Article

pubs.acs.org/EF

Analysis of Syngas Quality from Portuguese Biomasses: An

Experimental and Numerical Study
Valter Silva, Eliseu Monteiro, Nuno Couto, Paulo Brito, and Abel Rouboa*,,

INEGI-FEUP/ECT, Universidade de Tras-os-Montes e Alto Douro, Vila Real, Portugal

IPP-ESTG, Portalegre, Portugal

C3iCentro Interdisciplinar de Investigacao e Inovacao , Instituto Politecnico Portalegre, Portalegre, Portugal

MEAM Department, University of Pennsylvania, Philadelphia 19020, Pennsylvania, United States

ABSTRACT: A comprehensive two-dimensional multiphase model was developed to describe the gasication of three large
available Portuguese biomasses in a pilot scale uidized bed gasier within the computational uid dynamics Fluent framework.
An EulerianEulerian approach was implemented to handle both the gas and the dispersed phases. The kinetic theory of
granular ows was used to evaluate the constitutive properties of the dispersed phase, and the gas-phase behavior was simulated
employing the k turbulent model. Devolatilization phenomenon was also modeled. Results from the numerical model were
later compared with experimental data also gathered for the three Portuguese biomasses. The simulated syngas compositions are
in good agreement with the experimental results with maximum deviations of 20% (considering all simulations and biomasses).
The eect of the gasication temperature and steam-to-biomass ratio on the syngas composition was evaluated as well as the cold
gas eciency, H2/CO ratio, CH4/H2 ratio, and carbon conversion. From this analysis, it was possible to dene which biomass has
the greater potential to be used concerning the selected application. Among the three biomasses, it was concluded that coee
husks and vine-pruning residues show better H2/CO ratios, while the higher CH4/H2 ratios are obtained by using the forest
residues. The highest cold gas eciencies were obtained by using the vine-pruning residues. These data are crucial to describe
scenarios concerning the potential use of biomass as a source of energy in Portugal.

1. INTRODUCTION Table 1. Potential Energy from Pruning and Forest Residues

in Portugal3
Biomass utilization for fuel, products, and power is recognized
worldwide as a critical component for a nations strategic plan alcohol equivalent production
when it faces continued and growing dependence on imported biomass (millions of dm3) (106 ton/year) TWh/year
oil.1 Portugal has a good potential for generating energy from pruning 357.7 2.2
residues
biomass to partially replace coal or petroleum. The biomass
forest 6.5 11.6
resources that can be used for energy generation are diverse, residues
but in Portugal, an important contribution can be found from
agricultural-related residues. Coee husks, forest and vines forest residues produced in Portugal and the corresponding
pruning residues are largely available and have attractive low energy potential.
costs. Coee husks show an economical advantage, since they The total primary energy demand in Portugal amounted to
should not be pelletized. Additionally, the use of these residues 104 000 GWh in 2005.3 From Table 1, it can be noticed that
avoids their dumping and the corresponding environmental only forest and pruning residues can potentially produce 13 800
consequences. Typical values for the coee industry indicate GWh (about 13.3% of the total primary energy demand in
that about 20% of coee is converted into husks, solid waste, Portugal). Additionally, the power production from bioresour-
and parchment. It was veried that coee husks generated in ces was 9750 GWh in 2005 (including biomass, municipal solid
Portuguese industry have a caloric value of 15.100 MJ/kg. The waste, and biogas). Previous data show that both forest and
coee husk heating value and composition in terms of volatile pruning residues can play an important role in the Portuguese
matter and ultimate analysis shows that they have acceptable energy scenario.
energy properties considering typical biomasses.2 Portugal is These residues require adequate treatment or recovery to
one of the main wine producers in the world generating about minimize environmental impact and increase their correspond-
700 million dm3 of wine per year. There is a great interest by ing economic value. A large variety of technologies has been
Portuguese entities to study the best ways to valorize the developed in recent decades to deal with this problem. Among
residues and subproducts generated by this industry. The the proposed technologies, those oriented toward energy
potential energy derived from pruning residues is depicted in recovery, including combustion and gasication of biomasses,
Table 1. About 35% of Portuguese territory is composed of
forest. It is estimated that economic activities related to the Received: February 5, 2014
Portuguese forestry generates about 113 thousand jobs and Revised: July 28, 2014
about 200 million euros/year. Table 1 shows the amounts of Published: July 28, 2014

