Two dimensional numerical computation of a circulating fluidized bed biomass gasifier

A cross section area of the gasifier m2 Vi∗ the ultimately attainable yield of volatile matter for the gaseous component i kg/kg biomass C gas concentration mol/m3 Cp specific heat J/mol K

Contents lists available atSciVerse ScienceDirect Computers and Chemical Engineering

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 / c o m p c h e m e n g

Two dimensional numerical computation of a circulating fluidized

bed biomass gasifier

Department of Mechanical Engineering, Faculty of Engineering, Akdeniz University, 07058 Antalya, Turkey

a r t i c l e i n f o

Article history:

Received 20 December 2011

Received in revised form 10 July 2012

Accepted 21 September 2012

Available online 3 October 2012

Keywords:

Fluidized bed

Simulation

Biomass

Gasification

a b s t r a c t

A two dimensional model for an atmospheric CFB biomass gasifier has been developed which uses the particle based approach and integrates and simultaneously predicts the hydrodynamic and gasification aspects Tar conversion is taken into account in the model The model calculates the axial and radial distribution of syngas mole fraction and temperature both for bottom and upper zones The proposed model addresses both hydrodynamic parameters and reaction kinetic modeling Results are compared with and validated against experimental data from a pilot scale air blown CFB gasifier which uses different types of biomass fuels given in the literature Developed model efficiently simulates the radial and axial profiles of the bed temperature and H2, CO, CO2 and CH4 volumetric fractions and tar concentration versus gasifier temperature The minimum error of comparisons is about 1% and the maximum error is less than 25%

© 2012 Elsevier Ltd All rights reserved

1 Introduction

In order to have environment friendly hydrogen, it must be

pro-duced by renewable methods A number of ways and a variety

of resources for producing renewable hydrogen are being

inves-tigated Of all the renewable resources, biomass holds the greatest

promise for hydrogen production in the near future (Mahishi

& Goswami, 2007) Bio-chemical and thermo-chemical processes

are used for the recovery of energy from biomass Bio-chemical

process involves bio methanization of biomass Thermo-chemical

processes are combustion, pyrolysis and gasification Gasification is

economical at all capacities from 5 kWe onwards Therefore, there

is a constant and consistent interest in the production of energy

from biomass through gasification (Kirubakaran et al., 2009)

Gasi-fication is a robust proven technology that can be operated either

as a simple, low technology system based on a fixed-bed gasifier,

or as a more sophisticated system using fluidized-bed

technol-ogy (McKendry, 2002) In the past decades, significant efforts have

been directed towards the development of biomass gasifiers to

replace traditional combustion systems (Brown, Gassner, Fuchino,

& Marechal, 2009)

Fluidized-bed gasifiers provide excellent mixing and gas/solid

contact, causing high reaction rates and conversion efficiencies

Further, there is the possibility of addition of catalysts to the bed

material to influence product gas mole fraction and reduce its tar

content (Schuster, Löffler, Weigl, & Hofbauer, 2001)

∗ Corresponding author Tel.: +90 532 397 30 88; fax: +90 242 310 63 06.

The circulating fluidized bed (CFB) is a natural extension of the bubbling bed concept, with cyclones or other separators employed

to capture and recycle solids in order to extend the solids residence time The riser of a CFB gasifier operates in either the turbulent or fast fluidization flow regime CFB gasification is now undergoing rapid commercialization for biomass Fundamental and pilot stud-ies are, nevertheless, required for scale-up, as well as to fill gaps in understanding the underlying principles (Li et al., 2004)

Design and operation of a gasifier requires understanding of the effect of various operational parameters on the performance of the system The simulation of the gasifier can provide a quantita-tive tool for gaining insight into and understanding the integrated process It is very useful for the analysis, evaluation, and design

of the process Researchers have done a lot of work with regard

to modeling of fluidized beds in biomass gasification Hydrody-namics, heat transfer, and reaction kinetics play crucial role on the gasification performance of a CFB biomass gasifier Hydrodynamic models based on the fundamental laws of conservation of mass, momentum, energy, and species conversion have enabled us to give better understanding of the fluidized beds and to be useful

to enhance the process performance (Vejahati, Mahinpey, Ellis,

& Nikoo, 2009) As the computational capacity increased, com-putational fluid dynamics (CFD) had become an advanced tool in modeling hydrodynamics, and it is now considered as a standard tool for the simulation of single-phase flows However, CFD still needs verification and validation for modeling multiphase flow sys-tems such as fluidized beds Further improvements regarding the flow dynamics and computational models may be required to make CFD more suitable for fluidized bed reactor modeling and scale-up (Nguyen, Ngo, et al., 2012; Vejahati et al., 2009)

0098-1354/$ – see front matter © 2012 Elsevier Ltd All rights reserved.

