Energy Conversion and Management 65 (2013) 783–792

Contents lists available at SciVerse ScienceDirect

Energy Conversion and Management

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

Numerical and experimental investigation of downdraft gasification of wood chips

I. Janajreh ⇑, M. Al Shrah
Masdar Institute of Science and Technology (MI), Mechanical Engineering Program, P.O. Box 54224, Abu Dhabi, United Arab Emirates

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

Article history: Biomass is widely perceived as a potential renewable energy source. Thermo-chemical conversion tech-
Available online 7 November 2012 nologies including gasification, co-firing, and pyrolysis are of primary interest due to their higher conver-
sion efficiency and throughput when compared with the low temperature digestion and fermentation for
Keywords: lignocellulose and wood-based feedstock. In this paper, a small scale, air blown, downdraft gasification
Gasification system is operated using wood to investigate its conversion efficiency. Wood chips of 0.5 cm thickness,
Downdraft gasifier 1–2 cm width, and 2–2.5 cm length constitute the feedstock to the downdraft gasifier that is assembled
CFD
and instrumented at Masdar Institute’s Waste-to-Energy laboratory. The experimental investigation of
Wood pellets
the temperature field inside the gasifier is followed by high fidelity numerical simulation using CFD to
model the Lagrangian particle coupled evolution. The numerical simulation is conducted on a high reso-
lution mesh accounting for the solid and gaseous phases, k–e turbulence, and reacting CFD model. The
temperature distribution and the evolution of species are computed and compared with the experimental
results and with the ideal equilibrium, zero dimensional case.
Ó 2012 Published by Elsevier Ltd.

1. Introduction These reactors are distinguished by the concurrent flow of feed-

The world reliance on biomass energy has increased to 14% of
total energy consumption. This figure is higher than that of coal
(12%) and is competing with gas (15%) and direct electricity
(14%) energy consumption [1]. Although biomass is primarily con-
sumed in third-world countries as an energy source, many indus-
trialized countries are re-considering the involvement of biomass
in their energy systems due mainly to its CO2 neutrality and
renewable nature [2]. Wood as a biomass feedstock is character-
ized by high volatile and low sulfur and ash contents thereby, max-
imizing its potential to be used as a feedstock for energy
generation [3–5]. Fig. 1 shows the position of wood and other bio-
mass types composition as compared to coal on the O/C–H/C graph
known as Van Kreveln diagram.
One of the promising thermo-chemical conversion technologies
that can be used to convert biomass to useful energy forms suited
for small to medium size throughput is downdraft gasification. It is
an ideal choice for small-scale gasification processes where the
maximum electricity generated is about 500 kW [6]. Chemical
and physical phenomena that occur during gasification of wood
are essentially similar to typical carbonaceous material thermal
decomposition. This includes drying, devolatilization (which in-
cludes release of volatile flammable gases, flaming combustion of
the volatiles, and glowing combustion of fixed carbon), heat con-
duction, fissuring, shrinkage, and fragmentation of solid particles
[5].
⇑ Corresponding author. Tel.: +971 2 698 8170; fax: +971 2 698 8121.
E-mail address: ijanarjeh@masdar.ac.ae (I. Janajreh).
stock and gasifying agent. Four primary gasifier zones can be
tracked based on the dominant reactions occurring in each one.
From top to bottom, these are: Drying, pyrolysis (volatile release),
combustion, and reduction/gasification zones [7]. A schematic of
the reactor used in this study is shown in Fig. 2. Reactor parame-
ters of primary importance that influence its performance are the
size of the reduction zone (located under the air nozzles towards
the bottom of the gasifier) and its stabilization location, thus con-
trolling the residence time of the biomass particles within the
reactor. The reduction zone signifies the location of the bulk of
the gasification reactions including char reduction reactions and
tar cracking and hence it controls to a great extent the yield of
syngases which in turn define the gasification efficiency. Wood
char reactions are essentially slow and require relatively long res-
idence time compared to the gas-phase reactions thus imposing a
minimum limitation on the reduction zone size [6]. The gasifier is
also characterized with low producer-gas tar content due to the
relatively high temperatures (greater than 1000 °C) within the
reduction zone [3,4,8]. A downside to these types of gasifiers is
the low overall efficiency due to the high sensible heat swept by
the producer gas which leaves at temperatures of 900–1000 °C
adding the convective surface heat loss to the surrounding [8]. In
addition, the high particulate matter that carried by the producer
gas due to the inhomogeneous oxygen mixing and its consequent
temperature fields within the gasifier. Other limitations include
feedstock moisture content which is limited to a maximum of
25% by mass, ash content up to 30%, and the complexity of scale
up [7].

