Advanced Powder Technology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Advanced Powder Technology journal homepage: www.elsevier.com/locate/apt
Original Research Paper
Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk Cong-Binh Dinh a, Chun-Chung Liao b, Shu-San Hsiau a,⇑ a b
Department of Mechanical Engineering, National Central University, Jhongli City, Taoyuan County 32001, Taiwan, Republic of China Department of Mold and Die Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 80778, Taiwan, Republic of China
a r t i c l e
i n f o
Article history: Received 29 July 2016 Received in revised form 30 September 2016 Accepted 12 October 2016 Available online xxxx Keywords: Fast pyrolysis Fluidized-bed reactor Hydrodynamics Heat transfer Eulerian multi-fluid model
a b s t r a c t The present study investigates the hydrodynamics and heat transfer phenomena that occur during the biomass fast pyrolysis process. A numerical approach that combines a two-dimensional Eulerian multi-fluid model and the kinetic theory of granular flow has been applied to simulate the gas-solid flow in a bubbling fluidized-bed reactor. In this study, rice husk and quartz sand with specified properties were used as biomass and inert material, respectively. Our model was first validated the feasibility using previous findings, then an extensive parametric study was conducted to determine the effects of the major variables, especially the size of rice husk particles, on the flow distribution and the heat transfer between the phases. The concept of standard deviation attributed to the dispersion of solid volume fraction was used to calculate the intensity of segregation. The simulated results indicated that the mixing of binary mixture was strongly affected by different sizes of rice husk particles. The heat transfer occurring inside the fluidized bed was described by the distribution of solids temperature, the variation of surface heat flux and heat transfer coefficient. Both heat transfer quantities were observed to be dominant in the dense bed regions as they strongly depend on the solids concentration in the fluidized bed. The increasing inlet gas velocity promoted the mixing of solid particles, thus resulted in the effective heat transfer from wall to particles and between the particles. Ó 2016 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.
1. Introduction When the need for reducing fossil fuels consumption and carbon dioxide emission is considered, modernized bioenergy systems are expected to be important contributors to future sustainable energy systems in both developed and developing countries. Thermo-chemical processes, such as combustion, pyrolysis and gasification, are currently the most common techniques by which various energy products can be produced from raw biomass feedstock for different applications. As bio-oil attributed to high energy density could be stored and transported to anywhere, fast pyrolysis has been widely studied in recent years. Biomass fast pyrolysis is known as the rapid thermal decomposition of organic materials
Abbreviations: 2-D, two-dimensional; AMG, algebraic multigrid; BFB, bubbling fluidized bed; CFD, computational fluid dynamics; FBR, fluidized bed reactor; HTC, heat transfer coefficient; KTGF, kinetic theory of granular flow; MFM, multi-fluid model; SIMPLE, semi-implicit method for pressure-linked equations. ⇑ Corresponding author. Fax: +886 3 425 4501. E-mail address:
[email protected] (S.-S. Hsiau).
at elevated temperatures in the absence of oxygen to produce syngas, bio-oil and bio-char [1]. Among various types of reactors chosen for fast pyrolysis process, fluidized bed reactor (FBR) is a very popular choice of design as they offer many advantages, such as simple construction and operation, effective fluid-solid contact, high heat transfer rates, and the ability of handling various materials [2]. Fluidized beds often involve in the mixtures of solid particles with different sizes, shapes, and densities, which usually tend to separate during fluidization. The mixing/segregation behavior of these mixtures in a fluidized bed is greatly important for both industries and research since it has specific influences on bed expansion, chemical reactions, heat transfer and mass transfer characteristics [3,4]. Various papers in the literature have been reported in that subject from different aspects of fluidization conditions and particle properties. Rowe and Nienow [5] were among the first researchers who investigated the segregation of binary mixtures of different particle densities and sizes in a bubbling fluidized bed (BFB). They proposed the terms ‘‘flotsam” and ‘‘jetsam” to represent the particles occupying the upper and the lower of the bed, respectively [6]. The
http://dx.doi.org/10.1016/j.apt.2016.10.013 0921-8831/Ó 2016 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
2
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
Nomenclature Cp d D ess g hi hsurface hgs H k k M n Nu p Pr q00 Q gs Re S t T V, v
specific heat capacity, J/(kgK) particle diameter, m bed width, m restitution coefficient, dimensionless gravitational acceleration (=9.81 m/s2), m/s2 specific enthalpy, J/kg surface heat transfer coefficient, W/(m2K) volumetric heat transfer coefficient, W/(m2K) bed height, m thermal conductivity, W/(mK) turbulent kinetic energy, m2/s2 mixing index, dimensionless number of nodes, nodes Nusselt number, dimensionless pressure, Pa Prandtl number, dimensionless surface heat flux, W/m2 quantity of interphase heat transfer, W/m2 Reynolds number, dimensionless segregation index, dimensionless time, s temperature, K velocity, m/s