A cross section area of the gasifier (m2)

Vi∗ the ultimately attainable yield of volatile matter for

the gaseous component i (kg/kg biomass)

C gas concentration (mol/m3)

Cp specific heat (J/mol K)

db0 initial bubble size (m)

dbm the limit size of bubble expected in a bed (m)

Ei apparent activation energy for component i (kJ/mol)

hp heat exchange coefficient between particle and the

emulsion phase (J/m2s K)

k0,i per-exponential factor (1/s)

kbe exchange coefficient between bubble and emulsion

phase per unit volume of bubble phase (1/s)

rj reaction rate of j reaction (mol/m3s)

Tp temperature of biomass particle (K)

u0 superficial velocity (m/s)

ub bubble velocity (m/s)

umf minimum fluidization velocity (m/s)

Vi instantaneous yield of volatile matter for the

gaseous component i (kg/kg biomass)

z axial coordinate of the reactor (m)

Greek letters

˛s specific particle surface area (m2/m3)

εb void fraction of the bubble phase

εmf void fraction in the dense phase at minimum

flu-idization conditions

ij stoichiometric coefficient of component i of reaction

j

Subscripts

Dealing with gas–solid hydrodynamics, two different

approaches are generally used to apply CFD modeling to the

gas–solid fluidized beds: (1) Eulerian–Lagrangian model (so-called

Lagrangian model) and (2) Eulerian–Eulerian model (Eulerian

model) (Gungor & Eskin, 2007; Nguyen, Ngo, et al., 2012)

Lagrangian models solve the Newtonian equations of motion for

each individual particle, taking into account the effects of particle

collisions and forces acting on the particle by the gas (Gungor &

Eskin, 2007) The Lagrangian model is normally limited to a

rela-tively small number of particles because of computational expense

(Taghipour, Ellis, & Wong, 2005) Eulerian models consider all

phases to be continuous and fully interpenetrating The equations

employed are a generalization of the Navier–Stokes equations for

interacting continua Regarding the continuum representation of

the particle phases, Eulerian models need additional closure laws

to describe the rheology of particles An extension of the classical

kinetic theory of gases to the dense particle flow is most commonly

used (Reuge et al., 2008) The Eulerian model makes it possible to

be applied to multiphase flow processes containing a large volume fraction of solid particles (Huilin, Yurong, & Gidaspow, 2003) Researchers such asGungor and Eskin (2007),Gungor (2008a) andJiradilok et al (2008)paid attention to modeling and simulation

of the hydrodynamic characteristics of fluidized bed systems They studied particles and gas flow behaviors in the riser section of a CFB using the kinetic theory for the particulate phase

Recently a 2D CFD simulation is carried out to study hydrody-namics of a cold-mode dual fluidized bed gasifier including the riser and gasifier using a commercial CFD code (Fluent Inc., USA) (Nguyen, Ngo, et al., 2012) Experiments were also conducted on a pilot-scale DFB in the cold mode The solid circulation rate and solid holdup obtained from CFD simulation are compared with those measured by experiment In addition, hydrodynamics of the hot mode is predicted at a given temperature profile along the riser and gasifier measured by experiment

Modeling and simulation of biomass gasification may be also divided into three categories: (1) thermodynamic equilibrium models (Pröll & Hofbauer, 2008; Shen, Gao, & Xiao, 2008), (2) kinetic rate models (Corella & Sanz, 2005; Petersen & Werther, 2005), and (3) neural network models (Brown, Fuchino, & Maréchal, 2006)

In the kinetic rate models, initial conditions and kinetic parame-ters are not well known because of a variety of feedstock (Corella

& Sanz, 2005) The neural network models as a kind of black-box models have achieved high prediction accuracy However, it is hard

to obtain physical meaning from these models, and the scale-up and adaption abilities of the neural network models are restricted The kinetic models predict the progress of product composition with respect to the residence time in a gasifier, whereas the equilibrium models provide the maximum yield of a desired product which

is achievable from a gasification system (Li et al., 2004) Although kinetic rate models are considered as a rigorous approach, equilib-rium models are valuable because they can predict thermodynamic limits which are used to design, evaluate and improve the pro-cess (Karmakar & Datta, 2010) The equilibrium models have been used for preliminary study on the influence of the most important process parameters

