Accepted Manuscript

Numerical Analysis of Wood Biomass Packing Factor in a Fixed-bed Gasification

Process

William A. González, Juan F. Pérez, Sergio Chapela, Jacobo Porteiro

PII: S0960-1481(18)30063-6

DOI: 10.1016/j.renene.2018.01.057

Reference: RENE 9662

To appear in: Renewable Energy

Received Date: 23 August 2017

Revised Date: 22 December 2017

Accepted Date: 17 January 2018

Please cite this article as: William A. González, Juan F. Pérez, Sergio Chapela, Jacobo Porteiro,
Numerical Analysis of Wood Biomass Packing Factor in a Fixed-bed Gasification Process,
Renewable Energy (2018), doi: 10.1016/j.renene.2018.01.057

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 Numerical Analysis of Wood Biomass Packing Factor in a Fixed-bed Gasification

2 Process
3
4 William A. González a, Juan F. Pérez a,*, Sergio Chapela b, Jacobo Porteiro b
5
6 a Grupo de manejo eficiente de la energía (GIMEL), Departamento de ingeniería
7 mecánica, Facultad de ingeniería, Universidad de Antioquia; Calle 67 No. 53-108,
8 Medellín, Colombia
9 b E.T.S. Ingenieros Industriales, Universidad de Vigo, Rúa Maxwell s/n, Campus Lagoas-

10 Marcosende, 36310 Vigo, Pontevedra, Spain.

11
12 Abstract
13 The biomass gasification process in fixed bed was studied by means of computational
14 fluid dynamics (CFD) numerical analysis. The aim was to evaluate the effect of the
15 biomass packing factor on the thermochemical process. The fuel-wood used was
16 Jacaranda Copaia in various shapes: chips, cylinders, and cubes with packing factors
17 (PF) of 0.38, 0.48, and 0.59, respectively. The mathematical model is a transient 2D CFD
18 model, which was developed through the implementation of User Defined Functions in
19 ANSYS-Fluent. The model was extended to simulate the gasification process by
20 expanding the chemical kinetic mechanism and by adapting the stages of pyrolysis,
21 oxidation, and reduction. The model was validated with experimental data. The average
22 relative error between experimental and numerical data was 5.45%. By means of the
23 sensitivity analysis, it was found that with an increase in the packing factor from 0.38 to
24 0.59, the absorption of radiative heat transfer increases by 27% leading to increase the
25 solid temperature in the reaction front, but due to a lower penetration of radiation, the
26 drying and pyrolysis reaction rates decrease. But nevertheless, the higher solid
27 temperature with packing factor favors the convective solid-gas heat transfer in the drying
28 stage.
29 Keywords: biomass; packing factor; fixed bed gasification; CFD; numerical analysis;
30 heat transfer
31
32 1. Introduction
33 Biomass is a renewable energy source and its energy use in fixed bed gasifiers is an
34 important process for power generation and cooking systems [1]. The comprehension of
35 thermal, physical, and chemical phenomena involved in the biomass to gas (BTG)
36 transformation enables improved reactor efficiency and reduced pollutant emissions [2].
37 The gasification process includes mass transfer mechanisms associated with drying,
38 pyrolysis, oxidation, and reduction stages, and energy transfer mechanisms, such as
39 convection and radiation. Therefore, acquisition of experimental data with complex

*Corresponding author. Phone: (+57 4) 2198552. Fax: (+574) 2110507. e-mail: juanpb@udea.edu.co
(Juan F. Pérez).

1
 ACCEPTED MANUSCRIPT

1 parameters during the experimental stage is challenging [3]. Modeling helps to describe
2 system behavior, enabling identification of complex phenomena, such as incident
3 radiation, solid-gas heat transfer, diffusive parameters, and the kinetic interaction of
4 particles in solid-gas reactions [4,5].
5 There is an ongoing line of work in the scientific community that study the conversion of
6 biomass in fixed beds by means of CFD models using commercial software. The CFD
7 models require a phenomenological description of the biomass conversion to gaseous
8 fuels. This interaction between the solid and gas phases has been implemented through
9 User Defined Functions (UDFs), where sub models created in C++ were adapted; these
10 sub-models are embedded into the code and therefore are also solved by the CFD software
11 [2,6,7].
12 CFD models of biomass combustion and/or gasification in fixed beds have been used to
13 study the phenomenology involved during thermochemical conversion, various important
14 operation parameters of the process, and chemical kinetic mechanisms, and physical
15 properties of biomass have been analyzed. Several authors developed dimensional models
16 for both combustion [2,3,8] and gasification [9,10] in fixed beds, with their work aiming
17 to study the influence of the heterogeneous properties of the bed during combustion, and
18 to compare temperature fields and gaseous concentrations in the two-dimensional domain
19 during gasification. Other authors used numerical models for validating such parameters
20 as combustion time, furnace temperature, combustion gases emissions (including NOx),
21 carbon content in the ashes, and combustion total efficiency [11]. Regarding fixed bed
22 combustion, several studies experimentally measured flame front velocity, process speed
23 movement (which is dependent on the air supply velocity), biomass heating value, and
24 particle size in order to obtain a transient behavior of the local temperature, oxygen
25 consumption rate, and heat transfer phenomena using model predictions [12]. Other
26 authors have also validated gas composition for different air flows in downdraft
27 gasification, as well as the temporal and spatial evolution of temperature, comparing
28 model results with experimental ones available in the literature [13]. For downdraft
29 gasifiers, the Lagrangian model to predict the temperature field inside the reactor, and
30 comparing the theoretical temperature distribution with experimental data has been
31 proposed by Janajreh et al [14]. Other studies have focused on the analysis of the effect
32 of the geometry and configuration of the combustion chamber over combustion
33 efficiency, emissions and process temperature [6]. Regarding biomass shape and size,
34 there are thermochemical conversion models used to predict the intra-particle temperature
35 gradient, mass loss rate, particle size and density [7]. Physicochemical and geometrical
36 properties of biomass have also been studied during the operation of reactors and burners,
37 with several authors presenting numerical models where two types of geometries can be
38 evaluated, including cylindrical particles (horizontal and upright) and spherical ones [7].
39 Other authors have focused on the analysis of the biomass gasification process through
40 the evaluation of different types of biomass with several particle geometries, considering

2
 ACCEPTED MANUSCRIPT

1 physicochemical properties, such as density, void fraction, sphericity, surface/volume

