Moving porous bed of suction gasifier is modeled as one-dimensional 1-D with finite control volumes CVs, where conservation of mass, momentum and energy is represented by fluid flow, heat tr
Trang 1Modeling and simulation of a downdraft biomass gasifier 1 Model development and validation
Mech Engg Dept., D.C.R University of Science & Technology, Murthal, Sonepat 131 039, Haryana, India
a r t i c l e i n f o
Article history:
Received 29 December 2009
Received in revised form 27 September
2010
Accepted 3 October 2010
Available online 29 October 2010
Keywords:
Modeling
Simulation
Biomass gasification
Equilibrium
Kinetics
Suction gasifier
a b s t r a c t
An ‘EQB’ computer program for a downdraft gasifier has been developed to predict steady state perfor-mance Moving porous bed of suction gasifier is modeled as one-dimensional (1-D) with finite control volumes (CVs), where conservation of mass, momentum and energy is represented by fluid flow, heat transfer analysis, drying, pyrolysis, oxidation and reduction reaction modules; which have solved in inte-gral form using tri-diagonal matrix algorithm (TDMA) for reaction temperatures, pressure drop, energet-ics and product composition Fluid flow module relates the flow rate with pressure drop, while biomass drying is described by mass transfer 1-D diffusion equation coupled with vapour–liquid-equilibrium rela-tion When chemical equilibrium is used in oxidation zone, the empirically predicted pyrolysis products (volatiles and char) and kinetic modeling approach for reduction zone constitutes an efficient algorithm allowing rapid convergence with adequate fidelity Predictions for pressure drop and power output (ifier) are found to be very sensitive, while gas composition or calorific value, temperature profile and gas-ification efficiency are less sensitive within the encountered range of gas flow rate
Ó 2010 Elsevier Ltd All rights reserved
1 Introduction
Thermochemical conversion of woody biomass under restricted
supply of oxidant is among the most promising non-nuclear forms
of future energy Besides utilizing a renewable energy sources, the
technology also offers an eco-efficient and self sustainable way of
obtaining gaseous fuel usually called producer gas It can be used
in either premixed burners (dryers, kilns, furnaces or boilers) for
thermal applications or in direct feeding of high efficiency internal
combustion engines/gas turbines for mechanical applications
After adequate cleaning up and reforming, the generated gas can
also be used for feed high temperature fuel cells or for production
of hydrogen fuel[1] For electric power generation applications, the
motive power from prime mover such as IC engine or gas turbine
can be connected to an electric generator to produce electric
en-ergy Applications of IC engines have proved to be the most
effi-cient and least expensive decentralized-power-generation
systems at lower power range Research efforts have been
ex-panded worldwide to develop this technology cost-effective and
efficient in lower power range
Recent progression in numerical simulation techniques and
computer efficacy become the effective means to develop more
ad-vanced and sophisticated models in order to provide more accurate
qualitative and quantitative information on biomass gasification
In the present work, the objective is not merely to develop a theo-retical model of a downdraft gasifier system, but also to develop an efficient algorithm that allow rapid convergence and adequate accuracy of predictions Presently, the gasification modeling tech-niques include the application of thermodynamic equilibrium, chemical kinetics, diffusion controlled, diffusion–kinetic approach and CFD tools None of approaches have clear advantage over the others Pure equilibrium approach has thermodynamic limitations, instead of its inherent advantage of being generic, relatively easy to implement and rapid convergence, even though, researchers have successfully demonstrated the application of equilibrium chemis-try in downdraft gasifiers Zainal et al.[2]reported an interesting model for biomass gasifier describing the equilibrium calculations considering water–gas shift and methane–char reactions Melgar
et al.[3]combine chemical and thermal equilibrium in order to predict gas composition and Baratieri et al [1] presented an equilibrium model based on minimization of Gibbs energy using Villars–Cruise–Smith (VCS) algorithm They validated the predic-tions successfully Later, Sharma[4]has compared the theoretical predictions of reduction zone using equilibrium, kinetic modeling and experimental data For optimum performance, Sharma has identified a critical length for the reduction zone (where all char gets converted) At a more sophisticated level, Ratnadhariya and Channiwala[5], suggested that separate thermodynamic modeling can be approached to different zones of a downdraft gasifier On the other hand, non-equilibrium formulations such as kinetic rate
0196-8904/$ - see front matter Ó 2010 Elsevier Ltd All rights reserved.
⇑Tel.: +91 09416722212; fax: +91 01302484004.
