentrations of the reacting gaseous species.
ksreff ¼ þ εksr dp ð29Þ
1 1 Reforming reactions of methane and tars and the water gas shift reac-
þ
ksa ksr dp tion are also taken into account. Relevant reactions are given in
Table 3 where the subscript d stands for drying, p pyrolysis, c combus-
ks ¼ ksreff þ ks0 ð30Þ tion, g gasification, and sr steam reforming. The reaction rates and the
parameters are given in Table 4.

3.2. Chemical reactions

The source terms of the conservation equations represent drying,
pyrolysis, oxidation, and reduction. Drying process is described as a
first order reaction with Arrhenius rate, assuming that drying process
is controlled by heat transfer not diffusion. This is in contrast to the lit-
erature where rate for water is calculated based on the difference be-
tween the water vapor pressure at the wood temperature and partial
pressure of water at the grid point. Note that in this approach the vapor-
ization rate is independent of the amount of water in the wood, thus,
3.3. Solution technique
The large number of partial differential equations and the auxiliary
equations encompassing each species is solved using finite difference
technique. The gasifier has been divided into equal steps with size of
Δz and the temporal step size Δt is adjusted through the program in
order to reach convergence. The concentrations, temperatures, pres-
sures and velocities are calculated. At each step chemical reaction
rates are calculated first followed by transport equations and

Table 2
Specific heat capacity parameters.

Specie a b c d
−3 −3
Wood* 1.031 · 10 3.86 · 10
Moisture* 4.18
Oxygen 25.48 1.520 · 10−2 −0.7155 · 10−5 1.312 · 10−9
Hydrogen 29.11 −1.916 · 10−3 0.4003 · 10−5 −0.8704 · 10−9
Carbon monoxide 28.16 1.675 · 10−3 0.5372 · 10−5 −2.222 · 10−9
Carbon dioxide 22.26 5.981 · 10−2 −3.501 · 10−5 7.469 · 10−9
Methane 19.89 5.024 · 10−2 1.269 · 10−5 −11.010 · 10−9
Steam 32.24 1.923 · 10−3 1.055 · 10−5 −3.595 · 10−9
Tar 7.3662 0.3178 −1.7325 · 10−4 4.2604 · 10−8
Nitrogen 28.90 −1.571 · 10−3 0.80841 · 10−5 −2.873 · 10−9
Char* 1.39 3.86 · 10−3
Air 28.11 1.967∙10−3 0.4802 · 10−5 −1.966 · 10−9
Wall* 0.45 0.28 · 10−3 −2.91 · 10−7 1.340 · 10−10
150 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154

Table 3
Some of the important relevant reactions in the reactor.

Process Chemical reaction Ref.

Rd H2 O l →kd H2 O v [19]
Rp1 W→kp1 vc C þ vt1 T 1 þ vco CO þ vco2 CO2 þ vH2 H2 þ vCH4 CH4 þ vH2 O H2 O [20]
Rp2 T 1 →kp2 v0 t2 T 2 þ v0 co CO þ v0 co2 CO2 þ v0 H2 H2 þ v0 CH4 CH4 [14]
Rc1 ðT 1 Þ CH1:522 O0:0228 þ 0:867O2 →kc1 CO þ 0:761H2 [21]
Rc2 CH4 þ 1:5O2 →kc2 CO þ 2H2 O [21]
Rc3 2CO þ O2 →kc3 2CO2 [21]
Rc4 2H2 þ O2 →kc4 2H2 O [21]
Rc5 ðT 2 Þ CH1:522 O0:0228 þ 0:867O2 →kc5 CO þ 0:761H2 [14]
Rc6 CHα Oβ þ γO2 →kc6 ð2γ þ β  α2  1ÞCO2 þ α2 H2 O [14]
Rg1 CHα Oβ þ CO2 →kg1 2CO þ βH2 O þ ðα2  βÞH2 O [22,23]
Rg2 CHα Oβ þ ð2  α2 þ βÞH2 →kg2 CH4 þ βH2 O [22,23]
Rg3 CHα Oβ þ ð1  βÞH2 O→kg3 CO þ ð1  β þ α2 ÞH2 [22,23]
Rsr1 ðT 2 Þ CH1:522 O0:0228 þ 0:978H2 O→ksr1 CO þ 1:74H2 [14]
Rsr2 CH4 þ H2 O→ksr2 CO þ 3H2 [24]
Rwg CO þ H2 O→kwg CO2 þ H2 [24]