2 ratio [13] and particle diameter [6].
3 The literature also presents the characterization and analysis of the flame front and
4 process velocity movement as functions of the input experimental settings, such as air
5 mass flow, biomass heating value and particle size. Those studies analyze the transient
6 evolution of the reaction front for all solid to gas conversion stages [13], studying mass
7 and heat transfer phenomena (including radiation, convection, and conduction
8 phenomena) [12].
9 In general, CFD numerical models enable the prediction of the reactor/plant behavior to
10 the variation of several operating conditions [11], where the analyses of the channeling
11 generated in the bed is highlighted. Channeling has been blamed of generating higher
12 concentrations of nitrous oxides and unburned products under combustion regimes [8].
13 This work presents a fixed bed biomass combustion model that has been extended to
14 tackle gasification conditions. The effect of the packing factor on the gasification process
15 behavior was evaluated. The validating parameters are temperature fields, flame front
16 velocity, biomass consumption rate, syngas volumetric flow, low heating gas,
17 equivalence ratio, cold gas efficiency, and producer gas composition. Afterwards, a
18 sensitivity analysis of convection and radiation heat transfer mechanisms was performed
19 in function of biomass packing factor, enabling a better description of the phenomenology
20 involved in the process with different physical properties of wood biomass. The model
21 used in this work allows to study different types of biomass, such as energy crops, forest,
22 agricultural waste, and others, as renewable feedstock for thermochemical processes.
23 These solid biofuels are characterized by different densities (bulk and particle) [15]. The
24 packing factor (PF) quantifies the relationship between these two parameters of
25 significant importance for the biomass gasification process in fixed bed [16].
26
27 2. Materials and methods
28 According to the state of the art, the numerical analyses of the effect of biomass packing
29 factor (PF=Average bulk density/Particle density) on fixed bed gasification/combustion
30 by means of models or experimental tests are scarce. The PF parameter is a fundamental
31 variable that affects gasification performance due to the random packing of biomass in a
32 fixed bed [17]. Therefore, the packing factor is the process parameter (the operating
33 condition as a function of biomass physical properties) considered in this study. Three
34 values for the PF were studied and selected to perform the sensitivity analysis. The three
35 geometries of wood biomass are chips, cylinders and cubes, with a PF of 0.38, 0.48 and
36 0.59, respectively.
37 The performance of the biomass gasification process as a function of PF has been
38 evaluated in this work by analyzing its effect on heat transfer mechanisms of the reaction
39 front (solid-gas convective heat transfer and radiative heat transfer in solid phase) among
40 other key variables, such as solid temperature field, flame front velocity, biomass
41 consumption rate, gas velocity, and fuel/air equivalence ratio. These parameters are the
42 results obtained from the CFD model.

3
 ACCEPTED MANUSCRIPT

1 The experimental setup shown in Figure 1 was designed and built for determining the
2 process propagation velocity during biomass gasification in a batch type gasifier [18].
3 This reactor is known as a top-lit updraft reactor, or reverse downdraft fixed-bed reactor;
4 the experimental setup is described in section 2.1. The experimental results were taken
5 from a previous work presented by Lenis et al. [1] and were used to validate the CFD
6 model as described in section 3.
7 The model accuracy can be quantified through the root mean square error (RMSE) (see
8 Eq. (1)). Other authors have used this parameter to determine the errors of gasification
9 models with regard to the experimental data [19,20].
10
n

RMSE =
∑ (Xreference,i ‒ Xmodel,i)2 (1)
i=1
n
11
12 2.1. Fuel physical and chemical characterization
13 The biomass used in the experimental tests considered as reference to validate the CFD
14 model under gasification conditions is Jacaranda Copaia wood, this is a fast-growing
15 wood native from Latin America, which reaches yields between 25-35 m3/ha/year with
16 harvested time around 15 years [1]. Table 1 shows the biomass physical and chemical
17 characterization for different shapes of wood fuels. Moreover, the volumetric airflow that
18 has been setting in each experimental test is presented. The complete experimental study
19 is shown in detail by Lenis et al. [1].
20
21 Table 1. Physical and chemical characterization of Jacaranda Copaia woods (±
22 standard deviation)
Properties Wood biomass shape
Chips Cylinders Cubes

Air flow [slpm] 40.93±0.67 40.68±0.71 40.83±0.69

Dimensions [mm±0.1 mm] 6-13 Ø10 x L8 10 x 10 x 8
Average bulk density [kg/m3] 130.07±0.57 209.61±2.38 198.59±1.55
Particle density [kg/m3] 344.21±16.01 434.43±15.92 338.52±11.19
Particle average hardness [Shore D] 33.75±3.08 43.75±2.41 33.82±3.62
Packing factor [--] 0.38±0.02 0.48±0.02 0.59±0.02
Equivalent radius [mm] 5.65±4.2 6.35±4.4 6.8±2.3
Sphericity [--] 0.76±0.11 0.85±0.09 0.79±0.02

4
 ACCEPTED MANUSCRIPT

Moisture [%] 5.4±0.02 9.4±0.1 9.4±0.1

Surface area/volume [mm-1] 0.69±0.02 0.55±0.01 0.56±0.01
Proximate analysis [% wt. d.b.]
Volatile matter 89.92
Fixed carbon 8.67
Ash 1.41
Ultimate analysis [%wt. d.a.f.]
C 49.88
H 7.35
O 41.08
N 1.52
S 0.17
Substitution formula [d.a.f] CH1.769O0.617N0.026S0.001
LHVdb [kJ/kg] 18212.95
1
2 2.2. Experimental setup
3 The experimental setup used for validating the data obtained by simulation consists of a
4 laboratory autothermal fixed bed gasifier that operates under atmospheric pressure and
5 temperatures between 800 °C and 1000 ºC under statistical repeatability conditions [17].
6 The main dimensions of the reactor are 102 mm internal diameter, 185 mm external
7 diameter, and a bed length of 400 mm. The gasifying agent during the process is air,
8 obtained from the lab’s compressed air line, which is controlled by pressure regulators
9 and valves, and measured by a Honeywell AWM500 flow meter. The gasifier gas outlet
10 has a conditioning unit (drying and filtering) to measure the syngas composition (O2, N2,
11 CO2, CO, H2 and CH4) using Agilent 3000 Micro GC gas chromatography.
12 The gasifier is fed with biomass of different shapes, and it is lit from the top; this reactor
13 is also known as an inverse reactor or top-lit up-draft (TLUD) reactor. The gasifying agent
14 (air) is supplied from the bottom, causing a downward propagating reaction front. The
15 temperature profile is measured by nine K type thermocouples with a separation between
16 each one of 30 mm; these are inserted up to the center of the reactor as the fuel is fed.
17 Figure 1 shows a scheme of the experimental setup. Characteristic variables, such as the
18 flame front progress from top to bottom, are measured. It is possible to assert that tests
19 are conducted under unsteady state. This is because the gasifier walls are heated as the
20 flame front advances [17]. A full description of this experimental configuration and the
21 parameters derived thereof are described in detail in [4,21,22].
22