In their review studyGomez-Barea and Leckner (2010)stated that,Sanz and Corella (2006)have presented a whole model for CFB biomass gasifiers Such model is 1D and for steady state The model has a semirigorous character because of the assumptions that had

to be introduced by lack of accurate knowledge in some parts of the modeling It must be noted that most common fluidization models for fluidized bed gasifier are 1D models, but 3D models (Petersen

& Werther, 2005) also fit into this category Therefore, no matter

if the fluidization model is formulated in one, two or three dimen-sions, it still needs input from fluid-dynamic knowledge computed

by ‘external’ correlations CFD for fluidized bed gasifiers are rela-tively new, and in spite of offering promising expectation, much has to be added Because of the considerable computational times required for CFD computations, especially when chemical reac-tions are involved, fluidization models are still the most common approach (Gomez-Barea & Leckner, 2010)

Ngo et al (2011) investigate the biomass gasification with the steam agent in a bench-scale CFB gasifier and develop a quasi-equilibrium three-stage gasification (qETG) model for the prediction of process performance in dual CFB The qETG model

is divided into three main stages: (1) pyrolysis of volatiles in biomass, (2) solid–gas reactions between biomass char and gasi-fying reagents (carbon dioxide or steam) in the fluidized bed, and (3) gas-phase reactions among the gaseous species in the free board of the gasifier At each stage, empirical models are estab-lished based on the experimental data to calculate the gaseous components Especially, the deviation from equilibrium reaction

is taken into account in the third stage by a non-equilibrium fac-tor The model is first validated by the experiment data conducted

in the bench-scale CFB gasifier with pine woodchips, and the data

taken from the literature The effects of gasification temperature

and steam to fuel ratio on product gas composition and yield

were also experimentally investigated for steam gasification of

pine woodchips in a bench-scale CFB gasifier with external heat

supplier

In order to avoid complex processes and develop the simplest

possible model that incorporates the principal gasification

reac-tions and the gross physical characteristics of the reactor, have

developed models using the process Simulator Aspen Plus Aspen

Plus is a problem-oriented input program that is used to facilitate

the calculation of physical, chemical and biological processes If

more sophisticated block abilities are required, they can be

devel-oped as FORTRAN subroutines (Arnavat, Bruno, & Coronas, 2010)

Recently,Ramzan, Ashraf, Naveed, and Malik (2011)have

devel-oped a steady state simulation model for gasification using Aspen

Plus The model can be used as a predictive tool for optimization

of the gasifier performance The gasifier has been modeled in three

stages In first stage moisture content of biomass feed is reduced

In second stage biomass is decomposed into its elements by

spec-ifying yield distribution In third stage gasification reactions have

been modeled using Gibbs free energy minimization approach In

the simulation study, the effect of the operating parameters like

temperature, equivalence ratio (ER), biomass moisture content and

steam injection on syngas composition, high heating value (HHV)

and cold gas efficiency has been investigated

The most recent study was conducted byNguyen, Ngo, et al

(2012)and it concluded a three-stage steady state model

devel-oped for biomass steam gasification in a dual CFB to calculate the

composition of producer gas, carbon conversion, heat recovery, cost

efficiency, and heat demand needed for the endothermic

gasifica-tion reacgasifica-tions The model was divided into three stages including

biomass pyrolysis, char–gas reactions, and gas-phase reaction At

each stage, an empirical equation was estimated from

experimen-tal data to calculate carbon conversion and gaseous components

The parametric study of the gasification temperature and the steam

to fuel ratio was then carried out to evaluate performance criteria

of a 1.8 MW DFB gasifier using woodchips as a feedstock for the

electric power generation (Nguyen, Ngo, et al., 2012)

Tar conversion is usually not modeled at all or modeled as one

or two lumped species reacting by oxidation, thermal cracking, or

reforming with H2O Fields for further research have been identified

as devolatilization and conversion of tar and char are recognized as

the processes that require major modeling efforts Formulation of

a model naturally occupies the main attention of a modeler

How-ever, validation is a necessary step before a model can be safely

applied Unfortunately, very poor detailed experimental data are

available on this argument, and then it is difficult to verify the

general validity of the proposed mathematical models (Cao, Wang,

Riley, & Pan, 2006; Pfeifer, Rauch, & Hofbauer, 2004)