E-mail address: avdhesh_sharma35@yahoo.co.in
Contents lists available atScienceDirect Energy Conversion and Management
j o u r n a l h o m e p a g e : w w w e l s e v i e r c o m / l o c a t e / e n c o n m a n
Trang 2and/or diffusion controlled model including CFD tools are more
accurate, no doubt, but are detailed and computationally more
intensive It takes time for convergence by a few orders of
magni-tudes[6] Non-equilibrium approaches use char conversion as a
surface phenomena describing by char reactivity and global
reac-tions of char–gas and gas–gas reacreac-tions An effective global rate
constant may be defined to account for both diffusion and kinetics
of these reactions Wang and Kinoshita[7]modeled the kinetics of
the heterogeneous and homogeneous reactions of char conversion
in reduction zone for a given residence time and bed temperature,
while Giltrap et al.[8]used the reaction kinetics parameters
re-ported by Wang and Kinoshita in order to develop a model of the
gas composition and temperature for char reduction zone of a
downdraft gasifier Babu and Sheth[9]further modified these
reac-tion rates using a variable char reactivity factor to predict the
re-sults agreeing with experimental data Later, Gao and Li [10]
presented the downdraft gasifier model by combining a pyrolysis
model (based on Koufopanos scheme) and reduction model
field in time and space
The overall pressure drop across the gasifier system is an
impor-tant parameter It monitors not only the health of a suction gasifier
but also the volumetric efficiency of engine and hence the engine
power output The pressure drop across the conventional packed
bed depends on system geometry, medium porosity, permeability
and physical properties of working medium Unlike, in gasifiers
the bed maintains widely varying temperature specifications,
par-ticle size distribution and bed porosity Such study on pressure
drop through a downdraft biomass gasifier bed is limited in open
literature Sharma [11], measured the pressure drops across the gasifier bed at various particle size arrangements in cold and hot flow, and at various locations of a 20 kWe open top downdraft gas-ifier in addition to temperature profile, gas composition, calorific value These data has been used in this work
In fact, the selection of level and modeling approach (viz chem-ical equilibrium, chemchem-ical kinetics or diffusion controlled) depends
on statement of the problem and therefore may vary considerably from one case to another Since, the objective of the present work is not to invoke the highest level or most sophisticated gasifier
mod-el, yet it is an attempt to develop an efficient algorithm that enable rapid convergence without affecting the validity Such comprehen-sive work, in fact, is missing in the archival literature This compre-hensive work, therefore, presents the modular treatment (allowing scope of further improvement at module level) to fluid flow, heat transfer, biomass drying, pyrolysis, and oxidation and reduction reactions processes to form a powerful tool for simulation of suc-tion (downdraft) gasifier Here biomass drying has been described via thermal equilibrium, where mass transfer determines the rate
of moisture removal from wet biomass particles Devolatilization rate and pyrolysis products is described by single pseudo-first or-der reaction and empirical model, chemical equilibrium for rapid convergence in oxidation zone (>800 °C), and kinetic scheme for reduction zone (<800 °C), constitutes an efficient algorithm for suc-tion biomass gasifier allowing considerable saving in iterative time without degradation in accuracy of predictions Such gasifier algo-rithms are desired when combining a gasifier model with a gas– engine model towards simulation/optimization of gasifier–engine system Predictions of model and its subroutines have been
Nomenclature
k thermal conductivity/rate constant
Dm_ solid mass conversion
[C] species concentration
D diffusion coefficient/diameter
_
m= _n mass/molar flow rate
Ru universal gas constant
Dtres residence time in CV
r
_
diameter ratio of annulus
Dh hydraulic diameter (m)
hhv high heating value (kJ/kg)
_
Q heat flow or release/absorbed
Y mass fraction or ratio
Greek letters
l dynamic viscosity (kg m1s1)
erad radiative emissivity
n correction factor for annulus Subscripts/superscripts
mfd modified fully developed preheat preheating zone
vol/v volatile
s solid (biomass, char, ash)
DBp mass percentage in dry biomass
Trang 3validated, which are showing realistic performance including
pres-sure drop trends in cold and hot flow condition
2 Mathematical formulation
The packed bed of the biomass gasifier has been modeled as a
porous medium in which fluid flow rate increases in the direction
of flow due to thermo-chemical conversion of solid particles
con-stituting the bed The flow of air and biomass consumption in
the gasifier is related by the phenomena of fluid flow, heat transfer,
and thermo-chemical processes, viz., preheating, drying, pyrolysis,
combustion and reduction reactions For simplification, these
ther-mo-chemical processes are described by five separate zones in
addition to annular jacket zone as shown inFig 1, however, the
ac-tual dividing lines between these zones evolve as solution
pro-ceeds The upper end of the oxidation zone starts at the elevation
of the tuyers, and all the other zones are determined by their
respective temperatures as the solution evolves
For modeling, these zones are further subdivided into a number
of CVs for analysis, where each CV has been characterized by the
average values of parameters such as temperature, particle size,
fluid flow rate, reactor diameter, etc In all CVs, the solid particles
are considered spherical, with uniform diameters In the drying
and preheating zones, there is no shrinkage in particle size due
to drying and preheating process However in pyrolysis, oxidation
and reduction zones, feedstock undergoes chemical reactions
lead-ing to change in the particle size, thus, the diameter is allowed to
vary from one CV to another The revised particle size is computed
from the assumption of constant intrinsic density;
mp
mp;initial
¼d
3
p
Here, mpand mp,initialare the mass of particle in the current CV and
mass of particle at the start of pyrolysis process Bulk porosity in the
gasifier bed varies with particle diameter and thus can be
deter-mined from the correlation of Chen and Gunkel[12]as
2.1 Fluid flow
Initial phase for gasifier model development is fluid flow
mod-eling, which is used to quantify the apportionment of air inflow
from the open top and through the air tuyers, and the pressure
drop through the gasifier bed as a function of flow rate The tuyres
are straight pipes of circular cross-section, the pressure drop can be computed from the Darcy–Weisbach equation The entrance and developing flow effect through the tuyers has been modeled in terms of average entry length pressure drop parameter Kdev, fitted
to the data of Schmidt and Zeldin in Ref.[13]as given inTable 1 The pressure drop through the gasifier bed (maintaining widely varying temperature specifications, particle size distribution and bed porosity) has been obtained using Ergun correlation[14]for complete flow regime Pressure drop through concentric annulus
is modeled from modified Darcy–Weisbach friction-factor in terms
of effective Reynolds number and effective (annulus) diameter as reported in[15] The details of equations describing the flow resis-tance across the tuyers, porous bed and annulus are given in
2.2 Heat transfer For heat transfer analysis, the approach of Sharma et al.[16]has been followed in the present work Here, fuel bed is assumed to be isotropic; solid and gases are considered to be in local thermal equilibrium These assumptions are justified for fixed bed gasifiers operating under steady state conditions, since residence time of solids in the CV is two to three order magnitude higher than that
of gases This module describes the formulation of energy interac-tion for the heat inflows and outflows due to advecinterac-tion of fluid and solids, heat loss through insulated wall, internal thermal interac-tion between adjacent CVs and the quantity of heat generated or consumed in each CV in order to compute the reaction tempera-ture of each CV In developing heat transfer module, the heat gen-erated/absorbed during drying, pyrolysis, oxidation or reduction is prescribed as input These would subsequently be determined by the modules of the respective sub-processes (cf Eq (6) inTable 2)
Preheating zone
Drying zone
Pyrolysis zone
Oxidation zone
Reduction zone
Ash
Producer gas
Annular Jacket regeneration zone
Air (Tuyers)
Gas
Table 1 Fluid flow: pressure drop equations.