0196-8904/$ - see front matter Ó 2012 Published by Elsevier Ltd.

http://dx.doi.org/10.1016/j.enconman.2012.03.009
784 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792

Nomenclature

A pre-exponential factor mi transport species

CFD computational fluid dynamics m0p initial particle mass
CGE cold gas efficiency O/C oxygen to carbon ratio
Di,m diffusion coefficient S/ source term
DPM discrete phase method Sot turbulent Schmidt number
DSC differential scanning calorimeter Tevp vaporization temperature
e turbulent dissipation rate TG thermo gravimetric
E activation energy u fluid phase velocity
ER equivalence ratio v stoichiometric coefficient
fv volatile fraction U dependent variable
HHV high heating value g reaction rate
H/C hydrogen to carbon ratio
K reaction constant

Several scientists approached downdraft gasification technol-
ogy using wood as a feedstock and studied the various parameters
that affect the producer gas quality including moisture content of
wood chips, oxidizer (air) flow rate, chip size, location of air inlet
nozzle and reduction zone length (and thereby particle residence
time) [4]. Hsi et al. [3] experimentally studied the effect of airflow
rate, air preheating temperature, and moisture content of wood
pellets on the producer gas composition and its heating value using
an air-blown fixed-bed stratified downdraft reactor. The conver-
sion rate of wooden pellets increased with increasing cold airflow,
while the heating value of the producer gas reached a maxima at
15 Nm3/h then decreased as a result of dilution by cold air. Pre-
heating the air to 573 K did not affect significantly the heating va-
lue of producer gas. Whereas, the moisture content of wood pellets
was found to significantly lower the conversion rate of the pellets
and the heating value of the producer gas. The optimum mean
higher hating value (HHV) with a moisture content of 18% and cold
air was found to be 5.2 MJ/Nm3. A one dimensional model pre-
dicted the axial temperature profile and species distribution was
in agreement with experimental findings.
Zainal et al. [4] studied the effect of equivalence ratio in down-
draft gasifiers on conversion of furniture wood and wood chips.
They found that the H2 production increased linearly with the
equivalence ratio (in terms of gasification), whereas CO, CH4, and
the calorific value of producer gas reached a maxima and then de-
clined as the equivalence ratio further increased due to the short-
age of air as expected. Zainal et al. also explained the temperature
fluctuations in the combustion zone by referring to the effect of the
interrupted contact between the ignited, glowing wood pellets and
the thermocouple on one hand and the relatively colder voids on
another. More importantly is the phenomenon of bridging that Zai-
nal et al. observed. This phenomenon occurs when relatively large
shredded chips create a bridge that obstructs the continuous flow
of the wood pellets and thus causes high localized temperatures
inside the gasifier. Their work showed that bridging was reduced
by using a 60o angled-throat which is higher than the repose angle
of wood chips (45o) at ambient temperatures. This angle provides
smooth gravitational flow of wood through the combustion zone
and facilitates the cracking of tar without bridging.
Kumar et al. [5] studied the effect of fuel particle shape and size
on devolatilization time using Casuarina wood in a fluidized bed
combustor. Various shapes like disc, cylinder, rod, sphere, and
cube, with 5–30 equivalent sphere diameter and aspect ratio 0.2–
10 were tested. The fluidized bed reactor (130 mm diameter) used
a 550 lm sand grains for the bed material and its operating tem-
perature was 650–850 °C. It was found that the devolatilization
time increases with increasing feedstock particle size and density,