temperatures. All mass transfer equations are solved simultaneously
using ode45 Matlab routine which uses explicit Runge–Kutta Method
of Order 4 with Variable Step Size Control. However, energy equations
are solved with ode15i Matlab routine which uses variable order meth-
od. Please note that in the implicit solution technique the nonlinear
terms could be taken care of by iteration.
The following is presented as examples of finite differencing tech-
nique used. The mass flow balance equation was transformed to ODE
from PDE, the species mass balance (W =ρAc) in general form is





∂ðWÞ
W ziþ1; t U ziþ1; t  W zi1; t U ziþ1; t
zi; t ¼  þAc ðzi ÞX reac zi; t
∂t
2Δz
þ Ac ðzÞX inj zi; t
The derivative matrix is defined as An , n for ∂ðWÞ
∂z
ðz; tÞ ¼ A Wðz; tÞ
where A is the coefficient matrix and W (z, t) is the vector form of the
variable W with n components. Similarly the final vector form of the
continuity equation at one point in time is
 
∂W
¼ A1…n;1…n ðW 1…n ∙U 1…n Þ þ Ac1…n ∙X reac1…n
∂t 1…n
þ Ac1…n ∙X inj 1…n
2. Determine n, the number of points in axial direction.
3. Compute cross sectional area for each point it z-direction from the
measured 3-D shape data of the reactor.
4. Set temporal step size Δt.
5. Process the initial data (for concentrations, fluxes etc.).
6. Select finite difference order and build derivative matrices.
7. Compute the reaction rates at all the grid points.
8. Solve concentrations with Runge–Kutta 4th order method.
9. Compute auxiliary equations (set of fluxes, pressure, heats of
reactions, specific heats etc.).

10. Compute time-weighted average of heats of reactions (for every
20th iteration).
11. Solve temperature equations with ode15i (for every 20th iteration).
ð31Þ 12. Boundary check for temperatures (upper and minimum)
13. Check convergence, if necessary adjust the temporal step size and
go to step 7.
14. Advance time by Δtfinal as tnew = t + Δtfinal.
15. Set all the variables calculated as the previous time values (initial
value for the next time step).
16. Save data at prescribed time intervals.
17. Return to item step 7 if tnew b tmax (maximum time for the process
duration).
ð32Þ 18. Stop.

The solution algorithm in a compact form is presented below

1. Read the data.

Table 4
Relevant reaction parameters for kj =Aj exp(-Ej/RT.).

Process Chemical reaction rate Aj Ej (kJ/kmol) Ref.

Rd kdCH2Ol 5.56 ⋅ 106 8.79 ⋅ 104 [25]