homogeneity or mixing behavior of binary mixture of solids has generally been characterized by the mixing index or segregation index. To date, various expressions of mixing index have been proposed by many authors [5–8]. A large number of experiments have been made to characterize the segregation/mixing of solid mixtures in gas-solid fluidized beds [9–13]. Oliveira et al. [13] investigated the effect of particle size on the hydrodynamics of binary mixtures composed of different types of biomass and sand. They reported that the higher diameter ratios of sand/biomass led to more noticeable bed segregation or the reduction in the fluidization quality. Besides experimental researches, various numerical studies on the hydrodynamics of multiphase FBRs have been performed [3,4,7,14–17]. The major emphasis of those studies was put on elucidating the segregation and mixing mechanisms of solid particles in dense gas fluidized beds. Gibilaro and Rowe [14] were the pioneers who developed a simple model (G-R model) describing the particle segregation in a binary mixture of solids. Accordingly, Bilbao et al. [15] adapted the G-R model to predict the hydrodynamics of a non-steady fluidized bed consisting of sand and straw. Sun et al. [16] applied a multi-fluid model (MFM) based on the kinetic theory of granular flow (KTGF) to investigate the flow behavior of a binary mixture of sand/rice husk particles in a BFB. They found that the superficial gas velocity, the particle size and the mass fraction of sand particles had considerable influences on the segregating behavior of rice husk particles. Sharma et al. [17] used both two-dimensional (2-D) and three-dimensional Eulerian MFMs to describe the hydrodynamics of a biomass/bio-char mixture in a BFB. They found that an increasing of superficial gas velocities resulted in the better mixing of the solid phases in the fluidized bed. In relation to momentum transport, the heat transfer in fluidized beds has also been an important aspect of concern in this field. Three main thermal processes in gas-solid fluidized beds referred to wall-to-bed, gas-to-particle and particle-to-particle heat transfer, have been widely studied over the years for different cases, using different approaches. Schmidt and Renz [18] used both the Eulerian approach and the KTGF to predict the fluid dynamics
Greek letters volume fraction, dimensionless b drag force coefficient, dimensionless Dp pressure drop, Pa e turbulent dissipation rate, m2/s3 q density, kg/m3 l dynamic viscosity, shear viscosity, Pas = Ns/m2 u specularity coefficient, dimensionless s shear stress, N/m2
a
Subscripts 0 initial value at t = 0 bio biomass phase bulk bulk phase g gas i general index mf minimum fluidization s solid sand sand phase w wall
and the influence of bubbles on the heat transfer in a gas-solid fluidized bed. Their results showed a strong relation between the local distribution of solid volume fraction and heat transfer coefficient (HTC). Armstrong et al. [19] conducted an extensive parametric study for different restitution coefficients, particle sizes and inlet velocities in a heated wall bubbling fluidized bed (BFB) using the two-fluid Eulerian model coupled with the KTGF. Two drag models, namely Gidaspow model and Syamlal-O’Brien model, were compared to determine its effects on the particle distribution and the wall-to-bed heat transfer. In addition, heat transfer and hydrodynamics of an unsteady gas-solid flow at different superficial gas velocities were described by the Eulerian MFM, incorporated with the KTGF and the standard ke turbulence model [20–22]. A series of investigations based on the dynamics of biomass particles in BFB reactor were performed by using a combination of an Euler approach for the nitrogen gas and sand phases and a Lagrange approach for the biomass particle [2,23,24]. Accordingly, the fluid-particle interaction and the influences of biomass particle size, shape and heat transfer conditions on pyrolysis of biomass in a laboratory-scale FBR were reported. Because of the lack of comprehensive understanding, the design of industrial FBRs for the biomass pyrolysis was primarily based on empirical correlations and experiments in laboratory and pilotscale units. Whereas the measurement of important physical variables in either reactive or non-reactive fluidized beds is still a challenge due to the dense flow of particles and extremely high temperature, a well-defined multiphase flow model should be an alternative technique for investigating the role of the key parameters in our processes of concern. The main objective of this study was to provide a better understanding of biomass pyrolysis with an emphasis on the solid particles behavior and associated heat transfer mechanisms that occur during pyrolysis process. In addition, an extensive parametric study was carried out for a variety of the major influences, such as inlet gas velocity, drag models, mixture composition, and biomass/sand size ratio to determine their effects on the flow distribution and hence the heat transfer performance. These works were mainly based on
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
3
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
Fig. 1. The reactor component and corresponding geometry used in the simulations.