1.8 Van Krevelen diagram

Biomass
1.6
1.4 Peat
Atomic H/C Ratio

1.2 Lignite
Wood
1.0 Coal Lignite
0.8 Cellulose

0.6 Incr
eas
0.4 eH
eat
ing
valu
0.2 Anthracite e

0.0
0 0.2 0.4 0.6 0.8
Atomic O/C Ratio

Fig. 1. Van Kreveln diagram illustrating the relevant composition and heating value
of wood compared to other fuels. Fig. 2. Schematic of the downdraft gasifier.
 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792
sphere had the longest time for a given volume, the disc and rod
had the shortest time, while cylindrical shapes had the intermedi-
ate time. In fact, it was shown that the diameter of the particle has
a larger effect on the devolatilization time compared to its length
because the volume is proportional to the length and to the square
of the diameter. For rods, increasing the length of the particle be-
yond 50 mm for a constant diameter of 10 mm does not affect
the devolatilization time because the volume to surface area ratio
becomes a function of only diameter and hence it is constant for a
value of 10 mm diameter. Similarly, for discs, at larger diameters
the volume becomes only a function of length which is constant
and hence the devolatilization time is constant as well. The impor-
tance of studying the devolatilization time of wood lies behind the
fact that wood contains up to 80% of its weight volatiles and this
significantly contribute to the energy released, emissions and
NOx formation during combustion [5]. However, [5] highlighted
that the ratio of the volume to surface area makes the prediction
of devolatilization time as a function of particle dimensions less
trivial, because given the increase in particle volume does not nec-
essarily imply an increase in the devolatilization time (due to the
increase of the volatile content) if the surface area also increased
and hence the heat transfer rate leading to a compensated effect.
2. Experimental set up
A schematic of the assembled downdraft gasification system is
presented in Fig. 2 which is based on Gekgasifier Inc. design with a
total capacity of 20 kw [16]. The reactor is cylindrical with an inter-
nal diameter of 27 cm and a height of 48 cm. It consists of an auto-
matic feed hopper system, swirl burner, and rotary grate connected
to syngas cyclone and packed bed filter for ash collection, in which
the experiment is done in a semi-batch gasifier. The initial ignition
Fig. 3. Wood chips used in the experiment.
Fig. 4. Thermocouple arrangement and location inside the gasifier used in the experiment.
785
of the reactor was made using 4 kg batches of charcoal consisting
of relatively large chunks of coal and then followed by the wood
chips of 0.5 cm thickness, 1–2 cm width, and 2–2.5 cm length as
sampled in Fig 3.
The downdraft gasifier is assembled and instrumented at Mas-
dar Institute’s Waste-to-Energy laboratory. The temperature field
inside the gasifier was measured using eight K-type thermocouples
positioned inside the gasifier at different heights and at two radial
locations. Seven thermocouples were placed at varying heights at
the center of the gasifier, and three thermocouples were placed
at varying heights at 2.5 in offsetting from the center. The configu-
ration of the thermocouples is shown in Fig. 4 which were bundled
into two sets and mounted and sealed by thermal mortar.
Data from the thermocouples was monitored and recorded at a
0.1 Hz sampling rate using a CR3000 Micrologger. The gasifier was
ignited by a propane torch at the ignition port. The energy gener-
ated from the partial coal combustion drives the endothermic reac-
tions of gasification. Vacuum pressure in the gasifier of 10 mm H2O
is achieved by an air compressor connected with the outflow gases.
Ignition time took around 2 min. A white, vapor-like effluent was
observed at the outlet port indicating the onset gasification of coal.
Gas sample was collected at the reactor exit leading to the cyclone
in a 1-Liter Tedlar bag for future gas chromatography analysis.
3. Computational model
The downdraft gasifier is modeled using the finite volume code
[9] coupled with a conjugate heat transfer with the bulk metal sep-
arators and insulation. The computation is carried out on a two
dimensional axi-symmetric geometry of the gasifier. The segre-
gated implicit solver was used to solve the transport equations.
The following assumptions are made to simplify the physics of
the problem without loss of accuracy:
1. Flow is axi-symmetric, i.e. two dimensional.
2. Flow is steady.
3. No-slip condition applies on reactor walls and separator/insula-
tion bulk material.
4. The particles have uniform distribution and are spherical in
shape but with a much smaller size (0.1 mm) than what is been
used experimentally.
5. Discrete phase model is used, given the small particle size and
compared to the reactor volume.
786 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792

Second order discretization scheme was implemented for all
calculated variables. The SIMPLE algorithm is used for velocity–
pressure coupling. Turbulence is modeled by the k–e with standard
wall function.
4. Geometry and mesh construction
where Di,m is the diffusion coefficient. Sct is the turbulent Schmidt
number which is a ratio of the eddy viscosity lt to the eddy diffu-
sivity Dti,m. These transport equations incorporate a reaction source
term Ri in addition to the discrete phase interaction source Si. The Ri
term is governed by the stoichiometric reaction below:
X
N kf ;r X
N

An axi-symmetric model is built for the downdraft gasifier with

v 0i;r Si () v 00i;r Si ð4Þ
kb;r
i¼1 i¼1
the dimensions and generated mesh displayed in Fig. 5. It is fitted
with a baseline structural multi-blocking mesh consisting of The ith species production/destruction due to the reaction r is
53,155 rectangular cells. Refinement near the air nozzle outlet written as:
and in the combustion and reduction zones at the gasifier bell !
was emphasized to allow for better capturing of the steep flow gra-
  Y
N
g Y
N
v 0
dients in these highly turbulent regions.
Ri;r ¼ Mi;r v 00
i;r v 0
i;r kf C j;r;rj  kb C j;rj;r ð5Þ
j¼1 j¼1

5. Governing equations where k here is the Arrhenius reaction constant, and (C)j is the mo-
lar concentration of jth specie raised to stoichiometric coefficients v
The continuous phase is governed by the following equation: and reaction order g, and Mi is the molecular weight of species i.
  The discrete solid particle phase is solved in a Lagrangian frame
@ @ @ @/
ð/Þ þ ðui /Þ ¼  C/ þ S/ ð1Þ of reference. This phase consists of spherical particles of 100 lm in
@t @xi @xi @xi source
Timerate adv ectiv e diffusion diameter dispersed in the continuous phase. Their trajectory is pre-
dicted by integrating the force balance on the particle. This force
where q is the density and S/ is the source terms due to the dis-
balance equates the particle inertia with the forces acting on the
persed/discrete phase interaction. / is the dependent variable cor-
particle and can be described as:
responding to density (q), the density velocity multiple (q ui), and
the temperature (T), representing the conservation of mass,
d~
uP
momentum, and energy, respectively. It can also represent turbu- ¼ F D ð~
u ~ g ðqP  qÞ=qP
uP Þ þ ~ ð6Þ
dt
lent scalars, i.e. turbulent kinetic energy (k) and turbulent dissipa-
tion rate (e). These two equations in steady state flow regime are where FD (u  up) is the drag force per unit particle mass; u is the
written as: fluid phase velocity; up is the particle velocity; q is the fluid density,
   