Rp1 kp1ρw 1.52 ⋅ 103 6.28 ⋅ 104 [26]
Rp2 kp2ρt1 4.26 ⋅ 106 1.07 ⋅ 105 [27]
Rc1 kc1TgCt10.5CO2 9.2 ⋅ 106 8.02 ⋅ 104 [21]
Rc2 kc2TgCCH40.5CO2 9.2 ⋅ 106 8.02 ⋅ 104 [21]
Rc3 kc3CcoCO20.25CH2Ov0.5 1.3 ⋅ 1011 1.26 ⋅ 105 [21]
Rc4 kc4CH2CO2 1.0 ⋅ 1011 4.20 ⋅ 104 [21]
Rc5 kc5TgCt20.5CO2 9.2 ⋅ 106 8.02 ⋅ 104 [21]
Rc6 Apvp(1/km + 1/kc6)CO2 5.67 ⋅ 107 1.60 ⋅ 105 [28]
Rg1 Apvp(1/km + 1/kg1)CCO2 1.0 ⋅ 107 2.17 ⋅ 106 [29]
Rg2 Apvp(1/km + 1/kg2)CH2 1.0 ⋅ 104 2.17 ⋅ 106 [29]
Rg3 Apvp(1/km + 1/kg3)CH2Ov 1.0 ⋅ 107 2.17 ⋅ 106 [29]
Rsr1 ksr1CT2CH2Ov 0.3 ⋅ 109 0.30 ⋅ 105 [13]
Rsr2 ksr2CCH4CH2Ov 0.3 ⋅ 109 0.30 ⋅ 105 [13]
C CO2 C H2 kwg : 2.78 1.26 ⋅ 104
Rwg kwg ðC CO C H2 Ov − Þ [30]
kwg;e kwg,e : 0.0265 −3.29 ⋅ 104
Fig. 2. Temperature distribution inside the gasifier.
 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154
4. Results and discussion
Several experiments were conducted and the experimental data are
used to validate the model. A representative data set is presented. After
the operation reaches the steady state the behavior of solid and gas
temperatures in the gasifier is presented Fig. 2. The model was started
by setting all the temperatures to 300 K. The initial bed density is
416 kg/m3. Then a fixed heat flux is given to the gasifier for 5 min via
a butane torch in order to stimulate the gasifier. The solid temperature
was seen to increase slowly and simulation was continued until the
gasifier has reached the steady-state. At this time the wall, exiting gas,
and solid temperatures were almost the same. Solid and gases were
introduced to the reactor from the top at 300 K as well. Injected air
passes through the exiting hot gas before entering the combustion
zone from the bottom through 5 J-tubes, so the injected air temperature
approaches the core temperature due to extensive heat gain.
At the combustion zone (peak in Fig. 2) the thermal energy spreads
with an effective conductivity which contains radiation component as
well. Dehumidification and pyrolysis take place prior to combustion.
The heat transported is used for endothermic reactions. It can be seen
that there is a small temperature difference between gas and solid
phases. Due to ignition of volatile materials, gas temperature is expected
to be higher than solid temperature at the combustion zone as seen.
The reaction of char with hydrogen and steam or oxygen is observed
at the reduction zone following the combustion zone. Since most of the
oxygen is consumed at the combustion zone char-steam gasification is
more dominant. Additionally, major portion of water–gas-shift reaction
takes place in this zone. One can see that there is a certain amount of
difference between the experimental and modeling results in the
reduction zone. This is due to the fact that the reduction zone could
not be measured exactly due to the geometry of the gasifier. The
thermocouple was buried 3–3.5 cm within the char where the char
and accompanying ash have lower thermal conductivities resisting the
supply of heat from above to compensate for the losses in this section
due to the clean-up gate and an opening for shaking. Thus the experi-
mental temperature measurements seem lower in comparison to the
model prediction. A careful observation of the reduction zone in Fig. 2
shows that there is a logarithmic decrease in temperature. As the char
is consumed during the reduction reactions, the surface area per unit
volume increases, and total surface reactions become higher. Since the
endothermic char-steam reactions are dominant as expected, the tem-
peratures decrease accordingly (logarithmically). Additionally, in-
creased surface area accelerates the process of thermal equilibration of
the solid and gas. Since shrinking core model is used, the pellet size is
allowed to decrease to that of a pellet consisting of ash only. Once this
minimum pellet size is reached no further char reactions take place.
Fig. 3. Transient behavior of temperature at injection point.
151
Model and experimental results are shown on Fig. 3. It can be seen
that there is a good agreement between the model and the experiment
for the transient behavior of the temperature inside the gasifier. For the
first 500 s, ignition process is simulated. Butane torch temperature
reaches 1300 K. The temperature of the injected air and the heat flux
has been calculated using the experimental data. This is quite important
because the heating temperature has a pronounced effect on the pro-
cess. The resonation of the temperatures in the experimental and
model values is due to the fact that air is drawn with vacuum. In the
mathematical model the pressure inside is calculated continuously
and air is drawn in to compensate the vacuum difference. Thus the air
injected is proportional to the vacuum difference: Qi0 = C∙(Po − Pi).
Even very simple vacuum systems are comprised of many conductance
elements, some are connected in series like this case, and C refers to
overall conduction which can be computed or experimentally
measured.
The evolution of the gaseous species is shown in Fig. 4 as volume
fraction. The behavior of the gas concentrations could be followed
along the gasifier (0.55 m for this set-up) as well as along the path