Table 1 Simulation parameters. Properties
Values
Comments
Bed width, W Static bed height, H0 Biomass density, qbio Biomass particle diameter, dbio Biomass specific heat capacity, Cp,bio Biomass thermal conductivity, kbio Sand density, qsand Sand particle diameter, dsand Sand specific heat capacity, Cp,sand Sand thermal conductivity, ksand Gas density, qg Gas viscosity, lg Gas specific heat capacity, Cp,g Gas thermal conductivity, kg Inlet gas velocity, V0,g Initial void fraction, a0,g Restitution coefficient, ess Specularity coefficient, u
0.042 m 0.1 m 380 kg/m3 440; 1540; 2640; 3520; 4400 lm 1200 J/(kg K) 0.036 W/(m K) 2650 kg/m3 440 lm 835 J/(kg K) 0.35 W/(m K) 0.437 kg/m3 3.42 105 kg/(m s) 1115.52 J/(kg K) 0.0535 W/(m K) 0.5; 0.7 m/s 0.4 0.9 0.05
Fixed value Fixed value Rice husk 05 different cases Fixed value Fixed value Quartz sand Fixed value Fixed value Fixed value Nitrogen (773 K) Nitrogen (773 K) Nitrogen (773 K) Nitrogen (773 K) Superficial velocity Fixed value Both solid phases Fixed value
an Eulerian MFM, using commercial CFD code ANSYS FLUENT version 13.0.0, together with some supporting tools. 2. Methodology 2.1. Model description A laboratory-scale bubbling FBR for biomass fast pyrolysis was designed and fabricated to study gas-solid flows and heat transfer characteristics. The reactor vessel was a stainless steel cylinder with 4.2 cm in diameter and 34 cm in height. Accordingly, the sim-
References
[25,26] [17,25] [26] [26] [27] [23] [23] [23] [28] [28] [28] [28] [17,23] [21,23] [29]
ulations were implemented in a 2-D model based on actual sizes of the reactor component, as shown in Fig. 1. In this reactor, the biomass fuel (rice husk) and inert bed material (quartz sand) were fluidized by a fluidizing agent (nitrogen gas) to undergo the fast pyrolysis process. The model simply considered the fluid flow and heat transfer behaviors without chemical reactions occurring between the phases in a fluidized-bed pyrolyzer. Several trial runs were conducted to select appropriate bed height, gas superficial velocities, particle properties and other parameters. Eventually, a final set of parameters used in the simulations is described in Table 1.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
4
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
2.2. Mathematical model
gas
to
sand
*
In a typical biomass pyrolyzer, there exist three phases such as gases (fluidizing nitrogen and producer gas), solid biomass particles (virgin biomass and pyrolysis char), and inert bed media (sand). These phases have completely different properties which cannot be described by the classical two-fluid model. For this reason, a comprehensive MFM with numerical equations for one gas phase and two solid phases was used in this study, as briefly described below. Here, we considered a binary mixture of spherical smooth biomass and sand particles, and assumed a closed system with no mass transfer and reaction between the phases. The CFD simulation approach chosen in this work was based on the combination of an Eulerian MFM and the KTGF in which equations were written under incompressible, transient and turbulent flow conditions. Accordingly, the conservative equations, i.e., continuity, momentum (for gas and solid phases) and fluctuation energy equation (for granular or solid phases) were numerically solved by the finite volume method using the commercial CFD code ANSYS FLUENT 13.0.0 in a Cartesian coordinate system. 2.2.1. Multiphase flow governing equations The MFM is described as the pseudo-fluid interaction of solid phases with the gas phase in which the presences of all phases are attributed to the volume fractions denoted by ai. The concept of volume fraction here refers to the space occupied by each phase, and then the conservation of mass and momentum are individually satisfied for each phase. Each computational cell is occupied by interpenetrating phases, so that the sum overall volume fractions is unity:
ag þ
X
as ¼ 1
ð1Þ
s¼bio;sand
2.2.1.1. Conservation equations. Continuity equations for the gas and solid phases (conservation of mass)
@ðag qg Þ * þ r ðag qg v g Þ ¼ 0 @t * @ðas qs Þ þ r ðas qs v s Þ ¼ 0 @t
ð2Þ ð3Þ
Momentum equations for the gas and solid phases (conservation of momentum) + Gas phase: *
@ðag qg v g Þ * * g þ ag q þ r ðag qg v g v g Þ ¼ ag rp þ r s g @t *
*
*
g þbgs ðv g v s Þ
ð4Þ
+ Solid phases: *
* * @ðabio qbio v bio Þ þ r ðabio qbio v bio v bio Þ ¼ abio rp rpbio @t * * * * bio þ abio q * þrs bio g þbg-bio ðv g v bio Þ þ bbio-sand ðv bio v sand Þ
ð5Þ *
* * @ðasand qsand v sand Þ þ r ðasand qsand v sand v sand Þ ¼ asand rp rpsand @t * * * * * sand þ asand q þrs sand g þbg-sand ðv g v sand Þþbsand-bio ðv sand v bio Þ
ð6Þ *
*
*
*
where, bg-bio ðv g v bio Þ, bg-sand ðv g v sand Þ are the drag forces between the gas and solid phases (gas to biomass particles and
particles,
respectively);
*
*
*
bbio-sand ðv bio v sand Þ,
bsand-bio ðv sand v bio Þ are the drag forces between the solid phases (biomass to sand and vice versa). According to Newton’s second law, the change of momentum is equal to the net force acting on a domain. In a gas-solid fluidized bed, the forces consist of the viscous force, body force, static pressure force, solid pressure force and interphase drag forces. Energy equations for the gas and solid phases (conservation of energy) The energy equations take into account the thermal conduction within the phases, the heat exchange between gas and solid phases, the viscous dissipation and the term involving the expansion work of the void fraction. Here for simplicity, we neglected the heat source induced by chemical reaction and radiation. + Gas phase:
X @pg * * @ g : r* ðag qg hg Þ þ r ðag qg v g hg Þ ¼ ag v g r qg þ Q gs þs @t @t s¼bio;sand ð7Þ + Solid phase:
X * * * @ @p ðas qs hs Þ þ r ðas qs v s hs Þ ¼ as s þ s Q sg s : rv s r q s þ @t @t s¼bio;sand ð8Þ where hi is the specific enthalpy of the ith phase, hi ¼ *
R Ti T ref
*
C p; i dT i ; q i
is the heat flux of the ith phase, q i ¼ r ðai ki rT i Þ; Q gs and Q sg are the intensity values of heat exchange between the gas and solid phases. Here, the heat exchange between the phases must comply with the local balance conditions, Q gs ¼ Q sg and Q gg ¼ Q ss ¼ 0. The heat exchange between the gas and solid phases are the functions of the temperature difference between the gas and solid phases:
Q gs ¼ hgs ðT g T s Þ; W=m2
ð9Þ
where hgs are the volumetric heat transfer coefficients (HTC) between the gas and solid phases. They are related to the particle Nusselt number using the following equation:
hgs ¼
6kg ag as Nus 2
ds
; W=ðm2 KÞ
ð10Þ
where kg is the thermal conductivity of the gas phase and hgs should tend to zero whenever one of the phases is not present within the domain. In this study, we used the Nusselt number correlation by Gunn [32] which is applicable to a porosity range from 0.35 to 1.0 and a Reynolds number of up to 105 (0.35 6 ag 6 1.00 and Res 6 105) [30].