5
 ACCEPTED MANUSCRIPT

1
2 Figure 1. Experimental setup
3
4 3. Model description
5 The fixed bed biomass gasification approach was developed by considering separately
6 the solid and gaseous phases. The model presented in this work is an extended version of
7 the model developed for biomass combustion in fixed beds. All the solid parameters are
8 formulated as Eulerian scalars that are embedded into the code by User Defined Scalars
9 (UDS). The interaction between phases is modeled as sources in the corresponding
10 transport equations, which are introduced in the code by User Defined Functions (UDFs)
11 [2,3,6,7]. The solid fuel transformation in a gaseous fuel considers the main stages
12 involved in the gasification process, such as drying, pyrolysis, oxidation, and reduction
13 [4]. The species of solid phase (biomass) are solid fraction, moisture density, dry biomass
14 density, char density and solid phase temperature [2]. The geometry of solid biomass is
15 also considered to estimate the parameters involved in the solid-gas mass and heat transfer
16 mechanisms. The model has been developed under the following hypothesis:
17 1) The bed is considered fixed during the time-step, i.e., any movement of the solid due
18 to the feeding process or due to compaction of the bed, takes place at the end of the
19 time-step and before the next one.
20 2) Biomass particles are considered thermally thin [4,23].
21 3) The volume of biomass does not vary during the drying and pyrolysis processes
22 (shrinking density regime) [3,4].
23 4) The volume (particle size) of the biomass particle is affected by the oxidation and
24 reduction reactions according to the shrinking core regime [24].
25 5) The moisture evaporation temperature is 100ºC and the process is thermally
26 controlled [3,24].
27 6) The gases are considered as an ideal gas mixture [4,24].
28 7) Immediate outflow, i.e., there is no diffusion limitation for the gases from the interior
29 of the particles to their surface.

6
 ACCEPTED MANUSCRIPT

1 8) The solid-gas reaction (oxidation and reduction) starts after the devolatilization
2 process ends [3].
3 The CFD model presented in this work uses the SIMPLE algorithm for the governing
4 equations [3,8]. The momentum and energy special discretization were solved through a
5 second order upwind method, while radiation and pressure discretization were solved
6 using PRESTO [25]. The radiation heat transfer was modelled using a modification of the
7 Discrete Ordinate Model (DOM) where the solid phase is included [2,3]. The turbulence
8 was modelled by the k-epsilon model with Enhanced Wall Treatment (EWT).
9 The model presents a conservative equation for each scalar variable of the solid phase.
10 Table 2 shows the conservative equations used by the model (Eqs. (2)-(6)). The energy
11 conservation of the solid phase is described by Eq. (2). This is a two dimensional and
12 unsteady differential equation. The mass conservation of solid fraction, moisture density,
13 dry wood density, and char density were modeled by equations (3) to (6). The source term
14 of the energy balance for the solid phase is presented in (Eq.(7)). These energy sources
15 involve the energy consumed by drying, pyrolysis, and reduction, and the energy released
16 by oxidation. Moreover, the convective and radiative heat transfer mechanisms, and the
17 heat losses are considered in this energy source. The complete description of the model
18 was presented in detail by Gómez et al. and Collazo et al. [2,3].
19
20 Table 2. Solid phase conservation equations
Solid phase energy conservation
∂(ερpCpTs) (2)
= ∇(ks,eff·∇Ts) + Ss
∂t
Solid fraction
''' (3)
∂ε ωchar
=‒ ε
∂t ρp
Moisture density
∂(ερmoist) (4)
'''
=‒ ωmoistε
∂t
Dry wood density
∂(ερwood) (5)
'''
=‒ ωwoodε
∂t
Char density
∂(ερchar) (6)
∂t
( '''
)
'''
= ωG,char ‒ ωC,char ε
Solid energy Source
reac reac conv rad loss (7)
Ss = S s + Schar + S s + S s + S s

21

7
 ACCEPTED MANUSCRIPT

1 Table 3 shows the different source terms of the energy equation in solid phase (Eq. (7)).
reac
2 Eq. (8) is the absorbed energy during the drying and devolatilization processes, Schar term
3 is represented in the Eq. (9). This source term accounts for the energy released or
4 consumed due to char reactions (oxidation and reduction). The convective heat exchange
5 is opposite to the gas phase energy source (Eq. (10)). Through this parameter the
6 convective solid-gas heat transfer is calculated. Eq. (11) is the approach to the radiation
7 source term in the solid phase. The calculated energy loss due to the loss of solid mass in
8 a cell over time is expressed in Eq. (12).
9 The chemical reactions of the previous model, biomass combustion under fixed bed
10 conditions, considers the drying, pyrolysis, oxidation of char, tar, methane, hydrogen, and
11 carbon monoxide dissociation [2,3]. However, the kinetic rates of theses reactions had
12 been adapted looking that the model estimate the producer gases composition with good
13 accuracy, the kinetic rates of the reaction mechanism considered in this model are
14 presented in Tables 5 and 7. To extend the model for simulating the biomass gasification
15 process under fixed bed conditions, the kinetic mechanism was extended to include the
16 reduction of steam, carbon dioxide, and hydrogen with char in the heterogeneous
17 reactions (see Eqs (14)-(16)). Moreover, the methane reforming and the water gas shift
18 reaction have been included in the homogeneous reaction mechanism, see Eqs. (17)-(22).
19 The new kinetic mechanism was adapted to model the different sub-processes involved
20 in the fixed bed gasification process [4].
21
22 Table 3. Energy source terms
reac ''' ''' reac (8)
S s =‒ ωmoisε LHmoist ‒ ωwood fgas ε LHdev + Schar ε
reac
Schar
(9)
= (
kglobAvCO MC[(2φ ‒ 1)∆HCO + 2(1 ‒ φ)∆HCO] + kglobAvCCO MC∆Hg1 + kglobAvCH OMC∆Hg2
ox g1 g2
2 2 2 2

)k s/g
conv conv (10
S s =‒ S g = hAv(Tg ‒ Ts)
)

2 4 (11
( )
4π αsn σT s
rad
Ss = ∫ 0
αsI(r,s) ‒
π
dΩ )

S
loss
s = (ωmois
'''
+ ωwood fgas + ωchar)ε(CpTs)
''' ''' t ‒ ∆t (12
)
23
24 Table 4. Chemical reactions
Process Chemical reaction
Drying km
Rm: H Ol → H2Ov (13)
2

8
 ACCEPTED MANUSCRIPT

{
Pyrolysis kp1
→ gas
kp2
Rp1 ‒ Rp3: Biomass = → tars (14)
kp3
→ char

Heterogeneous char reactions kox

Rox: C + φO2 → 2(1 ‒ φ)CO + (2φ ‒ 1)CO2 (15)
kg1
Rg1: C + CO2 → 2CO (16)
kg2
Rg2: C + H2O → CO + H2 (17)