Tuyers
DP tuy ¼ ðP atm P in;tuy Þ þ ðP in;tuy P exit;tuy Þ
¼qV2in
2 þ f L
D þ K de v
in
2 ¼ 1 þ f L
D þ K de v
2
(3)
Pressure drop parameter [13]
K de v ¼ exp 0:3 2:9 10 3 = Lde v
ReD
1:43 10 6 = Lde v
ReD
(3a)
L de v
ffi 4.4Re 1/6
Porous Bed: Ergun equation [14]
DP i ¼150ð1eb;i Þ 2 l ðT i Þl i
q ðT i Þ e 3 b;i d 2 p;i A T ð _ m f Þiþ1:75ð1eb;i Þl i
q ðT i Þ e 3 b;i d p;i A 2 ð _ m f Þ 2
Concentric annulus [15]
DP an ¼ K de v þ f mfddLD eff1
pg
2 qpgA 2 an
(5)
n ¼ ð1 r
_
Þ 2 ð1 r_2Þ ð1 r_4 Þþð1 r_2 Þ 2 = lnð r_Þ where r_¼ r o =ri (5b)
Table 2 Heat transfer equations.
Energy equation P
solid ð _ m i Cp i Þinþ P
gases ð _ m i Cp i Þin
T in þ _ Q v ap þ _ Q pyr þ _ Q oxid
þ _ Q red þ P
jk Q _ dif ;jk ¼ P
solid ð _ m i Cp i Þoutþ P
gases ð _ m i Cp i Þout
T out
(6)
Effective thermal conductivity [16]
k eff ¼2ks k f C 1 ðln C 2 þC 1 Þ
k g ðln C 2 þC 1 Þk s C 1 þks d 2
ct
d 2 þ 4rXd p T3andX¼eb =1 þeb ð1 e rad Þ
2 e rad ð1 e b Þ (7) Thermal resistance for ith zone
_
Qdif ; jk¼DTjk
R si ; where R si = R t(i,bed) + R t(i,ins) + R t(i,o) (8) where jk = up, down, side
Trang 4using enthalpy of formation of reactants and products The transfer
of energy between adjacent CVs due to fluid and solid particles
mo-tion is accounted for by the mass flow rate, temperature of the fluid
and solid flows and all heat transfer interactions including the
ra-dial outward (heat loss) from the bed to surroundings have been
modeled using thermal resistance as shown inFig 2 The details
of equations representing the heat transfer module for each CV
are given inTable 2
The total resistance to radial heat loss to the surroundings in the
ith zone of the gasifier bed is given as the sum of resistances due to
granular bed, insulation and the outer surface of the reactor (cf Eq
(8)) In the preheating zone, there is an additional resistance due to
the annular jacket The axial heat transfer of the porous gasifier bed
has been modeled by considering advection of solid (biomass/char)
and fluid (air/gas) streams, while conductive and radiative heat
fluxes at boundaries of each CV have been modeled in terms of
effective thermal conductivity, Keff, following Sharma et al.[16]
The keffmodel needs inputs in terms of bed temperature, particle
size and bed porosity at current location Here, bed porosity varies
with current particle size and modeled using Eq.(2), while
emissiv-ity of char particles is fixed at 0.75
2.3 Thermochemical processes
Modeling of the biomass thermo-chemical conversion
phenom-ena: preheating, drying and pyrolysis, and chemical reactions:
oxi-dation and reduction in a downdraft gasifier has been presented to
predict the rate of heat generation/absorption in each CV and
out-flow products
2.3.1 Biomass drying
The mechanism of moisture transfer to woody biomass includes
diffusion through the fluid film around the solid particles and
dif-fusion through the pores to internal adsorption sites The actual
process of physical adsorption is practically instantaneous, and
equilibrium can be assumed to exist between the surface and the
fluid envelope As moist biomass particles came into contact with
air having low humidity level, the particles tend to lose moisture
to the surrounding air until equilibrium is attained For modeling,
following assumptions are made:
1 No shrinkage in particle due to moisture evaporation
2 Temperature gradient in moist biomass particles is neglected
3 Equilibrium can be assumed to exist between the surface and
the fluid envelope
4 Drying is allowed to continue through pyrolysis zone as well as
oxidation and reduction zones as well
The local thermal equilibrium between the gaseous and solid
media is assumed in each control volume, which makes it implicit
that heat transfer between the solid and gases is much faster than
the mass transfer Thus, mass transfer determines the rate of
mois-ture removal from the biomass particles to the gases/air flowing
around them The analytical solution for one-dimensional mass
diffusion in a spherical particle of wood[17]is used in this work Equations representing the drying process with coefficients are listed inTables 3 and 4
2.3.2 Pyrolysis of biomass
In downdraft gasifier, the pyrolysis process is modeled at slow heating rate to predict pyrolytic yields (viz., volatile composition and char) and devolatilization rate as a function of temperature and residence time The biomass particles shrink on pyrolysis giv-ing char and ash Followgiv-ing assumptions are invoked:
Char and biomass particles are non porous
Char yields from cellulose, hemicellulose and lignin considered
to be pure carbon
Char yield in the gasifier is insensitive to pyrolysis temperatures encountered in the pyrolysis zone
The complex constituents of volatiles are assumed to be decom-posed into CO, H2, CO2, H2O, tar (heavy hydrocarbons) and light hydrocarbons (mixture of methane and ethylene)
The whole process of thermal decomposition of dry biomass can
be represented by a single equation as:
Dry biomass ðDBÞ !kdryChar
þ Volatiles ðCO; H2; CO2; H2O; Methane-Equivalent & TarÞ
ð14Þ Fig 2 Single CV used in heat transfer module with all thermal interactions.
Table 3 Equations representing to moisture evaporation.