Nus ¼
hgs ds 1=3 ¼ ð7 10ag þ 5a2g Þ 1 þ 0:7Re0:2 s Pr kg 1=3 þ ð1:33 2:4ag þ 1:2a2g ÞRe0:7 s Pr
where, Res ¼
qg ds j~ v g ~ vsj lg
and Pr ¼
ð11Þ
C p;g lg . kg
Heat transfer between the heated wall and the bulk phases (gas and solid phases) [30]
hsurface ¼
q00 ; W=ðm2 KÞ T w T bulk
ð12Þ
where hsurface is the wall-to-bed HTC or surface HTC; q00 is the surface heat flux; T w is the wall temperature; and T bulk is the bulk phases temperature.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
5
2.2.1.2. Constitutive relations. Conservation equations of mass, momentum, and granular temperature can be solved with suitable constitutive equations and boundary conditions. These constitutive equations specify how the physical parameters of the phases interact with each other. In addition, various equations of the momentum transfer between the phases, the KTGF for granular phases, turbulent model, and heat transfer model were applied for this study. They can be found in ANSYS FLUENT Theory’s guide [30], and several related studies [20,23,25,31,32].
Table 2 Boundary conditions used for the simulation.
2.2.2. Mixing/segregation index The concept of standard deviation was employed to calculate the intensity of segregation (or segregation index). Accordingly, the standard deviation was represented the varying amount of the dispersion of solids’ volume fraction from its mean value. The segregation index for quantifying the segregation degree could be defined as [33]
sonably chosen for our problem. The computational domain was divided into a finite number of control volumes. The velocity components were calculated for the points lying on the faces of the control volume (cell edges), whereas all other variables were calculated at the center of each control volume (cell center). The conservation equations were solved using the phase coupled SIMPLE algorithm and a second-order implicit scheme for transient formulation with a time step size of 0.001 s. The SIMPLE algorithm uses a relationship between velocity and pressure corrections to enforce mass conservation and to obtain the pressure field. In other words, a pressure-correction equation is built based on total volume continuity and then pressure and velocity are corrected to satisfy the continuity constraint. Consequently, the pressure-correction equation and the equations formed by velocity components of the phases are solved using the algebraic multigrid (AMG) scheme in a segregated manner. The energy and turbulence equations are then solved based on the obtained results of the phases. Herein, the set of resulting algebraic equations were solved iteratively. The relative error between two consecutive iterations was specified by using a convergence criterion of 103 for each scaled residual component. A second-order spatial discretization method was used to improve the accuracy of the simulations [30]. In order to obtain statistically reasonable results, all the simulations were run for 60 s in total, and the time-average results were taken from 10 s to 60 s by which a dynamic steady state had been obtained. Accordingly, all the results of solid volume fractions and other quantities were collected and computed as time-mean values. The raw results from FLUENT post-processing were then processed by other codes and tools (Matlab, OriginPro, MS Excel, Paint, etc.) to obtain the final desired results. In addition, gridindependence test should be done as a sufficient condition for getting acceptable results. It was indicated from some previous studies in the analysis of proper grid sizes used in DEM-CFD simulations that a reliable prediction of the fluid dynamics and bubble behaviors in a fluidized bed was still obtained with a coarse grid size of 5–10 times higher than the particle diameters [34–36]. Moreover, the characteristics of the fluidized bed was only evaluated at mesoscale structure with a small enough grid size to avoid the time consuming of grid-independent study. Accordingly, a sensitivity analysis was conducted using two grid sizes of 1 mm and 2 mm. It was found that there was insignificant difference between the test results of two grid sizes. Therefore, a 2-mm mesh was used in this study to achieve acceptable results with less computational cost.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pn 2 i¼1 ðas;i as Þ S¼ n1
ð13Þ
s is the average where as;i is the solid volume fraction of ith node, a solid volume fraction in the whole bed and n denotes the number of nodes (n = 3758 in this study). The segregation index ranges from 0 to 1, where S = 1 indicates the completely segregated state and S = 0 represents the fully mixed state. In this sense, the mixing index, denoted by M, could be determined by the relation of, M = 1–S. 2.3. Model implementation 2.3.1. Initial and boundary conditions In all simulations, the initial values of the variables (ag, abio, asand, Tg, Tbio, Tsand, vg, vs, k, e, etc.) were assigned for the whole computational domain. Initially, the velocities of the two solids and nitrogen gas were set at minimum fluidization condition. An initial bed height of given volume fractions of biomass and sand was specified at the beginning of each run. The pressure was assumed to be atmospheric pressure at the outlet. The values of turbulence intensity were properly determined for gas phase at inflow boundary condition. Basically, the fluidizing gas of nitrogen was supplied from the bottom of the reactor at a stable temperature of 773 K. The wall temperature was assumed to maintain at a constant of 873 K needed for the required heat of fast pyrolysis process. In addition, it was found from some first trial results that the biomass and sand particles at least needed to be pre-heated to the temperatures of 373 K and 823 K, respectively. All heat sources contributed to heat the bed materials to the condition of fast pyrolysis process, i.e., the bed temperature attained approximately 773 K. Accordingly, it was assumed that the biomass particles were initially fed into the hot sand bed with a given solid volume fraction, and then the entire bed was pre-fluidized for a few seconds until it achieved a uniform-mixed condition. For the aforementioned reasons, the initial void fraction, the initial temperatures of biomass particles, sand particles, inlet nitrogen gas and walls were set to 0.4, 373 K, 823 K, 773 K and 873 K, respectively. It is noted that the minimum fluidization velocity of sand was calculated and used as a reference for choosing the inlet gas velocity at the bottom of the bed (since Umf,sand > Umf,rice-husk). Some main descriptions for boundary conditions used in our model are shown in Table 2. 2.3.2. Solution procedure All our works were performed in 2-D to reduce computational effort. A pressure-based solver with segregated algorithm was rea-