@k @uj @ui @uj @ lt @k and qp is the density of the particle. The trajectory equations are
qui ¼ lt þ þ  qe solved by stepwise integration over discrete time steps. Integration
@xi @xi @xj @xi @xi rk @xi
    ð2Þ of Eq. (6) yields the velocity of the particle at each point along the
@e l e @uj @ui @uj @ lt @ e qe2
qui ¼ C 1e t þ þ  C 2e trajectory, with the trajectory itself predicted by
@xi k @xi @xj @xi @xi re @xi k
The right hand terms represent the generation, the diffusion dx
¼ uP ð7Þ
and the destruction, respectively. In these equations, lt is the tur- dt
2
bulent or eddy viscosity lt ¼ fl C l q k =e where f and C along with Equations similar to (6) and (7) are solved for each coordinate
C1e, C1e, rj re are empirical constants. Similarly, the transportation direction to predict the trajectories of the discrete phase. The tra-
of species mi is written as: jectories of the discrete phase particles are computed as well as
@ @ @ @mi the heat and mass transfer to and from them. Solid particles are
ðqmi Þ þ ðqui mi Þ ¼ ðqDi;m þ lt =Sct Þ þ R i þ Si ð3Þ subjected to stepwise conversion laws. Inert heating law is applied
@t @xi @xi @xi
while the particle temperature is less than the vaporization tem-

Fig. 5. Domain topography (in inch) and mesh details where black represents insulations.
 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792 787

perature. Devolatilization law is applied to the combusting particle Moisture content is of primary importance because of its poten-
mass (mp) when the temperature of the particle reaches the vapor- tial effect on the temperature of the gasifier and hence the heating
ization temperature, Tvap. It is written as: value of syngases produced due to the significant latent heat of
h water evaporation [3]. Wood used in this simulation has a low den-
dmp   i
 ¼ AeðE=RTÞ mp  1  fv0 m0p ð8Þ sity of 400 kg/m3, and a high heating value of 18.7 MJ/kg as deter-
dt mined by the Bomb calorimeter Par600 experiments. Several
where fv, and mop are the volatile fraction and initial mass, respec- complex events occur upon gasification of biomass and the reliabil-
tively. It remains in effect while the mass of the particle, mp, ex- ity of the CFD results is contingent upon accurately modeling these
ceeds the mass of the non-volatiles in the particle. The heat events. The thermal analysis of wood chips DSC results in mass loss
transfer to the particle during devolatilization process governs the evolution and heat balance shown in Fig. 6. Where drying, devola-
contributions from convection, radiation, and the heat consumed tilization, and combustion of wood particles is demonstrated as a
during devolatilization. It is written as: function of temperature. Nitrogen initially was used as a purge
gas whereas air was used as an oxidant at a flow rate of 100 ml/
dT p dmp  
mp cp ¼ hAp ðT 1  T p Þ þ hfg þ ep Ap r T 4R  T 4p ð9Þ min. Samples weighing several milligrams (5–10 mg) were heated
dt dt at a constant heating rate of 5 °C/min up to 180 °C in order to re-
where cp, hfg A, and r are specific heat, latent heat of evaporation, lease their moisture in a nitrogen-only atmosphere. Following
particle surface area and Stefan constant, respectively. After the vol- the release of the moisture, air is injected at 100 ml/min. Heating
atile component of the particle is completely evolved, a surface continued until the sample weight stabilizes ensuring completion
reaction begins, which consumes the combustible fraction of the of reaction at temperature of 600–900 °C.
particle until the combustible fraction is consumed. Heat, momen- The first event of the thermal analysis corresponds to the mois-
tum, and mass transfer between the solid fuel particles and the gas ture release which ceases at the onset of volatile release of about
will be included by alternately computing the discrete phase trajec- 200 °C. This is in agreement with previous reported values found
tories and the continuous phase equations. in [10,11]. This event corresponds to the drying zone in the down-
draft gasifier where heat generated from the exothermic combus-
tion and gasification reactions at the center and bottom of the
6. Feedstock conversion modeling gasifier is used to supply the necessary sensible and latent heat
for the release of moisture. The rate of evaporation is directly pro-
In principle, modeling wood gasification follows similar guide- portional to the particles surface area, the internal diffusivity and
lines generally used for the more conventional coal feedstock. moisture content of the particle, the temperature difference be-
Wood composition is described by its ultimate and proximate tween the particle and the gasifying agent, and the relative humid-
analysis. A typical composition which was experimentally ob- ity and velocity of gases inside the gasifier [7].
tained using the organic elemental analyzer FLASH 2000 and the The second event of the thermal analysis represents the devol-
combined thermogravimetric differential scanning calorimetry atilization process where the volatiles are released at an onset tem-
(DSC) analyzer SDT-Q600 is given in Table 1. perature as low as 200 °C and continues till approximately 350 °C.
This phenomenon was also observed by [10] and similar tempera-
ture intervals were recorded. The main wood material that is
Table 1
Wood composition in mass%. decomposed into various gases and tar during this event is the
hemicelluloses leaving behind a high yield of char [10,11]. The
Proximate analysis Ultimate analysis
air is consumed rapidly at the surface of the particles due to the
Fixed carbon 31.46 Carbon 49.59 fast kinetics of volatile reactions and hence heterogeneous/surface
Volatile 56.39 Hydrogen 6.28
reactions are minimal at this zone [7]. In this model, the volatile
Moisture 9.71 Oxygen 43.74
Ash 2.44 Nitrogen 0.39 content of char is assumed to decompose further into CO, H2,
Sulfur 0.00 H2O, and N2 by virtue of the fuels’ proximate, ultimate, and heating
value. The scheme developed by ANSYS [9] is followed to

