Modelling of fluidized bed gasification processes

Modelling of fluidized bed gasification processes

12 Modelling of fluidized bed gasification processes A. G ó m e z - B a r e a, University of Seville, Spain DOI: 10.1533/9780857098801.2.579 Abstract...

1MB Sizes 2 Downloads 425 Views


Modelling of fluidized bed gasification processes A. G ó m e z - B a r e a, University of Seville, Spain DOI: 10.1533/9780857098801.2.579 Abstract: The modelling of fluid bed gasifier (FBG) units is the subject of this chapter. Approaches applied for reactor modelling, from equilibrium models to computational fluid-dynamic models, are described. The difference between gasification of biomass and coal is emphasized, underlining the impact of the fuels on the design and operation of the gasifier. The conversion of single fuel particles, char, and gas is examined to understand how sub-models are implemented in the conservation equations. Various types of commercial biomass and waste FBG units are simulated with the models presented to show their ability in predicting the performance of the gasifier. Key words: gasification; fluidized bed, model, coal, biomass.



Gasification is an important route for conversion of solid fuel into useful products. Gasification in fluidized beds offers advantages, since fluidized beds are capable of being scaled up to medium and large scale, overcoming limitations found in smaller, fixed-bed designs (Highman and van der Burgt, 2008). Often, the bed temperature is limited to prevent agglomeration and solid conversion is limited in stationary fluidized beds (non-circulating units). Also, if the temperature is not high enough in the gasifier, the presence of tar in the product gas makes downstream gas cleaning necessary, penalizing the process efficiency. When gasifying biomass, requiring a relatively clean gas, tar removal can become a technical problem (Gómez-Barea and Leckner, 2009b). Models can be helpful for optimization of gasifier design and operation with different fuels. The modelling may be undertaken with different aims: the field of interest ranges from preliminary design of an industrial process to complex simulation of a unit. The tools available for modelling of FB gasifier (FBG) reactors are the more or less simplified equations for conservation of mass, energy and momentum, complemented by boundary conditions, constitutive relationships, and terms expressing the sources and sinks of the system. To determine the latter, rate laws for the conversion processes are needed, chemical as well as physical. Thermodynamic data are 579 © Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

useful for estimation of properties and thermal data as well as of reaction products by equilibrium assumptions. A few reviews have been devoted to survey gasification in FB on the reactor level for coal (Gururajan et al., 1992; Moreea-Taha, 2000) and biomass and waste (Buekens and Schoeters, 1985; Nemtsov and Zabaniotou, 2008). A great deal of review work is also available on modelling of FB combustion (see Chapter 11), sharing many aspects with FB gasification, for instance, in fluid-dynamics, devolatilization, oxidation of volatiles, and in char conversion and comminution processes. There are differences, though, such as in the mode of conversion of the char particles and in the amount of heat transferred to surfaces.

12.1.1 Gasification of solid fuels The gasifier design and operation depend greatly on the fuel properties. In the following, distinction is made between biomass gasification and coal to understand the main differences between the two extreme types of fuel (Gómez-Barea and Leckner, 2009b). Biomass conversion is similar to that of coal in the sense that biomass can be regarded as a young coal. However, there are also significant differences between biomass and coal, as well as between one biomass and another. Compared to coal, biomass fuels have a higher content of oxygen and volatiles, and the nature of the ash differs substantially from that of coal ash. Fuels with high volatile content like biomass (with up to 80% on dry and ash-free fuel (daf)) are characterized by a rapid conversion into a gaseous product. This can be compared with coal, whose volatile content ranges from 5% for anthracite to about 40% for low-rank sub-bituminous coal and lignite. The remainder after devolatilization is char. Biomass fuels produce small amounts of char (10–20% of the mass of the original daf fuel). This char is more porous and reactive and easier to gasify than coal, which, on the other hand, can produde char yields up to 90% of the mass of the original daf fuel. The moisture content of biomass can be much higher than that of coal, and the release of moisture leads to an even higher fraction of volatile matter with respect to the solid fraction remaining. High moisture content of the fuel lowers the temperature of a gasifier, setting an upper limit of moisture content for satisfactory operation. The moisture contained in biomass together with the lower heating value (20 MJ/kg daf compared with 33 MJ/ kg for coal) related to the higher oxygen content of biomass fuels, affects the gasification process in lowering the heating value of the product gas. Therefore, the throughput of a biomass gasifier has to be increased in order to produce a gas with a power output that is equivalent to that produced by coal. The chemical composition of the ash affects the melting behaviour of the

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


ash and the gasification rate. Biomass ash may have a comparatively low melting point compared to coal. The inorganic constituents are critical for the tendency for fouling and slagging. Alkali and alkaline earth metals, in combination with other fuel elements, such as silica and sulfur, facilitated by the presence of chlorine, are responsible for many undesirable reactions in gasification and combustion. In FBG in contrast to entrained-flow gasifiers, a large portion of the bed material is ash (especially for coal, since biomass ash can be only a small part in biomass FBG) which is normally removed as a solid. For this reason FBG operate at temperatures below the softening point of the ash, which is typically in the range of 950–1100°C for coal and 800–950°C for biomass. This limits the operation of systems with biomass of high-alkali content (i.e. straw) or waste (olive-oil residue). Enhancement of tar yield is an important aspect of biomass gasification, operating at low temperature. Moreover, the nature of the tar is more difficult to convert in biomass compared to coal. These characteristics have a notable effect on gas cleaning and application. Physical properties, such as density and particle size distribution also differ substantially between fuels. They are the key properties for reactor design, including feed system and other ancillary devices. The fibrous character, particularly in vegetable biomass, is an important aspect to take into account as it may have a strong impact on handling. Bulk density differs widely between biomass and coal and between different types of biomass. Other factors make gasification of coal different from biomass, such as the amount of sulphur, nitrogen, chlorine and trace metals. The composition can vary widely but, in general, the content of sulphur is lower in biomass but the content of free alkali metals (free means not bound in the mineral substance of the ash, such as in coal) is higher than in coal.

12.1.2 Types of gasifiers There are three types of gasifier: fixed or moving bed, fluidized bed, and entrained flow. Among these designs there are variations, such as spouted bed, draught tube, internally circulating fluidized bed gasifier, etc. Moving-bed gasifiers (also called fixed-bed gasifiers) are characterized by a bed in which the fuel moves slowly downwards by gravity as it is gasified. The designs are basically updraft (countercurrent) or downdraft (cocurrent). In updraft gasifiers, the fuel bed moves downwards and the gasification agent flows from the bottom upwards (updraft). The outlet temperature is generally low, increasing efficiency and reducing oxygen comsuption. Updraft gasifiers can be used for a variety of biomasses and wastes and are less sensitive to biomass size. It is widely used for coal, especially anthracite and coke since these fuels yield a gas with a limited amount of tar (Highman and van der Burgt, 2008). On the other hand, in cocurrent gasifiers, the fuel and

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

gasification agent flow in the same direction and the gas leaves the reactor near the hottest zone, which makes the tar concentration much lower than in updraft gasifiers. Downdraft gasifiers are rarely used for coal, but are used in small-scale biomass devices for production of distributed electricity (Gómez-Barea and Leckner, 2010). Fluidized bed gasifiers have a number of advantages over fixed beds, especially with regard to mixing, reaction rates, and the possibility of being built in sizes far above those of fixed-bed gasifiers. There are main types according to the gross movement of solids in the reactor: bubbling (BFBG) where solids remains stationary forming a fluidized bed with limited entrainment of particles, and circulating (CFBG) with a net recirculation of solids to the bed. Recycling of fines, a solution that can be applied even if the bed is called bubbling, leads to a greater efficiency of carbon conversion by increasing the conversion time of particles. CFBG is taller and provided with a continuous solids recycling system for re-injection of particles into the bed (particle separator, return leg and seal). It operates with higher superficial velocities, typically in the range of 2–5 m/s, whereas the velocity in the BFBC is in the range of 1–2 m/s. The entrained-flow gasifier with liquid ash removal (slagging gasifiers) converts solid fuels with high efficiency, yielding a gas free from tar, thus removing the two main drawbacks of FBG systems. While it has been used over the years for coal gasification (Highman and van der Burgt, 2008), there are a few drawbacks related to the application with biomass due to the economical disadvantage with particle size reduction of biomass and the inherently small size of equipment caused by transport limitations of the quantities of biomass that can be delivered to a plant (Gómez-Barea and Leckner, 2009b). Gasification can be grouped into two concepts, depending on the way the heat for gasification is provided to the gasifier: autothermal and allothermal gasification. In autothermal or direct gasification, the heat is released by partial oxidation of the fuel in the gasifier itself. The partial oxidation can be carried out using air or oxygen. Steam can also be added to these oxidants. Air gasification produces a low heating value gas (5–8 MJ/Nm3) suitable for nearby boiler, engine or turbine operation. Oxygen gasification produces a medium heating value gas (10–18 MJ/Nm3) as the product is virtually free from nitrogen. Allothermal or indirect gasification uses steam as gasification agent obtaining the heat necessary for gasification from a source outside the gasifier itself. This concept allows generating gas of medium heating value (14–18 MJ/Nm3), rich in hydrogen, without the need for oxygen. Two concepts of indirect gasification are possible, depending on whether the heat is supplied from internal or external sources. In external indirect gasification, the heat is delivered from an external source like in plasma or solar gasification.

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


In indirect internal gasifiers, the energy originates from the process itself, either by recirculation of gas or char. Recirculation of gas is into the gasifier, not an interesting option as this will dilute the product gas with nitrogen. The char-indirect gasifier, consisting of two separate reactors, has been developed as follows: an FB steam devolatilizer produces the syngas and an FB combustor burns the residual char and provides the necessary heat to the gasification reactor. The continuous circulation of solids between the two reactors makes it possible to maintain the process. This type of gasifier is currently offered commercially for biomass (Paisley and Overend 2002; Rauch et al., 2004). A model of this unit is presented in Section 12.5.3.


Qualitative description of the main conversion processes

12.2.1 Fuel conversion Fuel fed to a gasification reactor undergoes a series of conversion processes, listed in Table 12.1 and sketched in Fig. 12.1. The gas fluidizing the bed is, in general, a mixture of steam, oxygen, nitrogen and carbon dioxide. Initially, Table 12.1 Main reactions in the biomass gasification process Stoichiometry

Standard heat Name of reaction (kJ/mol)

Biomass → char + tar + H2O + light Endothermic gas (CO +CO2 + H2 + CH4 + C2 + N2 + ...)