2014 American Chemical Society 5766 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

have attracted much interest.4 There are several methods to EulerEuler-based CFD model to describe the trend of
convert the biomass feedstock to energy, but it is expected that biomass decomposition versus temperature with a kinetic
biomass gasication will become the dominant technology.5 model based on superimposed hemicellulose, cellulose, and
Biomass gasication is very attractive because the obtained lignin reactions. It was concluded that biomass particle size and
gases are intermediates in high-eciency power production.5 supercial gas velocity inuenced tar yield and residence time
Additionally, the environmental performance is one of the considerably with a xed bed height. Gungor and Yildirim14
greatest strengths of gasication technology, which often is computed the axial and radial distribution of syngas mole
considered a sound response to the increasingly restrictive fraction and temperature across an atmospheric CFB biomass
regulations applied around the world. On the other hand, the gasier by using a two-dimensional (2-D) model that uses the
use of gasiers instead of combustors shows clear advantages, particle-based approach and integrates and simultaneously
namely, the generation of a syngas with improved quality to be predicts the hydrodynamic and gasication aspects. The
used in the production of Fischer-tropsch liquids, fuel cells, authors claimed that the developed model computes
clean fuel combustion, and cheaper CO2 underground successfully the radial and axial proles of the bed temperature
sequestration. Biomass gasication also brings high expectations and H2, CO, CO2, and CH4 volumetric fractions and tar
toward the generation of hydrogen. concentration versus gasier temperature with errors lower
Hydrogen is one of the key energy sources due to its than 25%. These works compare the numerical results versus
potential to provide energy for industry, residences, trans- experimental results gathered in laboratory setups. Indeed,
portation, and mobile applications. Indeed, hydrogen can there is a lack of experimental data where the numerical results
replace petroleum-based fuels in the transportation sector and can be conrmed with pilot scale or industrial setups, namely,
can even supply remote stations that cannot be fueled by using uidized bed technology. The scaleup is still one of the
conventional distribution grids.6 Additionally, hydrogen major problems to be handled considering the biomass and coal
production can be environmentally friendly when green gasication.15 This is related to the complex nonlinear
technologies such as biomass gasication are employed. The hydrodynamic behavior, chaotic nature, and complicated
worldwide commercial production of hydrogen is approx- reactions that strongly impact the gassolid ow behavior.16
imately 40 million tonnes.7 However, most of the traditional Additionally, the lack of data is also related to the expense of
technologies for hydrogen production are fossil-fuel based, 80 experiments for obtaining valuable information on uidized bed
85% of the worlds total production being derived from natural gasiers in pilot scale or industrial dimensions. This implies an
gas via steammethane reforming,8 which is a very energy- urgent need for both theoretical as well as experimental
intensive process that releases large amounts of carbon dioxide approaches to obtain more helpful data and information.
and other greenhouse gases. The use of uidized bed technology in biomass gasication
Hydrogen can be obtained by other processes than steam units makes it possible to use relatively smaller gasiers and
reformation of methane, namely, water electrolysis, biological larger capacities than with a xed bed. The higher eciency of a
processes using microorganisms, or by thermocatalytic uidized bed comes from its higher specic reaction area when
reformation of hydrogen-enriched organic compounds.9 compared to a xed bed. Fluidized beds have also proven to
Among the aforementioned technology options, biomass permit more versatile units in terms of feed materials of
gasication is identied as the most ecient and economic dierent characteristics and origins.17
route for hydrogen production.10 This paper aims to present a numerical methodology within
A gasication system includes a set of phenomena such as the framework of the commercial CFD code (Fluent) applied
uid ow, heat transfer, and complex chemistry that can only be to the gasication of Portuguese agro-industrial residues. The
solved by applying several governing mathematical equations, use of coee husks, forest and vine-pruning residues is studied
mostly based on conservation equations, that is, mass, heat, and concerning the generation of hydrogen and the quality of
momentum. This level of complexity can be achieved using syngas based on dierent indices including carbon conversion,
computer uid dynamic (CFD) tools. However, only a few H2/CO ratio, CH4/H2 ratio, and cold gas eciency (CGE).
studies regarding the use of CFD tools have been reported to The numerical simulation results were compared and validated
simulate the uidized bed biomass gasication. Two main versus a set of runs using these residues that were gasied in a
approaches are considered to describe the solid phases: the bubbling uidized pilot scale unit.
EulerianEulerian approach (a continuum approach for both
phases) and the EulerianLagrangian approach (a continuum 2. MATERIALS AND EXPERIMENTAL SETUP
approach for the uid phase and a discrete particle method for
The gasication runs were performed using the three biomass
the solid phase). Xie et al.11 developed a three-dimensional feedstocks. The proximate and ultimate analyses of these feedstocks
numerical model to compute the syngas composition by using are shown in Table 2.
forestry residues as feedstock and using the Eulerian The experiments were performed in a pilot-scale gasication plant,
Lagrangian approach as numerical method. The experimental which is based on an up-ow uidized bed gasier operating at up to
gasication experiments were carried out in a lab-scale pine 850 C, under a total pressure below 1 bar, and at a maximum pellet
gasier. The EulerianLagrangian approach implies a huge feeding rate of 70 kg/h. Figure 1 displays a diagram of the biomass
computational eort, and more versatile computing solutions gasication plant used in the experiments.
can be found using an EulerianEulerian approach. Xue and The main components of the unit are the following: (a) a biomass
Fox12 developed an EulerianEulerian computational uid feed system with two inline storage tanks that allow discharging the
biomass into the reactor using an Archimedes screw at a variable and
dynamics model to simulate wood gasication using air as a controllable speedsthe two storage tanks act as buers to avoid air
uidization agent in a lab-scale uidized bed gasier (FBG). entering through the feeding system; (b) a tubular uidized bed
Simulations provided detailed information on several processes reactor, 0.5 m in diameter and 2.5 m in height, internally coated with
such as the dynamic particle processes, char elutriation, and gas ceramic refractory materials; biomass enters the reactor at a height of
composition at the reactor outlet. The same authors13 used an 0.4 m from the base, and preheated air enters the reactor from the base

5767 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777

Energy & Fuels Article

Table 2. Humidity, Density, Net Heat Value, and Ultimate respectively, with CP81069, CP81071, CP81072, CP81073, and
and Proximate Analyses of Coee Husks, Forest and Vine- CP81025 Varian GC columns), using helium as carrier gas.
Pruning Residues
3. MATHEMATICAL MODEL
forest coee vine-pruning
biomass properties residues husks residues A numerical model was built according to experimental data
elementary analysis (%)
gathered from the Portalegre pilot-scale setup. To simulate the
N 2.4 5.2 2.6
gasication process a multiphase (gas and solid phase) model
C 43.0 40.1 41.3
from a Fluent database was used. In the comprehensive 2-D
H 5.0 5.6 5.5
numerical model, the gas phase is treated as continuous, and
O 49.6 49.1 50.6
the solid phase is described through an Eulerian granular
humidity (%) 11.3 25.3 13.3
model. Interactions between both phases were modeled as well,
density (kg/m3) 650 500 265
since both phases exchange heat by convection, momentum
net heat value 21.2 20.9 15.1
(given the drag between gas phase and solid phase), and mass
(MJ/kg biomass) (given the heterogeneous chemical reactions).
proximal analysis (%)18 3.1. Mass Balance Model. The continuity equation for the
ash 0.2 2.5 3.1 gas and solid phases are given by
volatile matter 79.8 83.2 83.6