The performance of CFB biomass gasifier may greatly depend on

the movement of solids and gas in the riser Modeling of solids and

gas mixing can identify the best arrangement for design and

oper-ation of the gasifier (Gomez-Barea & Leckner, 2010) The primary

goal of this study has been to improve previous biomass

gasifica-tion kinetics and hydrodynamic models for CFB biomass gasifiers

In this study, a two dimensional model for an atmospheric CFB

biomass gasifier has been developed which uses the particle based

approach and integrates and simultaneously predicts the

hydro-dynamic and gasification aspects Tar conversion (tar formation

and thermal tar cracking) is taken into account in the model The

model calculates the axial and radial distribution of syngas mole

fraction and temperature both for bottom and upper zones The

proposed model addresses both hydrodynamic parameters and

reaction kinetic modeling The model results are compared with

and validated against experimental data from a pilot scale air blown

CFB gasifier which uses different types of biomass fuels given in the literature

2 Model

The two-phase fluid dynamics is of great importance for the design and operation of the CFBs Because of containing com-plex gas–solid flow and gas-phase reactions, modeling of CFBs is rather difficult The fluid dynamics of this gas–solid two-phase flow is very complex and strongly dominated by particle-to-particle interactions Furthermore, the numerous homogeneous and het-erogeneous catalytic gas-phase reactions and their kinetics for the description of the gasification phenomena and the tar formation and destruction are not completely known The present CFB model can be divided into three major parts: a submodel of the gas–solid flow structure; a reaction kinetic model for gasification; and a con-vection/dispersion model with reaction

2.1 Hydrodynamic structure

In the present study, gasifier hydrodynamic is modeled taking into account previous work (Gungor, 2008a) The model addressed

in this paper uses a particle-based approach that considers 2D motion of single particles through fluids According to the axial solid volume concentration profile, the riser is axially divided into the bottom zone and the upper zone

Most of the models in the literature do not completely take account of the performance of the bottom zone, consider the bot-tom zone as well-mixed distributed flow with constant voidage, and use generally lumped formulation (Gungor, 2008a) In lumped formulation, the contribution of the net flow has no meaning, since solids and gas are lumped into a single component, so no distinc-tion is made between the gas and solids However, in CFB biomass gasifier, the reacting gas environment in the wall and core has been found to be different (Gomez-Barea & Leckner, 2010; Li et al., 2004) Another point of consideration is that the particle size distribu-tions in the wall layer and the dilute zone of the transport zone are known to be different So, an extension by consideration of two phases in the freeboard (instead of lumping the gas and solid as

in the model developed above), formulated with an explicit dis-tinction between gas and solids and with some exchange of gas, could be necessary In addition, different particle size distributions

in the wall and core zones might need to be accounted for ( Gomez-Barea & Leckner, 2010) From this point of view, in this study, the bottom zone is modeled in detail as two-phase flow that is sub-divided into a solid-free bubble phase and a solid-laden emulsion phase A single-phase back-flow cell model is used to represent the solid mixing in the bottom zone A two-phase model is used for gas phase material balance In the upper zone core-annulus solids flow structure is established It is assumed that the particles move upward axially and move from core to the annulus region radially Thickness of the annulus varies according to the bed height In the annulus region, the particle has a zero normal velocity The pressure drop through the bottom zone is equal to the weight of the solids

in this region and is considered only in the axial direction In the upper zone, pressure drop due to the hydrodynamic head of solids is considered in the axial direction while pressure drop due to solids acceleration is also considered in the axial and radial directions The solids friction and gas friction components of pressure drop are considered as boundary conditions in momentum equations for solid and gas phases, respectively in the model Solids friction

is defined as the frictional force between the solids and the wall, whereas the gas friction is the frictional force between the gas and the wall The hydrodynamic model takes into account the axial and radial distribution of voidage and velocity, for gas and solid phase,

Table 1

Hydrodynamic parameters of the bubbling fluidized bed.

¯



⎧

⎨

⎩

H

0.73

2

0.0791

g

v

pressure drop for gas phase, and solids volume fraction and particle

size distribution for solid phase The hydrodynamic parameters of

the CFB are listed inTable 1 The conservation of mass and

momen-tum equations and the constitutive relations used in hydrodynamic

model are given inTable 2 Further details on the model are given

elsewhere (Gungor, 2008a)

2.2 Kinetic model

The overall process of biomass gasification in the bubbling

flu-idized bed can be divided into four steps The first step is drying,

where the moisture of biomass evaporates The second step where

volatile components in biomass evaporate is called

devolatiliza-tion In the model, volatiles are entering the gasifier with the fed

biomass particles It is assumed that the volatiles are released along

the riser at a rate proportional to the solid mixing rate The degree

of devolatilization and its rate increase with increasing

temper-ature (Li & Suzuki, 2009) This is followed by pyrolysis, the step

where the major part of the carbon content of biomass is converted

into gaseous compounds Biomass pyrolysis generates three

dif-ferent products in difdif-ferent quantities: gas, tar, and char In the

kinetic model it is assumed that the biomass decomposed directly

to each product i by a single independent reaction pathway (Ji,

Feng, & Chen, 2009; Radmanesh, Chaouki, & Guy, 2006) The rate of

formation of a product i in yield Viat time t is given by

dVi

where k0,i and Ei are the pre-exponential factor and the

appar-ent activation energy for componappar-ent i, respectively The quantity

Vi∗is the ultimately attainable yield of component i.Table 3lists

the parameter values for each species The parameters are adopted

from the literature (Ji et al., 2009)