Biomass devolatilization



Char combustion C + ½ O2 → CO C + O2 → CO2

–111 –394

Partial combustion R2 Complete combustion R3

Char gasification C + CO2 → 2CO C + H2O → CO + H2 C + 2H2 → CH4

+173 +131 –75

Boudouard reaction R4 Steam gasification R5 Hydrogen gasification R6

Homogeneous volatile oxidation CO + ½O2 → CO2 H2 + ½O2 → H2O CH4 + 2O2 → CO2 + 2H2O CO + H2O ↔ CO2 + H2

–283 –242 –283 –41

CO oxidation Hydrogen oxidation Methane oxidation Water gas-shift reaction

R7 R8 R9 R10

Partial oxidation Dry reforming Steam reforming Hydrogenation Thermal cracking

R11 R12 R13 R14 R15

Tar reactions (tar assumed CnHm) Endothermic CnHm + (n/2)O2 → nCO + (m/2)H2 (except R11) CnHm + nCO2 → (m/2)H2 + (2n)CO (200–300) CnHm + nH2O → (m/2 + n)H2 + nCO CnHm + (2n – m/2)H2 → nCH4 CnHm → (m/4)CH4 + (n – m/4)C

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

Gasification agent


Primary conversion (generation)

Secondary conversion Tar 1 Tar 2 Light gas 2 Char 2

Light gas 1

Char 1 Heat • Thermal and chemical conversion • Shrinkage • Primary fragmentation

• • • • • • •

Reforming Cracking Oxidation Polymerization Char conversion Secondary fragmentation Attrition

12.1 Scheme of reactions of the primary conversion process during devolatilization. The primary products have been indicated ‘1’, while ‘2’ is used for the secondary products.

the fuel particle is dried, devolatilized (R1), (primary pyrolysis), yielding char and volatiles. Subsequently, volatiles (R7–9) and char (R2–3) may be oxidized, and finally, char may be gasified by carbon dioxide and steam (R4–6). Fuel particles shrink, and primary fragmentation may occur immediately after the injection of the fuel into the bed. Secondary fragmentation and attrition of char take place together with char conversion. After primary decomposition, various gas–gas and gas–solid reactions take place: secondary conversion, during which the tar may oxidize (R11), reform (R12 and R13), and further react by cracking (R15), dealkylation, deoxygenation, aromatization and formation of soot by polymerization. Primary and secondary tar conversion processes can be homogeneous and heterogeneous, occurring inside as well as outside of a particle. The tar conversion can be catalysed by solids added to the bed (dolomite, olivine, etc.) or simply by the carbonaceous surfaces in the devolatilizing particles. Simplification of the process into primary (generation) and secondary conversion (see Fig. 12.1) is usually made in modelling on the basis of the conversion times of the stages mentioned above (Gómez-Barea and Leckner, 2010). The rates of char gasification with H2O and CO2 are orders of magnitude lower than those of the primary pyrolysis (few minutes to gasify a char particle vs. a few seconds to devolatilize a fuel particle). Conversion of volatiles is more rapid than char. Therefore, when pyrolysis takes place in an atmosphere containing steam and O2, O2 is consumed preferentially

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


by combustion of the volatiles (Yan et  al., 1997). Due to the fast release of volatiles, the gasification agent does not significantly penetrate into the particle during devolatilization and the rate and yield of devolatilization are quite insensitive to the composition of the surrounding gas; the bed simply provides the heat for the process. Once the volatiles are emitted from the particle, they mix with the surrounding gas and react by secondary reactions, mainly reforming and oxidation, resulting in secondary gas, char and tar, indicated by ‘2’ in Fig. 12.1. There are, at least, two levels of description in the analysis of fuel conversion in an FB: the particle and the reactor levels. The reactor level includes the bottom bed with its bubble and emulsion phases and the freeboard zone, represented by a core-annulus structure. The particle level contains release of gases, gasification of char and gas phase reactions. These processes are included in the source terms of the conservation equations and are treated by sub-models during execution of numerical calculations.

12.2.2 Fluid-dynamics The conversion of fuel in an FBG is related to the effective time for reaction of the fuel, char and gas, as well as on the local conditions of mixing in the reactor. Formation of bubbles, bypassing of gas, entrainment of material and other factors influence the reaction time (Gómez-Barea and Leckner, 2010). The key operating parameter in an FB is the superficial velocity of the gas, which affects mixing and entrainment. Higher superficial velocity improves solids mixing but increases entrainment of material, decreasing the time of reaction of fuel and char particles in the reactor time (Gómez-Barea and Leckner, 2010). At a given gas velocity, char particles more likely circulate within the bed, while devolatilizing fuel particles may float on the bed’s surface. In Fig. 12.2 there are two limiting cases illustrating the influence of mixing of a devolatilizing particle on the tar content in the outlet gas and the way of conversion of char in the reactor. Figure 12.2(a) corresponds to an under-fed FBG, when the vertical mixing of fuel particles is rapid compared to the time of devolatilization. The devolatilization takes place on the top of the bed, where the fuel particles remain floating. The gas obtained from a case like this has a pyrolytic nature, i.e., its composition is similar to that obtained in pyrolysis tests. At the bottom of the bed, the amount of volatile matter is expected to be small, and hence, the fluidization gas meets the hot char, which is oxidized to some extent, depending on the local temperature. The remaining char is converted by gasification with steam and CO2. In Fig. 12.2(b) the devolatilization is rapid, and most of the volatiles are released in the bottom zone of the bed. The presence of combustible matter, i.e. H2, CO and CH4, at the bottom consumes the oxygen in a short height (Yan et al., 1997, 1999). Then, due

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion D Fuel t = t0



Fuel particle

L t1


t3 t4

Fuel t = t0

Fuel t = t0



(a) High segregation

(b) Low segregation

12.2 Motion of a biomass particle during devolatilization in the dense bed of an FBG. (a) and (b) correspond, respectively, to high and low ratios of the rates of vertical mixing and devolatilization. Times t1, t2 ,... t• represent successive instants during the devolatilization process.

to the absence of O2 in most reactor zones, the char can only be converted by H2O and CO2. A qualitative judgement of the tendency of a particle to be carried away can be made by the terminal velocity of a single particle ut (Corella and Sanz, 2005). In circulating FBG, large devolatilizing fuel particles move towards the top of the riser, increasing the tar yield of the gas as discussed above with regard to Fig. 12.2. In a bubbling bed, in contrast, the fuel particles are likely to remain in the bed (or on its surface) most of the devolatilization time due to lower superficial velocity (lower entrainment). Entrainment of fuel particles during devolatilization does not play an important role in a bubbling unit, but entrainment of finer char particles may be severe. The size of the gasifier may also influence the conversion process. In large commercial devices, characterized by bed aspect ratios of unity or lower (wide beds), the mixing of the volatiles generated around the feed port may not be fast enough and changes of concentration of gas and tar in the horizontal direction are found, affecting the tar and volatile conversion through the bed (Petersen and Werther, 2005). Ascending plumes with highly concentrated pyrolysis gas are likely to be formed, yielding gas with high tar concentration. Maldistribution of volatiles and solids near the port may

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


affect the performance of an FBG and needs to be analysed for scaling-up as we will discuss in Section 12.5.2.


Types of reactor models

A comprehensive model of FBG is based on a description of the gas–solid processes by applying mass, species, heat, and momentum balances, including formulation of the source and sink terms, boundary conditions and constitutive relationships for each phase. Many models have been published with different names, classifications and categories depending on the purpose at the time (Gómez-Barea and Leckner, 2010). The three groups of model sorted according to the simplification adopted to solve the fluid-dynamics are: computational fluid-dynamics models (CFDM), fluidization models (FM) and black-box models (BBM). CFDM solves Eq. [12.1], making certain hypotheses for the interaction between the phases. FM directly assumes that the bed consists of various phases (most often two) or regions with a predefined topology, allowing transport of mass and heat between them. The momentum equation is not solved, and the fluid-dynamic pattern is described by semi-empirical correlations, establishing the dynamics of bubbles and particles in the bed, which give proper closures for chemical reactor modelling. Finally, BBM consist of overall balances over the FBG. In some cases equilibrium is assumed, whilst empirical relationships are used on other occasions.

12.3.1 The general model The closest representation of the real process is a balance on the transported variables formulated and solved for each phase k (gas and solids and their i components): density, velocity, and enthalpy (rk, rk,i, uk, and hk, in general terms, j). The balance of the conserved variables j over a fixed element (Eulerian formulation) of reactor volume can be written in the following form (somewhat simplified, especially in the case of momentum), applicable to any reactor type: the accumulation of j is due to the net difference between the rates of change by convection and dispersion and to generation and consumption, S, per unit volume: ∂j k + div(u k j k ) = div( div(D Dj k grad grad j k ) + Sj, k ∂t


The boundary conditions for mass, species and enthalpy include fuel feed points, gas inlet ports and non-permeability of the reactor walls (Enwald et al., 1996). The solution of a mathematical model of an FBG like that in Eq. [12.1], requires a long time and is numerically complex.

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

12.3.2 Computational fluid-dynamic models The essential distinction between CFDM and other kind of models is that the fluid mechanics are dealt with in detail in CFDM, whereas the source terms are similar in all models. The momentum equations are solved for the gas and the solids phases. The gas phase is described by a continuum approach, adopting a Eulerian framework and modelled similar to singlephase flow with an additional term, accounting for the interaction with the solid phase. The solid phase is described by two distinct approaches: if the solid phase is treated as a continuum, a Eulerian framework is applied to describe the motion of the solids; if the particles are individually tracked, the equation of motion of the particles is used. The two approaches are named Eulerian–Eulerian (EEM) and Eulerian–Lagrangian (ELM) models. In the ELM, the discrete element/particle method (DEM/DPM) is commonly applied, inspired by molecular dynamics. EEM, on the other hand, is solved by the so-called two-fluid model based on the assumption that the gas and particulate phases form two inter-penetrating continua (Enwald et al., 1996). The necessary closure relations are not simple: several equations with semiempirical parameters have to be solved simultaneously. The kinetic theory of granular flow is used in the two-fluid model TFM to simulate particle collision for closure. There are relatively few CFD models applied to FBG compared to the other class of models. Existing work includes CFD models for coal in bubbling (Yu et al., 2007) and circulating (Grabner et al., 2007). In contrast, entrained-flow CFD models of gasification are abundant for coal because the solids flow is more disperse and this application is computationally less expensive. There are also many published models applying CFD to simulate FB boilers burning biomass and wastes. Some efforts have also been made to model pyrolysis in FB by CFD. A recent review on CFD modelling, applied to thermochemical conversion of biofuel, has been published (Wang and Yan, 2008).