xed carbon 20.0 14.3 13.3 (gg ) + (gg vg) = Sgs
t (1)

through a set of diusers, with a ow of about 70 m3/h (the schematic

(ss ) + (ss vs) = Ssg
of the uidized reactor are depicted in Figure 2); (c) a gas-cooling t (2)
system consisting of two heat exchangers; the rst exchanger cools the
syngas to about 570 K using a cocurrent air ow that enters the unit, The biomass feed reacts with the oxygen steam and carbon
and the second heat exchanger cools the syngas to 420 K by a forced dioxide to change solid phase into gas phase, so S, which is
ow of air coming from the exterior; (d) a cellulosic bag lter that dened as the mass source term due to heterogeneous
allows the removal of carbon black and ash particles produced during reactions, can be expressed as follows:
the gasication process; lter cleaning is done with pressurized syngas
injection; black carbon is collected at bottom of the bag lter and Ssg = Sgs = Mc cR c (3)
stored in a proper tank; (e) a condenser where liquid condensates are
removed by cooling the syngas to room temperature on a third tube The ideal gas behavior was considered for computing the gas-
heat exchanger. Gasication runs were performed using coee husks, phase density:
forest and vine-pruning residues at 790 and 815 C. n
Syngas analysis was performed in a Varian 450-GC gas chromato- 1 RT Yi
graph with two thermal conductivity detectors that allow the detection =
g p i=1
Mi (4)
of H2, CO, CO2, CH4, O2, N2, C2H6, and C2H4 (equipped,

Figure 1. Biomass gasication semi-industrial plant at Polytechnic Institute of Portalegre, Portugal. The main components of the unit are described
in the text.

5768 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777

Energy & Fuels Article

Figure 2. Schematic of the bubbling uidized bed gasier.

The solid-phase density was assumed as constant.

( k ) + (kui)
3.2. Momentum Equations. The momentum equation of t xi
the gas phase is
= + t + G k + G b YM + S k
xj k (7)
(gg vg) + (gg vgvg)
t
= gpg + g g + (vg vs) + gg + SgsUs
(5) () + (ui)
t xi
The momentum equations of solid phase can be written as 2
= + t + C1 (G k + C3G b) C 2 + S
xj xj k k

(ss vs) + (ss vsvs) (8)
t
= sps + s g (vg vs) + ss + SsgUs (6) 3.4. Granular Eulerian Model. In the granular Eulerian
model, stresses in the granular solid phase are obtained by the
3.3. Turbulence Model. The standard k model in analogy between the random particle motion and the thermal
ANSYS FLUENT has become the workhorse of practical motion of molecules within a gas accounting for the inelasticity
engineering ow calculations in the time since it was proposed of solid particles. As in a gas, the intensity of velocity uctuation
by Launder and Spalding.19 It is a semiempirical model, and the determines the stresses, viscosity, and pressure of the granular
derivation of the model equations relies on phenomenological phase. The kinetic energy associated with velocity uctuations
considerations and empiricism. The selection of this turbulence is described by a pseudothermal temperature or granular
model is appropriate when the turbulence transfer between temperature, which is proportional to the norm of particle
phases plays a predominant role as in the case of gasication in velocity uctuations.
uidized beds. The conservation equation for the granular temperature,
The turbulence kinetic energy k and its rate of dissipation obtained from the kinetic theory of gases, takes the following
are obtained from the following transport equations: form:
5769 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

3 ( ) considers both the Arrhenius and the eddy-dissipation reaction

s s s + ( v
s ss s
rates. The water gas shift reaction and the CO, H2, and CH4
2 t combustion reactions were considered, respectively:
= ( PI
s + s ): (vs) + (k a(s)) a + ls CO + H 2O CO2 + H 2 (14)
(9)
CO + 0.5O2 CO2 (15)
The diusion coecient for granular energy was computed
by using the following equation due to Syamlal et al.20 H 2 + 0.5O2 H 2O (16)
12 2 CH4 + 2O2 CO2 + 2H 2O (17)
ks = 15ds /4(41 33)
s s s


1+ (4 3)sgo,ss
5
The Arrhenius rates for each one of these reactions can be
16 expressed as follows, respectively:
+ (41 33)sgo,ss
15 (10)
16 000
where = 1/2(1 + ess); d is the biomass particle diameter; s is rCO combustion = 1.0 1015 exp C C 0.5
T CO O2 (18)
the solid phase.
The granular energy dissipation can be computed by using 3430 1.5
rH2combustion = 5.159 1015 exp T CO2C H21.5
the expression derived by Lun et al.21 When the granular ow T
has a smaller volume fraction than the maximum possible value, (19)
a solid pressure is considered for the pressure gradient term in
the momentum equation, which includes a kinetic term and a 15 700 1
rCH4combustion = 3.552 1014 exp T CO2CCH4
particle collision term: T
(20)
ps =
s s
+ 2s (1 + ess)s 2go,sss (11)
s