Diffusion equation [17]
X in X eqb
X out X eqb ¼ 8
where b ¼4Ddif t res
d 2 ;t res ¼Mb;CV
_
m b
Simpson [18] relationship
Xeqb¼ 1800
W 1KhKh þ K1 Khþ2K 1 K 2 K 2 h 2 1þK 1 Khþ2K 1 K 2 K 2 h 2
where
W = 349 + 1.29 (T 273) + 0.0135 (T 273) 2
K = 0.805 + 0.000736 (T 273) 0.00000273 (T 273) 2
K 1 = 6.27 0.00938 (T 273) 0.000303 (T 273) 2
(11)
K 2 = 1.91 + 0.0407 (T 273) 0.000293 (T 273) 2
Relative humidity ratio
h ¼ xair
x air;sat ¼ xair
pv ;sat mw w =p a mw air (12) Antoine equation [19]
log10ðpv;sat Þ ¼ A B
TþC
(13)
Table 4 Coefficients for Antoine equation for saturation vapour pressure [19]
Trang 5On heating, these constituents become unstable and decompose
into char and volatiles Furthermore, the volatiles break-up into
various lighter hydrocarbons For describing the volatile
composi-tion and char yield during slow pyrolysis of the biomass, the
pres-ent work follows the approach of Sharma et al [20], where the
thermal degradation of biomass constituents has been described
by individual decomposition scheme of cellulose, hemicellulose
and lignin Model uses mass fractions of cellulose (Ycel),
hemicellu-lose (Yhc) and lignin (Ylg) in biomass as input information given in
ele-mental balance knowing the mass fractions, chemical formulas
and molecular masses of cellulose, hemicellulose and lignin The
rate of devolatlization of biomass during slow pyrolysis process
can be described by a single pseudo-first order reaction as given
by Eq (15) inTable 6
Each of the three constituents of dry and ash-free biomass, viz.,
cellulose, hemicellulose and lignin are considered to break up into
a fixed fraction of char and volatiles as described by Eqs (16) and
(17) inTable 6 These fractions of char from these three
constitu-ents along with their chemical formula are presented inTable 7
Six species are considered to be part of the volatiles, viz., CO,
CO2, H2, H2O, C1.16H4(ME) and C6H6.2O0.2(tar) following[24] Thus,
the process of pyrolytic decomposition of dry and ash free biomass
C6HHBOOBcan be represented as:
C6HHBOOB¼ C1HcharOcharþ _nv1CO þ _nv2CO2þ _nv3H2
þ _nv4H2O þ _nv5C1:16H4þ _nv6C6H6:2O0:2 ð23Þ
2.3.3 Oxidation chemistry in gasifier bed
The pyrolysis products get oxidized in short supply of oxygen in
the oxidation zone (near air tuyers) of a gasifier Owing to the
widely varying reaction equilibrium constants and the reaction
time scales, some of the reactions might not be attaining
rium in the oxidation zone, and hence the solution of full
equilib-rium equations to compute oxidation process in the gasifier would
both be erroneous and numerically difficult In the present work,
therefore, a heuristic approach is adopted Oxidation of the
pyroly-sis products is allowed to consume the available oxygen in a
se-quence of descending order of reaction rates as described below:
1 Oxidation of hydrogen (Reaction (R1) in Table 8) completes itself first
2 If oxygen remains, light hydrocarbons are oxidized to H2O and
CO (R3)
3 Oxidation is fast, and is assumed to happen instantaneously whenever oxygen is available
4 Products of oxygen are assumed to attain equilibrium in each CV
5 If more oxygen remains, tar (R4) and char (R5) share the oxygen
in the proportion of their reaction rate constants at the temper-ature of the CV under consideration to get oxidized to CO The principal chemical reactions taking place in the oxidation zone along with their rate expressions are listed in Table 8 Although these expressions are not used in the present computa-tions, they have been used only to guide the sequence of oxidation reactions described above
If _nV k stands for the molar flow rate (mol/s) of species k, then after completely consuming all the H2in the gaseous phase (Reac-tion (R1) inTable 8), the O2that would remain _nV O2;1¼ _nV O2– _nV H2=2
If oxygen remains ð _nV O2;1>0Þ, light hydrocarbon or methane-equiv-alent gets oxidized to CO and H2O, therefore, _nVO2;2¼ _nVO2;1 1:58 _nVCO If more oxygen remains ð _nVO2;2>0Þ, simultaneous consumption of tar (Reaction (R4)) and char (Reaction (R5) in
by these reactions has been accounted for by considering the ratio
of the two reaction rates r* = kchar/ktar, where the reaction rates are obtained fromTable 8 Two cases can be discussed: one, when there
is enough oxygen to oxidize all the tar and a proportionate quantity
of char; and second, there is less oxygen than what is required to oxidize tar completely Oxygen remains after tar oxidation if _nVO2;2>ð1 þ rÞð4:45 _nV tarÞ Here, 4:45 _nV tar mol/s of O2is used up to oxidize tar and the remainder for char: thus, for every mole of char oxidized, r*moles of char are also oxidized (cf Reaction (R5)) In case _nVO2;2<ð1 þ rÞð4:45 _nV tarÞ, all oxygen is consumed In this case, the molar rate of tar oxidation is _nV O2;2=½4:45ð1 þ rÞ, and the tar that exits the zone is thus _nV tar _nV O2;2=½4:45ð1 þ rÞ Correspondingly, rate of char oxidation is ½ _nV O2;2r=4:45ð1 þ rÞ mol=s This gives the moles of char oxidized in the current CV If oxygen remains all of
it is then used to oxidize CO in a likewise fashion
Turns [27] quoted that for fuel-rich combustion, the water shift equilibrium equation can be safely applied, therefore we can write
_nVCO2_nVH2= _nVCO: _nVH2O¼ KpðTiÞ ¼ expðDG0ðTiÞ=RuTiÞ ð24Þ whereDG0ðTiÞ ¼ g0
COðTiÞ þ g0
H 2 OðTiÞ g0
CO2ðTiÞ g0
H2ðTiÞ Here,DG0(Ti) is the standard-state Gibbs function changes at atmospheric pressure The Gibbs function g0for each species can
be calculated using Eq.(42) 2.3.4 Modeling reduction chemistry in gasifier bed Reduction of the oxidation zone products are primarily domi-nated by heterogeneous reactions of solid–char (R6)–(R8) and homogeneous reactions of gas–gas (R9) in complete absence of
Table 5
Proportion of cellulose, hemicellulose and lignin in hardwood [21]
Type of wood Cellulose (Y cl ) Hemicellulose (Y hc ) Lignin (Y lg )
Table 6
Equations representing to pyrolysis model.