Zones
Type
Descriptions
Inlet
Velocityinlet
Outlet
Pressureoutlet Stationary wall
Uniform distribution was assumed for the gas phase. / Thedirection of the gas flow was normal to the surface. / Solidparticles were assumed impenetrable through the bottom wall. At atmospheric pressure (fixed value)
Wall
Constant wall temperature / No slip for gas phase / Partial slip for solid phases
3. Results and discussion 3.1. Generalized flow behaviors of binary mixture in the fluidized bed 3.1.1. Concentration of the solid phases The concentration variation of biomass/sand particles along the bed height is due to the dynamic equilibrium state between the mixing and segregation of particles occurring during the fluidization. Fig. 2a and b respectively illustrates the biomass and sand
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
6
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
bed. The bed of solid particles considered as the obstruction to the gas flow reduces the pressure in the bed. The graph of pressure drop versus time is the representation of drag force variation on solid particles. An increase in drag force increases pressure drop through the granular bed. The bed pressure drop reaches its maximum at the onset of the gas injection due to the large drag force. However, when the granular bed reaches the equilibrium state, the pressure drop decreases and then fluctuates about an average value. The pressure drop fluctuation may be attributed to bubbles’ forming, stretching, coalescing and going through the fluidized bed. Fig. 3 shows the fluctuations of pressure drop between the inlet and outlet of the reactor versus time for different initial volume fractions of biomass. The results indicate that the pressure drop increases with decreasing the initial volume fraction of biomass. This trend can be explained by the bed density with the dominance of sand particles. In other words, the lower biomass percentage or the higher sand percentage results in the greater bed pressure drop. 3.1.3. Velocity vectors The mixing of solid particles is promoted by the solids motion around the rising bubbles. The mixing/segregation mechanism can be analyzed by determining the velocity vectors of solid species in the fluidized bed. For this purpose, the velocity vector profile of biomass versus simulation time for Syamlal-O’Brien drag model is shown in Fig. 4. It can be observed that biomass particles circulate in the bed due to the movement of gas bubbles along the bed height. In the upper part of the bed, the biomass particles tend to move downwards due to collapsing of gas bubbles. In general, particles mainly move upwards at the core region and fall down in the near-wall region referred to as core-annular flow pattern. As biomass particles fall, they may collide with other rising particles to form some particle clusters which vigorously move upwards and downwards. These movements of both biomass and sand particles in the bed result in the strong mixing and the appearance of some vortices in the field of velocity vector. We also see that the solid velocity near the upper-bed region is higher than that in the lower part (Fig. 4). Fig. 2. A comparison of biomass and sand concentrations along the reactor height using different drag models and inlet gas velocities. (a) rice husk, (b) sand.
3.1.4. Static temperature of solid phases The distributions of static temperature of biomass and sand are shown in Fig. 5a and b, respectively. At the beginning of fluidiza-
concentrations along the reactor height using different inlet gas velocities and drag models (Syamlal-O’Brien model and Gidaspow model). By comparing the distributions of solids’ volume fraction from these two figures, we can see that the biomass particles mainly accumulate in the middle and upper regions of the bed (Fig. 2a), whereas sand particles sink to the bottom of the reactor (Fig. 2b). It can be noted that this segregation occurring during the settling of the granular bed is inevitable. In addition, the solid volume fractions in the freeboard are almost zero as no particles’ presence is observed in this region. Moreover, with increasing inlet gas velocity, the distributions of biomass volume fraction become more uniform along the bed height. At lower inlet gas velocities, no significant discrepancy is observed for different drag models. However, the results show some differences between two drag models at higher gas velocities. This indicates the significant effect of gas velocity in drag functions, and thus in solid concentrations. In addition, the different degrees of bed expansion during the fluidization are attributed to the change in bubbling behavior induced by different gas velocities, drag models and other factors. 3.1.2. Pressure drop Bubble motion behavior and solid concentration in the fluidized bed are the crucial parameters causing pressure fluctuations in the
Fig. 3. The averaged fluctuations of bed pressure drop versus time for different initial volume fractions of biomass a0,bio (V0,g = 0.5 m/s, Gidaspow drag model).