Homogeneous reactions kc1

Rc1: C H6 + 4.5O2 → 6CO + 3H2O (18)
6
kc2
Rc2: CH4 + 1.5O2 → CO + 2H2O (19)
kc3
Rc3: 2CO + O2 → 2CO2 (20)
kc4
Rc4: 2H2 + O2 → 2H2O (21)
kg3
Rg3: CH4 + H2O → CO + 3H2 (22)
kwg
Rwg:CO + H2O → CO2 + H2 (23)
1
2 The moisture evaporation temperature is 100ºC, but the model considers that in the
3 already dried external layers, heating continues. Consequently, when the biomass particle
4 starts to dry, a portion of the heat gained is used for drying the inner layers while the rest
5 is invested in the overheating of the outer layers. This effect modelled by Eq.(24), in
6 which only part of the heat gained by the particle () is employed for moisture
7 evaporation. The pyrolysis is considered a conversion of the dry biomass into gas, tar,
8 and char. The kinetic reaction of this process is modelled by three Arrhenius rates (Eq.
9 (25) and Table 5).
10
ρpCp ∂Ts
'''
ωmoist = τ , T ≥ Tevap (24)
LHmoist ∂t s
11
3

( )
Ei
'''
ωp = ‒ ρwood ∑ Aiexp ‒
Ru T (25)
i=1

12

9
 ACCEPTED MANUSCRIPT

1 Table 5. Kinetics of pyrolysis reactions

ki Ai Units of A E [J·mol ]
‒1 Ref.
i i
kp1 1.44·10
4
s
‒1 88000 [26]
kp2 4.13·10
6
s
‒1 112700 [26]
kp3 7.38·10
5
s
‒1 106500 [26]
2
3 The char consumption reaction was modelled through three heterogeneous reactions (Eq.
4 (26) and Table 4). The global constants of heterogeneous reactions are described in Eq.
5 (27). The parameter φ is considered as the char oxidation parameter (Eq. (28)). Kinetics
6 reaction rates of the char oxidation are expressed in Table 6.
7
'''
ωc = k
glob glob glob (26)
ox AvCO2MC + k g1 AvCCO2MC + k g2 AvCH2OMC

8
glob 1
k i = 1 1 (27)
+ m
ki k
i
9
2 + 4.3exp ( ‒ 3390 T) (28)
φ=
2(1 + 4.3exp ( ‒ 3390 T))
10
11 Table 6. Kinetics of heterogeneous char reactions
ki Ai Units of Ai E [J·mol ]
‒1 Ref.
i
kox 1.7·Ts m·s
‒1 74830 [27,28]
kg1 3.42·Ts m·s
‒1 129700 [27,28]
kg2 5.7114·Ts m·s
‒1 129700 [28]
12
13 The volumetric source terms in the solid-gas conversion are shown in Table 7. Eq. (29)
14 represents the H2O source in the drying process. The source of volatile species in the
15 devolatilization is expressed by Eq. (30). Eqs. (31)-(35) show sources terms of the
16 different species involved in the adapted heterogeneous reactions.
17
18 Table 7. Volumetric source terms in the solid-gas reactions
'''
SH O,moist = ωmoisε (29)
2

( '''
Si,vol = γi ωwood ‒ ωg,char ε
'''
) (30)
SCO ((2φ ‒ 1)kglob
2,char
=
ox
AvCO MCO ‒ kglobAvCCO MCO )ε
g1
2 2 2 2
(31)
SCO,char = ((2φ ‒ 1)kglobAvCO MCO + 2kglobAvCCO MCO + kglobAvCH )
ox g1 g2
MCO ε (32)
2 2 2O

10
 ACCEPTED MANUSCRIPT

g2
SH =‒ kglobAvCH MH ε (33)
2O,char 2O 2O
g2
SH = kglobAvCH MH ε (34)
2,char 2O 2
ox
SO =‒ φkglobAvCO MO (35)
2,char 2 2

1
2 The kinetic rates of the homogeneous reactions are shown in Table 8. The reaction
3 constants behave on the Arrhenius form. The turbulence and kinetic interaction in the
4 gaseous phase are calculated by the Finite-rate/Eddy-dissipation (FR-ED) model [3].
5
6 Table 8. Kinetic rate of homogeneous reactions
Reaction 3 Ref.
Kinetic rate [kmol/m /s]
Rc1 8
''' 9
ωc1 = 1.3496·10 exp ( 𝑅𝑢𝑇
CC H C O
6 6 )
1.256·10 ‒ 0.1 1.85
2
[2]

Rc2 7
''' 8
ωc2 = 2.9093·10 exp ( 8.023·10
𝑅𝑢𝑇 )
0.5
TgCCH CO
4 2
[29]

Rc3 8
''' 19
ωc3 = 7.0795·10 exp ( 1.6628·10
𝑅𝑢𝑇 )
0.25 0.5
CCOC O CH O
2 2
[29]

Rc4 7
''' 14
ωc4 = 1·10 exp ( 4.2·10
𝑅𝑢𝑇 )
CH CO
2 2
[29]

Rg3 8
''' 6
ωg3 = 3.015·10 exp ( 1.2552·10
𝑅𝑢𝑇
CCH CH O
4 2 ) [30]

Rwg 7
'''
ωwg = 2780exp ( 1.26·10
𝑅𝑢𝑇 )
CCOCH O
2
[2]

7
8 4. Results and Discussion
9 The aim of this work is twofold. First, to extend the CFD model initially developed for
10 simulating combustion in a fixed bed to simulate the gasification process. Therefore, in
11 this section, a detailed model validation is presented. Second, to take advantage of the
12 model versatility and power calculation, this study evaluated the effect of the packing
13 factor of wood biomass on the fixed bed gasification process.
14
15 4.1. CFD model validation
16 4.1.1. Temperature field
17 Figure 2 shows the comparison of temperature profiles obtained experimentally (coded
18 with the suffix - exp), see Figure 2 a, c, and e, against those calculated by numerical
19 simulation (coded with the suffix - mod), see Figure 2 b, d, and f.

11
 ACCEPTED MANUSCRIPT

1 The slight overestimation of the maximum temperature calculated by the model, shown
2 in Figure 2, is probably due to the bridging formation (channeling) taken by the air in
3 some places inside the bed. The air channels favor small combustion zones, therefore, the
4 temperature is increased [8,22]. Nevertheless, the model proved versatile and produced
5 results similar to the experimental results. Numerical results in agreement to those
6 reported for fixed bed biomass combustion are highlighted [12,13].
7 If we compare the experimental and model temperature profiles as functions of time, and
8 we measure the elapsed time between the rise of thermocouple #1 and #9, the average
9 discrepancy is 13%. This is a demonstration of the good agreement between the empirical
10 and simulated values of some key parameters such as flame front velocity, the maximum
11 temperature inside the bed, among others (see section 4.1.3.)
12
1200 1200
1000 Temperature [ºC] 1000
Temperature [ºC]

800 800

600 600

400 400

200 200

0 0
0 500 1000 1500 0 500 1000 1500
Time [s] Time [s]
a) PF=0.38 - exp b) PF=0.38 - mod
1200 1200
1000 1000
Temperature [ºC]