1510 CCO2C H2
The radial distribution function allows the dierent levels of rwater gas shift = 2780 exp C C
T CO H2O 0.0265 exp 3968
compressibility. It works as a correction factor that gives the T ( )
probability of collisions when the granular phase goes to denser (21)
states. Ansys Fluent provides empirical relations for the radial
The eddy dissipation reaction rate can be expressed using the
distribution function when there is one solid phase.
following equation:
3.5. Chemical Reactions Model. 3.5.1. Devolatilization.
Devolatilization is a process where moisture and volatile Y p yp

matters are driven out from biomass by heat. No devolatiliza- reddy dissipation = i , rM w, iA minminR R
, B N

k R , rM w, R i i , RM w, i
tion process is included in Ansys/Fluent considering the
EulerianEulerian method. To develop a reliable gasication (22)
model, it is necessary to include a devolatilization model in the The minimum value of these two contributions can be
Fluent code. dened as the net reaction rate.
Biomass is thermally decomposed into volatiles, char, and 3.5.3. Species Transport Equations. The local mass fraction
ash, in agreement with the following equation: of each specie Y is computed by using a convection-diusion
biomass char + volatiles + steam + ash (12) equation as follows:
Note that ash appears for illustration only; the simulation does
(Yi ) + (Yi ) = Ji + R i + Si
not consider ash content in the solid phase. t (23)
In the present work, the volatile matter is composed of the where Ji is the diusion ux of species i due to concentration
following species: gradients, Ri is the net generation rate of species i due to
volatiles 1CO + 2CO2 + 3CH4 + 4 H 2 (13) homogeneous reaction, and Si is a source term related to the
species i production from the solid heterogeneous reaction.
The biomass mixture composition is determined based on the The diusion ux was computed as a function of the turbulent
proximate and elemental analysis shown in Table 2. Schmidt number.
Since there are no data on the exact distribution of the 3.5.4. Heterogeneous Reaction Rate. Char is the solid
volatiles in biomass, the single-rate model developed by devolatilization residue. Heterogeneous reactions of char with
Badzioch and Hawsley22 was adopted. The single-rate model the gas species such as O2 and H2O are complex processes that
produced moderate and reliable devolatilization rates with little involve balancing the rate of mass diusion of the oxidizing
computational eort. Additionally, a demoisturization equation chemical species to the surface of biomass particles with the
was also considered. These two pseudo-heterogeneous surface reaction of these species with the char. The overall rate
reactions were modeled with a single reaction rate in agreement of a char particle is determined by the oxygen diusion to the
with the Arrhenius law. particle surface and the rate of surface reaction, which depend
3.5.2. Homogeneous Gas-Phase Reactions. The modeling on the temperature and composition of the gaseous environ-
of the homogeneous gas-phase reactions should consider both ment and the size, porosity, and temperature of the particle.
the kinetic and the turbulent mixing rate eects.23 The gas The commonly simplied reaction models consider the
phase is conditioned by the eect caused by the chaotic following overall reactions, char combustion (eq 24), H2O
uctuations of the solid particles. This turbulent ow leads to (eq 25) and CO2 char gasication (eq 26):
velocity and pressure uctuations on the gaseous species.
Fluent provides the nite-rate/eddy-dissipation model, which C + O2 CO2 (24)

5770 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777

Energy & Fuels Article

C + H 2O CO + H 2 (25) 6k pqpNuq
hpq =
d p2 (32)
C + CO2 2CO (26)
3.7. Numerical Procedure. The problem under consid-
r
eration is solved by using Fluent, a nite volume method-based
The heterogeneous reactions are inuenced by many factors,
CFD solver. Mesh was built using GAMBIT software, and the
namely, reactant diusion, breaking up of char, interaction of
reactor schematic is depicted in Figure 3. Quadrilateral cells of
reactions, and turbulence ow. To include both diusion and
kinetic eects the kinetic/diusion surface reaction model
(KDSRM)24,25 was applied. This model weights the eect of
the Arrhenius rate and the diusion rate of the oxidant at the
surface particle. The diusion rate coecient can be dened as
[(Tp + T) 2]0.75
D0 = C1
dp (27)

where dp means particle diameter.

The Arrhenius rate can be dened as follows:
( E )
rArrhenius = Ae RTp
(28)

The nal reaction rate weights both contributions and is

dened as follows:
Figure 3. Schematic of the mesh and corresponding bubbling uidized
dmp RTZox D0rArrhenius bed gasier.
= A p
dt M w ,ox D0 + rArrhenius (29)
uniform grid spacing were used. The grid is chosen to be
This model was included in the CFD framework by using the suciently ne to capture the steep gradients in the vicinity of
UDF tool. the inlet and outlet of the gasier. Several simulations were
At biomass and air inlets the mass ow rate was prescribed, performed considering dierent sized quadrilateral cells. The
while at the outlet the atmospheric pressure level was assigned. optimal size was found when no meaningful dierences were
At the walls the materials (steel and insulation) and obtained considering syngas composition, temperature, velocity
corresponding thicknesses and heat conduction coecients and turbulence proles. The selected size is about 12 times
were prescribed, as well as the ambient temperature conditions larger than the particle size, which, as found in literature,26 can
(25 C). capture eectively the hydrodynamics in uidized bed reactors.
All materials (gas species, solid bio mass particles) were The second order upwind scheme is used to discretize the
assigned appropriate properties from standard thermodynamic convective terms in the momentum and energy equations. The
tables. The properties of the gas species (density , viscosity , SIMPLE scheme is used for solving the pressurevelocity
thermal conductivity k, and specic heat capacity Cp) were coupling. The system of algebraic equations is solved using the
allowed to vary with local main phase temperature, and the GaussSeidel point-by-point iterative method in conjunction
mixture value was calculated from its local composition and with the algebraic multigrid method solver. Relative con-
available Fluent laws (ideal gas law for and mass-weighted vergence criterion of 108 for residuals of the continuity and
mixing law for , k, and Cp). momentum equations and of 1014 for residual energy equation
3.6. Energy Conservation. To describe the energy were prescribed in this work. To nd a stable solution, gas
conservation the following energy conservation equation must solid ow is previous solved without using the chemical
be solved for each phase: reactions. After nding the established ow pattern, the
chemical reactions were included, and the full system was
(qq hq )
+ (qq uq hq ) solved.
t
(pq ) 4. DISCUSSION AND RESULTS
= q + q : (uq )
qq + Sq 4.1. Experimental Runs and Model Validation. To the
t best of our knowledge there are limited data regarding syngas
n
obtained from biomass substrates using a pilot scale thermal
+ (Q pq + m pq hpq m pq hpq) gasication plant. The majority of the literature shows syngas
p=1 (30) compositions obtained in a laboratory or small-scale gasier.
The rate of energy transfer is assumed as a function of the Large-scale systems are expensive, and their inherent logistical
temperature dierence between phases: diculties are easily understandable. However, there is a large
dierence between the laboratory and semi-industrial results
Q pq = hpq (Tp Tq) (31) due to phenomena of complex nonlinear hydrodynamic
behavior, the chaotic nature and complicated reactions that
Provided convection is the main heat transfer mechanism strongly impact the gassolid ow behavior. Additionally, there
within the reactor, the heat transfer coecient is related to the are no studies comparing the quality of syngas based on
Nusselt number of phase q by dierent indices for a gasication process by using biomasses
5771 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