Rate of devolatilization [22]
dMv ol
dt ¼ k pyr Mvol ¼ 7:0 107ðs 1 Þ expð1560=TÞM DB Yvol (15)
D _ mvol;i ¼ dMvol
dt
i ¼ ðDt res Þ i dmv ol
dt
i
Char yield [20]
Y char,ash-free = Y cl Y char + Y hc f char + Y lgcchar (16)
Empirical mass ratios [20]
Y CO=CO 2 ¼ e1:8447896þ7730:317T þ 5019898
Y ME=CO 2 = 5 10 16 T 5.06
(20) Heat of pyrolysis [20]
Dhopyr¼ hof
DB Y char h o
f
char Y v ol P k¼6 Y k h o
f
k
(21)
Table 7 Fractional char yields from biomass constituents.
Biomass constituents
Fractional char yield
et al [21] Chemical
formula
C 6 H 10 O 5 C 6 H 10 O 5 C 9 H 7.95 O 2.4 (OCH 3 ) 0.92 Grobski
et al [23]
Trang 6oxidants These reduction reactions are inherently slower than the
oxidation reactions by several orders of magnitude, thus,
equilib-rium may not be established in the reduction region At moderately
high temperatures (<800 °C), the equilibrium products may deviate
from reality, thus, kinetic or non-equilibrium models are more
suitable and accurate[28] In the present work, therefore, a steady
state kinetic model for reduction reactions has been employed
fol-lowing[4,6] Kinetic model predicts the un-reacted char and final
gas composition For modeling of reduction chemistry in reduction
zone, following assumptions were made:
1 Reduction reactions are slow reactions, and are treated using
the kinetics of these reactions
2 All char is consumed by the end of reduction zone
3 The average diameter of the ash particle is 5 mm
The reaction rates of global reduction reactions (R6)–(R9) can
be described by the departure of the reactant concentrations from
their equilibrium values and their values of pre-exponential factors
Ajand activation energies Ejfor reactions j = 1 4 are given by
Wang and Kinoshita[7] CRFis the char reactivity factor, which
rep-resents the reactivity of char (or number of active sites on the char
surface) and is a key parameter in simulation of fixed bed
gasifica-tion As char burn-off proceeds, the char size decreases and char
porosity increases, the gas would encounter more active sites
The higher CRF, the process becomes more fast Giltrap et al.[8]
rec-ommended a constant value of 1000 for the char reactivity factor
(CRF) In the present work, the same value of char reactivity factor
has been included in order to account for the active sites present
on char surface (cf.Table 9) The symbol Pkis the partial pressure
of gaseous species k of the reduction zone Keq,jis the equilibrium
constant for reaction j
The net rate of production of the kth species (Rtk) thus can be
evaluated in terms of the above reaction rates: for instance,
RtCO= 2r1+ r2+ r4; RtH2= r2 2r3+ 3r4, etc These Rtk values of
kth species can be used to compute outflow species concentration
for known inflow concentration of each species and volume of each
CV (VCV) as:
3 Solution procedure For fluid flow module, assuming suitable guess of biomass con-sumption rate, the airflow rate can be calculated using global mass balance of produced gas, total air, wet biomass and ash For a given input of gas flow rate at gasifier exit and airflow rate, Eq.(3)for the pressure drop through the tuyers and Eq.(4)for pressure drop in gasifier bed are related in terms of air/gas flow rates through each
CV Fluid flow rates through these CVs are also related to consump-tion of solid substrate (e.g dry biomass, moisture in biomass, char and ash) by the intrinsic mass balance for each CV Thus, the sum of pressure drops across the preheating, drying, and pyrolysis zones
in terms of fluid flow rate through them can be related to pressure drop across the tuyers as:
DPpreheatþDPdryþDPpyro¼DPtuy ð26Þ Above Eq.(26)in conjunction with Eqs.(3) and (4), gives ratio of air coming from the open top and through the tuyers This ratio influ-ences the reaction temperature profile in the bed and thus the chemistry of gasification In the second stage, which corresponds
to heat transfer module, here the energy Eq.(6)in conjunction with
simulta-neously using tri diagonal matrix algorithm (TDMA) with known values of heat generation/absorption in different zones When tem-perature specifications in each CV are known, the actual mass con-version and heat released or absorbed in each CV has been obtained using thermochemical phenomena sub-models
For preheat and drying zone, equilibrium mass fraction of mois-ture in wood, Xeqb, in each CV is computed using vapour–liquid equilibrium relationship, while the knowledge of residence time and diffusivity gives Xout, the moisture mass fraction of the biomass leaving the CV is calculated using mass transfer one-dimensional diffusion Eq.(9)in conjunction with Eqs.(10) and (13), the quan-tity of moisture evaporated from the wood particles and heat of vapourization can be quantified The pyrolysis products including char and volatile components are obtained using elemental bal-ances for C, H and O and empirical mass ratios as a function of tem-perature as written by Eqs (18) and (20) inTable 6 Once outlet products is known this gives heat of pyrolysis, which serves input
to heat transfer module
Table 8
Chemical reactions in oxidation zone.
exp(E CO /R u T)[C CO 2 ][C H 2 ] 1.5
1.63 10 9
C 1.16 H 4 +1.58O 2 ? 1.16CO +2H 2 O k ME = A CH 4 exp(E CH 4 /R u T)[C O 2 ] 0.8
[C CH 4 ] 0.7
1.585 10 9
C 6 H6:2 +4 45O 2 ? 6CO + 3.1H 2 O ktarffi kHC= AtarTP0:3
A exp(E tar /R u T)[C O 2 ] 1
[C HC ] 0.5 2.07 10 4
a
C 1.16 H 4 (light hydrocarbon or methane-equivalent).
b
C 6 H6:2o0:2 (heavy hydrocarbon) represents the methane and tar respectively.