12.3.3 Fluidization models Most comprehensive models published in connection with FBG belong to the category FM. The reason is that FM is a compromise between BBM and CFDM: FM avoids the details of complex gas–solid dynamics but still maintains the fluid-dynamic effects by assuming a multiphase pattern in the bed. This is done by introducing two (sometimes three) regions or phases. The flow pattern of the regions is described by semi-empirical correlations. The term ‘phase’ differs from the merely thermodynamic meaning. It is just a region with a predefined configuration (particle concentration, state of gas and solids mixing, etc). Most common FM for FBG are one-dimensional

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


models, but three-dimensional models also fit into this category. Therefore, no matter if the FM is formulated in one, two or three dimensions, it still needs input from fluid-dynamic knowledge computed by ‘external’ correlations. Early models treated FB reactors as if the gas and solids were mixed, avoiding the multiphase nature of the bed. The multiphase description in FB was effectively introduced by the two-phase theory of fluidization, where a division between an emulsion and a bubble phase was assumed based on observations. The momentum balance is not explicitly solved, but Davidson’s model (Davidson and Harrison, 1963) is used together with further semiempirical relationships and further assumptions to estimate the division of gas flow between the phases, the fraction occupied by bubbles in the bed, the porosity and velocity of gas in the dense phase, and the velocity and size of the bubbles (Souza-Santos, 2004; Gómez-Barea and Leckner, 2010). Modifications and simplifications have been introduced over the last decades by applying FM to various reaction systems. Often FM are classified by the names of the authors: Davidson–Harrison (DHM), Kunii–Levenspiel (KLM), but also by the main features of the model, Kato and Wen (bubble assemblage model, BAM), Fryer and Potter (countercurrent backmixing model, CCBMM). Discussion of different approaches applied to FBG appears in Gómez-Barea and Leckner (2010). Originally, FM were formulated and applied to catalytic systems. For non-catalytic systems, the original KL model was applied for FBG of coal (Yoshida and Kunii, 1974), the DHM has been applied up to date in most FB coal (Weimer and Clough, 1981; Saffer et al., 1988; Ma et al., 1988) and biomass (Jiang and Morey, 1992) gasification models. The CCBMM has been applied to describe the movement of char in a bubbling FBBG (Radmanesh et al., 2006). Despite the successful application of FM in the modelling of FBG systems, the pseudo-empirical nature of the description of the fluid-dynamics makes this approach difficult to extrapolate to conditions that differ from those where the correlations were obtained, mostly the type of solid particles, superficial velocity and reactor diameter.

12.3.4 Thermodynamics-based models An equilibrium model, EM, assumes that equilibrium is attained in the outlet stream. Two approaches can be adopted for equilibrium modelling: stoichiometric and non-stoichiometric. The equilibrium formulations can be homogeneous or heterogeneous, the latter if there are solid species in the outlet stream, for instance, solid-phase carbon. Besides equilibrium data for the species considered, the only input needed in EM is the elemental composition of the fuel, which is readily obtained from the ultimate analysis (Desrosiers, 1979). Most EM for biomass gasification systems aim at predicting the composition

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

of the outlet stream, involving the main gaseous species: CO, CO2, H2, H2O, CH4, O2, N2 and temperature (Desrosiers, 1979). The bed temperature can be calculated if the heat loss is estimated, or if the bed temperature is imposed, the heat loss is calculated by the model. The presence of heavy hydrocarbons or tar can be taken into account, represented by compounds such as, e.g., benzene or naphthalene, although this is rarely done in EM. The solid carbon is not usually considered in EM since its concentration is virtually zero if equilibrium is attained. Pseudo-equilibrium models, described below, have been formulated to deal with the solid carbon (Gómez-Barea et al., 2007). Assuming that there is no O2 in the exit gas and that all nitrogen entering is in the form of N2, a simple EM to predict the six main gaseous species (CO, CO2, H2, H2O, CH4, N2) and T can be formulated by considering: the atomic balances for C, H, O; two equations given by the equilibrium of the water-gas-shift reaction (WGSR) and steam reforming of methane (SRMR), and a heat balance (the heat loss is estimated or assumed to be a small fraction of the energy in the input fuel). If the temperature T is assumed, the heat balance is not coupled with the rest of the equations, but it separately serves to calculate the heat loss, and the solution becomes a single non-linear equation. It has been concluded that EM overestimates the yields of H2 and CO, underestimates that of CO2, and predicts an outlet stream free from CH4, tars and char (Jand et al., 2006). An extension is to consider that the methane in the outlet stream is known and equal to that produced during devolatilization, i.e. it is not converted in the bed. Empirical yields of methane (kg methane/ kg of dry fuel) are obtained from devolatilization experiments for each fuel. Exclusion of CH4 from the calculations eliminates one equilibriun restriction in the EM, leading to a simpler model containing six unknowns (CO, CO2, H2, H2O, N2 and T). Another extension can be made by introducing a factor representing the approach-to-equilibriun of the WGSR. The factor is empirically adjusted based on observations: the equilibrium of the WGSR is attained only at high enough temperature, and gas residence time, and especially when there is a catalyst in the bed. The model comprising EM with the two extensions significantly improves predictions of the gas composition of FBG giving reasonably good agreement with experimental results. However, two empirical parameters have to be given, based on the particular system being modelled. A step forward improving the above models is made by applying the procedure schematized in Fig. 12.3, which shows the essential idea behind classical pseudo-EM. It allows solid carbon, methane and tar to be contained in the outlet gas, and the corresponding quantities of carbon and hydrogen are discounted from the input fuel. The remaining fuel elements and the gasification agent react to attain equilibrium. The outlet gas is then obtained by summing the gas components given by the equilibrium and the carbon,

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes



Element decomposition (C, H, O, N)




Chemical equilibrium

CH4 Tar

Gasifier output

Producer gas

12.3 Conversion processes in a gasifier. Basis for pseudo-equilibrium models: implementation accounting for non-equilibrium factors (adapted from Li et al., 2001, 2004).

methane and tar taken off initially. This type of pseudo-EM needs three input parameters: the concentrations of CH4, solid carbon and tar in the outlet stream. It is necessary, therefore, to estimate these species in the reactor. Usually, these estimates are made by experiments in the same plant where the model will be applied, or they are somehow estimated according to experience. This fact makes pseudo-EM quite sensitive to such estimates, and this type of model is less predictive. A variety of sophisticated pseudo-EM have been published, based on the scheme (or some variant) of Fig. 12.3, recently reviewed in Gómez-Barea and Leckner (2010). In general, pseudoEM leads to more or less precise estimates of the gas composition (different from tar) but they do not generally predict the char and tar contents in the outlet gas. For proper prediction of these quantities more advanced models are necessary.

12.3.5 Comparison of models FM is the best developed model to date for FBG, consisting of a comprehensive theoretical treatment, linked with experimental observations made during the last decades. CFD for FBG are relatively new, and in spite of offering much promise, further developments are needed. Finally, models based on thermodynamics (TM), corrected by some empirical relations, i.e. pseudoequilibrium models, allow reasonable predictions for the gas phase (different from char, tars and light hydrocarbons) with minimum input data. For reliable

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

prediction of carbon conversion, FM is necessary. For simulating hardware details, CFDM are required.


Fluidization modelling

An FM consists of conservation equations applied to the various regions of the reactor, simplified by an assumed topology. The flow pattern in each region is established by semi-empirical fluid-dynamic modelling to determine the distribution of gas species and reacting particles, the size and velocity of bubbles, and other relevant variables. A one-dimensional isothermal model for the bed and freeboard is formulated below. Such a formulation describes with sufficient detail the parameters to be estimated and the main equations to be solved. The bed temperature is input, and the temperatures of gas and solids are assumed to be equal. Heat balances could be formulated to give a step forward in the capability of prediction of a gasifier. However, isothermal reactor models are often sufficient for practical prediction since the temperature of gas and solids is nearly uniform in most zones of the FBG. Non-isothermal effects are spatially located in some specific parts of the reactor, like near the feed port, secondary injection of air, etc. However, a one-dimensional model does not work in that case. Further discussion on various non-isothermal models and extensions are discussed elsewhere (Gómez-Barea and Leckner, 2010).

12.4.1 Conservation equations Equation [12.1] is applied for species i in any zone of the reactor. Assuming a steady-state, one-dimensional system and neglecting diffusion, Eq. [12.1] can be simplified as: d (uri ) = Si dh


In the bottom bed (index ‘B’), the mass balances for gas species in the bubble (index ‘b’) and emulsion (index ‘e’) at height h are, according to Eq. [12.2]: 1 dFi,i,b = e r b + f b gg b–e,i b–e gg,i ,i AB dh


1 dFi,i,e = e (1 e e e (1 – e b )rgg,i + S (1 – e e )(1 – e b )s m rgs,m b–e,i gs,m,i gs ,m,i – fb–e ,m,i m AB dh


The variables in Eqs [12.3] and [12.4] are function of height h. The mass flow rates of gas i in the bubble and emulsion are Fi,b = ubri,bAb and Fi,e = ueri,eAe. The cross-section areas in the bubble phase, Ab, and the emulsion, © Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


Ae, are related to the reactor (bottom zone ‘B’) cross section, AB, by: AB = Ab + Ae. Figure 12.4 shows the geometry of a differential volume ABdh in the bottom zone with the main fluid-dynamic parameters, schematically represented as a bubble phase and a particle phase, the latter divided into particles and void. The chemical reaction terms are given by e/b rgg ,i

n e/b rgg

= S R j,iin ii,j MM i




rgse ,m ,m,i ,i = S R j,m,iin ii,j MM i j=1

n rch

= Rpy,fue MM Mi ,f l,iin i,i, py mm i + S R j,char j,j,char,i char,i ,in i,j i, j M



The main homogeneous reactions are the oxidation of volatiles, the watergas-shift reaction and the reforming of hydrocarbons (Reactions R7–R15 in Table 12.1). The main heterogeneous reactions are the devolatilization (R1 in Table 12.1) and char-gas (mainly O2, CO2 and H2O) reactions (Reactions

F0,h + dh

Fb,h + dh

Fe,h + dh





Emulsion dh

Fng,h ee ue

eb ub


Fe,h h h=0


12.4 Definitions of fluid-dynamic parameters in a volume element of the bottom bed.

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

R2–R6 in Table 12.1). The term fb-e,i on the right-hand side of Eqs [12.3] and [12.4] arises from the mass exchange of species i between the bubble and emulsion phases, driven by the difference in concentration of species i and by a net flow (Fng in Fig. 12.4) resulting from the generation of gas by devolatilization and char reactions in the emulsion (all solids are assumed to be in the emulsion) as well as the increase in molar flow due to the homogeneous reactions in the emulsion. Several solids m can be part of the bed, for instance inert bed material, char and catalyst: rm is the volume fraction occupied by solid m (m3 solid m/m3 solids). The m-th solid subjected to gas-solid reaction with the rate rgs,m,i takes into account the generation of the gaseous species i. The mass balance for solids m in the bottom bed, where the solids are perfectly mixed, is Fm, in,B – Fm, out,B out =



(A (1 – e )(1)(1 – e )s B





e Si rgs,m dh gs,m,i ,m,i d


The flow of solids m entering the bottom bed, Fm,in,B consists of the solids feed and the recycling stream, if there is one. The flow of solids leaving the bed, Fm,out,B consists of the drainage (overflow or mechanical removal) and the net flow to the freeboard. In order to calculate the mass fraction of a particle of the m,l class at steady state in the bottom bed, a population balance for each solid m is formulated (losses = gains): Fm,l,in + Gm,gain = Fm,l,out + Gm,l,loss + ¬m,l