Fig. 6. DSC results of wood particle under gasification conditions.

788 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792

Fig. 7. Volatiles species prediction scheme.

Fig. 8. Temperature variation measured throughout the experiment by thermocouples placed through the gasifier center (above) and at 2.5’’ (6.35 cm) offset from the center
(below).
 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792 789

determine the devolatilization species. The corresponding down- Table 2

draft gasifier region for this event is the pyrolysis zone located Gas-phase reactions kinetics.

above the throat of the reactor. The last event in the TG thermal Kinetic parameters Aj, Ej (kJ/mol) Reference
analysis represents the gasification of the char into H2, CO, and AV2 = 1017.6 3 1
((m mol )  0.75 s 1
) [15,6]
CH4 through reactions with hot gases and water vapor produced EV2 = 166.28
along the gasifiers’ upper zones [10]. The heat release curve in AV3 = 1e11 (m3 mol1 s1) [15,6]
Fig. 8 shows that maximum reactivity occurs between 300– EV3 = 42
AV4 = 0.0265 [6]
500 °C due to the reduction of hot gasses by char resulting in the EV4=65.8
production of ash. A similar conclusion was found in [10]. Inside
the downdraft gasifier, this event corresponds to the combustion
zone close to the gasifier air nozzles near its throat and the reduc-
tion zone at the bottom of the gasifier. Table 3
Solid-particle reactions kinetics.

Kinetic Parameters Aj, Ej (kJ/mol) Reference

7. Reactions modeling
Ac1 = 5.67e9 (s1) [6]
Ec1 = 160
7.1. Water evaporation
Ac2 = 7.92e4 (m3 mol1 s1) [6]
Ec2 = 218
Water droplets representing the wet combustion of wood due Ac3 = 79.2 (m3 mol1 s1) [6]
to its moisture content undergoes inert heating, vaporization (boil- Ec3 = 218
ing), and inert heating of vapor. Classical laws found in [9] are use Ac4 = 7.92e4(m3 mol1 s1) [6]
Ec4 = 218
in this model. Water is thus a source for the gas continuity equa-
tion and water species transport equation.

7.2. Volatile release CO þ H2 O $ CO2 þ H2 ðR-V4Þ

The volatile release kinetics were used from [12] since large (R-V2-R-V4) chemical kinetics (pre-exponential factor and activa-
wood particles were studied. The size of the particle greatly affects tion energy) are summarized in Table 2.
the char conversion rate and gasification efficiency, higher conver-
sion is achieved at lower particle size [13]. The work of [12] re-
7.5. Particle surface reactions
ported devolatilization kinetics for a 1 cm size wood particles of
108 s1 for the pre-exponential factor, and 140 kJ/mol K for the
A multiple surface reactions model is used to model the wood
activation energy at the specified devolatilization temperature.
particles combustion and gasification reactions. The reactions con-
sidered involve char combustion, and char reduction of CO2, H2,
7.3. Volatiles decomposition
and H2O. Their kinetic data is summarized in Table 3:

A qualitative and quantitative knowledge of the devolatilization CðSÞ þ O2 ! CO2 ðR-C1Þ

species is essential for realistic modeling of the coupled thermo-
chemical conversion of wood into syngas [14]. Therefore, the vola-
tiles, which are represented by a single molecule of the form Cn0 Hm- CðSÞ þ O2 ! 2CO ðR-C2Þ
0 O 0 N 0 S 0 , are decomposed according to the following reaction (R-
x y z
V1) which holds a strict balance of mass and heat: CðSÞ þ 2H2 ! CH4 ðR-C3Þ
C0:268 H1:456 O0:639 N0:006 ! 0:268 CO þ 0:371 H2 O þ 0:357 H2
CðSÞ þ H2 O ! CO þ H2 ðR-C4Þ
þ 0:003 N2 ð10Þ
The formula of the volatiles essentially represents the elemental
mass balance of the volatile content of wood. The ANSYS [9] devel- 7.6. Continuous and discrete phase coupling
oped scheme for volatiles species prediction was followed. The de-
tails of the scheme which is based on phenomenological The coupling of momentum, heat, and mass transfer between
assumptions of the distribution of the volatiles elemental analysis the solid and gas phases is accounted for by the two-way coupling
over CO, CH4, H2, H2O, H2S, and tar while mass is balanced is shown Cloud model in [9]. In this model the Lagrangian-based particle tra-
in Fig. 7. jectory is tracked as a cloud of particles about a mean trajectory.
The mean trajectory is then calculated by solving the ensemble-
7.4. Gas-phase reactions averaged equations of motion for all particles represented by the
cloud.
Because of the wide gap between the gas-phase and the solid
particle reactions in terms of their kinetics, the finite rate/eddy dis-
sipation is used to model the interaction between the chemistry Table 4
Axi-symmetric model boundary and operating conditions.
and turbulence so that chemically controlled and eddy (mixing)
controlled reactions are accurately modeled. Global reactions rep- Wood particles inlet temperature (°C) 50
resenting CO combustion (R-V2), H2 combustion (R-V3), and Air inlet temperature (°C) 167
Turbulent intensity (%) 10
water–gas shift (R-V4) are accounted for:
Particle specific heat (kJ/kg K) 2.5
Wood particle density (kg/m3) 400
2CO þ O2 ! 2CO2 ðR-V2Þ
Wood particle diameter (mm) 0.1
Wood HHV (MJ/kg) 18
2H2 þ O2 ! 2H2 O ðR-V3Þ
790 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792

7.7. Boundary and operating conditions

The boundary and operating conditions are based on the exper-

imental operation and literature data [6,12,15] for wood gasifica-
tion in downdraft reactors. These conditions are summarized in
Table 4.
Equivalence ratio (ER) for gasification requires an air-deprived
atmosphere, thus the ER for ideal gasification ranges between
0.19–0.43 [4]. For this model, the experimental air and feedstock
feed rates correspond to an ER of 0.24 given that the stoichiometric
mass flow rate ratio of air to fuel is 6:1 for wood [16]. The external
wall was modeled with a mixed convection and radiation condi-
tion simulating the non-adiabatic conditions of the experiment.
The insulation section of the gasifier was modeled using the phys-
ical constants of fiberglass.

8. Results and discussion

8.1. Experimental results Fig. 10. Temperature profile along the gasifier center at t = 20 min.

Temperature variation at each thermocouple location was re-

corded during the experiment as shown in Fig. 8. The centerline
temperature of Fig. 8 starts fairly high for the lowest thermocouple
(over 400 °C) due to lingering heat from a previous run. Ignition is
performed at around 5 min, where the temperature at the lowest
thermocouple spikes up due to heat from the propane torch. After
the ignition is completed and the torch is removed, the tempera-
ture in all thermocouples begins to rise steadily until time equals
20 min, where the gasifier reaches steady state and the tempera-
ture fluctuates around the same value. At around 30 min, the flame
extinguished indicating complete gasification of the wood batch as
shown in Fig. 9. As the chemical reaction died out, the temperature
in the gasifier slowly decreased until recording was stopped.
Using the recorded temperature data, the temperature profile at
20 min at the center of the gasifier is plotted in Fig. 10. This figure
shows strong agreement between the results of this work and the
CFD results presented next. It is evident that the highest tempera-
ture is located near the air nozzle which corresponded to slightly
above 1000 °C at this moment of time (20 min).

8.2. CFD modeling results
The axial temperature field at 4 cm away from the gasifier ver-
tical axis is shown in Fig 11. The corresponding temperature points
used to construct this plot are shown along the gasifier height and
are denoted by circles. By examining the typical wood TG diagram
Fig. 9. Inside look into the gasifier showing the consumption of wood after flame
extinction.
Fig. 11. CFD axial temperature field at 4 cm away from the axis of the gasifier
showing the corresponding location of the computed points along the gasifier
height.
given in Fig 6, the height and location of the different zones within
the gasifier can be approximated.
Point A represents the end of the drying zone at about 185 °C
(as predicted by TG curve and [7]). Point B represents the end of
the devolatilization zone at about 825 °C where volatiles are re-
leased right above the air nozzle outlet. The TG curve shows a sig-
nificat heat release followed by a heat sink in the devolatilization
zone and the same trend is observed in Fig 6 with respect to tem-
perature. The heat release can be due to exothermic reactions of
very light voaltiles which were released at lower temperatures.
Whereas the heat sink can be related to the endothermic nature
of devolatilization reacton. Realistically, a complex rather than a
linear process occurs and simulataneous exothermnic and endo-
thermic reactions take place resulting in an overall net effect on
the reactor temperature. As expected, the heighest temperature
is located in front of the air nozzle where combustion takes place
at high concentrations of oxygen. This corresponds to the combus-
tion zone which extends from about 825–1000 °C at the thraot of
the gasifier. Finally, the hot gases are reduced by the remaining
unreacted char in the reduction/gasification zone at approximately
1000 °C.
 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792 791

Fig. 12. CFD particle temperature, char concentration, and volatile source as it travels down the gasifier. The species distribution of produced syngases CO and H2 are shown
in Fig. 13 below.