Table 9
Reduction reactions, their reaction rates and constants.
R u T
P CO 2 PCO
K eq;1
R u T
P H 2 O PCO PH2
K eq;2
1.517 10 7
121.62
R u T
P 2
H 2 PCH4
K eq;3
r 4 ¼ A 4 exp E 4
R u T
P CH 4 P H 2 O PCO P 3
H2
K eq;4
Trang 7For oxidation zone, using temperature specifications from heat
transfer module, the value of Kp determined in terms of standard
state of Gibbs function change for water gas shift reaction Using
Kp value in Eq.(24)and the atomic balances, the final composition
of gases leaving the oxidation zone can be determined The heat
re-leased in the oxidation zone has been computed from the enthalpy
of formation of the reactants and products Finally, the char
con-sumption and gas composition through the reduction zone can
be obtained solving kinetic rate Eqs (R6)–(R9) for known reaction
temperature profile In reduction zone each CV has been
subdi-vided into 100 subdivisions to ensure adequate accuracy of
ele-mental balances
The equilibrium constants Keq,jfor jth reaction are evaluated at
the temperature of the CV from standard state Gibbs functions of
the gaseous species k, gofrom Eq.(42) The polynomial fits for
stan-dard state enthalpy and entropy used to compute the Gibbs
func-tions as a function of temperature are obtained from NASA fits
on JANAF Tables data[27] Similarly, heat absorption in reduction
zone has been obtained using heats of formation of the reactants
and products The thermo-physical properties of working
sub-stances in terms of temperature are listed inTable 10, the values
of constants used inTable 10are obtained from their respective
references The consumption of char in reduction zone depends
mainly on feedstock composition and equivalence ratio of the
gas-ifier, the temperature of reduction zone The equivalence ratio of
the gasifier was controlled by the airflow rate The ratio of air to
biomass was adjusted so that the char flow rate at gasifier exit
be-comes zero
4 Model predictions and validation
A 20 kWe open top downdraft biomass gasifier developed in
In-dian Institute of Technology, Bangalore has been chosen The
experimental data of Sharma[11], generated on the same
configu-ration has been used in the present work for validation or testing of
various modules and overall gasifier model
4.1 Validation or testing of modules constituting the gasifier model The modules that constitute the gasifier model have been vali-dated against the experimental data or tested for qualitative trends The predictions of fluid flow module for pressure drop in cold flow have been validated against the experimental data of Sharma[11]for given particle size distribution and flow rate at the gasifier exit Since the pressure drop is a strong function of par-ticle size, the two sets of experimental data has been used in the present work; one set for freshly charged gasifier with nearly uni-form sized particles, while second set for extinguished gasifier (bed with decreasing particle size downwards in the direction of gas flow) Simulations are performed: (i) for uniform distribution of particle diameter (ii) for spatially varying particle size distribution,
as given by Eqs.(1) and (2) Results from the simulations are com-pared with those from the experiments inFigs 3 and 4for an initial particle size in the range between 34 and 42 mm The predictions, for same range of particle sizes are in reasonable agreement with measured values of pressure drop for the case of extinguished gas-ifier, while for freshly charged gasgas-ifier, the predictions deviate
Table 10
Property data.
Thermal conductivity
k char = 1.4 10 6 T 2
k mixture ¼
P k
k¼1 v k k k ðmw k Þ 0:333
P k
k¼1 v k ðmw k Þ 0:333
Specific heat
Cp k = a k + b k T + c k T 2
+ d k T 3
+ e k T 4
Cpmixture¼ P k
Viscosity [30] , [32]
lk (T) =lk (T a )(T/T a ) n
(36)
lH2O= 7 10 12
T 2
+ 5.1 10 8
T 6.04 10 6
(37)
lTar lBenzene = 1.3404 10 11 T 2 + 3.5844 10 8 T 2.2588 10 6 (38)
lmixtureðTÞ ¼ P k
k¼1 P I v k l k
I¼1 v k ukI
(39) whereukI ¼ð1þðlk = l I Þ 0:5 ðmw I =mw k Þ 0:25 Þ 2
2:828ð1þðmw k =mw I ÞÞ 0:5
Enthalpy
h 0
f ;mixture ¼ P
k Ykh 0
Heating value [33]
Gibbs function [6]
0 1 2 3 4 5 6 7 8 9 10
Air flow rate (g/s)
Experimental data db=34mm db=42mm
Fig 3 Comparison with experimental data(freshly charged gasifier) for uniformly
Trang 8slightly at higher flow rates This may be due to the fact that the
particles are not perfectly spherical and due to uncertainty
associ-ated with particles (size) constituting the freshly charged bed
The heat transfer module uses the heat released/absorbed in
each zone as the input to predict the temperatures in each zone
Since the heat released/absorbed in an actual gasifier is closely
coupled with all other parameters, it was not possible to validate
the heat transfer part in isolation against experiments Therefore,
well tested model (tested for qualitative trends) of Sharma for heat
transfer[16], has been followed in the present work The drying
model is tested in the preheat zone (of length 1 m) in the gasifier
for the effects of zone temperature and particle diameter for
qual-itative trends as shown byFigs 5 and 6.Fig 5shows the trends for
moisture loss distribution along the testing bed for four isothermal temperatures i.e., 350, 400, 500 and 600 K The results show that as temperature increases, the biomass dries up quickly within the short length along the testing bed, as expected In order to study the effect of particle size on moisture evaporation; five levels of average particle size i.e 10, 20, 30, 40 and 50 mm are considered
in this analysis (Fig 6) Predicted results shows faster biomass dry-ing with decrease in particle diameter, as expected
A well tested pyrolysis sub-model of Sharma et al.[20]is used