The sum of Eq. [12.8] for all sizes l yields Eq. [12.7]. The meaning of each term in Eq. [12.8] can be described as follows: Fm,l,in is the contribution from the feed streams and the recirculation or recycling stream. Fm,l,out is the loss due to forced withdrawals and to entrainment at the surface of the bed. The fraction m,l of particles in the entrainment flow of a BFBG (1D model), is equal to that in the bottom bed xm,l,b as long as perfect mixing is assumed in the bed. In a CFBG, a distinction is made because both the particle size distribution and the composition of the downward wall layer flow Fw differ from those in the upward flow in the core Fc and also from those in the bed. Gm,gain and Gm,loss are the gain of particles of the m,l class due to the attrition and fragmentation of particles from superior levels (size > l) and the loss of particles of the m,l class to inferior levels (size < l). The rate of production of fines by attrition and fragmentation is given on pages 603–604. ¬m,l is the consumption by chemical reaction of particles of m,l class. This term depends on the conversion model of the particle and, besides the influence of the operating conditions, it may strongly depend on the size and other physico-chemical properties of the solid. The mass conservation of the components in the freeboard is

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes

1 d (Fi, F ) = e r F + S (1 F (1 – e F )s m rgs F gg,i gg,i gs,m,i ,m ,m,i m AF dh



where a single-phase model for the gas (plug flow) and an axial distribution of the solids in the freeboard, eF(h) is assumed. Note that Fi,F in Eq. [12.9] indicates the net flow of i in the upward direction, including the i travelling as both gas and as solid. The mass flows of solids tend to decrease at higher positions, accounted for by modelling of the entrainment and elutriation at different heights, giving eF(h). This one-dimensional single-flow model for the freeboard is often enough to describe the conversion in bubbling FBG. In CFBG a somewhat more complex, but still simplified, treatment is usually made by dividing the freeboard into regions as described below, based on the core-annulus model. This makes it possible to calculate the distribution, eF(h) as well as the solids flow in the freeboard by giving a one-dimensional formulation of a two-dimensional flow situation (sometimes called 1.5D models), and Eq. [12.9] can be used for both types of FBG. The model of the functions e(h) in the bottom bed and freeboard of FBG is formulated in Section 12.4.2. Boundary conditions are given by the composition and flow rates of the fluidization agent and of the solids fed to the reactor. The source terms of the reacting particles, i.e. chemical reactions and gas–solid transport coefficients, are discussed in detail in Section 12.4.3.

12.4.2 Fluid-dynamic modelling This section establishes the distribution of solids and gas, as well as other fluid-dynamics properties needed to close the conservation equations [12.3], [12.4], [12.7] and [12.9]. Modelling of the bottom zone In the bottom zone, the parameters to be estimated are: the (volume) fraction occupied by gas ex in the bottom bed, the fraction occupied by bubbles eb in the bottom bed, the fraction of gas in the emulsion phase, ee, the velocity of gas in the emulsion ue, the bubble velocity ub and bubble size db. At every height h, part of the gas flows through the emulsion phase and the rest forms bubbles. The overall voidage in the bed e is e = eb + (1 – eb)ee


The bubble velocity is estimated (Davidson and Harrison, 1963) by ub = uv + ubr, where ubr is the single bubble velocity in an infinitely free bed, which can be calculated accordingly, and uv is the visible bubble flow defined as uv = eb ub, which is equal to uv = (u0 – umf) according to the original two-

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

phase theory. Once umf are calculated (see Kunii and Levenspiel, 1991, for details of a variety of methods and empirical correlations), emf is given by applying the Ergun equation. The bubble size can be calculated with empirical correlations (Mori and Wen, 1975). The total fluidization flow supplied to the bottom bed, represented by the superficial velocity u0, is divided into three parts: the flow in the emulsion phase, ue, the visible bubble flow, uv and the throughflow, utf : u0 = ue + uv + utf. That is, a gas flow (uv + utf) does not pass through the emulsion phase. Taking into account these relations and that uv = eb ub, the following relation for eb can be found by:

eb =

1 = 1 ubr ubr 1+ 1+ uv (u0 – ue – utf )


ue is usually assumed to be equal to emf, so to determine eb from Eq. [12.11], the visible flow uv has to be determined, which is made by correlations (Johnsson et al., 1991). A more direct, also empirical, way to determine e is through a bed expansion factor fbex: fbex =

1 – ee 1–e


fbex is calculated by empirical correlations, like that obtained by fitting measurements of various commercial BFB coal gasifiers (Babu et al., 1978). Finally, the coefficient of mass exchange between bubble and emulsion, kb-e, is calculated by a combination of convective and diffusion processes resulting in the most accepted correlation (Grace, 1986) given by kbe =

1/2 (2umf ) Ê 12ˆ (Dg e m mff ub ) +Á ˜ db Ëp¯ db3/2


Modelling of freeboard The key concept of freeboard modelling is the quantification of the entrainment of particles from the stationary bottom zone in the form of a steady solids flow through the freeboard and further through a circulation loop back to the bed. The state in the freeboard can be regarded either as flux of particles (G), density of suspension (r), or as voidage (e = (1 – r)). The entrainment rate of particles at height h, that is, the flux of particles G, can be empirically represented by (Wen and Chen, 1982): G = G∞ + (Gx – G∞) exp (– a(h – Hx))


where Gx, is the entrainment flux of particles at the surface of the dense bed (h = Hx) and G∞ is the particle flux in an imaginary long column, © Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


whose height is higher than the transport disengaging height. There are few correlations for direct estimation of Gx (Kunii and Levenspiel, 1991). G∞ is estimated by empirical correlations for the elutriation rate coefficient, E∞ (E∞ equals the saturation carrying capacity, i.e. entrainment at heights above the transport disengaging height, i.e. G∞ = E∞). Numerous experiments have been conducted to determine the elutriation rate constant for different particle sizes i, Ei,∞ (Geldart, 1986; Oka, 2004; Souza-Santos, 2004). For a distribution of particles in the bed, E∞ is calculated as E∞ = ∑Ei,∞ xi,b, where xi,b is the mass fraction of particles i in the bed. Compilation of correlations for Ei,∞ under different operating conditions are available (Geldart, 1986). There is a large disagreement between experimental data and correlations for specific elutriation fluxes, and these empirical correlations should be used carefully within the experimental conditions for which they have been derived. In accordance with the flux, a similar relationship expresses the particle density at height h (or loading r = G/up) as

r = r∞ + (rx – r∞) exp (– a(h – Hx))


where rx = rp(1 – ex) is the cross-section average suspension density at the surface of the bed h = Hx and r∞ is the density above the transport disengaging height. In CFB, Eq. [12.15] does not agree with measurements in large-scale units (Leckner, 1998). A more rigorous model can be derived by dividing the freeboard into a splash zone, a transport zone and an exit zone, giving the following solids distribution along the freeboard (Johnsson and Leckner, 1995)

r = rbx exp(– a(h – Hx)) + r0 exp(– K(h – H))


rx is the bottom-bed density and rox is calculated by rox = ro exp(– K(Hx – H)). The decay constants a and K have been determined empirically (Johansson et al., 2007). Figure 12.5 shows how r (Eq. [12.16]) is obtained by summing up the two contributions. The first part of Eq. [12.16] is the contribution of the splash zone (which is present in a wide riser with sufficient bed inventory), caused by the particles thrown up by the movement of the bed, similar to the classical form proposed by Lewis and others, typical for the bubbling bed, shown in Eq. [12.12]. The second part of the Eq. [12.16] is the contribution from the riser flow above the splash zone extrapolated to the surface of the bed. If the transport mechanism is neglected in Eq. [12.16], the second term vanishes and rx = rbx. In this case Eq. [12.16] equals Eq. [12.23], which is valid since r∞ is negligibly small for both the splash zone and for the transport zone, as illustrated in Fig. 12.5. Hence, Eq. [12.16] is a general empirical representation of the fluid dynamics of the freeboard in an FB, valid for both circulating and bubbling FB units, whereas Eq. [12.15] is a simplification that takes into account only one mechanism of

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion h = H

r Height, h ro exp(–K(h – H )) rbx exp(–a(h – Hx))

h = Hx



rbx rx = rbx + rox Solids density, r

12.5 Solids density profiles in an FB (thick black line is given by Eq. [12.16]).

backmixing, usually applicable for bubbling units. In BFB the decay in solids concentration above the bed is a direct function of the superficial velocity as au0 = constant (Kunii and Levenspiel, 1991). In Johansson et al. (2007) measurements on average suspension density were compared for various large-scale CFB boilers and the agreement of Eq. [12.16] with Eq. [12.15] was shown. Despite this experimental evidence, Eq. [12.15] instead of Eq. [12.16] is still sometimes used for freeboard modelling of CFB units. To sum up, the axial density profile in the freeboard of a CFB can be represented by Eq. [12.16] and that of a BFB by the first part of Eq. [12.16] or by Eq. [12.15]. The decay constants are obtained by correlations or by measurements. ro can be estimated from the circulating flux G assuming dilute transport (zero slip velocity) or, alternatively, from pressure measurements along the riser. Finally, in the equations, the bottom bed parameters rx and Hx are obtained by the bottom bed model presented above or by direct measurements.

12.4.3 Modelling fuel conversion processes The model of mass and energy transport in the reactor contains source terms that handle the production or consumption of fuel, char or gas. A brief overview on models to predict these source terms is presented in this section. Detailed

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


treatment on particle conversion processes can be found in Gómez-Barea and Leckner (2010) and Chapters 7 and 8 in this book. The scheme to follow the gaseous species, tar and char, along with the conversion of a particle, and with the processes in the reactor was given in Fig. 12.1. This leads to a distinction between primary and secondary conversion. Generation (primary conversion) includes devolatilization inside a particle in the absence of oxygen, whilst the subsequent conversion (secondary conversion) consists of extra-particle processes: oxidation, cracking and reforming of the gases. Due to the fast release of volatiles, the gasifying agent surrounding the particles has only a small impact on the devolatilization itself, which then occurs in an inert atmosphere as pyrolysis. In contrast, the volatiles released from a particle are converted in the bulk gas, where the presence of oxygen and steam (just to mention the main reactants), leads to oxidation and reforming reactions. Correspondingly, the conversion of a char particle is also treated in two steps. Devolatilization The primary decomposition of the fuel into char, tar, light gases and water is given by R1 in Table 12.1. A model aims at predicting the rates and yields of the species. Ideally, the C-O-H composition of tar and char and their chemical structures should be predicted. R1 in Table 12.1 is not a single reaction, but it represents several physical and chemical transformations; consequently, it cannot be presented by a single rate. The most important classes of model have been categorized as basic, distributed activation energy and structural models (Souza-Santos, 2004). Basic models represent pyrolysis by a single reaction (global model) or by a combination of series and parallel reactions, usually treated as first-order and independent reactions. The global model is the simplest representation of pyrolysis kinetics, where the only parameters to be adjusted by experiments are the kinetic coefficient kpy (and reaction order if different from unity) and the coefficient of distribution between primary volatiles. Many publications deal with kinetic parameters for the calculation of kpy for various fuels, having a great variation between the representations, probably because a number of physical and chemical factors are incorporated in one expression. Despite this fact, this type of expression has been widely used in reactor models due to its simplicity making it computationally tractable, needing only a small set of input data. Improvement is achieved by considering a combination of series and parallel reactions and the method of distributed activation energies but, in general, these add complexity and the kinetics of these reactions are not well known, especially at the high heating rate applied in FBG. The models mentioned predict the amounts of tar, gas and char released during pyrolysis, but the yield of the main gas species is not predicted. Instead,