As mentioned above, how to prevent the tar formation will be

the key for the biomass gasification Tar is defined, according to

the International Energy Agency’s tar protocol, as organic

compo-nents/contaminants with molecular weight greater than benzene

The chemical formula for tar is CHO The parameters (x; y) are

temperature and heating rate dependent In this work, phenol is used to represent the tar from primary pyrolysis as discussed in the paper byGerun et al (2008) It is well-known that the in-bed additives or catalysts deeply affect the kinetics of the tar elimina-tion On the other hand, the thermal cracking of tar, also called secondary pyrolysis, has a significant effect on the final gas mole fraction, because more than half of the primary pyrolysis products accounts for tar The reactions and the reaction kinetics for the tar cracking are presented inTable 4(Ji et al., 2009)

In the last step, the char (char = 1− total devolatilization) is partly gasified with steam and converted into gaseous products The amount of unreacted char is a function of gasification con-ditions, such as temperature and biomass particle residence time

in the gasifier All homogeneous and heterogeneous reactions and their reaction rates using in the model are given inTable 5

2.3 Particle based approach The importance of particle based approach is clearly explained

by Sommariva, Grana, Maffei, Pierucci, and Ranzi (2011) They stated that the selection of particle size used needs a particular attention, due to the variability of product yields depending on par-ticle size These differences could be attributed mainly at a different biomass composition, even if also intra-particle resistances, which strong depend on particle shape, could play a definite role

It is a well-known fact that small particle size biomass signif-icantly increases the overall energy efficiency of the gasification process, but it also has a negative effect on the gasification plant cost It has been estimated that for a 5–10 MWe gasification plant, about 10% of the output energy is required for the biomass particle size reduction On the other hand, an increase in biomass particle size reduces the pre-treatment costs, but the devolatilization time increases, and thus the gasifier size increases (Mahishi & Goswami,

2007)

The non-uniformity of the biomass particles will influence gasi-fication reaction rate However, due to intense mixing caused by the fluidized sand, temperature longitudinally does not vary much

εi

in

˙n j

εi

out

˙n j

ε i

˙ R g,j

˙ J g,j

εp,i

in

˙m j

ε p,i

out

˙mj

εp,i

˙ R s,j

˙ J s,j

εi

εi

vp,i

vp,i

εp,i

εp,i

εi

εi

vp,i

vp,i

εp,i

εp,i

2 

 ∂u

2 

 ∂

2 

 ∂u

1 u

2 

 ∂

 ∂u

 ∂

CD

24 p

3 CD

εi

εi

1 |u−p

CD

cv

cv

cv

c p

c p

c p

 ˙ Q − wall

 ∂u

2 +

 ∂u

2

 +

 ∂u

2

 ∂

2 +

 ∂

2

 +

 ∂

2

Table 3

biomass)

and are almost similar, indicating that the irregular shapes and size of biomass particles do not effect the temperature (Alauddin, Lahijani, Mohammadi, & Mohamed, 2010) On the other hand,Lv

et al (2004)observed that the producer gas yield, LHV and carbon conversion were improved as the biomass particle size decreased

It was explained that small biomass particles contribute to large surface area and high heating rate which in turn produce more light gases and less char and condensate Therefore, the yield and composition of the producer gas improved while using the small particle biomass Yet another explanation is that for small particle sizes the pyrolysis process is mainly controlled by reaction kinet-ics; as the particle size increases, the product gas resultant inside the particle is more difficult to diffuse out and the process is mainly controlled by gas diffusion (Chaiprasert & Vitidsant, 2009) Similar results were obtained byJand and Foscolo (2005)who studied the effect of wood particle size (5–20 mm) in a FB

Since the particle size distribution is known to have a strong influence on the hydrodynamics and gasification performance, the model also considers the particle size distribution and the attrition phenomena Particles in the bottom zone include particles coming from the solid feed and recirculated particles from the separator Particles in the model are divided into n size groups in the model and mean particle diameter of different-sized particles considers

as follows:

dp=n 1

In the fluidized beds, particle attrition takes place by surface abrasion, i.e particles of a much smaller size break away from the original particle The upper limit size of the fines produced is in the range 50–100␮m (Wang, Luo, Li, Fang, & Ni Cen, 1999) The attrition rate for the bottom zone is calculated as follows (Wang

et al., 1999):