to predict the species concentration in volatile matter and char yield at known pyrolysis temperature It uses input of the percent-ages of three major constituents – cellulose, hemicellulose and lig-nin in biomass and fraction of char due to the breakup of each of these three constituents from Tables 5 and 7 For validation of the oxidation module, the oxidation of volatiles alone has been considered The products of oxidation of volatiles predicted by the present model have been compared with equilibrium code of Olikara and Borman as given in Ref.[27], which uses input in the form of CNHMOLNKand equivalence ratioU Volatiles are consid-ered to have the chemical formula of C1.3H3O1.4 Char oxidation has been excluded from the validation part since the code of
Olika-ra and Borman is meant only for those reactions which are ex-pected to reach equilibrium.Fig 7shows the comparison of CO,
H2and CO2contents in the products of oxidation as predicted by the present model with the predictions of the code of Olikara and Borman, for an equivalence ratioU= 1.85 The comparison is found to be quite good These figures also show the variation in the content of these species with the reaction temperature With increase in temperature, the CO content increases while H2 and
CO2 decrease, as expected For reduction environment, a well tested kinetic model for reduction reactions has been used[4,6] 4.2 Validation of gasifier model
After the validation and testing of above modules individually,
it is also essential to validate the overall gasifier model after cou-pling of these modules The gasifier model predicts the pressure drops, biomass consumption rate, airflow rates, gas composition and its calorific value for a given value of producer gas flow rate and size of the feedstock particles being fed from the top For validation, the experimental data of Sharma [11] on the
20 kWe downdraft gasifier has been used at wide range of pro-ducer gas flow rate In his experiments, Sharma used sun dried Kikar wood (Acacia), chopped in cubic shape with average size
36 mm having average moisture content in the range of 11–13%
on dry basis Simulations are also performed for the similar operat-ing condition for gasification of hardwood feedstock However,
0
5
10
15
20
25
Air flow rate (g/s)
Experiments db=34mm db=42mm
Fig 4 Comparison with experimental data (extinguished gasifier) for spatially
varying of particle size distribution of, T bed = 300 K, cold flow.
0
0.02
0.04
0.06
0.08
0.1
0.12
Distance along preheating zone (cm)
T=350K T=400K T=500K T=600K
Fig 5 Effect of drying zone temperature on moisture loss profile, d p = 4 cm.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Distance along preheating zone (cm)
dp=1cm dp=2cm dp=3cm dp=4cm dp=5cm
5 6 7 8 9 10 11 12 13 14
Temperature in oxidation zone (K)
Composition (%vol) CO: Present work CO: ’PER’ model
H 2 : Present work H 2 : ’PER’ model
CO 2 : Present work CO 2 : ’PER’ model
Fig 7 Comparison of predicted product composition of oxidation model with those
Trang 9since in experiments, the particle size even at the gasifier inlet
var-ies considerably, the choice of a constant particle size at gasifier
in-let can have a strong bearing on comparison of simulation with the
experimental values Thus, for comparison with experimental data,
the simulation results have been plotted in particle size range from
36 to 50 mm The pressure drop across the gasifier predicted by the
model for various producer gas flow rates is compared with the
experimental values in Fig 8 This deviation could be owing to
the uncertainty in the particle size of the feedstock It is observed
that the predicted pressure drops are in agreement with the
mea-sured data within the experimental uncertainty Predicted
temper-ature profile in the gasifier bed at gas flow rate of 7 g/s has been
compared with experimental data as shown inFig 9 As expected,
the maximum temperature in oxidation zone predicted by the
model is 1217 K at the gas flow rate of 7 g/s A good agreement
of predicted and measured temperature profile across the bed
can be clearly observed
experimental data of Sharma Predictions for CO and H2percentage
in gas increase gently with gas flow rate The theoretical trends for
CO and H2composition are in good agreement with experimental
measurements of Sharma[11] The calorific value of the gas from
prediction and experiments is compared inFig 11, and good
agree-ment is obtained Since the experiagree-mental data is limited to a little
range of gas flow rate, the predictions have been extended to 30 g/
s, in order to demonstrate the predictable capability of the above
model at higher flow rates Model predicts a very small percentage
of CH4below 0.4%, water vapour varies from 10% to 12% and tar
content was theoretically absent in the resulting gas for the wide range of gas flow
4.3 Gas flow rate Some trends of pressure drop, temperature profile, dry gas com-position and calorific value against gas flow rate have been dis-cussed in previous section (cf Figs 8–11) In this section, the trends of temperature profiles across gasifier bed for different val-ues of gas flow rates; cold gasification efficiency and gasifier power output for wide range of gas flow rate are studied as shown in
0
5
10
15
20
25
30
35
Producer gas flow rate (g/s)
dp=36mm dp=50mm Experiments
Fig 8 Comparison of the predicted pressure drop with experimental data, spatially
varying particle size, hot flow.
250
450
650
850
1050
1250
1450
Distance from open top (cm)
Experiments Predictions
Fig 9 Comparison of predicted temperature profile with experimental data,
5 10 15 20 25 30
Gas flow rate (g/s)
Experiments (H 2 ) Predictions(H 2 )
Fig 10 Comparison of predicted CO and H 2 composition in producer gas with experiments.
1000 1500 2000 2500 3000 3500 4000 4500 5000
Gas flow rate (g/s)
Experiments Predictions
Fig 11 Comparison of predicted calorific value of gas with experiments.