Fig. 13. CFD contours of mole fraction of CO (left) and H2 (right).

Table 5
Fig. 12 illustrates the physical and chemical evolution of a single Producer gas composition for CFD and ideal equilibrium case with the corresponding
CGE and average reactor temperature.
wood particle using its temperature (°C), char concentration (kg/
m3), and volatile source (kg/s) along the gasifier height. As indi- Gas component CFD Equilibrium (Ideal)
cated previously, the volatiles are readily released upon the parti- calculations mole%
Non-adiabatic Adiabatic
cles injection due to the high temperatures within the gasifier and
the relatively high volatile content of wood. Consequently, the con- Mole% Mass% Mole% Mass%
centration of char increases as the particle’s temperature drops CO 38.23 51.23 51.93 57.59
owing to the endothermic devolatilization process. At the pyrolysis H2 23.24 2.24 10.65 0.85
CO2 0.85 2.05 8.01E04 0.14
zone, the last remaining un-evolved volatiles are released as indi-
H2O 11.79 10.72 3.56E04 2.58–04
cated by the small jump in volatile source at about 0.2 m below the CH4 0.11 0 1.79E4 8.61E5
hopper feeding point as indicated in Fig. 12. Char is consumed as it N2 25.76 33.66 37.29 41.37
travels down the reactor and the temperature of the particle in- O2 0 0 0 0
creases. An abrupt temperature increase of the particle is observed CGE 70% 89%
Carbon conversion% 100 100
near the air nozzle located at the dashed black arrow. This jump in
Reactor average temperature (°C) 1140 1080
temperature signifies the hot combustion zone near the gasifier
792 I. Janajreh, M. Al Shrah / Energy Conversion and Management 65 (2013) 783–792
throat. Notice that char consumption slope changes fiercely at the
intersection point with the particle temperature where the jump in
temperature starts due to the combustion reactions.
The outlet gas composition calculated using CFD is summarized
in Table 5 and compared with the ideal, adiabatic, equilibrium
case. The equilibrium calculations are based on the same equiva-
lence ratio for the feedstock of the CFD model. Three reactions
namely: Water gas, boudouard, and methanation reactions consti-
tute the governing equations in addition to the four elements bal-
ance (C, H, O, N), heat balance, and Dalton equation for the total
molar fraction of producer gas are also used. The producer gas tem-
perature, ratio per mole of feed, and the yield of seven gases (CO,
H2, CH4, H2O, CO2, O2, and N2) are then calculated.
Cold gas efficiency (CGE) is calculated based on the LHV of the
produced H2, CO, and CH4 compared to that of the feedstock [13].
The significantly higher CGE for the ideal case is directly related
to the higher yields of syngases.
It is noticed that the non-adiabatic conditions of the CFD model
results in heat losses to the environment which are not captured in
the ideal case. Consequently, the maximum temperature computed
by CFD is far less than that of the ideal case. The main reasons are
neglecting the convective heat losses, and under-simplifying the
chemistry kinetics in the equilibrium model.
The low experimentally recorded temperatures highlight one of
the challenges of small scale gasification namely, heat losses. Due
to the smaller reactor diameter, the surface to volume ratio is high
resulting in large heat losses for small scale gasifiers compared
with Mega Watt-scaled ones. The higher the heat losses, the lower
the temperature inside the gasifier, and hence the endothermic
reaction rates are lowered resulting in less yield of syngases and
less gasification efficiency.
9. Conclusion
The potential of using small scale biomass for gasification is
investigated from a thermo-chemical conversion point of view. A
commercial downdraft gasifier scaled for 10–20 kW batched with
wood chips of medium size (0.5 cm thickness, 1–2 cm width, and
2–2.5 cm length) was operated and the temperatures at ten differ-
ent locations along gasifier were recorded.
The experiment provided adequate understanding of the limit-
ing factors and the challenges facing the implementation of small
scale downdraft gasifiers for biomass gasification. It was shown
that the density in terms of the heating value and moisture content
of the feedstock significantly affects reactivity. The consumption of
the batch of wood was relatively fast in the order 25 min suggest-
ing the importance of using a higher fixed carbon content (i.e.
dense) feedstock or the availability of continuous feeding rather
than semi-batched.
The lowest thermocouple stretching 12 in (30.48 cm) into the
19 in (48.26 cm) height gasifier (located close to the air nozzle)
measured temperatures slightly higher than 1000 °C as expected.