Ra= ka(U0− Umf)Wb

dp

(3) For the upper zone, attrition rate is defined in terms of gas and solid velocities:

Ra= ka(u−v)Wb

where kais the attrition constant and is obtained varying in the range 2–7× 10−7with a superficial gas velocity of 4–6 m/s and a

circulating solids mass flux from 100 to 200 kg/m2s (Wang, Luo,

Ni, & Cen, 2003) In the model, the attrition constant value is taken

as 3× 10−7for the biomass particles in the model calculations in

both bottom zone and upper zone (Scala & Chirone, 2006) In the model, the attrition constant value is taken as 1.9× 10−7 for the

attrition constant of the inert bed particles (Gungor, 2008a) Weight fraction of particles after attrition is considered as fol-lows:

xa=ka(ud−v)

pi

(5)

Table 4

RT



RT



RT



RT



2 C 0.2

T



RT



6 H6C 1.85

T



In the model, particles are considered as spherical Particles are

discretized into 10 groups totally The particle size distribution

depends on attrition in the bed As mentioned above the Sauter

mean diameter is adopted as average particle size (Eq.(2))

3 Numerical solution

The model allows dividing the calculation domain into m× n

control volumes, in the radial and the axial directions and in the

core and the annulus regions respectively In this study the

calcu-lation domain is divided into 8× 50 control volumes in the radial

and the axial directions and in the core and the annulus regions

respectively With the cylindrical system of coordinates, a

symme-try boundary condition is assumed at the column axis At the walls,

a partial slip condition is assumed for the solid and the gas phases

Tsuo and Gidaspow (1990)had successfully applied the two-fluid

model with effective solid viscosity based on a solid stress

modu-lus to describe core annular flow behavior in a riser For two-phase

flow, two friction coefficients are obtained, one for the gas and one

for the solid Modified Hagen–Poiseuille expression is used for wall

friction factor of gas phase and Konno’s correlation is used for wall

friction factor of solid phase in the model (Table 1) (Gungor, 2008b;

Huang, Turton, Park, Famouri, & Boyle, 2006) The temperature has

been evaluated by a thermal balance along each of the control

vol-umes which the fluidized bed has been divided (Table 1) In the

gasifier, temperatures of product gas, bed material and biomass

particles are assumed to be equal The product gases of biomass

gasification are H2, CO, CO2, CH4, H2O, C10H8, and C6H6; tar is taken

into account as C6H6O in the model Particles are spherical and of

uniform size and the average diameter remain constant during the

gasification, based on the shrinking core model The set of

differen-tial equations governing mass, momentum and energy for the gas

and solid phases are given inTable 2, and are solved with a

com-puter code which is written in FORTRAN and should be modular

to allow users to update component modules easily as new

find-ings become available The combined Relaxation Newton–Raphson

methods are used for solution procedure The backward-difference

methods are used for the discretization of the governing equations

Flow chart of the numerical solution for biomass gasification is

shown inFig 1

The inputs for the model are the dimensions; biomass feed rate

and particle size, biomass properties, air ratio, steam to biomass

ratio, air to biomass ratio, and the superficial velocity The

simula-tion model calculates the axial and radial profiles of product gases,

gasifier temperature and tar concentrations in the gasifier

4 Model validation

The 2D hydrodynamic model presented in a previous paper

(Gungor, 2008a) has been used to predict the hydrodynamic

behav-ior of CFB biomass gasifier Firstly, hydrodynamic model simulation

performance is tested against four published data sets (Abdullah, Husain, & Yin Pong, 2003; Andreux, Petit, Hemati, & Simonin, 2008; Karmakar & Datta, 2010; Lee et al., 2010) with regard to the bed pressure drop and the solid mass flux variation by the opera-tional bed velocity, and the axial pressure drop profile and the solid holdup along the bed height Measurement conditions of the experimental data are given inTable 6

Secondly, developed 2D model of biomass gasification for CFB

is validated in this study The comparison data are obtained from

a pilot scale CFB biomass gasifier, which were published in the lit-erature (Li et al., 2004) To test and validate the model presented

in this paper, the same input variables in the tests are used as the simulation program input in the comparisons