followed by the exit gas including the jacket beyond 0.55 m on x-axis.
The air has 23% O2, 75% N2 and 2% H2O by weight at the top of the
gasifier. At the beginning no velocities have been defined; however
the increase in concentrations due to pyrolysis and volatilization has
been converted to pressure via equation of state and the products
formed are driven out with this increased pressure and concentration
difference. Pyrolysis and evaporation starts within first 2 cm of the gas-
ifier because of the temperature. The combustion of volatiles being exo-
thermic reactions increases the gas temperature followed by increase in
solid temperatures. Naturally as fresh air is injected to the system all
gases cool down. The air has similar composition as the one at the top
given above. The increase in water vapor content beyond this point is
due to both humidity of the air introduced and the combustion reac-
tions. The chemical composition of exit biogas is found using the
model as follows: 25.85% CO; 18.25% H2; 7.84% CO2; 7.17% H2O; 2.9%
CH4; the rest was N2.
The modeling involves a smooth flow of the solids in line with the
consumption at various locations keeping the reactor volume full. The
pneumatic eccentric shaking of the ash-grate section helps prevent
the bridging and dead zones; however, this is not perfect as compared
to the forced solid movement systems. Thus the experiment is affected
by bridging and channeling to some extent in contrast to the model.
These effects can be seen on the concentrations in Fig. 5 between
1750 and 2000 s. First 500 s simulate ignition process so the figure starts
after the ignition process and the experimental mean gas composition
Fig. 4. Evolution of the species through the entire gasifier.
152 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154
Fig. 5. Evolution of the species with time (experimentally).
was as follows: 26.07% CO; 14.71% H2; 10.4% CO2; 3.42% CH4; the rest
was N2.
In Fig. 6 the gas velocity and flow indicate that pyrolysis and volatil-
ization are effective until 0.25 m from the top of the reactor. The gas
velocity provides withdrawal of the product gases of these reactions. A
sudden increase in gas velocity between the 0.25 and 0.35 m levels is
a result of air injection and combustion reactions. After 0.35 m level
although gas flow rate is fairly constant, the gas velocity changes in
line with the changing reactor diameter due to its geometry. Also note
that solid flow is a direct result of the cumulative consumption of the
solids starting from the initial pellet down to the prevalent pellet diam-
eter. Thus pyrolysis is effective until the 0.30 m level and between 0.3
and 0.45 m levels char gasification reactions are dominant.
Throated and stratified gasifiers are compared in Figs. 7 and 8. It can
Fig. 7. Flow rate analysis of different gasifier types.
caused by the hydrocarbons spending more time with water at higher
temperatures.
The comparison between the temperature distribution of throated
and stratified gasifier shows that peak and reduction temperatures are
similar. So the combustion and char reduction characteristics are the
same. Nevertheless the throated gasifiers are more efficient and give
more reliable and clean output.
Although we had had performed 6 different experiments and the
model was validated accordingly, additional validation was carried out
at several Equivalence Ratios using the data of Barrio et al. [31]. The
equivalence ratio (air/kg of wood) is defined to reflect the combined
effect of air flow rate, rate of wood supply and duration of the run
be seen that stratified gasifier has a sharp increase in the temperature Equivalence ratio 
but throated one has more favorable temperature distribution. With Flow rate of air Flow rate of air 
¼ = ð33Þ
the same core size which determines maximum solid consumption,
stratified gasifier output is 24.5% less than throated one. Exiting gas
flows around gasifier causes pre-heating so the temperature differences
between the gasifiers come from output ratio.
Although exit flow rates of methane and tar seem to be similar in
Fig. 7, considering the output ratio, actual concentrations of methane
and tar are quite smaller in throated gasifiers. Pre-heating of air circulat-
ing in the jacket increases temperature from the top toward the injec-
tion point; this then, increases tar and methane conversion. This is
Fig. 6. Axial profile of diameter, gas and solid velocities.
Flow rate of wood Flow rate of woodstoichiometric
The stoichiometric ratio of air flow rate to biomass consumption rate
is 5.22 m3 air/kg of wood. The comparison of our model and Barrio et al.
[31] results is shown in Fig. 9 (*points are from Barrio et al.). Note that
each data point in x-direction represents a different run. Although
their experiments were conducted with a straight reactor, whereas
ours with a throated one, the results were comparable.
Finally the model system was found to be responsive to the
changes in parameters. For instance the porosity cannot be
Fig. 8. Solid temperature behavior of different gasifier types.
 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154 153