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

empirical correlations developed for coal (Loison and Chauvin, 1964) and biomass species (Neves et al., 2011) have been used in reactor simulations. These correlations have been obtained for a particular set of operating conditions and type of fuel, not always similar to those to be simulated. Therefore, more fundamental data are desirable to increase the predictive capability of model simulations. Structural models have been developed to perform such predictions by accounting for the chemical constitution of the solid fuels. A critical comparison of these models applied to coal is made in Souza-Santos (2004), concluding that, with the exception of the Flash–Chain model, the composition of the gas released during pyrolysis is not predicted. Therefore, FBG models for coal have rarely used these models. A notable exception is the FBG models of Hamel and Krumm (2001) where a modified version of the Flash–Chain model is applied. A simulation of a coal FBG unit using the Hamel model is given in Section 12.5.1. In recent years efforts have been made to adapt structural models developed for coal to biomass fuels (Gómez-Barea and Leckner, 2010). In the kinetics models described above, the influence of mass and heat transfer is not explicitly accounted for. In FBBG, the size of fuel particles may influence the rate of devolatilization and the distribution of species. Devolatilization of fuel particles is caused by thermal degradation, where heat transport rather than kinetics is rate-determining, whilst the transport of mass is of secondary importance. Depending on the size of the fuel particles, internal or external transport processes or both may be decisive together with the rate of the conversion reaction. Detailed discussion of the methods for the calculation of heat and mass transfer coefficients in FBG is presented in Chapter 5 of this book and elsewhere (Leckner, 2006). The overall conclusion is that there is no published model that combines a structural kinetics model with diffusion effects for FB conditions, so the basic input for FBG modelling is based on empirical relations and experimental data to characterize the rate and the composition of the gases released during devolatilization. The essential idea behind the simplified models is to estimate the time of complete devolatilization (including drying in the case of wet fuels) by considering only the rate-limiting phenomena. The yields of char and volatiles and the composition of volatiles are not predicted, but these quantities are estimated separately by mass balances and empirical relations. This two-stage modelling approach is useful in applications to FBG if it is combined with some limiting cases for mixing and devolatilization already discussed in Section 12.2.2 (see Fig. 12.2). In simulations where the individual yields have to be predicated in each step, the simplified modelling is not valid and advanced fuel particle models have to be used. In these cases, however, the time of devolatilization can be estimated semi-empirically by direct measurement of the devolatilization time, including drying, for the type of fuel and range of particle size of interest as expressed by a correlation

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


with two coefficients a1 and a2 together with the characteristic dimension de of the fuel particle, t = a1dea2 (Ross et al., 2000). The constants have some physical meaning as can be seen from a derivation of the times of drying and devolatilization of thermally small particles or thermally large particles. Chemical conversion of char The char particles in an FBG are generated after rapid devolatilization of the fuel. The char consists mainly of ash and carbon. Chemical conversion of char occurs by reactions with O2 (R2 and R3 in Table 12.1), CO2 (R4), H2O (R5) and, in much minor proportion, with H2 (R6). At steady state, an FBG contains a distribution of char particles with different sizes and degrees of conversion. The rate of char conversion is influenced by variables, such as char temperature, partial pressure of the gasifying reactants and the products, particle size, porosity, and mineral content of the char, some of which vary with time due to chemical conversion and attrition. Variation between chars from biomasses and coal is significant, due to the differences in the nature of the fuels. Chars from biomass vary greatly in porosity, directionality and catalytic effects. Therefore, caution should be exerted in applying expressions from one char to another. The reactivity of lignocellulosic chars with O2, H2O and CO2 has been reviewed recently (Di Blasi, 2009). Extensive collection of char kinetics of various coals is also available (Smoot and Smith, 1985). The reactivity of a char sample with a given gas reactant at time t is defined as: dmc 1 dxc rm = – 1 = mc dt (1 – xc ) dt

[kg/(kg s)]


where mc and xc are the mass of carbon contained in the sample and its conversion at time t, this latter defined as xc = 1 –

mc mc0


By studying the effect of concentration r of the reactant gas (O2, H2O, CO2 or H2) and the temperature T in the laboratory, the rate of the char–gas reaction can be determined by Eq. [12.17]. For instance, assuming a kinetic law, often of nth order, r m = k mr n

[kg/(kg s)]


the kinetic coefficient km and the order of reaction n can be obtained by reactivity experiments. According to the classical plot of r vs. 1/T, some regimes can be distinguished (Laurendeau, 1978). If the experiments are conducted in the kinetic regime (Regime I), the concentration and temperature of the gas in all surface sites on the particle are the same as in the bulk gas and the

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

reactivity determined is the true reactivity. If the char particle is uniformly converted, the profiles of conversion are flat, and the global conversion of the particle xc equals the local conversion X at any interior site. The reactivity can also be defined per unit of reacting surface: rA = k Ar n

[kg/(m2 s)]


The rate coefficient kA (as well as km) is related to temperature via an Arrhenius expression of the form kA = A exp(–Ea/RT ) where A is a pre-exponential factor, Ea an activation energy and R the gas constant. Equation [12.20] expresses global nth order kinetics, valid for certain operating conditions. It is a simplified representation of a more complex surface process, usually described by the Langmuir–Hinshelwood equation (Laurendeau, 1978). To relate rm and rA, the total reacting surface area per unit of mass Am (m2/ kg) is introduced, rm = Am rA. A distinction is sometimes made between the internal (Am,i) and external (Am,e) areas of a particle, so that Am = Am,i + Am,e. The change in reactivity during conversion is described by the variation of Am, while kA is assumed to depend on temperature and concentration only. The surface changes continuously during conversion due to pore enlargement and pore coalescence. The form of variation depends on the char and operating conditions. A practical way to describe this variation is to relate Am to a reference state of conversion (‘0’), using a structural profile f(X), Am = Am0 f(X)



f(X) takes into account the variation of the total reacting surface in relation to the initial (or reference) surface. Semi-empirical expressions for f(X) have appeared in the literature for a variety of chars, valid for specific ranges of operating conditions. Expressions for f(X) have also been developed from structural models, describing the change of the pore system during conversion (Gómez-Barea and Leckner, 2010). For the development of particle models, it is sometimes useful to express the reactivity per unit of volume rm = rc0 (1 – X) rA, where rc0 is the initial density of the char. The reactivity of a single char particle may differ from the data obtained from a particular laboratory test, depending on the size of the particle and operating conditions. A model of a single particle is then required. Assuming a continuum description of the porous solid particle, the reactivity of a single char particle rm,p can be calculated by integration of the volumetric reactivity rv throughout the volume of the particle Vp: rm,p (xc (t )) =

rc0 Am0 k r n (1 – X X)) f (X) X dW mc,p ÚVp ((t) A


To obtain the profiles of gas reactant, temperature and conversion, r(W,t), T(W,t), and X(W,t), the conservation equation for the gas species involved in

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


the reaction and the energy equation have to be solved inside the char particle and its boundary layer, together with the carbon balance at any position within the char particle. This problem has been treated for a variety of situations for single-char particles (Morell et al., 1990). However, the solutions are not used in FBG models, probably because of numerical difficulties and the amount of input required. Therefore, tractable models such as those derived in Gómez-Barea and Leckner (2010) have been demonstrated to be simple enough for implementation in reactor codes. Extensive reviews on char particle models are available (Laurendeau, 1978; Gómez-Barea and Leckner, 2010; Di Blasi, 2009). The char is mainly converted in a direct FBG by CO2 and H2O and to a lesser extent by O2. Moreover, the temperature of a char particle is roughly uniform (different from FB combustors) and equal to the average bed temperature (van den Aarsen, 1985). Since the rate of char conversion with CO2 and H2O is slow, the reactants diffuse into the particles and the reactions take place inside the particle to a greater extent than during combustion. Therefore the interior surface changes significantly (internal area and degree of catalytic effect). In this case, a model accounting for the change in local carbon conversion, reaction area and gas concentration is essential. Advanced particle models have been published to calculate rm,p during gasification of char from coal (Morell et al., 1990) and biomass (Gómez-Barea et al., 2005), providing understanding of the process; however, they are not used for reactor simulation due to the complexity of the calculations. Instead, models of kinetics without transport effects are often employed in reactor calculations. Finally, the overall char reactivity in the bed has to be calculated taking into account the reactivities of individual char particles rm,p(xc) having different stages of conversion. A population balance may be needed, coupled with the physical process (entrainment and attrition described below) to evaluate the overall char gasification rate. Most FBG models assume that rm,b can be approximated by rm,p(xc,b), i.e. the reactivity evaluated at the average conversion in the bed, xc,b. The difference between rm,b and rm,p(xc,b) has been analysed in Caram and Amundson (1978) and Gómez-Barea and Leckner (2009a), concluding that a population balance should be included in most cases. Fragmentation and attrition Fuel size is reduced by shrinkage during devolatilization, by primary fragmentation, secondary and percolative fragmentation of char, and fines generation by attrition. The comminution can differ from one fuel to another, but it is difficult to infer attritability from fuel properties, so it has to be characterized by experiments (Chirone et al., 1991). Chapters 6 and 7 address

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

this problem in detail so only a brief description is given below. The attrition rate in an FB is often represented by: dmatt m = K att Du ch dt dch


∆u = u0 – umf the velocity difference in a BFB and the slip velocity u0 – ut in a CFB. Katt is the dimensionless attrition constant, determined experimentally for a variety of chars and solids and roughly independent of the type of fluidization. Also, friability coefficients Qm are taken from the literature for first estimates of the behaviour of the solids in the bed (Souza-Santos, 2004). Significant differences have been observed between biomass and coal and between different types of biomasses with respect to attrition and fragmentation. An investigation of FB combustion of three biomasses showed that the conversion of fines occurred essentially by percolative fragmentation. Inclusion of fragmentation and attrition in the population mass-balances given in Eq. [12.8] is rather complex, demanding detailed knowledge of the mass-flow rates at different particle sizes. Simplified treatment has been presented for FB combustion for char from coal (Arena et al., 1995) and biomass (Scala et al., 2006), as well as for char gasification in FBG (GómezBarea and Leckner, 2010). Secondary conversion of volatiles Secondary conversion of gas compounds includes the homogeneous combustion and reforming of volatiles as well as heterogeneous conversion of these on the surfaces of the bed solids such as char and catalyst. The conversion of char has been treated above separately due to its importance in the conversion of the fuel. The key points needed in order to deal with secondary conversion of volatiles are the rate of the reactions, which depends on chemical kinetics and gas mixing. There is a great amount of work on the kinetics of homogeneous reactions. Some authors have analysed the difference in the kinetics available in the literature, especially for the homogeneous oxidation of CO and H 2 during devolatilization of solid fuels (Souza-Santos, 2004; Gómez-Barea and Leckner, 2010). There is a great variation between the expressions for the same reaction. The homogeneous oxidation and steam reforming of CH4 are much slower than the oxidation of CO and H2, concluding that in the absence of catalytic effects, the concentration of methane is expected to remain constant after devolatilization and oxidation in an FBBG. This fact has been assumed in FBG models over the years to simplify the gasification process. However, the reforming of CH4, as well as the WGSR are catalysed by the solids in an FBG, for instance by coal ash (Chen