Temperature [ºC]

800 800
600 600
400 400
200 200
0 0
0 500 1000 1500 0 500 1000 1500
Time [s] Time [s]
c) PF=0.48 - exp d) PF=0.48 - mod

12
 ACCEPTED MANUSCRIPT

1200
1200
1000 1000
Temperature [ºC]

Temperature [ºC]
800 800
600 600
400 400
200 200
0 0
0 500 1000 1500 0 500 1000 1500
Time [s] Time [s]
e) PF=0.59 - exp f) PF=0.59 - mod

1 Figure 2. Temperature fields

2
3 Analyzing the calculated temperature fields for the three PFs (Figure 2 b, d and f), a higher
4 separation in the first three thermocouples is observed. This finding is due to the model
5 stabilization after ignition. In addition, the model results show two positive slopes in the
6 solid temperature (Figure 2 b, d and f); this behavior was not detected through the
7 experimental temperature measurement inside the bed (Figure 2 a, c and e). The first stage
8 occurs approximately between room temperature and up to about 150°C, whose average
9 slope is 30 °C/min. This sub-process corresponds to the biomass drying stage. While the
10 second stage that occurs between 150 and 200°C and up to the maximum temperature
11 (with a slope of around 440 °C/min). This energy released by the exothermic reactions
12 increases the process temperature. This phase has been identified in this kind of reactor
13 under gasification conditions by other authors, and it is named the pyrocombustion stage
14 [4,29], and refers to a situation where the volatile species is rapidly oxidized by the
15 oxygen available around the particles. Therefore, according to the good agreement
16 between the temperature fields, it is demonstrated that the CFD model can simulate the
17 gasification process in fixed bed reactors with good accuracy.
18
19 4.1.2. Producer gas composition
20 Figure 3 shows the experimental gas composition versus the one numerically estimated.
21 For the PFs 0.48 and 0.59, the dimensional proximity of the model with the CO and H2
22 composition is evident, while for chips (PF=0.38) these two species are overestimated by
23 the model. CH4 is underestimated for all the biomass geometries, while the CO2 is
24 overestimated by the model during the three simulation cases. The CO2 overestimation in
25 the model is related to the carbon mass balance and generation of some preferential air
26 paths, giving rise to small combustion zones. The CH4 underestimation is due to the
27 kinetic reaction selected for it, where some fraction of CH4 can react with steam to
28 produce H2 through the reactor length (see Eq. (21)). The syngas composition has been
29 presented by other authors with good dimensional proximity [13,14]. However, the
30 produced gas composition depends on the kinetic mechanisms of the reaction and the

13
 ACCEPTED MANUSCRIPT

1 kinetic rates used to simulate the thermochemical process. This finding explains why the
2 produced gas composition is difficult to be estimated by modelling [4,31]. In this work,
3 the good accuracy of the model regarding the producer gas composition is highlighted
4 because the average root mean square error varies between ±0.76%vol to ±8.33%vol for
5 the gaseous species (see Table 9).
6
CO-PF=0.38 CO2-PF=0.38 CH4-PF=0.38 H2-PF=0.38
CO-PF=0.48 CO2-PF=0.48 CH4-PF=0.48 H2-PF=0.48
CO-PF=0.59 CO2-PF=0.59 CH4-PF=0.59 H2-PF=0.59

25
Syngas composition, mod [%vol]

20

15

2
10
1

0
5 0 1 2

0
0 5 10 15 20 25
Syngas composition, exp [%vol]
7
8 Figure 3. Syngas composition- linear adjustment.
9
10 4.1.3. Process parameters
11 The model validation as a function of experimental (exp) and theoretical (mod) flame
12 front velocity (Vff) is presented in Figure 4 a. The model tends to follow the experimental
13 results, where the flame front velocity decreases if PF increases from 0.38 to 0.59. For
14 PF=0.38 the model tends to overestimate the propagation velocity, while for the two
15 others PFs (0.48 and 0.59) the model slightly underestimates this parameter in regard to
16 the experimental results. Nevertheless, the average relative error of the model regarding
17 this parameter is 3.39%, which indicates the proper accuracy of the model for simulating
18 the flame front propagation over the raw biomass under fix bed gasification conditions.
19 The flame front velocity is higher for chips (PF=0.38) because the surface-volume
20 relation of this particle is 21% higher than for cubes and cylinders (see Table 1).
21 Moreover, the low PF of chips (high bed void fraction) favors radiation penetration, which
22 leads to increased amounts of biomass involved in the process [20]. The biomass
23 consumption rate presents a similar trend to the one obtained from flame front velocity
24 (see Figure 4 a). This behavior is due to the fact that the biomass consumption rate is a
25 function of this parameter [18].

14
 ACCEPTED MANUSCRIPT

1 The syngas volumetric flow (Vgas) results are also contrasted in Figure 4 a. The model
2 provides a good prediction of the gas production because it shows a trend where the gas
3 flow increases if PF increases too. The response variables of the model are within the
4 standard deviation of the experiments and indicate that the adapted model appropriately
5 predicts these results.
6
100 60 ER,exp ER, mod 1200
mbms,exp mbms,mod LHVgas,exp LHVgas, mod
Vgas,exp Vgas,mod CGE,exp CGE, mod
90 Tmax,exp Tmax, mod
Vff,exp Vff,mod
50 1000
80
0,5xmbms [kg/h/m2] - 500xVgas [m3/s/m2]

100xER [--] - 10xLHVgas [MJ/Nm3]

70
40 800

60

0,5xCGE [%]
Vff [mm/min]

Tmax [ºC]
50 30 600

40
20 400
30

20
10 200

10

0 0 0
0.35 0.40 0.45 0.50 0.55 0.60 0.35 0.40 0.45 0.50 0.55 0.60
PF [--] PF [--]

a) Biomass consumption rate, gas b) Maximum temperature, equivalence ratio,

generation rate and flame front velocity syngas calorific value and cold efficiency of the
process
7 Figure 4. Process parameters as a function of PF.
8
9 Regarding the heating value of the syngas, the model slightly overestimates this
10 parameter (LHVgas) for PF=0.38, due to the theoretical overestimation of CO and H2 (see
11 Figure 4 b). For PFs of 0.48 and 0.59, following the trends of gaseous fuels (Figure 3),
12 the LHVgas is slightly underestimated by the model. Although the CO and H2 are well
13 predicted by the model, it underestimates the amount of CH4.
14 The Equivalence Ratio (ER) is a parameter that depends on the packing factor. Figure 4
15 b shows that ER increases if PF rises, this is due to the lower flame front velocity caused
16 by the high packing of the bed which reduces the biomass consumption rate [20]. For
17 cylinders (PF=0.48), it is highlighted that the model predicts the highest ER, which
18 suggest that the mathematical approach predicts coherently the effect of the particle and
19 bulk densities on the flame front velocity and biomass consumption rate.