Fig. 9. Concentration change with equivalence ratio.

measured easily when the reactor is running. So we have made a In addition to all the above, throated gasifiers present a high challenge
simple analysis of changing it at time 2000 s by + and − 20%. The adding more to an already complicated process. This has made it very dif-
mean responses at time 3000 s were − 5.3% and 2.8% in Tg; − 3.4% ficult to model and the computer solution became very difficult. However
and + 1.4% in Ts,; − 0.4% and + 3.4% in pressure. the transient behavior was successfully modeled and validated. The use of
realistic operating and design parameters and the successful solution of
the complex mathematical model contributed to the work in this area.
5. Conclusions
Nomenclature
A transient kinetic and transport model has been introduced for the Ac cross-sectional area m2
first time in a throated downdraft gasifier. The advantageous points of An,n coefficient matrix
previous models were taken and applied to the present downdraft Ap particle surface area m2
throated geometry and all the parameters were taken from the actual C concentration kmol/m3
experimental system and present gasifier in line with the experiments cp specific heat kj/kg/K
conducted. D reactor diameter
The main objective of this work was to produce a suitable and realistic dp particle diameter
model for a commonly used fixed bed downdraft type gasifier. The model G superficial mass flux kg/m2/s
has been produced and the validation was done for a 10 kW modified h heat transfer coefficient kW/m2/K
throated imbert gasifier. The model parameters, geometric and thermo- H specific enthalpy kJ/kg
physical properties were selected based on the actual system used. k thermal conductivity kW/m/K
Distinct from all other models not only the gasification process is km mass transport coefficient m/s
modeled; the gasifier has been modeled as a complete unit. Such a M molecular weight kg/kmol
model shows transient behavior of the system better. The modeling re- P pressure kPa
sults were in line with the experimental findings. In the absence of P particle
bridging and channeling the ranges of experimental gas concentrations Pr Prandtl number (cpgμg/kg)
were 35.07–20.77% CO; 18.30–13.66% H2; 13.68–5.95% CO2; 6.7–1.2% Q heat loss kW/m3
CH4. The model results were 25.85% CO; 18.25% H2; 7.84% CO2; 2.9% q” heat flux kW/m2
CH4; the rest was N2. The model results were in line with the results q”’ heat flux kW/m3
of experiments performed. The hydrogen concentrations, although R ideal gas constant, 8.314 KJ/kmol/K
within experimental limits, were a bit more than expected. Note that r reaction rate kmol/m3/s
no fitting and effectiveness parameters were used here. Re Reynolds number (dpG/μg)
Since many studies were conducted on mini scale lab equipment we S surface area to length ratio m2/m
chose to work on pilot scale unit. This allowed for better observation of Sc Schmidt number (μg/ρgD)
operating conditions, heat losses, bridging and channeling and mea- T temperature K
surement problems. At any rate, the gasifier contains quite a large num- T0 298 K
ber of unknowns, effective properties, lump sum quantities and t time s
reactions, and approximated events. Also the amount of tar and its con- u superficial velocity m/s
stituents present further difficulties. Especially tar analysis requires W mass flow per unit length kg/m
careful attention is required. W _ mass flow rate kg/m/s
154 O. Yucel, M.A. Hastaoglu / Fuel Processing Technology 144 (2016) 145–154