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
7
Fig. 4. Velocity vector profile of biomass particles during simulation time with a typical case (Syamlal-O’Brien drag model). Unit: m/s.
tion, the heat transferred from hot sand particles to rice husk particles increases significantly due to the higher specific heat capacity and thermal conductivity of the sand particles. The hot surface of reactor and hot gas also provide an additional heat to the bed of solid particles. As seen in Fig. 5a, the biomass temperature rapidly increases in time along with the bed expansion as a result of heat transfer from the heat sources. In contrast to the biomass temperature, the sand temperature gradually decreases as its heat is transferred to biomass particles (Fig. 5b). Once the two solids have reached an identical temperature after 2 s, the thermal equilibrium state is established and there is no more heat transfer between the two species. In the core region of the bed, the solid temperatures are relatively homogeneous, whereas they are higher in the wall region. At 60 s, the biomass temperature in the whole domain is fairly homogeneous, oscillating from 780 K to 820 K. It can be explained that as the gas bubbles rise along the heated walls, solid particles change in their positions, and thermal energy is diffused through the bed due to the strong mixing caused by the random motion of these bubbles. It can be noted that the biomass particles need almost 2–3 s to reach the reactive temperature required for pyrolysis, i.e., approximately 773 K (or 500 °C), as the residence time of fast pyrolysis is close to 2 s. Since solids concentrated in the freeboard are mainly emulsion phases, the surface areas between gas-emulsion contact and wallemulsion contact are relatively large. Therefore, the temperature gradients between gas-solids and wall-solids in the upper bed are greater than those in the lower bed. In addition, the thermal diffusion mainly occurs in the emulsion phases, and thus higher temperatures are expected in the regions of larger concentration of solid emulsion (splash region). 3.2. Effect of size ratio of biomass-sand mixture on mixing and heat transfer performance 3.2.1. Effect of size ratio of biomass-sand mixture on solid mixing Fig. 6a and b respectively shows the concentrations of biomass and sand along the reactor height for various size ratios of these two solids. The differences in density and size between the two different species contribute to the solid segregation in the bed. In general, due to the dominance of density segregation, the lighter component (rice husk) tends to float up to the top of the bed
(Fig. 6a), whereas the heavier component (sand) tends to sink to the bottom of the bed (Fig. 6b). Besides that, the smaller the size ratio of biomass-sand is used, the higher the bed expansion is obtained for both biomass and sand particles. It can be noted that the mixtures of the smaller size ratios seem to be easier in moving up to the top region of the bed. For different size ratios of the biomass-sand mixture, the segregation indexes are calculated using Eq. (13), and then its instantaneous values versus time are shown in Fig. 7. From this figure, the segregation index of sand increases with the increasing size ratio of the binary mixture. Initially, biomass and sand are uniformly mixed with the segregation index of zero. As time evolves, because of the fluidization, both solid species disperse towards different directions; and thus the segregation occurs. However, the degree of segregation is not much due to the good mixing nature of fluidized bed. The results show that the smaller the size of biomass particles, the better the mixing of the binary mixture. Also, it can be observed that the best mixing is obtained for the case of the lowest ratio (dbio:dsand = 1:1). A qualitative agreement can be found in experimental or/and modeling results of other previous studies [13,37]. This conclusion will be a helpful reference for using proper sizes of biomass feedstocks in pyrolysis process. The concept of granular temperature has been commonly used in the researches of particle motion behavior. The granular temperature is a measurement of random oscillations of the particles and is defined as one-third of the mean square particle velocity fluctuations [25,31]. Fig. 8a and b respectively describes the distributions of biomass and sand granular temperature along the reactor height for different mixture size ratios. The results show that the granular temperature of sand particles is lower than that of biomass particles. This may arise from the fact that heavier sand particles possess lower velocity in the fluidized bed. Because of the shear rate variation from collision of particles and the effect of bubbles collapsing, the granular temperatures of both solid species are relatively higher near the upper region of the bed than other regions. Basically, these observations are consistent with the results reported by other previous studies [3,16]. We also see that the distribution of biomass granular temperature becomes more uniform in the bed region with greater size ratios of the biomass-sand mixture (Fig. 8a). Especially, the sand granular temperature is still high in splash zone with the size ratio of 1:1 (Fig. 8b).
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
8
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
Fig. 5. Static temperature profiles of the two solid phases during simulation time with a typical case (a) rice husk, (b) sand. Unit: K.