Table 3. Experimental Operating Conditions and Syngas Analysis

experimental conditions forest residues coee husk vine-pruning residues
temperature (C) 815 815 790 815 790 790 790 790 815
admission biomass (kg/h) 63 74 63 28 28 41 25 55 55
air ow rate (Nm3/h) 94 98 98 75 72 80 52 40 40
syngas ow rate (Nm3/h) 106 94 100 106 88 116 107 108 102
syngas fraction (dry basis)
H2 8.2 8.4 7.6 12.4 7.6 7.5 5.1 10.4 12.7
CO 18.6 18.0 17.9 11.4 11.1 10.6 8.3 11.7 14.1
CH4 4.6 4.4 4.4 1.6 2.4 2.4 1.1 2.4 2.3
CO2 16.7 17.1 17.1 18.7 17.0 18.5 16.5 20.1 17.9
N2 48.0 48.2 49.2 52.3 54.2 55.2 56.4 51.2 49.1
syngas NHV
MJ/Nm3 5.16 5.02 4.93 3.34 3.20 3.07 1.99 3.46 4.02

largely available in Portugal. The present model was calibrated

and validated using experimental data and operating
information from a pilot-scale uidized bed gasier.
Table 3 shows the experimental operating conditions and the
syngas analysis in three dierent gasication runs for the three
dierent studied biomasses. The net heat values (NHV) were
computed considering the syngas basis composition deter-
mined by gas chromatography and using the standard
combustion heats of the compounds obtained at 25 C. An
increase in gasication temperature promoted the formation of
a syngas with higher hydrogen and carbon monoxide contents
(the major contributors to the caloric value of syngas) and,
consequently, higher NHV of the syngas. Generically, the
results obtained with the dierent biomasses show a relation
between the biomass caloric value and the caloric value of
the syngas obtained: higher biomass caloric value (Table 2)
results in higher caloric syngas production. In fact, it turns out
that the syngas formed with the highest caloric value was
obtained by gasication of forest residues, a more energetic
material. This relationship between the biomass caloric
content and the syngas NHV can be explained considering
rst that the biomass caloric value is related to the amount of
carbon and hydrogen present in the biomass, that is, an
increased amount of carbon and hydrogen lead to higher
caloric value, and second, a larger amount of these two
elements allows production of larger quantities of hydrogen and
carbon monoxide, the major contributors for the caloric value
of the syngas. Analysis of the dependence of producer gas
properties on conditions of biomass admission seems to show
that for all gasication runs studied there is a decrease in syngas
caloric value with increasing quantity of biomass admitted.
Because the syngas NHV is also dependent on the air ow rate,
dierent biomass feeding rates were simulated keeping the
other operating conditions constant to verify the former
assumption. It was observed that increasing the biomass feeding
rate leads to lower syngas NHV (for all the studied biomasses).
These results are in agreement with the literature.27 This is
consistent with the fact that a faster feeding of biomass to the
reactor reduces the extent of the reactions. From the
experimental gasication runs, it can be observed that the
vine-pruning residues lead to the highest generation of
hydrogen, while forest residues promote the highest content
Figure 4. Experimental and numerical molar syngas composition
of CO. (obtained from coee husks) considering the operating conditions
Figure 4 gives the experimental and numerical results for from (a) run 1, (b) run 2, and (c) run 3.
coee husks gasication runs described in Table 3. It can be
veried that the numerical results show a good agreement with
those obtained from the experimental runs carried out in a
5772 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