250 450 650 850 1050 1250
Distance from open top (cm)
mpg=6g/s mpg=9g/s mpg=12g/s mpg=17g/s mpg=21g/s
Trang 10Figs 12 and 13 The variations in temperature profiles for five
dif-ferent gas flow rates viz., 6, 9, 12, 17 and 21 g/s have been
com-pared in Fig 12 As expected, the maximum temperatures
(predicted) can be observed in oxidation zone The overall
temper-ature profiles at increasing gas flow rates are found to be
improv-ing A maximum temperature is found to be increasing from
1141 K to 1354 K for typical gas flow rate variation of 6–21 g/s
The gasification efficiency on cold basis can be described in terms
of the ratio of net heating value of gas at ambient (neglecting the
sensible heat) to the input energy intake by biomass feedstock
The heating values of biomass and product gas at the gasifier exit
can be obtained from literature[26,27,31]in terms of heating
val-ues of individual components With these heating valval-ues, the
gas-ification efficiency (cold basis) and gasifier power output can be
computed and results of cold gasification efficiency and gasifier
power output (kW) are plotted inFig 13 The cold gasification
effi-ciency is observed to be increasing from 72% to 74% with gas flow
rate variation from 6 to 25 g/s A steep increase in gasifier power
output (21–92 kW) can be observed (almost linear trend) for above
gas flow rate variation
Increase in gas flow rate improves the temperature profile
lead-ing transformation of the non-combustibles components (i.e CO2,
H2O) into combustibles (i.e CO, H2) and thus improving the
calo-rific value of the product gas, the cold gasification efficiency and
gasifier power output as well However, the temperatures in drying
and pyrolysis zone are lower at higher flow rates, and thus the
pressure drop in these regions may be less at higher flow rate
But in reduction zone, where maximum char conversion takes
place, the particle sizes are the smallest, has higher temperature
at higher gas flow rates This would add significantly to the
pres-sure drop The predicted trends agree with this expected
behaviour
5 Conclusions
A mathematical model EQB for a downdraft biomass gasifier has
been developed to predict the pressure drop, airflow rate from
open top and through the tuyers, biomass consumption,
tempera-ture profile and gas composition for given gas flow rate Model was
developed in three stages: first stage, fluid flow module is carried
out, where isothermal flow of air was considered through the
gas-ifier bed; second stage corresponds to heat transfer module, here
energy equation was solved to obtain the temperatures in each
CV with heat generation/absorption in different zones considered
as known; third stage, the physical and chemical phenomena take
place due to biomass drying, pyrolysis, oxidation and reduction reaction sub-process, and their energetics decide the heat genera-tion or absorpgenera-tion in each CVs The subroutines constituting the gasifier model have been validated or tested The fluid flow module has been validated in cold flow for constant particle size (freshly charged gasifier) as well as for variable (decreasing) particle size distribution in gasifier bed (due to thermochemical conversion) Mass transfer model for biomass drying have been tested in pre-heating zone and found working well for right trends of response
to particle size, rate of drying and prevailing temperature Equilib-rium based oxidation model is validated with the equilibEquilib-rium code
of Olikara and Borman and found to be robust and adequate for prediction of product composition, but predicts a steep tempera-ture rise within a single control volume where oxidation completes itself Finally, the gasifier model was validated against the experi-mental data with good agreement
For the range of gas flow rate encountered in this work, any improvement in the reaction temperature leads to better thermo-chemical transformation of biomass material into combustibles (i.e., CO, H2), thus, improving the gasifier performance in terms
of energy efficiency and power output The rise in gasifier temper-ature due to chemical reactions specially at high gas flow rate also strongly influences the gasifier pressure drop Furthermore, reduc-tion zone is recognized as the most sensitive region for remarkably high pressure drop, where highest char conversion leads to small-est particle sizes and high reaction temperatures as well specially
at higher gas flow rate
Chemical equilibrium for oxidation zone (where reaction tem-peratures proceeds beyond 800 °C establishing equilibrium) and empirically predicted pyrolysis products (volatiles and char) allow-ing faster convergence, while implementallow-ing kinetic modelallow-ing for reduction zone is helpful in restoring the accuracy of predictions (where reaction temperatures less than 800 °C and thus equilib-rium is far away from reality) This combination constitutes an effi-cient algorithm allowing rapid convergence with adequate fidelity When, objective is to couple a gasifier model with a gas engine model for predicting the performance of a gasifier–engine system model, the above algorithm of gasifier simulation may be a prefer-able choice
Acknowledgements Author is grateful to Prof M.R Ravi and Prof S Kohli, Indian Institute of Technology, Delhi for their valuable contribution in car-rying out of mathematical modeling and computational work References
[1] Baratieri M, Baggio P, Fiori L, Grigiante M Biomass as an energy source: thermodynamic constraints on the performance of the conversion process Bioresour Technol 2008;99:7063–73.
[2] Zainal ZA, Ali R, Lean CH, Seetharamu KN Prediction of performance of a downdraft gasifier using equilibrium modeling for different biomass materials Energy Convers Manage 2001;42:1499–515.
[3] Malgar A, Pérez JF, Laget H, Horillo A Thermochemical equilibrium modelling
of a gasifying process Energy Convers Manage 2007;48(1):59–67.
[4] Sharma AK Equilibrium & kinetic modeling of reduction reactions in the char bed of a downdraft (biomass) gasifier: a comparison Int J Solar Energy 2008;82:918–28.
[5] Ratnadhariya JK, Channiwala SA Three zone equilibrium and kinetic free
2009;34:1050–8.
[6] Sharma AK Simulation of a biomass gasifier–engine system PhD thesis, Indian Institute of Technology, Delhi, India; 2006.
[7] Wang Y, Kinoshita CM Kinetic model of biomass gasification Solar Energy 1993;51(1):19–25.
[8] Giltrap DL, McKibbin R, Barnes GRG A steady state model of gas–char reactions
in a downdraft biomass gasifier Solar Energy 2003;74:85–91.
[9] Babu BV, Sheth PN Modeling and simulation of reduction zone of downdraft biomass gasifier: effect of char reactivity factor Energy Convers Manage
65
67
69
71
73
75
77
79
Gas flow rate (g/s)
0 10 20 30 40 50 60 70 80 90 100
Conversion efficiency Gasifier power output
Fig 13 Effect of producer gas flow rate on gasification efficiency and gasifier power
output (kW).