3.2.2. Effect of size ratio of biomass-sand mixture on heat transfer performance The time-mean surface heat flux and HTC versus different size ratios of the two solids are simultaneously described in Fig. 9. With the increasing size ratio of biomass-sand, both surface heat flux and HTC decrease as biomass particles absorb less amount of heat transferred from hot surfaces. In other words, the transfer of heat from hot surfaces to smaller biomass particles is greater than to bigger ones. This may be due to the fact that the specific surface area of the smaller particles is larger than that of bigger particles at the same condition of initial volume fraction. In addition to heat transfer, the particle mixing is also an issue of concern. The variation of heat flux closely relate to the change of the bed porosity and the mixing of solid particles. According to our earlier results, the solid mixing is diminished by the increase of the size ratio of binary mixture. The segregation may restrict the contact between biomass and sand particles, and thus it leads to the reductions in the surface heat flux and HTC. Fig. 10 shows the variations of static temperature of biomass versus time for different size ratios of binary mixture. In general,
the greater the biomass/sand size ratio is, the slower the increment of biomass temperature is until the bed reaches the required temperature within the first few seconds of pyrolysis process. For the size ratio of 1:1, the biomass temperature rapidly increases to reach the pyrolysis temperature at about 1–2 s right after the onset of the fluidization. Due to the strong mixing and high heat transfer rate in a fluidized bed, the bed is heated quickly and uniformly as the fluid dynamics evolves in time. As expected, biomass particles can be heated to a relatively high bed temperature in a very short time, and then the increase of its temperature becomes slower as it reaches the bed steady state after 20 s. During the steady-state period, the variations of biomass temperature for different biomass/ sand size ratios are nearly similar to each other. Moreover, the preceding discussion can be supported by the fluctuation in time of total surface heat flux (Fig. 11). Initially, surface heat flux rapidly increases as the temperature difference between hot surfaces and biomass particles is still high. As time goes on, the temperature difference is gradually diminished as the biomass temperature approaches the temperature of hot surfaces leading to a reduction in surface heat flux.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
Fig. 6. Solid volume fractions along the reactor height for different size ratios of biomass-sand mixture. (a) biomass, (b) sand.
Fig. 7. Variations of segregation index versus time for different size ratios of biomass-sand mixture.
9
Fig. 8. Solid granular temperature along the reactor height for different size ratios of biomass-sand mixture. (a) biomass, (b) sand.
Fig. 9. Variations of surface heat flux and HTC versus different size ratios of biomass-sand mixture.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
10
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx
regions of the bed, whereas sand particles accumulated at the bottom of the reactor. Due to the strong mixing and high heat transfer rate, the whole bed was heated quickly and uniformly, leading to a fairly homogeneous distribution of biomass temperature in the bed region. The increasing inlet gas velocity promoted the mixing of solid particles, thus resulted in effective heat transfer between the particles and from gas to particles. It was found that the particle size and thermal boundary conditions of the concerned phases were among the most important parameters affecting gas-solid flow behavior and heat transfer performance in a fluidized-bed pyrolyzer. Although only the simulation results were presented, they could provide some important contributions for the improvement of the system performance. In general, the model is capable of predicting the flow behavior, solid mixing, bed temperature and heat transfer from the heated surfaces to the bed.
Acknowledgements Fig. 10. Variations of static temperature of biomass versus time for different size ratios of biomass-sand mixture.
This work was supported by The Ministry of Science and Technology of the Taiwan R.O.C., through Grant, MOST 104-3113-E008-001.
References
Fig. 11. Variations of total surface heat flux versus time for different size ratios of biomass-sand mixture.
Although experiment results were lacked in this study due to some limitations, our numerical results were validated by the agreement with previously published researches. In general, the proposed model along with the achieved results could be acceptable within the scope of this study.
4. Conclusions The hydrodynamics and heat transfer of biomass pyrolysis process in a 2-D fluidized-bed reactor have been investigated by applying CFD simulation. Unsteady behavior of a binary mixture of rice husk/sand has been simulated by using the continuum Eulerian MFM coupled with the KTGF. The distribution and mixing of the two solids were strongly influenced by the inlet gas velocity and the initial volume fraction of biomass. Better particle mixing could be obtained by increasing the inlet gas velocity and the percentage of biomass. The results indicated that biomass particles mainly distributed in the middle