pilot-scale bubbling uidized bed gasier. Considering all opposite eect for CO. CO2 and H2 acquire their exit values
simulations (for the three dierent biomasses), the maximum very quickly below the free board area. At the bottom of the
obtained error was about 20%. The simulations also predict the gasier there is a large amount of unconverted carbon, which
species molar fraction behavior at dierent runs with similar becomes reduced along the gasier height. On the other hand,
trends considering the increasing or decreasing of species at the volatile combustion also leads to increased amounts of CO2
dierent operating conditions. The numerical code was applied at higher positions in the gasier. The amount of CH4 is higher
successfully to the three biomasses. These errors are very in the inlet region of the gasier as a result of devolatilization.
reasonable for a complex system such as biomass gasication in N2 shows the higher amounts in the low region of the gasier
a pilot-scale reactor. Dierences can be found due to some and then decreases for upper zones. Similar pattern
simplifying assumptions followed by our developed model. distributions were found for forest and vine-pruning residues.
Proximate analysis and kinetics (some data are relative to coal The developed model was also applied to dierent operating
particles) were taken from literature. Ash and some light conditions, namely, testing the eect of the gasication
hydrocarbons were not considered. Also, a 2-D approach was temperature and the steam-to-biomass ratio (SBR) on the
followed. A more detailed devolatilization approach should be syngas molar fractions. Figure 6 depicts the syngas molar
attempted, and tar decomposition should also be treated. These fraction as a function of the gasication temperature. From
simplications should be properly treated in future papers. Figure 6 it can be seen that the increase of the gasication
4.2. Model Performance. The thermal biomass gas- temperature promotes the formation of a syngas with higher
ication process involves a set of complex chemical reactions hydrogen and carbon monoxide content and consequently
that lead to the formation of three fractions: syngas, ashes higher syngas NHV, as stated in the experimental data. Both
(chars), and condensates. The most important fraction, hydrogen and carbon monoxide seem to increase up to
amounting to more than 70% (by weight) consists of light asymptotic values. On the other hand, the CH4 and CO2
gases, namely, CO, H2, CH4, CO2, and N2. Figure 5 shows the contents follow an opposite trend. CH4 decreases as a function
contours of H2 and CO2, CH4, and N2 molar fractions in the of the temperature because the methane reaction formation is
exothermic. These results are in agreement with the literature.28
gas mixture (obtained from coee husks), respectively. The
According to Le Chateliers principle, higher temperatures favor
higher values for H2 are located immediately above the air
products in endothermic reactions. Therefore, the endothermic
inlets, while at a greater height for CO2. The watergas shift
reaction in eq 15 was strengthened leading to an increase in the
reaction is exothermic, and the decrease of temperature in the
hydrogen content.
upper zones makes the H2 values increase and makes the Figure 7 shows the syngas H2/CO molar ratio as a function
of the temperature. The main application for a syngas with high
H2/CO ratio is related to the fuel cell technology. In general,
syngas generated from biomass tends to have a relatively low
H2/CO ratio (<1). It can be observed that vine-pruning and
coee husks residues show the higher H2/CO ratios with values
ranging approximately from 0.7 to 0.8. On the other hand, and
due to the higher values of CO, forest residues show H2/CO
ratios nearer to 0.4 and approximately constant along the
selected temperature range.
For domestic purposes, it is preferable to obtain a syngas
with higher CH4/H2 ratios. Figure 8 depicts the syngas CH4/
H2 molar ratio as a function of the temperature for the three
studied biomasses. All the studied biomasses show a similar
pattern with a decrease of the CH4/H2 molar ratio as the
temperature increases. Once again, both coee husks and vine-
pruning residues show similar ratio values. The higher values
are found for forest residues.
The carbon conversion denes the fraction of carbon from
biomass converted to carbon in syngas components:
12 M
carbon conversion =
XC m (33)
where M is the total molar ow of carbon in the syngas
components, XC is carbon fraction in the biomass (obtained
from the ultimate analysis), and m is the biomass ow into the
gasier.
From Figure 9, it can be seen that higher carbon conversions
are obtained by using forest residues as substrate, with a
maximum value of 97.9% at 1073 K. Typical values in a
gasication reactor vary between 70% and 90%. On the other
hand, gasication using coee husks only leads to carbon
conversions up to 80%. From literature there are no clear
Figure 5. Countours of the (a) H2, (b) CO2, (c) CH4, and (d) N2 trends of carbon conversion evolution as a function of the
molar fractions in the gas mixture (run 1 for coee husks). temperature and SBR.29 Surprisingly, it was found that carbon
5773 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

Figure 7. Syngas H2/CO molar ratio as a function of the temperature

(operating conditions from run 1coee husks).

Figure 8. Syngas CH4/H2 molar ratio as a function of the temperature

(operating conditions from run 1coee husks).

Figure 6. Syngas molar fractions as a function of the gasication

temperature for (a) coee husks, (b) forest residues, and (c) vine-
pruning residues (operating conditions from run 1coee husks).

conversion is a very weak function of temperature. Indeed, data

found in the literature are extremely scattered. From our
numerical simulation, it was observed that the carbon
conversion changes only slightly with the temperature, and
depending on the substrate a slight increase or decrease is
found. These results conrm the main conclusions drawn from
the literature.
The CGE is dened as the percentage of the heating value of Figure 9. Carbon conversion as a function of the temperature
(operating conditions from run 1coee husks).
biomass converted into the heating value of the product gas
(HHV). It can be computed as follows:
From Figure 10, it can be veried that the CGE increases
gas yield HHV of product gas
CGE = with the gasication temperature and that the vine-pruning
HHV of fuel + heat addition (34) residues show the higher eciencies. High gas eciencies are
5774 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777
Energy & Fuels Article