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


et al., 1987) and bed material (Weimer and Clough, 1981; Yan et al., 1997). Secondary conversion of tar takes place on its way through the reactor. The main processes are thermal cracking, reforming and partial oxidation (R11–R15 in Table 12.1). The presence of an in-bed catalyst such as calcined limestone, dolomite, olivine, nickel-based catalyst, has influence on the rate of cracking and reforming reactions, affecting conversion and composition. The char itself is a tar catalyst and the char load achieved in the gasifier may also affect the secondary tar and its conversion. The modelling of tar has been focused on making a simplified representation of tar formation and conversion. All the kinetic data for estimation of reaction rates are global and none of them represents elementary steps. Much work has been published to obtain practical kinetics of tar formation and secondary decomposition (Devi et al., 2002). This can be sorted out in a variety of ways by taking a generic lump of tar (or a specified compound, or tar indicator/ model) or by considering a continuous representation of the tar mixture evolving with time. The first approach has been widely applied in earlier models taking one lump that represents the gravimetric tar to describe the overall tar concentration (van den Aarsen 1985). Also naphthalene has been selected in FBG of biomass because it is hard to destroy, giving a conservative estimate of the overall tar concentration (usually called gravimetric tar). Acetol, anisole and phenol have been chosen to represent the oxygenated compounds that constitute the primary tar, released during fuel devolatilization. Toluene has been used to describe the behaviour of the alkyl products class and naphthalene as an indicator compound for polyaromatic hydrocarbon (PAH). In addition to the selection of one or various tar lumps, the scheme of reaction and the interaction between the lumps or individual tar compounds have to be defined in the model. The simplest scheme is to assume overall kinetics of formation and decomposition of the tar lumps, giving a rough estimation. Despite various complex schemes being proposed to improve accuracy (Corella et al., 2002), tar conversion in FBG is still modelled in a simple and empirical way by considering the overall kinetics of thermal cracking, partial oxidation and reforming (Gómez-Barea and Leckner, 2010). Contaminants Most fuels contain additional material beyond the carbon, hydrogen and oxygen. Modelling of these minor compounds is necessary to consider the fate of sulphur and nitrogen because of the effect of the resulting compounds downstream of the gas production, for example, environmental emissions, catalyst poisons, and so on. Detailed modelling of emissions of contaminants

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

is beyond the scope of the present treatment and can be found elsewhere (e.g., Liu and Gibbs, 2003) and Chapter 9 of this book. Sulphur in the fuel is converted into H2S and COS, and the nitrogen into elemental nitrogen, NH3, and HCN. Most of the sulphur volatilizes during gasification, incorporating to the gas phase. Nitrogen is split between solid and gas phases in the order of 50%. Typical distribution ratios of S- and N-compounds in the gas phase are the following: H2S/COS = 10, NH3/HCN = 5–10 (Highman and van der Burgt, 2008). The observed decrease in the NH3 concentration in the exit gas with increasing gasifier temperature appears to result from the more rapid gas-phase equilibration of reaction N2 + 3H2 ´ 2NH3 at higher temperature.

12.4.4 Summary of model characteristics and suggestions for improvement Based on the critical survey presented here and in a recent review (GómezBarea and Leckner, 2009b), the main characteristics of the FM published to date and some key issues to deal with for the improvement of the existing FBG models are as follows: 1. Most models are one dimensional and steady state, and the fluiddynamics are based on the two-phase theory of fluidization with some modifications. 2. The freeboard is not modelled in many bubbling FBG models, but for circulating FBG it is taken into account as a core-annullus structure (1.5D model). 3. A common assumption is instantaneous devolatilization of the fuel. The composition of volatile species is not clearly reported by some authors. In some cases correlations are used from other biomass materials or even from coal. 4. A majority of modellers have not paid sufficient attention to char conversion, neglecting the effects of mass transfer in the particle, the change in reactivity during char gasification, and the distribution of conversion in the bed. 5. Fragmentation and attrition of fuel and char are not treated in FBG despite char attrition and entrainment being identified as the main source of inefficiency in FBG, especially in bubbling units. 6. Tar conversion is not usually modelled or modelled as one or two lumped species reacting by oxidation, thermal cracking or reforming with steam. 7. The effect of in-bed catalysts has not been modelled in a comprehensive way by any author. 8. Detailed treatment of non-isothermal effects in FBG is not applied in current FBG models. © Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


Much work is still needed to make FBG models useful for design and optimization. Practical consideration, such as detailed study of the local gas mixing and reaction in the gas and fuel feed ports, is needed. Particular attention has to be paid to hot spots close to the oxygen injection points and to the plumes of volatiles created near the fuel ports. The knowledge of the chemistry of tar generation and conversion has to be significantly improved, as well as the effects of temperature, solids concentration and gas composition in the bed on the tar reactions. In general, models have only been validated with the temperature and composition of the outlet gas in laboratory units, so validation of model hypotheses and the scale-up of results are difficult. Detailed consideration of the specific characteristics of the fuel to be gasified has to be implemented in the models. Extrapolation from one fuel to another may invalidate the model’s ability to explain measurements. The most complete models published to date are those of Hamel and Krumm (2001), Petersen and Werther (2005) and Souza-Santos (2004). The first two of these are analysed using examples in the following section.


Examples of simulations of fluidized bed gasifiers (FBGs)

Various examples of simulations are given to demonstrate the type of information delivered by the types of models discussed in this work and their prediction capability.

12.5.1 O2-steam gasification of coal in a pressurized coal FBG The model of Hamel (Hamel and Krumm, 2001) is used to illustrate an advanced published FM of FBG. This model includes various sub-models for fluid-dynamics, devolatilization and char conversion similar to those described in Section 12.4. The simulation predicts the axial concentration of the reactant gas and the temperature along a 2.5 MPa FBG, processing brown coal (Hamel and Krumm, 2001) (HTW process). Figure 12.6 shows the basis of the model: the discretization is made along the column with a topology which allows simulating both bubbling and circulating FBG. A model of a single isolated cell is shown in the figure with indication of the main processes accounted for. The results are presented in Fig. 12.7 where the profiles of the main gas compounds and temperature are plotted along the gasifier. It is shown that most changes occur in the bottom bed and the gas concentration profiles are smooth in the latter part of the freeboard, indicating that the reaction rates are slow or that composition is in equilibrium. Besides the prediction capability shown, a common problem of model validation is emphasized: the gas composition is only available at the exit, so it is difficult to validate the

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion Solid Solid flow flow to from cell Gas flow to cell i-1 cell i-1 i-1

Heat transfer

Gas exchange

Gas feed Cell i Bubble phase Gas

Emulsion phase Homogeneous Solid Gas reactions

Homogeneous reactions

Solid recirculation flux

Homogeneous reactions

Solid discharge

Cell i i -1

Cell + 1

Gas flow to cell i + 1

Solid feed

Solid flow Solid from cell flow to i + 1 cell i + 1

12.6 Schematic diagram of the Hamel model showing the cell model and the processes taken into account in each cell.

Height (m)



Measured Simulated O2 CH4 CO H2


CO2 H 2O



0.0 0.0


Height (m)


0.6 Measured 0.4



0.1 0.2 0.3 0.4 Concentration (mol/mol) (a)


0.0 0.0


1000 1100 1200 1300 Temperature (K) (b)

12.7 Concentration profiles of the main gaseous compounds (a) and temperature (b) along the gasifier.

gas concentration inside the gasifier. It could be shown that a pseudo-model (with a correction to estimate the conversion of char such as that in GómezBarea et al., 2008) could also fit the gas composition well at the exit, and then a much simpler model could be a good predictive tool too. The Hamel model, however, also allows predicting the char conversion by accounting for herogeneous reactions (kinetically limited) as well as entrainment and

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


elutriation of fine char. Moreover the model predicts reasonable well the measured maximum of temperature near the gas entrance.

12.5.2 Air gasification of sewage sludge in a circulating FBG A three-dimensional model was developed by Petersen and Werther (2005) to simulate the air gasification of sewage sludge in a circulating FBG. The model included detailed fluid-dynamic and fuel conversion sub-models to be able to account for local effects. Continuous radial profiles of velocities and solids hold-up were used to represent the solid and gas motion in the bed. The main inputs for the fluid-dynamic model were validated with available measurements from cold flow CFB units from the literature. A gasification reaction network with kinetic rate expressions obtained for sewage sludge gasification was included. The influence of the axial location and the number of feeding points was examined for CFBG of different sizes, providing understanding of the effects of lateral mixing of volatiles in large units, which is a critical aspect for scale-up and process optimization (Petersen and Werther, 2005). Figure 12.8 illustrates the distribution of volatiles (CH4, CO and H2) near the feed point in a CFBG of 0.5 m square cross section. The simulation results shown in Fig. 12.8(a), predict steep concentration gradients of volatile species (CH4, CO, H2) in the horizontal direction, caused by limited radial mixing of the fuel. In Fig. 12.8(b), the corresponding distribution in the vertical direction of the x,z-plane is given for a cross section in the central part of the bed. The insufficient radial gas dispersion is evident. As seen, the horizontal gas mixing is slower than the axial transport, and the gaseous components, devolatilized at the feed height, flow in streamers to higher regions. Incomplete mixing is visible up to a height of about 10–12 m. Simulations allowed the conclusion that, due to the very fast release of volatiles and the high volatile content in the fuel (sewage sludge), mixing of the gas around the feed port is not fast, and an ascending plume with highly concentrated pyrolysis gas is formed. Further work verified that large devolatilizing biomass particles rapidly move to the top of the riser (Kersten et al., 2003), contributing to the tar measured in the product gas. Besides the conclusion itself, it is emphasized in this example that detailed threedimensional simulation is needed to predict this kind of processes.