X vector form of generation from reaction and injection [8] C. Di Blasi, C. Branca, Kinetics of primary product formation from wood pyrolysis,
Ind. Eng. Chem. Res. 40 (2001) 5547–5556, http://dx.doi.org/10.1021/ie000997e.
z axial [9] C. Di Blasi, Modeling chemical and physical processes of wood and biomass pyroly-
sis, Prog. Energy Combust. Sci. 34 (2008) 47–90, http://dx.doi.org/10.1016/j.pecs.
Greek letters 2006.12.001.
[10] D.L. Giltrap, R. McKibbin, G.R.G. Barnes, A steady state model of gas-char reactions in
a downdraft biomass gasifier, Sol. Energy 74 (2003) 85–91, http://dx.doi.org/10.
vp particle density 1/m3 1016/S0038-092X(03)00091–4.
ϑ stoichiometric matrix [11] B.V. Babu, P.N. Sheth, Modeling and simulation of reduction zone of downdraft bio-
mass gasifier: effect of char reactivity factor, Energy Convers. Manag. 47 (2006)
ΔH heat of reaction kJ/kmol
2602–2611, http://dx.doi.org/10.1016/j.enconman.2005.10.032.
ε porosity [12] C.D.D. Blasi, Dynamic behaviour of stratified downdraft gasifiers, Chem. Eng. Sci. 55
μ viscosity kg/m/s (2000) 2931–2944, http://dx.doi.org/10.1016/S0009-2509(99)00562-X.
ρ density kg/m3 [13] F.V. Tinaut, A. Melgar, J.F. Pérez, A. Horrillo, Effect of biomass particle size and air su-
perficial velocity on the gasification process in a downdraft fixed bed gasifier. An ex-
perimental and modelling study, Fuel Process. Technol. 89 (2008) 1076–1089,
Subscripts http://dx.doi.org/10.1016/j.fuproc.2008.04.010.
[14] C. Di Blasi, C. Branca, Modeling a stratified downdraft wood gasifier with primary
and secondary air entry, Fuel 104 (2013) 847–860, http://dx.doi.org/10.1016/j.
a ambient fuel.2012.10.014.
b bed [15] Y.A. Çengel, M.A. Boles, Thermodynamics: An Engineering Approach, McGraw-Hill
c char Higher Education, 2006 (https://books.google.com.tr/books?id=5-hSAAAAMAAJ).
[16] A. Sen Gupta, G. Thodos, Direct analogy between mass and heat transfer to beds of
e exiting gas spheres, AICHE J. 9 (1963) 751–754, http://dx.doi.org/10.1002/aic.690090609.
ea exiting gas to ambient [17] P.D. Aerts, K.W. Ragland, Pressurized downdraft combustion of woodchips, Symp.
eff effective Combust. 23 (1991) 1025–1032, http://dx.doi.org/10.1016/S0082-0784(06)80360–7.
[18] J.C. Chen, S.W. Churchill, Radiant heat transfer in packed beds, AICHE J. 9 (1963)
g gas
35–41, http://dx.doi.org/10.1002/aic.690090108.
gw gas to wall [19] K.M. Bryden, K.W. Ragland, C.J. Rutland, Modeling thermally thick pyrolysis of wood,
gs gas to solid Biomass Bioenergy 22 (2002) 41–53, http://dx.doi.org/10.1016/S0961-
i index for species 9534(01)00060–5.
[20] C.D. Di Blasi, G. Signorelli, G. Portoricco, Countercurrent fixed-bed gasification of bio-
j index for reactions mass at laboratory scale, Ind. Eng. Chem. Res. 38 (1999) 2571–2581 (http://www.
m moisture scopus.com/inward/record.url?eid=2-s2.0-0032794107&partnerID=tZOtx3y1).
r radiative [21] K.M. Bryden, K.W. Ragland, Numerical modeling of a deep, fixed bed combustor, Fuel
Energy Abstr. 37 (1996) 207, http://dx.doi.org/10.1016/0140-6701(96)88901-6.
s solid (wood) [22] H. Yoon, J. Wei, M.M. Denn, A model for moving-bed coal gasification reactors,
s0 solid conductivity without radiation AICHE J. 24 (1978) 885–903, http://dx.doi.org/10.1002/aic.690240515.