values of SBR both carbon and methane are reformed leading

to the formation of H2 with its concentration increase. At SBRs
higher than 1, the hydrogen increases slowly, and after 1.5 there
is no increase eect. This implies that the surplus of steam is
not reacting and that adding more steam is not ecient. Similar
trends were observed for forest and vine-pruning residues. The
eect of the SBR on the dierent indices for the syngas quality
is depicted in Table 4.
The H2/CO ratio in volatile matter is constant, so this ratio
increase is related to the chemical reactions that are being
developed along the gasication process and favor the H2
generation and CO consumption (water/gas shift reac-
tion).The H2/CO ratio increases as a function of the SBR.
Higher SBRs lead to the generation of H2, and consequently
this ratio increases. Considering the CH4/H2 ratio, it was
veried that the SBR eect was very slight. The carbon
conversion shows a minimum at middle SBR values and then
starts again to increase slightly. Finally, the CGE decreases with
Figure 10. Cold gas eciency as a function of the gasication
temperature (operating conditions based in run 1coee husks).
the SBR.
The developed model predicts with reasonable accuracy the
syngas composition at dierent operating conditions on a pilot
suitable for integrated gasication combined cycles (IGCC) scale, showing the eective trend of each one of the light gases.
applications. Other syngas quality indices can also be computed. This is a
The SBR can be dened as the steam mass ow rate divided powerful tool for designing large-scale systems that may be
by the fuel mass ow rate (dry basis). The SBR was varied over dominant in a near future. The adopted simplications are
a range of values from 0 to 1.5 by holding the other variables identied and will be the subject of analysis in future papers.
constant. Figure 11 shows the syngas molar fractions as a
function of the SBR.
5. CONCLUSIONS
A multiphase 2-D model was developed under the Fluent
framework to simulate the gasication of three large available
Portuguese biomasses: coee husks, forest residues, and vine-
pruning residues. An EulerianEulerian approach exchanging
mass, energy, and momentum coupled with homogeneous and
heterogeneous reactions was used. The model was calibrated
and validated using experimental data and operating
information from a pilot-scale uidized bed gasier. The
experimental data obtained for large-scale systems are crucial to
better understanding the gasication phenomenon and to
better predict nal syngas compositions. Numerical results are
in good agreement with the experimental data obtained at three
dierent operating conditions and for all the studied biomasses.
The numerical model also predicted the eect of the
gasication temperature and SBR on the nal syngas
composition, as well as the CGE, carbon conversion, and H2/
CO and CH4/H2 ratios.
Figure 11. Syngas molar fractions as a function of the steam to Dierences were observed mainly due to some simplica-
biomass ratio for coee husks (other operating conditions from run tions that were included in the numerical simulation develop-
1coee husks). ment; specically, proximal analysis and reaction kinetic data
were taken from the literature, light hydrocarbons and ash were
It can be observed that, with the increase of SBR, the content not treated, and a 2-D geometry was adopted.
of hydrogen and CO2 showed an increasing trend, while that The dierent indices are very useful to select the best
from CO decreased, in agreement with literature.28,30 At low biomass as a function of the gasication application. Among the

Table 4. Steam-to-Biomass Ratio Eect on H2/CO Ratio, CH4/H2 Ratio, Carbon Conversion, and Cold Gas Eciency for All
the Studied Biomasses

experimental conditions forest residues coee husk vine-pruning residues

SBR 0 0.75 1.5 0 0.75 1.5 0 0.75 1.5
H2/CO molar ratio 0.36 0.48 0.56 0.76 1.30 1.60 0.78 1.19 1.32
CH4/H2 molar ratio 0.70 0.70 0.71 0.17 0.17 0.18 0.19 0.20 0.21
carbon conversion (%) 97.9 92.2 92.9 78.0 72.2 74.1 89.3 83.9 84.3
cold gas eciency (%) 72.1 67.2 65.1 63.0 52.9 50.0 76.6 67.2 65.1

5775 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777

Energy & Fuels Article

three biomasses, it was concluded that coee husks and vine- ls = energy exchange between the uid phase and the solid
pruning residues show better H2/CO ratios, which are phase
especially suitable for fuel cell integration purposes. Consider- ka = diusion coecient
ing domestic applications, a typical index is the CH4/H2 ratio ka(s) = diusion energy
where the highest values were obtained by using the forest (PsI + s ): (vs ) = generation of energy by the solid stress
residues. The highest CGEs were obtained by using the vine- tensor
pruning residues. The obtained data are crucial to describe a = collisional dissipation of energy
scenarios concerning the potential use of biomass as a source of = tensor stress
energy in Portugal. = viscosity

AUTHOR INFORMATION
Corresponding Author
c = stoichiometric coecient
Subscripts
g = gas phase
*E-mail: rouboa@utad.pt. Phone: 351 259350317. Address: s = solid phase
Quinta de Prados, Apartado 1013,5001-801 Vila Real. Fax: i = component

+351 259 350 356.
Author Contributions REFERENCES
The manuscript was written through contributions of all (1) Ciferno, P.; Marano, J. Benchmarking Biomass Gasication
authors. All authors have given approval to the nal version of Technologies for Fuels, Chemicals and Hydrogen Production; U.S.
the manuscript. Department of Energy, National Energy Technology Laboratory,
Notes 2002. Available from: http://www.netl.doe.gov/technologies/
The authors declare no competing nancial interest. coalpower/gasication/pubs/pdf/BMassGasFinal.pdf, last accessed

29th January, 2013.

ACKNOWLEDGMENTS (2) Jenkins, B.; Baxter, L.; Miles, T., Jr; Miles, T. Combustion
properties of biomass. Fuel Proc. Technol. 1998, 54, 1746.
We would like to express our gratitude to the Portuguese (3) Ferreira, S.; Moreira, N.; Monteiro, E. Bioenergy overview for
Foundation for Science and Technology (FCT) for the support Portugal. Biomass Bioenergy 2009, 33, 15671576.
given to Grant No. SFRH/BPD/71686 and to Project PTDC/ (4) Fiori, L.; Florio, L. Gasification and Combustion of Grape Marc:
AAC-AMB/103119/2008 and ALTERCEXA-POCTEC Pro- Comparison Among Different Scenarios. Waste Biomass Valorization
gram. 2010, 1, 191.