Schematic diagram of pilot scale CFB biomass gasifier is shown in Fig 2 “The gasifier employs a riser of 6.5 m high and 0.10 m in diam-eter, a high-temperature cyclone for solids recycle and ceramic fiber filter unit for gas cleaning Air was supplied as the oxidant and fluidizing agent after passing through a start-up burner near the bottom of the riser Hot gas leaving the burner and pre-heated air were mixed to preheat the bed and, if needed, to maintain the suspension temperature at the desired level The temperatures of both the primary and secondary air could be varied by adjusting the total air supply and the fraction of each stream The start-up burner preheated the gasifier to 400–550◦C before coal or biomass fuel could be fed to the riser to further raise the temperature to the desired level The system was then switched to the gasification model” (Li et al., 2004)

Feed particles underwent moisture evaporation, pyrolysis and char gasification primarily in the riser The fast fluidization flow regime was maintained at the operating temperature, with a typ-ical superficial velocity between 4 and 10 m/s, corresponding to

an air flow of 40–65 N m3/h, and solids feed rate of 16–45 kg/h for typical sawdust The solids throughput was estimated to be 0.7–2.0 kg/m2s Coarser particles in the gas were captured by a high-temperature cyclone immediately downstream of the riser The solids captured in the cyclone were recycled to the bottom of the riser through an air-driven loop seal Hot gas leaving the cyclone

at a temperature of 600–800◦C was cooled by a two-stage water-jacketed heat exchanger and a single-stage air preheater before entering the filter unit (Li et al., 2004)

Comparison data are obtained from gasification test results of four sawdust species of whose ultimate analyses and other relevant properties are given inTable 7 Each sawdust was dried before being charged to the hoppers Bed ash collected from a previous run was used as the starting bed material for each new run, with silica sand making up for loss of solids In some runs, fly ash collected from the outlet product stream was pneumatically re-injected into the bot-tom of the riser The air used for re-injecting fly ash was included when calculating the air ratio The carbon content of the bed mate-rials and re-injected fly ash was accounted for in the overall mass and energy balance

1 O2

 2 −

 CO

 2

 CO

2 kCc

O2

kcr

 −1

Ru

 (kg

2 s

kc

Ru

/kcd

kcd

·Dg

dp

·Rg

2 s

kg

dp

Dg

Re

p ε

1

O2

2 s)

k1

pH

pH

k3

pH

k1

3 exp

 3548−

9 exp

H2

3 s)

O2

H2

3 s)

3 s)

6 exp

 1510−

CCO

CH

CCO

CH

KEQ

 (mol/m

3 s)

KEQ

 3968−

CH

3 s)

The operating pressure in the system was maintained at

∼1:05 bar, slightly higher than atmospheric The air ratio, a defined

as the ratio of the actual air supply to the stoichiometric air required for complete combustion, is one such measure The tar yield is expressed as the mass of tar per unit volume of raw gas,

in g/N m3 The operating temperature was maintained in the range 700–850◦C, while the sawdust feed rate varied from 16 to 45 kg/h

It must be noted that, the CFB used in the experiments men-tioned above is small-scale pilot unit A more detailed description

of the experiment is given in the literature (Li et al., 2004) The considered parameters and computation conditions are given in Table 8

Finally, a sensitivity analysis is carried out by using two pub-lished data sets from the literature (Li et al., 2004; Yin, Wu, Zheng,

& Chen, 2002) Schematic diagram of pilot scale CFB biomass gasi-fier which was used inYin et al.’s (2002)experiments is shown in Fig 3 The proximate and ultimate analyses of biomass fuels used

in experiments are given inTable 7

5 Results and discussion

As for the hydrodynamic aspect of results derived from this study, the simulation results could be listed as follows In this study, the hydrodynamic model simulation results of the bed pressure drop and the solid mass flux variation by the operational bed veloc-ity, and the axial pressure drop profile and the solid holdup along the bed height are tested against four published data sets (Abdullah

et al., 2003; Andreux et al., 2008; Karmakar & Datta, 2010; Lee et al.,

2010)

The axial solid holdup distribution along the riser obtained from the hydrodynamic model simulation results for two different solid circulation rates is presented in comparison withLee et al.’s (2010) experimental data inFig 4 It must be noted that, in the hydro-dynamic model used in this study, for the axial profile of the solid fraction along the upper zone,Zenz and Weil’s (1958)expression which was further confirmed byWein (1992)has been used as given inTable 1 In that equation, the decay coefficient, ˛, which

is a parameter to express the exponential decrease of solid flux

or solid fraction with height is taken into account as described by Chen and Xiaolong (2006) To calculate the cross-sectional aver-age solids concentration, the relationship suggestedRhodes, Wang, Cheng, and Hirama (1992)is used in the model as given inTable 1