15
 ACCEPTED MANUSCRIPT

1 Similarly, the maximum process temperature increases if PF increases. This is due to the
2 fact the absorption of radiation by the solid phase is more intense when the bed void
3 fraction is lower; therefore, as the penetration of radiative heat transfer decreases and
4 leads to higher energy concentration in the reaction front that favors the increase of the
5 temperature. This trend corresponds coherently with the behavior of ER, where, biomass
6 consumption decreases from PF=0.38 to PF=0.59; therefore, the equivalence air-to-
7 biomass ratio increases, thus, the process temperature increases [13,16].
8 The cold gas efficiency reaches higher values when PF increases. This can be seen in
9 Figure 4 b. A good relationship between the experimental and theoretical parameters is
10 highlighted. The CGE increases due to the lower biomass consumption (low flame front
11 velocity) that leads to decreased biomass energy fed to the thermochemical process for
12 the low radiation penetration in the solid phase.
13 Regarding the model accuracy, it is possible to state that the model appropriately
14 simulates the physicochemical processes of biomass under gasification conditions in a
15 fixed bed. This is stated according to low range variation of the model with regard to the
16 experimental data calculated through the RMSE. This calculated error is shown in Table
17 9. The model can predict the trend of the response variables as a function of PF with a
18 low relative error for the global parameters of the process, the average relative error of
19 the model is 5.45%.
20
21 Table 9. Root mean square error of the CFD model vs Experimental results
Root mean square error [units]
Validation parameter
PF=0.38 PF=0.48 PF=0.59 Average
Syngas composition
CO [%vol] 2.64 0.27 0.79 1.23
CO2 [%vol] 7.93 8.95 8.10 8.33
CH4 [%vol] 1.16 0.44 0.67 0.76
H2 [%vol] 5.65 0.93 0.40 2.33
Vff [mm/min] 0.29 0.44 0.61 0.45
2
mbms [kg/h/m ] 2.30 5.59 7.24 5.04
3 2
Vgas [m /s/m ] 2.51 x 10 -3 2.22 x 10-3 1.75 x 10-3 2.16 x 10-3
LHVgas [MJ/Nm3] 0.53 0.29 0.10 0.31
ER [--] 4.70 x 10 -3 1.51 x 10-2 1.91 x 10-2 1.30 x 10-2
CGE [%] 4.50 1.58 1.26 2.45
22
23 4.2. Sensitivity analysis
24 Once the model has been validated, it can be used as a computational tool to perform a
25 sensitivity analysis, which may explain and describe some of the phenomena that cannot
26 be seen or measured during the experimental phase. Hence, in this section different
27 phenomena that were affected by the biomass packing factor in a fixed bed are analyzed
28 under gasification conditions. This study complements the previous work developed in

16
 ACCEPTED MANUSCRIPT

1 an experimental setup, where empirical results were analyzed based on conceptual criteria
2 of the process [1], while in this work, those parameters are quantified.
3 The effect of the three packing conditions has been analyzed for a fixed time of 500
4 seconds. To study the flame front through the reactor length, x=0.5 m represents the top
5 of the reactor, and x=0 m represents the gasifier bottom (grate). Herein, the solid-gas
6 convective heat transfer mechanism, absorption of radiative heat transfer in solid phase,
7 and the distribution of solid temperature as a function of PF is discussed. These
8 parameters define the flame front propagation under gasification conditions and quantify
9 differences in the process for several characteristics [32].
10
11 4.2.1. Convective solid-gas heat transfer
12 The convective solid-gas heat transfer between solid and gas as a function of PF is
13 presented in Figure 5. In all three cases of PF, solid transfers the highest energy to the gas
14 in the drying zones, while in the pyro-combustion stage the heat is transferred from gas
15 to solid phase [32]. Although for PF=0.38 (chips) there is a small stage in the oxidation
16 zone (near to x=0.36 m) where solid transfers to gas. This phenomenon for wood chips is
17 attributed to the formation of preferential paths which favor the highest temperatures in
18 some zones of the flame front, and can produce a higher solid temperature that leads to
19 transfer of heat from solid phase to gaseous phase at a specific point during the pyro-
20 combustion stage [8].
21 The solid-gas heat transfer in the drying zone for the case of chips (PF=0.38) is lower
22 than the other two geometries (cubes and cylinders). This is because chips have a higher
23 bed void fraction, as can be seen in the radiative heat transfer (see section 4.2.2), which
24 favors the penetration of radiation in the solid phase favoring high reaction rates due to
25 high thickness of the flame front. This high penetration of radiative heat in the solid phase
26 leads to low temperatures in the reaction front, and hence the heat transfer in the drying
27 phase decreases. The convergence of a high radiative penetration in solid phase with high
28 particle surface area/volume ratio and with low PF favors the activation of drying and
29 pyrolysis reactions. This effect causes the flame front to advance faster for chips than for
30 the other two PFs. The highest flame front velocity for chips is shown in Figure 5, where
31 the flame front is ahead in the case PF=0.38 (chips), located in x=0.33 m, while for
32 cylinders and cubes, the reaction front is placed behind, between 0.4 and 0.42 m. In the
33 pyro-combustion zone, the solid-gas heat transfer increases around 20% if PF increases
34 from PF=0.38 to PF=0.59; while for PF=0.48, the lowest heating transfer rates are
35 obtained during pyro-combustion due to the high bulk density, particle density and
36 biomass hardness of cylinder shape, as can be seen in Figure 5.
37

17
 ACCEPTED MANUSCRIPT

10

-10
Qsg [kW/m2]
-20

-30
Qsg-PF 0.38
Qsg-PF 0.48

-40 Qsg-PF 0.59

-50
0.25 0.30 0.35 0.40 0.45 0.50
x [m]- 500 s
1
2 Figure 5. Solid-gas heating transfer as a function of PF.
3
4 4.2.2. Radiative heat transfer
5 The absorption of radiative heat quantifies the amount of radiative energy absorbed by
6 the solid phase; its calculation model is represented by Eq. (11). Figure 6 shows how the
7 radiative heat absorption increases if PF increases. This is because the low bed void
8 fraction avoids the penetration of radiation in the bed causing that energy concentration
9 increases in the flame front, which leads to an elevated temperature of the bed (see the
10 2D distribution of solid temperature in Figure 7). The absorbed radiation in the solid phase
11 increases with the PF. The relative increase with regard to the chips (PF=0.38), is 18.77%
12 for the cylindrical geometry (PF=0.48), and 26.71% for cubes (PF=0.59). The highest
13 power absorbed by cubes (PF=0.59) is due to the lower bed void fraction which favors a
14 low radiation penetration [12], leading to low reaction front velocity and biomass
15 consumption rate (as discussed in section 4.1.2.). The effects on fixed bed gasification
16 when PF increases can be summarized as follows: the absorption of radiative heat transfer
17 increases, leading to low reaction rates of drying and pyrolysis; however, the temperature
18 in the reaction front increases favoring the production of gaseous fuels in the producer
19 gas (H2, CO, and CH4); therefore, the CGE increases. Similar trends based on the
20 empirical observations were reported by Lenis et al. [1].
21