sr radiative solid conductivity [23] T.K. Kayal, M. Chakravarty, G.K. Biswas, Mathematical modelling of continuous updraft
gasification of bundled jute stick — a low ash content woody biomass, Bioresour.
sw solid to wall Technol. 49 (1994) 61–73, http://dx.doi.org/10.1016/0960-8524(94)90174-0.
t total [24] J. Xu, G.F. Froment, Methane steam reforming, methanation and water–gas shift: I.
w wall Intrinsic kinetics, AICHE J. 35 (1989) 88–96, http://dx.doi.org/10.1002/aic.
690350109.
we wall to exiting flow
[25] C. Mandl, I. Obernberger, F. Biedermann, Modelling of an updraft fixed-bed gasifier
wo outer wall operated with softwood pellets, Fuel 89 (2010) 3795–3806, http://dx.doi.org/10.
1016/j.fuel.2010.07.014.
[26] A.F. Roberts, G. Clough, Thermal decomposition of wood in an inert atmosphere, Symp.
References
Combust. 9 (1963) 158–166, http://dx.doi.org/10.1016/S0082-0784(63)80022-3.
[27] A.G. Liden, F. Berruti, D.S. Scott, A kinetic model for the production of liquids from
[1] M. Asadullah, Barriers of commercial power generation using biomass gasification
the flash pyrolysis of biomass, Chem. Eng. Commun. 65 (1988) 207–221, http://
gas: a review, Renew. Sust. Energ. Rev. 29 (2014) 201–215, http://dx.doi.org/10.
dx.doi.org/10.1080/00986448808940254.
1016/j.rser.2013.08.074.
[28] T. Kashiwagi, H. Nambu, Global kinetic constants for thermal oxidative degradation
[2] S. Chopra, A. Jain, A review of fixed bed gasification systems for biomass, Agric. Eng.
of a cellulosic paper, Combust. Flame 88 (1992) 345–368, http://dx.doi.org/10.1016/
Int. CIGR IX (2007) 1–23 (http://hdl.handle.net/1813/10671).
0010-2180(92)90039-R.
[3] A.K. Sharma, Equilibrium and kinetic modeling of char reduction reactions in a
[29] M.J. Groeneveld, W.P.M. van Swaaij, Gasification of char particles with CO2 and H2O,
downdraft biomass gasifier: a comparison, Sol. Energy 82 (2008) 918–928, http://
Chem. Eng. Sci. 35 (1980) 307–313, http://dx.doi.org/10.1016/0009-2509(80)80101-1.
dx.doi.org/10.1016/j.solener.2008.03.004.
[30] V. Biba, J. Macak, E. Klose, J. Malecha, Mathematical-model for gasification of coal
[4] M.A. Hastaoglu, M.S. Hassam, Application of a general gas-solid reaction model to
under pressure, Ind. Eng. Chem. Process. Des. Dev. 17 (1978) 92–98, http://dx.doi.
flash pyrolysis of wood in a circulating fluidized bed, Fuel 74 (1995) 697–703,
org/10.1021/i260065a017.
http://dx.doi.org/10.1016/0016-2361(94)00010-O.
[31] M. Barrio, M. Fossum, J.E. Hustad, Operational characteristics of a small-scale strati-
[5] M.L. Hobbs, P.T. Radulovic, L.D. Smoot, Modeling fixed-bed coal gasifiers, AICHE J. 38
fied downdraft gasifier, Sixth Int. Conference on Technologies and Combustion for a
(1992) 681–702, http://dx.doi.org/10.1002/aic.690380506.
Clean Environment, vol. 3, Instituto de Combustão, Porto, Portugal 2001,
[6] F. Shafizadeh, S. Chin, P. Peter, Thermal Deterioration of Wood, Wood Technol.
pp. 1269–12676.
Chem. Asp. American Chemical Society 1977, pp. 57–81, http://dx.doi.org/10.
1021/bk-1977-0043.ch005.
[7] F. Thurner, Mann, Kinetic investigation of wood pyrolysis, Ind. Eng. Chem. Process.
Des. Dev. 20 (1981) 482–488, http://dx.doi.org/10.1021/i200014a015.