It is observed that the solid fraction is high at the bottom zone and is low at the upper zone, due to the particle accumulation

in the bottom zone of the riser during operation, which is also reported in the literature (Goo et al., 2008; Jiradilok et al., 2008; Nguyen, Seo, Lima, Song, & Kim, 2012b) The solid holdup along the riser is related to the stability of the solid circulation An increase

in the solid circulation rate results in an increase of axial solid holdup distribution along the riser, as seen in Fig 4 Since the gas flow is not sufficient to entrain all the solids entering into the riser at a high solid circulation rate, the solid particles begin to accumulate at the bottom zone of the riser which forms a dense phase The higher the solid circulation rate the larger the accumu-lation amount of the particles at the bottom zone of the riser An increase in the axial solid holdup along the riser is observed, as the solid circulation rate increases (Lee et al., 2010) As the fig-ure shows, the hydrodynamic model predicts reasonably well the axial solid holdup distribution for two different solid circulation rates

Fig 5shows the predicted and experimental values of the axial pressure drop for FCC particles for conditions ofTable 6 Gener-ally, the change in the pressure gradient with height in CFB riser is small In the riser, the pressure gradient is always negative because the gas phase losses pressure head to accelerate and to suspend

Fig 1 Flow chart for the numerical solution of the CFB biomass gasifier model.

Table 6

Measurement conditions of the experimental data referred to in this study.

type

Bed tem-perature T

Bed diameter D (m)

Bed height

H (m)

Superficial

(m/s)

Particle

(␮m)

Particle density 

Karmakar

and

Datta

(2010)

Table 7

Ultimate analysis of biomass fuels.

the particles The absolute values of the pressure gradient decrease

steadily with increasing distance from the riser entrance and then

gradually approach a constant value as clearly shown inFig 5 In the

model, calculation of total pressure drop also considers the

pres-sure drop due to distributor plate at the primary gas entrance in the

bottom zone The basic assumption is that the hydrostatic head of

solids contributes to the axial pressure drop The suspension

den-sity is related to the pressure drop through the axial distance which

also shows coherence with the above figure The high pressure drop

at the bottom zone is due to the effect of solid feeding in that zone

as clearly seen fromFig 5 The pressure drop then decreases along

the height of the riser due to the decrease in solid concentration

The model results are in fair agreement with experimental data of

Fig 5 Similar results are also observed in the studies ofNguyen, Ngo, et al (2012)andKarmakar and Datta (2010)

In the CFB gasifier, the solid circulation of the hot bed-materials plays a critical role, since the heat carried by the solid material heated from the combustor is supplied to the gasification endother-mic reactions For the given system, an increase in the solid circulation rate will reduce difference of temperatures between the gasification and combustion zones On the other hand, a higher solid flux between the riser and gasifier can convey more unre-acted char from the gasification to combustion zone which reduces the required amount of additional fuel (Kaiser, Löffler, Bosch, & Hofbauer, 2003) The solid circulation rate is affected by several operating parameters such as gas velocities to the loop-seal, the

Fig 3 Schematic of 1 MW rice husk gasification and power generation system (Yin et al., 2002 ).

Solid holdup (-)

0.0

0.4

0.8

1.2

1.6

2.0

U =3.06 m /s Experiment (G =30.96 kg/m s) Experiment (G =46.62 kg/m s) Model

Fig 4 Comparison of model solid holdup predictions withLee et al.’s (2010)

exper-imental data for different solid circulation flux values.

riser, and the gasifier However, it is controlled mainly by the

gas velocities in the riser (Goo et al., 2008) The effect of the bed

operational velocity on the solid mass flux is presented inFig 6

which also plotsKarmakar and Datta’s (2010)experimental results

The measurement conditions of experimental data used for the

comparison are shown inTable 6 The gas introduced into the

gasi-fier provides momentum to the upward transportation of the solid

particles So, an increase in bed operational velocity leads to the

increase of the solid flux across the gasifier which as a result leads

to an increase in the solid circulation rate As the figures display,

numerical results are in good agreement with experiments, both

Pressure drop (Pa/m) 0

1 2 3 4 5 6 7 8 9

Experiment Model

Fig 5 Comparison of model simulation results withAndreux et al.’s (2008) exper-imental data.

in form and magnitude where the maximum error values do not exceed 0.06 The hydrodynamic behavior is also confirmed in an experimental study reported byGoo et al (2008)andNguyen, Seo,

et al (2012)

Table 8

Operating parameters of the experimental data referred to in this study.

bark–spruce

mixture

Mixed