18
 ACCEPTED MANUSCRIPT

1
2 Figure 6. Radiation absorption heat (W/m2) – t = 500 s
3
4 Figure 7 presents the 2D distribution of solid temperature; this parameter is a useful datum
5 to design gasifiers [33]. For the PF=0.38 case, the flame front advances faster towards the
6 raw biomass, hence, the maximum temperature in the reaction front decreases, because
7 stages of drying and devolatilization are favored. In function of the 2D numerical results,
8 it can be seen a high concavity of the solid temperature for PF=0.38, explaining the high
9 propagation velocity of the process due to the higher bed void fraction of biomass in the
10 bed.
11

12
13 Figure 7. Solid temperature (K) – t = 500 s
14
15 5. Conclusions
16 The adaptation of the model to gasification conditions required modification of the
17 chemical kinetics of pyrolysis, oxidation, and reduction processes using the adequate

19
 ACCEPTED MANUSCRIPT

1 kinetic mechanism of the reactions for gasification. The numerical simulation of biomass
2 gasification has a dimensional prediction because the average relative error of all the
3 thermodynamic variables that were validated (flame front velocity, biomass consumption
4 rate, volumetric gas flow, gas heating value, equivalence ratio, and cold efficiency) are
5 below 6%. Thereby, the model can be used as computational tool for predicting or
6 diagnosing different systems under gasification conditions with air as the gasifying agent.
7 It is highlighted that the biomass gasification 2D CFD model generates results within the
8 experimental standard deviation of each validated variable, considering that the model
9 predicts the trends of those parameters.
10 The effect of biomass PF increase (from 0.38 to 0.59) on fixed bed gasification can be
11 summarized as follows: the absorption of radiative heat increases 26.71%, leading to low
12 reaction rates of drying and pyrolysis. Thus, the flame front velocity decreases from 21.06
13 mm/min to 11.76 mm/min; whereas, the maximum temperature in the reaction front
14 increases (from 927 °C to 1027 °C) favoring the production of gaseous fuels in the
15 producer gas (H2, CO, and CH4), and increasing the CGE from 31 % to 38 %.
16 The CFD model, which was adapted to gasification conditions, allows study of the
17 phenomenology involved during the thermochemical transformation of biomass to a
18 combustible gas. This fact allows making dimensional analyses of operational
19 parameters, gases concentrations and heat transfer mechanisms studies. Due to these
20 facts, this model can be used to simulate several types of fixed bed reactors with different
21 fuels and geometries.
22
23 Acknowledgments
24 The authors acknowledge the financial support of Universidad de Antioquia through the
25 project “Sostenibilidad 2017-2018.” In addition, this work was financially supported by
26 project ENE2015-67439-R of the Ministry of Economy and Competitiveness (Spain) and
27 the work of Sergio Chapela López has been supported by the grant BES-2016-076785 of
28 the Ministry of Economy, Industry and Competiveness (Spain).
29
30 Nomenclature
Ai Pre-exponential factor
Av Area-volume ratio (m )
‒1

Ci Molar concentration (kmol·m )

‒3

Cp ‒1 ‒1
Specific heat (J·kg K )
Di 2 ‒1
Diffusivity of the specie i (m s )
deq Equivalent diameter (m)
Ei Activation energy
‒2 ‒1
h Convection coefficient (W·m K )
Hi ‒1
Enthalpy of formation of the specie i (J·kg )
‒2
I Irradiation intensity (W·m sr)
‒1 ‒1
k Thermal conductivity (W·m K )

20
 ACCEPTED MANUSCRIPT

ki Kinetic constant of the reaction i

m ‒1
ki Mass transfer coefficient (m·s )
ks/g Constant of solid/gas distribution of the char reaction energy
‒1
LH Latent heat (J·kg )
Mi Molecular weight (kg·kmol )
‒1

n Refractive index ( ‒ )
Nu Nusselt number ( ‒ )
Pr Prandtl number ( ‒ )
r Position vector (m)
Re Reynolds number ( ‒ )
Ru Ideal gas constant
s Direction vector (m)
‒3
S Source term (W·m )
Sc Schmidt number ( ‒ )
Sh Sherwood number ( ‒ )
Si,j ‒3 ‒1
Source of the specie i from the component j (kg·m s )
t Time (s)
T Temperature (K)

Greek symbols
‒1
α Absorption coefficient (m )
ε Solid fraction ( ‒ )
3
ρ Density (kg·m )
‒2 ‒4
σ Stefan-Boltzmann coefficient (W·m K )
φ Char oxidation parameter ( ‒ )
''' ‒3 ‒1
ωi Generation or consumption rates (kg·m s )
Ω Solid angle (sr)

Subscripts
C Consumption
char Char
dev Devolatilisation
eff Effective
evap Evaporation
g Gas
G Generation
glob Global
moist Moisture
p Particle
s Solid
wood Wood