The gasifier needed 10–15 min to reach near steady state after
which localized fluctuations persist. One of the drawbacks of such
small scale gasifiers was found to be the inhomogeneous temper-
ature field as indicated by the large variation in the measured axial
temperatures.
An axi-symmetrical CFD model for the small scale downdraft
gasifier was set-up and the temperature field along the axial direc-
tion of the reactor near the air nozzle outlet was computed and
compared to the experimental results. Theoretically, the tempera-
ture profile representing the net effect of the endothermic and exo-
thermic reactions matched the expected trend and it also matched
the expected temperature evolution predicted by the experimental
thermogravimetry TG-DSQ plot for heat balance on a single wood
particle.
The average temperature computed by CFD was Higher com-
pared to that measured experimentally and comparable to the cal-
culated ideal one which corresponds to equilibrium conditions.
The significant heat losses faced in small scale gasifiers due to their
large surface to volume ratios was illustrated by comparing the
average temperature of the CFD, experimental and ideal cases.
The efficiency of biomass gasification is assessed using the cold
gas efficiency (CGE). The CFD computed CGE was found to be 19
points less of that calculated for the ideal case. Therefore, due to
the complexity of the flow inside the downdraft gasifier, equilib-
rium modeling does not capture the physics and chemistry inside
the downdraft gasifier compared to other types of gasifiers espe-
cially those characterized by their high temperatures such as the
entrained flow gasifiers.
To account for the particle size of the wood, devolatilization ki-
netic data from literature [12] for particles of 1 cm length were
used to offset the effect of the particle diameter (0.1 mm) modeled
using the discrete phase method (DPM) in ANSYS [9]. The corre-
sponding CFD and experimental temperature profiles suggest that
this assumption is reasonable.
Acknowledgments
The support of Masdar Institute is highly acknowledged as well
as the help of the Waste-to-Energy laboratory team members. The
contribution of Mohamed Al Hamahmy, Arnar Valmundsson, and
Rana Qudaih to the experimental work is also greatly
acknowledged.
References
[1] Senneca O. Kinetics of pyrolysis, combustion and gasification of three biomass
fuels. Fuel Process Technol 2007;88(1):87–97.
[2] Demirbas A. Potential applications of renewable energy sources, biomass
combustion problems in boiler power systems and combustion related
environmental issues. Prog Energy Combust Sci 2005;31(2):171–92.
[3] Hsi C, Wang T, Tsai C, Chang C, Liu C, Chang Y, et al. Characteristics of an
air-blown fixed-bed downdraft biomass gasifier. Energy Fuels 2008;22(6):
4196–205.
[4] Zainal ZA, Rifau A, Quadir GA, Seetharamu KN. Experimental investigation of a
downdraft biomass gasifier. Biomass Bioenergy 2002;23(4):283–9.
[5] Kumar R, Kolar A, Leckner B. Effect of fuel particle shape and size on
devolatilization time of Casuarina wood; 2004.
[6] Blasi CD. Dynamic behaviour of stratified downdraft gasifiers. Chem Eng Sci
2000;55(15):2931–44.
[7] Dogru M, Howarth CR, Akay G, Keskinler B, Malik AA. Gasification of hazelnut
shells in a downdraft gasifier. Energy 2002;27(5):415–27.
[8] McKendry P. Review paper: energy production from biomass (part 3):
gasification technologies. Bioresource Technol 2002;83:55–63.
[9] FLUENT ANSYS 13. Fluent user manual. ANSYS Inc.; 2010.
[10] Shanmukharadhya KS. Simulation and thermal analysis of the effect of fuel
size on combustion in an industrial biomass furnace. Energy Fuels 2007;21(4):
1895–900.
[11] Grioui N, Halouani K, Zoulalian A, Halouani F. Thermogravimetric analysis and
kinetics modeling of isothermal carbonization of olive wood in inert
atmosphere. Thermochim Acta 2006;440(1):23–30.
[12] Chan WR, Kelbon M, Krieger BB. Modelling and experimental verification of
physical and chemical processes during pyrolysis of a large biomass particle.
Fuel 1985;64(11):1505–13.
[13] Jayah TH, Aye L, Fuller RJ, Stewart DF. Computer simulation of a downdraft
wood gasifier for tea drying. Biomass Bioenergy 2003;25(4):459–69.
[14] Kumar M, Ghoniem A. Renewable energy from waste, thermochemical
pathways. MIT Report; 2009.
[15] Tinaut FV, Melgar A, Pérez JF, Horrillo A. Effect of biomass particle size and air
superficial velocity on the gasification process in a downdraft fixed bed
gasifier. an experimental and modelling study. Fuel Process Technol
2008;89(11):1076–89.
[16] GEK Wiki sandbox. Retrieved 11.11.10, from <http://wiki.gekgasifier.com>.