12.5.3 Air-steam gasification of wood in a dual-bed FBG The Battelle biomass gasification process (Paisley and Overend, 2002) is an example of dual FBG, an indirectly-heated FBG as discussed above. The process comprises two reactors, a gasification reactor in which the biomass

© Woodhead Publishing Limited, 2013

Fluidized bed technologies for near-zero emission combustion


11.5-14 9-11.5 6.5-9 4-6.5


0.15 0.30 0.45 x, m



0.15 0.30 0.45 x, m (a)







0.5 x, m



20-25 15-20 10-15 5-10

0.1 0





14-18 9-14 5-9 0-5



x, m

0 0.5

0.2 0.1

0 0 0.1 0.2 0.3 0.4 0.5 x, m

15 H2


Height above distributor, m











12-15 9-12 6-9 3-6



x, m

Height above distributor, m


0.4 0.3

0.3 12-16 8-12 4-8 0-4


y, m






0.5 CH4

Height above distributor, m


21-28 14-21 7-14 0-7

0 0.5


12.8 Distribution of the concentrations of CH4, CO and H2 over the cross section of a CFBG with 0.5 m square cross section burning sewage sludge: (a) horizontal cross section on the feeding level z (axial coordinate) = 1.5 m; (b) vertical cross section at y = 0.25 m (half width) (from Petersen and Werther, 2005).

is converted into gas and residual char and a combustion reactor that burns the residual char to provide heat for gasification. Heat transfer between reactors is accomplished by circulating bed material between the gasifier and the combustor. A model version of a pseudo-equilibrium model was developed to simulate the performance of a gasification plant processing 1965 t/day of wood at 12% moisture as reference (1730 t/day of dry wood). Besides the prediction of the composition and flow rate of the gas, the simulation was

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


made to compute the amount of solids circulation required to maintain the system in the autothermal mode without additional fuel gas. This is made to avoid excess solid recirculation which is not achieved in practice due to fluid-dynamic limitations. Figure 12.9 presents the conceptual scheme of the simulation, employing a pseudo-equilibrium scheme as explained above. Experimental yields of main species (CO, CO2, CH4, C2H2, C2H4, C2H6 and tar, assumed to be C10H8, and char) were taken as input from the measurements conducted in the 9 t/day test facility of Battelle Columbus Laboratory (BCL). They were implemented as correlations in the form of quadratic functions of temperature. Simulations were conducted by setting the steam-to-biomass ratio to 0.4 and the temperature of the gasifier near 900°C. The air to the riser is given as 20% over the stoichiometric needed for burning completely the amount of char entering the riser. The gasification pressure is 1.5 bar and the steam fed is at 4 bar and 150°C. The results show the flow rate and composition of the produced gas as well as the flow of bed material necessary to maintain the system under autothermal conditions in the various scenarios. Table 12.2 presents the main results. It is shown that similar composition and flow rate of syngas is obtained because the gasifier temperature is more or less the same for the three simulations presented in the table. However, the circulation of solids material between the reactors increases sharply when reducing the CO, H2, H2O, CO2, CH4, C2H6,…, tar Char Energy balance gasification bed


Energy balance combustor bed

CO, H2,H2O, CO2, CH4, C 2H 6, .., char, tar

Volatilization data as a function of temperature

Air Char combustion (chemical equilibrium)

Flue gas

Steam Heat demand

Heat generation

12.9 Model scheme of the dual CFBG simulated (quasi-equilibrium model).

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion Table 12.2 Effect of variation of temperature gasifier rise in CFBG on performance parameters (fuel: wood) Temperature riser (°C) Temperature gasifier (°C) Solids circulation (tn/h) Syngas (kg/h)

915 909

950 905

985 901







H2 %v/v




CO %v/v




CO2 %v/v




H2O %v/v




CH4 %v/v




temperature difference between the two reactors. It was concluded that the system could work with standard design up to 950°C, otherwise the solids circulation rate is difficult to achieve. Another option studied was to decrease the temperature of the gasification reactor (not shown here) to assess the change in gas composition. A sensitivity analysis was made to assess the effects of a change in wood moisture (Fig. 12.10(a)). The bed temperature and the cold gasification effciency decrease with the increase in wood moisture, then more char is formed to satisfy the increase in heat demand. The effect of the steam/biomass ratio on the gasifier performance is shown in Fig. 12.10(b). Steam/biomass ratio increases the cold gas efficiency and gas yields slightly decrease, indicating that in the gasification bed the devolatilization is the dominant process.



Modelling of solid fuel gasification in fluidized bed reactors was reviewed. The relevant phenomena were formulated, including fluid-dynamics and chemical conversion processes. The various approaches applied for reactor modelling, from thermodynamic to computational fluid-dynamic models, were described, illustrating their state of development and the usefulness of each approach depending on the aim of the model. A one-dimensional fluidization model was formulated indicating the main semi-empirical correlation parameters to be estimated and the procedure to solve the conservation equations coupled with the sub-models of the main conversion processes. Despite the wealth of knowledge available, fluidization models of FBG are still treated to a great extent in an empirical way, with few exceptions. Examples are given of various commercial FBG to illustrate the use of the different types of models. Thermodynamic-based models (TM) corrected by some empirical relations, i.e. pseudo-equilibrium models, allow reasonable prediction of the gas phase, needing a reasonable amount of input data. For reliable prediction

© Woodhead Publishing Limited, 2013

CGE of product yield (kmol/ ton biomass d.b.)




5 10 15 20 Biomass moisture content (% w.b.)



80 70 CGE CO H2 CH4 CO2 Char

60 50 40 30 20 10 0

79 78 77 76 75 74 73 72

0.2 0.15 0.1 0.05

Char yield gasifier bed (kg/kg biomass d.b.)

920 910 900 890 880 870 860 850

Sensitivity analysis of biomass moisture content


Cold gas efficiency

Gasification bed temperature (°C)

Modelling of fluidized bed gasification processes

0 0.3 0.5 0.7 0.9 Steam/biomass ratio (kg H2O/kg biomas d.b.)

12.10 (a) Effect of wood moisture content on cold gas efficiency and gasification bed temperature. No auxiliary fuel is used in the combustor bed. (b) Effect of steam/biomass ratio on cold gas efficiency (CGE), gas and char yield in the gasifier bed. No auxiliary fuel is used in the combustor bed.

of carbon conversion, FM is necessary. For simulating hardware details, CFDM are required. However, much experimental work is still required before detailed models can be validated with confidence.



Arena, U., Chirone, R., D’Amore, M., Miccio, M. and Salatino, P. (1995), ‘Some issues in modelling bubbling and circulating fluidized-bed coal combustors’, Powder Technol, 82, 301–16. Babu, P., Shah, B. and Talwalker, A. (1978), ‘Fluidization correlations for coal gasification materials – minimum fluidization velocity and fluidized bed expansion ratio’, AIChE Symp Ser, 74, 176–86. Buekens, A.G. and Schoeters, J.G. (1985), ‘Modelling of biomass gasification’, in: Overend, R.P., Milne, T.A., Mudge, K.L., (eds) Fundamentals of Thermochemical Biomass Conversion. London: Elsevier Applied Science Publishers, 619–89. Caram, H.S. and Amundson, N.R. (1978), ‘Fluidized bed gasification reactor modeling 2. Effect of the residence time distribution and mixing of particles’, Ind Eng Chem Proc Des Dev, 18, 80–96.

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

Chen, W.J., Sheu, F.R. and Savage, R.L. (1987), ‘Catalytic activity of coal ash on steam methane reforming and water-gas shift reactions’, Fuel Proc Technol, 16, 279–88. Chirone, R., Massimilla, L. and Salatino, P. (1991), ‘Comminution of carbons in fluidized bed combustion’, Prog Energy Combust Sci, 17, 297–326. Corella, J. and Sanz, A. (2005), ‘Modeling circulating fluidized bed biomass gasifiers. A pseudo-rigorous model for stationary state’, Fuel Proc Techn, 86, 1021–53. Corella, I., Toledo, J.M. and Aznar, M.P. (2002), ‘Improving the modeling of the kinetics of the catalytic tar elimination in biomass gaification’, Ind Eng Chem Res, 41, 3351–6. Davidson, J.F. and Harrison, D. (1963), Fluidised Particles. Cambridge: Cambridge University Press. Desrosiers, R. (1979), ‘Thermodynamics of gas-char reactions’, in: Reed, T.B., (ed.), A Survey of Biomass Gasification. Golden, Co: Solar Energy Research Institute. Devi, L., Ptasinski, K.J. and Janssen, F.J.J.G. (2002), ‘A review of the primary measures for tar elimination in biomass gasification processes’, Biomass and Bioenergy, 24, 125–40. Di Blasi, C. (2009), ‘Combustion and gasification rates of lignocellulosic chars’, Prog Energy Combust Sci, 35, 121–40. Enwald, H., Peirano, E. and Almstedt, A.E. (1996), ‘Eulerian two-phase flow theory applied to fluidization’, Int J Multiph Flow, 22 (Suppl. 1), 21–66. Geldart, D. (1986), ‘Particle entrainment and carryover’, in: Geldart, D. (ed.) Gas Fluidization Technology. New York: Wiley. Gómez-Barea, A. and Leckner, B. (2009a), ‘Evaluation of char reaction rate in a fluidised bed gasifier: from reactivity determination to reactor simulation’, Proc 17th European Biomass Conference, Hamburg, Germany. Gómez-Barea, A. and Leckner, B. (2009b), ‘Gasification of biomass and waste’, in: Winter, F. (ed.) Handbook of Combustion, Vol 4. New York: Wiley. Gómez-Barea, A. and Leckner, B. (2010), ‘Modeling of biomass gasification in fluidized bed’, Progr En Comb Sc, 36, 444–509. Gómez-Barea, A., Ollero, P. and Arjona, R. (2005), ‘Reaction-diffusion model of TGA gasification experiments for estimating diffusional effects’, Fuel, 84, 1695–1704. Gómez-Barea, A., Thunman, H., Leckner, B., Campoy, M. and Ollero, P. (2007), ‘Prediction of gas composition in biomass gasifiers’, 2nd Int Congress of Energy and Environment Engineering and Management (IICIIEM2007). Badajoz, Spain. Paper ER-030. Gómez-Barea, A., Leckner, B. and Campoy, M. (2008), ‘Conversion of char in CFB gasifiers’, in: Werther. J., Nowak, W., Wirth, K.-E. and Hartge, E.-U. (eds) Proc 9th Int Conf on Circulating Fluidized Beds, Hamburg, Germany, 727–32. Gómez-Barea, A., Nilsson, S., Vidal-Barrero, F. and Campoy, M. (2010), ‘Devolatilization of biomass and waste in fluidized bed’, Fuel Proc Tech, 91, 1624–33. Grabner, M., Ogriseck, S. and Meyer, B. (2007), ‘Numerical simulation of coal gasification at circulating fluidised bed conditions’, Fuel Proc Technol, 88, 948–58. Grace, J.R. (1986), ‘Fluid beds as chemical reactors’, in: Geldart D. (ed.) Gas Fluidization Technology, Chichester: Wiley, 285–339. Gururajan, V.S., Agarwal, P.K. and Agnew, J.B. (1992), ‘Mathematical modelling of fluidized bed coal gasifiers’, Trans IChemE, 70, 211–38. Hamel, S. and Krumm, W. (2001), ‘Mathematical modelling and simulation of bubbling fluidised bed gasifiers’, Powder Technol, 120, 105–12. Highman, C. and van der Burgt, M. (2008), Gasification. Amsterdam: Gulf Professional Publishers/Elsevier Science.