Superscripts
conv Convection

21
 ACCEPTED MANUSCRIPT

loss Losses
reac Reaction
rad Radiation
1

2 References
3 [1] Y.A. Lenis, J.F. Pérez, A. Melgar, Fixed bed gasification of Jacaranda Copaia
4 wood: Effect of packing factor and oxygen enriched air, Industrial Crops and
5 Products. 84 (2016) 166–175.
6 [2] M.A. Gómez, J. Porteiro, D. Patiño, J.L. Míguez, CFD modelling of thermal
7 conversion and packed bed compaction in biomass combustion, Fuel. 117 (2014)
8 716–732.
9 [3] J. Collazo, J. Porteiro, D. Patiño, E. Granada, Numerical modeling of the
10 combustion of densified wood under fixed-bed conditions, Fuel. 93 (2012) 149–
11 159.
12 [4] J.F. Pérez, A. Melgar, F. V Tinaut, Modeling of fixed bed downdraft biomass
13 gasification : Application on lab-scale and industrial reactors, International Journal
14 of Energy Research. (2014) 319–338..
15 [5] D. Baruah, D.C. Baruah, Modeling of biomass gasi fi cation : A review, Renewable
16 and Sustainable Energy Reviews. 39 (2014) 806–815.
17 [6] H. Khodaei, Y.M. Al-Abdeli, F. Guzzomi, G.H. Yeoh, An overview of processes
18 and considerations in the modelling of fixed-bed biomass combustion, Energy. 88
19 (2015) 946–972.
20 [7] R. Mehrabian, S. Zahirovic, R. Scharler, I. Obernberger, S. Kleditzsch, S. Wirtz,
21 V. Scherer, H. Lu, L.L. Baxter, A CFD model for thermal conversion of thermally
22 thick biomass particles, Fuel Processing Technology. 95 (2012) 96–108.
23 [8] S. Hermansson, H. Thunman, CFD modelling of bed shrinkage and channelling in
24 fixed-bed combustion, Combustion and Flame. 158 (2011) 988–999.
25 [9] A.K. Olaleye, K.J. Adedayo, C. Wu, M.A. Nahil, M. Wang, P.T. Williams,
26 Experimental study , dynamic modelling , validation and analysis of hydrogen
27 production from biomass pyrolysis / gasification of biomass in a two-stage fixed
28 bed reaction system, Fuel. (2014).
29 [10] M.A. Masmoudi, M. Sahraoui, N. Grioui, K. Halouani, 2-D Modeling of thermo-
30 kinetics coupled with heat and mass transfer in the reduction zone of a fixed bed
31 downdraft biomass gasifier, Renewable Energy. 66 (2014) 288–298.
32 [11] Y. Bin Yang, R. Newman, V. Sharifi, J. Swithenbank, J. Ariss, Mathematical
33 modelling of straw combustion in a 38 MWe power plant furnace and effect of
34 operating conditions, Fuel. 86 (2007) 129–142.
35 [12] D. Shin, S. Choi, The Combustion of Simulated Waste Particles in a Fixed Bed,
36 Combustion and Flame. 180 (2000) 167–180.
37 [13] S. Mahapatra, S. Kumar, S. Dasappa, Gasification of wood particles in a co-current
38 packed bed: Experiments and model analysis, Fuel Processing Technology. 145
39 (2016) 76–89.
40 [14] I. Janajreh, M. Al Shrah, Numerical and experimental investigation of downdraft
41 gasification of wood chips, Energy Conversion and Management. 65 (2013) 783–
42 792.
43 [15] J. Cai, Y. He, X. Yu, S.W. Banks, Y. Yang, X. Zhang, Y. Yu, R. Liu, A. V.
44 Bridgwater, Review of physicochemical properties and analytical characterization

22
 ACCEPTED MANUSCRIPT

1 of lignocellulosic biomass, Renewable and Sustainable Energy Reviews. 76 (2017)

2 309–322.
3 [16] S. Mahapatra, S. Dasappa, Experiments and analysis of propagation front under
4 gasification regimes in a packed bed, Fuel Processing Technology. 121 (2014) 83–
5 90.
6 [17] Y.A. Lenis, A.F. Agudelo, J.F. Pérez, Analysis of statistical repeatability of a fixed
7 bed downdraft biomass gasification facility, Applied Thermal Engineering. 51
8 (2013) 1006–1016.
9 [18] J.F. Pérez, A. Melgar, P. Nel, Effect of operating and design parameters on the
10 gasification / combustion process of waste biomass in fixed bed downdraft
11 reactors : An experimental study, Fuel. 96 (2012) 487–496.
12 [19] M. Vaezi, M. Passandideh-fard, M. Moghiman, M. Charmchi, On a methodology
13 for selecting biomass materials for gasification purposes, Fuel Processing
14 Technology. 98 (2012) 74–81.
15 [20] J.F. Pérez, P.N. Benjumea, A. Melgar, Sensitivity analysis of a biomass
16 gasification model in fixed bed downdraft reactors : Effect of model and process
17 parameters on reaction front, Biomass and Bioenergy. 83 (2015).
18 [21] Y.A. Lenis, L.F. Osorio, Effects Fixed Bed Gasification of Wood Species with
19 Potential as Energy Crops in Colombia : The Effect of the Physicochemical
20 Properties, Energy Sources. (2013) 37–41.
21 [22] Y.A. Lenis, J.F. Pérez, Gasification of Sawdust and Wood Chips in a Fixed Bed
22 under Autothermal and Stable Conditions, Energy Sources. (2015) 37–41.
23 [23] Y. Bin, C. Ryu, A. Khor, N.E. Yates, V.N. Sharifi, J. Swithenbank, Effect of fuel
24 properties on biomass combustion . Part II . Modelling approach — identification
25 of the controlling factors, Fuel. 84 (2005) 2116–2130.
26 [24] M.A. Gómez, J. Porteiro, D. Patiño, J.L. Míguez, Fast-solving thermally thick
27 model of biomass particles embedded in a CFD code for the simulation of fixed-
28 bed burners, Energy Conversion and Management. 105 (2015) 30–44.
29 [25] M.A. Gómez, D. Patiño, R. Comesaña, J. Porteiro, M.A. Álvarez Feijoo, J.L.
30 Míguez, CFD simulation of a solar radiation absorber, International Journal of
31 Heat and Mass Transfer. 57 (2013) 231–240.
32 [26] K.M. Bryden, K.W. Ragland, C.J. Rutland, Modeling thermally thick pyrolysis of
33 wood, Biomass and Bioenergy. 22 (2002) 41–53.
34 [27] H. Thunman, B. Leckner, F. Niklasson, F. Johnsson, Combustion of Wood
35 Particles — A Particle Model for Eulerian Calculations, 2180 (2002) 30–46.
36 [28] K.M. Bryden, K.W. Ragland, Numerical Modeling of a Deep , Fixed Bed
37 Combustor, Energy & Fuels. 10 (1996) 269–275.
38 [29] C. Di Blasi, Dynamic behaviour of stratified downdraft gasifiers, Chemical
39 Engineering Science. 55 (2000) 2931–2944.
40 [30] H. Liu, B.M. Gibbs, Modeling NH 3 and HCN emissions from biomass circulating
41 fluidized bed gasifiers q, Fuel. 82 (2003) 1591–1604.
42 [31] R. Johansson, H. Thunman, B. Leckner, Sensitivity Analysis of a Fixed Bed
43 Combustion Model, Energy & Fuels. (2007) 1493–1503.
44 [32] F. V Tinaut, A. Melgar, J.F. Pérez, A. Horrillo, Effect of biomass particle size and
45 air superficial velocity on the gasification process in a downdraft fixed bed gasifier
46 . An experimental and modelling study, Fuel Processing Technology. 89 (2008)
47 1076–1089.
48 [33] K. Jaojaruek, Mathematical model to predict temperature profile and air – fuel

23
 ACCEPTED MANUSCRIPT

1 equivalence ratio of a downdraft gasification process, Energy Conversion and

2 Management. 83 (2014) 223–231.

24
 ACCEPTED MANUSCRIPT

− The effect of wood packing factor (PF) on gasification was studied by CFD analysis.
− If PF increases, the absorption of radiative heat transfer increases.
− If PF increases, the reaction rates of drying and pyrolysis decreases.
− If PF increases, the temperature in the reaction front increases.
− If PF increases, the production of H2, CO, and CH4, and efficiency increase.