(5) Krtay, E. Recent advances in production of hydrogen from

NOMENCLATURE biomass. Energy Convers. Manage. 2011, 52, 177889.
(6) Spiegel, C. Designing & Building Fuel Cells; McGraw-Hill; 2007.
A, B = calibration constants (7) The hydrogen economy: a non-technical review; United Nations
C1, C2, C3 = constants Environment Programme, 2006.
Cp = specic heat capacity (8) Rosen, M. A. Thermodynamic investigation of hydrogen
D0 = diusion rate coecient production by steammethane reforming. Int. J. Hydrogen Energy
ess = restitution coecient 1999, 16, 20717.
go,ss = radial distribution function (9) Levin, D.; Chahine, R. Challenges for renewable hydrogen
Gk = generation of turbulence kinetic energy due to the production from biomass. Int. J. Hydrogen Energy 2010, 35, 496269.
mean velocity gradients (10) Ahmed, T.; Ahmad, M.; Yusup, S.; Inayat, A.; Khan, Z.
Gb = generation of turbulence kinetic energy due to Mathematical and computational approaches for design of biomass
gasification for hydrogen production: A review. Renewable Sustainable
buoyancy Energy Rev. 2012, 16, 230415.
hpq = heat transfer coecient between the uid phase and (11) Xie, J.; Zhong, W.; Jin, B.; Shao, Y.; Liu, H. Simulation on
the solid phase gasification of forestry residues in fluidized beds by Eulerian
k = thermal conductivity Lagrangian approach. Bioresour. Technol. 2012, 121, 3646.
Mc = molecular weight (12) Xue, Q.; Fox, R. Multi-fluid CFD modeling of biomass
Mw,i = molecular weight of i component gasification in polydisperse fluidized-bed gasifiers. Powder Technol.
p = gas pressure 2014, 254, 187198.
qq = heat ux (13) Xue, Q.; Dalluge, D.; Heindel, T.; Fox, R.; Brown, R.
qth = specic enthalpy Experimental validation and CFD modeling study of biomass fast
R = universal gas constant pyrolysis in fluidized-bed reactors. Fuel 2012, 97, 757769.
(14) Gungor, A.; Yildirim, U. Two dimensional numerical
Rc = reaction rate computation of a circulating fluidized bed biomass gasifier. Comput.
Sk = user-dened source terms Chem. Eng. 2013, 48, 234250.
Sq = source term due to chemical reactions (15) Knowlton, T.; Karri, R.; Issangya, A. Scale-up of fluidized-bed
S = user-dened source terms hydrodynamics. Powder Technol. 2005, 150, 7277.
T = temperature (16) Eaton, M.; Smoot, D.; Hill, C.; Eatough, N. Components,
U = mean velocity formulations, solutions, evaluation, and application of comprehensive
v = instantaneous velocity combustion models. Prog. Energy Combust. Sci. 1999, 25, 387436.
Y = mass fraction (17) Bridgewater V.; Evans D. An assessment of thermochemical
Y M = contribution of the uctuating dilatation in conversion systems for processing biomass and refuse; Energy technology
compressible turbulence to the overall dissipation rate support unit (ETSU) on behalf of the Department of Trade, ETSU B/
T1/00207/REP; 1993.
Other Symbols (18) Wilson, L.; John, G.; Mhilu, C.; Yang, W.; Blasiak, W. Coffee
= volume fraction husks gasification using high temperature air/steam agent. Fuel Process.
= density Technol. 2010, 91, 133037.

5776 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777

Energy & Fuels Article

(19) Launder, B.; Spalding, B. Lectures in Mathematical Models of

Turbulence; Academic Press: London, England, 1972.
(20) Syamlal, M.; Rogers, T. J. MFIX documentation: Vol. 1. Theory
Guide; National Technical Information Service: Springeld, VA, 1993.
DOE/METC-9411004, NTIS/DE9400087.
(21) Lun, C.; Savage, S. B.; Jeffrey, D.; Chepurniy, N. Kinetic theories
for granular flow: Inelastic particles in couette flow and slightly
inelastic particles in a general flow field. J. Fluid Mech. 1984, 140, 223
256.
(22) Badzioch, S.; Hawsley, P. G. W. Kinetics of Thermal
Decomposition of Pulverized Coal Particles. Ind. Eng. Chem. Process
Des. Dev. 1970, 4, 52130.
(23) Yu, L.; Lu, J.; Zhang, X.; Zhang, S. Numerical simulation of the
bubbling fluidized bed coal gasification by the kinetic theory of
granular flow. Fuel 2007, 86, 72234.
(24) Baum, M.; Street, P. Predicting the behavior of coal particles.
Combust. Sci. Technol. 1971, 3, 23143.
(25) Field, M. A. Rate of combustion of size-graded fractions of char
from a low rank coal between 1200K-2000K. Combust. Flame 1969, 13,
237252.
(26) Gelderbloom, S. J.; Gidaspow, D.; Lyczkowski, R. W. CFD
simulations of bubbling/collapsing fluidized beds for three geldart
groups. AIChE J. 2003, 49, 84458.
(27) Miao, Q.; Zhu, J.; Barghi, S.; Wu, C.; Yin, X.; Zhou, Z. Modeling
Biomass gasification in Circulating Fluidized Beds: Model Sensitivity
analysis. Int. J. Energy Power Eng. 2013, 2, 5763.
(28) Turn, S.; Kinoshita, C.; Zhang, Z.; Ishimura, D.; Zhou, J. An
experimental investigation of hydrogen production from biomass
gasification. Int. J. Hydrogen Energy 1998, 23, 64148.
(29) Timmer K. Carbon conversion during bubbling uidized bed
gasication of biomass. Ph.D. Thesis, Iowa University, 2008.
(30) Silva, V.; Rouboa, A. Using a two-stage equilibrium model to
simulate oxygen air enriched gasification of pine biomass residues. Fuel
Process. Technol. 2013, 109, 11117.

5777 dx.doi.org/10.1021/ef500570t | Energy Fuels 2014, 28, 57665777