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes


Jand, N., Brandani, V. and Foscolo, P.U. (2006), ‘Thermodynamic limits and actual product yields and compositions in biomass gasification processes’, Ind Eng Chem Res, 45, 834–43. Jiang, H. and Morey, R.V. (1992), ‘A numerical model of a fluidized bed biomass gasifier’, Biomass and Bioenergy, 3, 431-47. Johansson, A., Johnsson, F. and Leckner, B. (2007), ‘Solids back-mixing in CFB boilers’, Chem Eng Sci, 62, 561–73. Johnsson, F. and Leckner, B. (1995), ‘Vertical distribution of solids in a CFB-furnace’, Int Conf on FBC, 13, 266–73. Johnsson, F., Andersson, S. and Leckner, B. (1991), ‘Expansion of a freely bubbling fluidized bed’, Powder Technol, 68, 117–23. Kersten, S.R.A., Prins, W., van der Drift, A. and van Swaaij, W.P.M. (2003), ‘Experimental fact-finding in CFB biomass gasification for ECN’s 500 kWth pilot plant’, Ind Eng Chem Res, 42, 6755–64. Kunii, D. and Levenspiel, O. (1991), Fluidization Engineering. 2nd edn. Amsterdam: Elsevier. Laurendeau, N.M. (1978), ‘Heterogeneous kinetics of coal char gasification and combustion’, Prog Energy Combust Sci, 4, 221–70. Leckner, B. (1998), ‘Fluidized bed combustion: mixing and pollutant limitation’, Prog Energy Combust Sci, 24, 31–61. Leckner, B. (2006), ‘Heat and mass transfer’, in: Crowe, C. (ed.) Multiphase Flow Handbook, Boca Raton, FL: CRC Press, Chapter 5.2. Li, X., Grace, J.R., Watkinson, A.P., Lim, C.J. and Ergüdenler, A. (2001), ‘Equilibrium modelling of gasification: a free energy minimization approach and its application to a circulating fluidized bed coal gasifier’, Fuel, 80, 195–207. Li, X., Grace, J.R., Lim, C.J., Watkinson, A.P., Chen, H.P. and Kim, J.R. (2004), ‘Biomass gasification in a circulating fluidized bed’, Biomass and Bioenergy, 26, 171–93. Liu, H. and Gibbs, B.M. (2003), ‘Modeling NH3 and HCN emissions from biomass circulating fluidized biomass gasifiers’, Fuel, 82, 1591–604. Loison, R. and Chauvin, R. (1964), ‘Pyrolyse rapide du charbon’, Chemie Indust, 91, 269–75. Ma, R.P., Felder, R.M. and Ferrell, J.K. (1988), ‘Modelling a pilot-scale fluidised bed coal gasification reactor’, Fuel Proc Tech, 19, 265–90. Moreea-Taha, R. (2000), Modeling and Simulation for Coal Gasification, IEA Greenhouse Gas R&D Programme. Morell, J.I., Amundson, N.R. and Park, S.K. (1990), ‘Dynamics of a single particle during char gasification’, Chem Eng Sci, 45, 387–401. Mori, S. and Wen, C.Y. (1975), ‘Estimation of bubble diameter in gaseous fluidized beds’, AIChE J, 21, 109–15. Nemtsov, D.A. and Zabaniotou A. (2008), ‘Mathematical modelling and simulation approaches of agricultural residues air gasification in a bubbling fluidized bed reactor’, Chem Eng J, 143, 10–31. Neves, D., Thunman, H., Matos, A., Tarelho, L. and Gómez-Barea, A. (2011), ‘Characterization and prediction of biomass pyrolysis products’, Progr En Comb Sc 37, 611–30. Oka, S.O. (2004), Fluidised Bed Combustion, New York: Marcel Dekker. Paisley, M.A. and Overend, R.P. (2002), ‘The SylvaGas process from Future Energy Resources – a commercialization success’, in: Palz, W., Spitzer, J., Maniatis, K.,

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

Kwant, K., Helm, P. and Grassi, A. (eds) Proc 12th European Conference on Biomass for Energy, Industry and Climate Protection, Amsterdam, 975−8. Petersen, I. and Werther, J. (2005), ‘Three-dimensional modeling of a circulating fluidized bed gasifier for sewage sludge’, Chem Eng Sci, 60, 4469–84. Radmanesh, R., Chaouki, J. and Guy, C. (2006), ‘Biomass gasification in a bubbling fluidized bed reactor: experiments and modeling’, AIChE J, 52, 4258–72. Rauch, R., Hofbauer, H., Bosch, K., Siefer, I., Aichernig, C., Tremmel, H., Voigtlaender, K., Koch, R. and Lehner, R. (2004), ‘Steam gasification of biomass at CHP plant in Güssing – status of the demonstration plant’, Proc 2nd World Conf on Biomass for Energy, Industry and Climate Protection, Vol. II, 1687-90. Ross, D.P., Heidenreich, C.A. and Zhang, D.K. (2000), ‘Devolatilisation times of coal particles in a fluidised-bed’, Fuel, 79, 873-83. Saffer, M., Ocampo, A. and Laguerie, C. (1988), ‘Gasification of coal in a fluidized bed in the presence of water vapor and oxygen; an experimental study and a first attempt at modeling the reactor’, Int Chem Eng, 28, 46–61. Scala, F., Chirone, R. and Salatino, P. (2006), ‘Combustion and attrition of biomass chars in a fluidized bed’, Energy Fuels, 20, 91–102. Smoot, L.D. and Smith, P.J. (1985), Coal Combustion and Gasification. New York: Plenum Press. Souza-Santos, M.L. (2004), Solid Fuels Combustion and Gasification. New York: Marcel Dekker. van den Aarsen, F.G. (1985), Fluidised bed wood gasifier. Performance and modelling, PhD Dissertation, Twente University Publication, University of Twente. Wang, Y. and Yan, L. (2008), ‘CFD modeling of a fluidized bed sewage sludge gasifier for syngas’. Asia-Pacific J Chem Eng, 3, 161–70. Weimer, A.W. and Clough, D.E. (1981), ‘Modeling a low pressure steam-oxygen fluidized bed coal gasification reactor’, Chem Eng Sci, 36, 549–67. Wen, C.Y. and Chen, L.H. (1982), ‘Fluidized bed freeboard phenomena entrainment and elutriation’, AIChE J, 28, 117-28. Yan, H.M., Heidenreich, C. and Zhang, D.K. (1997), ‘Mathematical modelling of a bubbling fluidised-bed coal gasifier and the significance of net flow’, Fuel, 77, 1067-79. Yan, H.M., Heidenreich, C. and Zhang, D.K. (1999), ‘Modelling of bubbling fluidised bed coal gasifiers’, Fuel, 78, 1027-47. Yoshida, K. and Kunii, D. (1974), ‘Complex reaction in fluidised beds – simulation of gasification’, J Chem Eng Jpn, 7, 34–39. Yu, L., Lu, J., Zhang, X. and Zhang, S. (2007), ‘Numerical simulation of the bubbling fluidized bed coal gasification by the kinetic theory of granular flow (KTGF)’, Fuel, 86, 722–34.


Appendix: notation

A Am a c d p E a

pre-exponential factor, various units; area, m2 area per unit of mass, m2/kg decay constant, m–1 concentration, kmol/m3 solid particle size, m activation energy, kJ/kmol

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes

E ∞ F fbex g G H H x H u h hm hb-e kb-e k b k K Katt m MM n p R Rj,i ¬ r rgg,i rgs,m,i rm/A/v rm,p rm,e/s S T t u,u x c X


elutriation rate (elutriation flux above the transport disengaging height), kg/(m2s) mass flow rate, kg/s, function bed expansion factor, – acceleration of gravity, m/s2 mass flux, kg/(m2 s) riser height, m height of the bottom zone, m lower heating value, kJ/kg vertical coordinate, m; heat transfer coefficient, kW/m2K; specific enthalpy, kJ/kg solid-to-gas mass transfer coefficient, m/s bubble to emulsion heat transfer coefficient, kW/(m2K) bubble to emulsion mass transfer coefficient, s–1 back-flow ratio, –; transfer coefficient, – kinetic coefficient, various units decay constant in transport zone, m–1; reaction coefficient attrition constant, – mass, kg; class of particle, – molecular mass, kg/kmol order of reaction, – distribution function universal gas constant, 8.315 ¥ 10–3 kJ/(molK) rate of reaction of species i in reaction j, kmol/(m3s) reaction rate for a size class m, l, kg/s rate of reaction, kg/(m3s) net rate of production of i in gas-gas reactions per unit of gas volume, kg/(m3s) net rate of production of i in gas-solid reactions per unit of k-solid volume, kg/(m3s) char reactivity per unit of mass (s–1), surface (kg/(m2s)) or volume (kg/(m3s)) char reactivity per unit of mass of a single particle, s–1 char reactivity per unit of mass of a single particle evaluated at emulsion/surface, s–1 source of a transported property per unit of volume, unit of transported property/(m3s) temperature, K time, s velocity, m/s global conversion, – local conversion, –

© Woodhead Publishing Limited, 2013


Fluidized bed technologies for near-zero emission combustion

12.8.1 Greek letters Gm.l e ni,j r s m j y

rate of attrition of particle m from at level l, kg/s volume concentration or porosity, m3/m3 stoichiometric coefficient of i species in reaction j mass concentration or density, kg m–3 volume fraction occupied by solids m, m3/m3; factor, – transported property; gas produced by gasification of char, kg gas/ kg char; shrinkage factor, – dimensionless visible bubble flow, –

12.8.2 Subscripts A B b bex b-e br c ch dev e ent fs F f g g-g g-s in i,j,k l m mf ng 0 out rec p py r

related to surface bottom zone, bed (bottom bed) bubble bed expansion bubble-emulsion rising bubble core, carbon (char) char devolatilization emulsion, external entrainment feed stream freeboard fuel gas gas–gas gas–solid inlet, entering the system counters for various entities particle size level of a particle size distribution type of solid, mass transfer minimum fluidization net (gas) flow reference value, initial, at the top of the riser (h = H 0 ), superficial outlet, leaving the system recycling stream solid particle pyrolysis reaction

© Woodhead Publishing Limited, 2013

Modelling of fluidized bed gasification processes

v t T th ws x ∞

vertical, visible (applied to velocity and flow), volume terminal (applied to velocity) total throughflow withdrawal stream at bed surface (height of bottom zone) above the transport disengaging heigth, at bulk conditions


black box model bubbling fluidized bed bubbling fluidized bed gasifier circulating fluidized bed circulating fluidized bed gasifier computational fluid-dynamic model fluidized bed fluidized bed gasifier/gasification fluidized bed biomass gasifier fluidization model water-gas-shift reaction

© Woodhead Publishing Limited, 2013