[1] Q. Xiong, S.C. Kong, A. Passalacqua, Development of a generalized numerical framework for simulating biomass fast pyrolysis in fluidized-bed reactors, Chem. Eng. Sci. 99 (2013) 305–313. [2] K. Papadikis, A.V. Bridgwater, S. Gu, CFD modelling of the fast pyrolysis of biomass in fluidised bed reactors. Part A: Eulerian computation of momentum transport in bubbling fluidised beds, Chem. Eng. Sci. 63 (16) (2008) 4218– 4227. [3] L. Huilin, H. Yurong, D. Gidaspow, Y. Lidan, Q. Yukun, Size segregation of binary mixture of solids in bubbling fluidized beds, Powder Technol. 134 (2003) 86– 97. [4] M. Coroneo, L. Mazzei, P. Lettieri, A. Paglianti, G. Montante, CFD prediction of segregating fluidized bidisperse mixtures of particles differing in size and density in gas–solid fluidized beds, Chem. Eng. Sci. 66 (2011) 2317–2327. [5] P.N. Rowe, A.W. Nienow, Particle mixing and segregation in gas fluidized beds, Powder Technol. 15 (1976) 141–147. [6] A.W. Nienow, N.S. Naimer, T. Chiba, Studies of segregation/mixing in fluidized beds of different size particles, Chem. Eng. Sci. 62 (1987) 52–63. [7] B.G.M. van Wachem, J.C. Schouten, C.M. van den Bleek, R. Krishna, J.L. Sinclair, CFD modeling of gas-fluidized beds with a bimodal particle mixture, AIChE J. 47 (2001) 1292–1302. [8] G. Olivieri, A. Marzocchella, P. Salatino, Segregation of fluidized binary mixtures of granular solids, AIChE J. 50 (12) (2004) 3095–3106. [9] M.J.V. Goldschmidt, J.M. Link, S. Mellema, J.A.M. Kuipers, Digital image analysis measurements of bed expansion and segregation dynamics in dense gas fluidised beds, Powder Technol. 138 (2–3) (2003) 135–159. [10] H.P. Cui, J.R. Grace, Fluidization of biomass particles: a review of experimental multiphase flow aspects, Chem. Eng. Sci. 62 (2007) 45–55. [11] Y. Zhang, B. Jin, W. Zhong, Experimental investigation on mixing and segregation behavior of biomass particle in fluidized bed, Chem. Eng. Process. 48 (2009) 745–754. [12] F. Fotovat, J. Chaouki, J. Bergthorson, The effect of biomass particles on the gas distribution and dilute phase characteristics of sand-biomass mixtures fluidized in the bubbling regime, Chem. Eng. Sci. 102 (2013) 129–138. [13] T.J.P. Oliveira, C.R. Cardoso, C.H. Ataíde, Bubbling fluidization of biomass and sand binary mixtures: minimum fluidization velocity and particle segregation, Chem. Eng. Process. 72 (2013) 113–121. [14] L.G. Gibilaro, P.N. Rowe, A model for a segregating gas fluidised bed, Chem. Eng. Sci. 29 (1974) 1403–1412. [15] R. Bilbao, J. Lezaun, M. Menéndez, M.T. Izquierdo, Segregation in straw/sand mixtures in fluidized beds in non-steady state, Powder Technol. 68 (1991) 31– 35. [16] Q.Q. Sun, H.L. Lu, W.T. Liu, Y.R. He, L.D. Yang, D. Gidaspow, Simulation and experiment of segregating/mixing of rice husk-sand mixture in a bubbling fluidized bed, Fuel 84 (2005) 1739–1748. [17] A. Sharma, S. Wang, V. Pareek, H. Yang, D. Zhang, CFD modeling of mixing/ segregation behavior of biomass and biochar particles in a bubbling fluidized bed, Chem. Eng. Sci. 106 (2014) 264–274. [18] A. Schmidt, U. Renz, Numerical prediction of heat transfer in fluidised beds by a kinetic theory of granular flows, Int. J. Therm. Sci. 39 (2000) 871–885. [19] L.M. Armstrong, S. Gu, K.H. Luo, Study of wall-to-bed heat transfer in a bubbling fluidised bed using the kinetic theory of granular flow, Int. J. Heat Mass Transf. 53 (2010) 4949–4959.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013
C.-B. Dinh et al. / Advanced Powder Technology xxx (2016) xxx–xxx [20] M. Hamzehei, H. Rahimzadeh, Experimental and numerical study of hydrodynamics with heat transfer in a gas-solid fluidized-bed reactor at different particle sizes, Ind. Eng. Chem. Res. 48 (6) (2009) 3177–3186. [21] M. Hamzehei, H. Rahimzadeh, G. Ahmadi, Computational and experimental study of heat transfer and hydrodynamics in a 2D gas-solid fluidized bed reactor, Ind. Eng. Chem. Res. 49 (11) (2010) 5110–5121. [22] Mahdi Hamzehei, Study of heat transfer and hydrodynamics in the fluidized bed reactors, in: Aziz Belmiloudi (Ed.), Heat Transfer - Mathematical Modelling, Numerical Methods and Information Technology, InTech, 2011, pp. 331–382. [23] K. Papadikis, S. Gu, A.V. Bridgwater, CFD modelling of the fast pyrolysis of biomass in fluidised bed reactors. Part B: heat, momentum and mass transport in bubbling fluidised beds, Chem. Eng. Sci. 64 (5) (2009) 1036–1045. [24] K. Papadikis, S. Gu, A.V. Bridgwater, Computational modelling of the impact of particle size to the heat transfer coefficient between biomass particles and a fluidised bed, Fuel Process. Technol. 91 (2010) 68–79. [25] D. Gidaspow, Multiphase Flow and Fluidization: Continuum and Kinetic Theory Description, Academic Press, San Diego, 1994. [26] Y. Zhang, A.E. Ghaly, B. Li, Physical properties of rice residues as affected by variety and climatic and cultivation conditions in three continents, Am. J. Appl. Sci. 9 (11) (2012) 1757–1768. [27] Wikipedia, Silicon dioxide, available at http://en.wikipedia.org/wiki/ Silicon_dioxide (Accessed on February 5th, 2016). [28] T.L. Bergman, A.S. Lavine, F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, seventh ed., John Wiley & Sons Inc., USA, 2011.
11
[29] H. Zhong, J. Gao, C. Xu, X. Lan, CFD modeling the hydrodynamics of binary particle mixtures in bubbling fluidized beds - effect of wall boundary condition, Powder Technol. 230 (2012) 232–240. [30] ANSYS FLUENT Theory’s Guide and User’s Guide. Version 13.0.0. ANSYS Inc., 2013. [31] C.K.K. Lun, F.B. Savage, D.J. Jeffrey, N. Chepurnity, Kinetic theories for granular flow: inelastic particles in Couette flow and slightly inelastic particles in general flow field, J. Fluid Mech. 140 (1984) 223–256. [32] D.J. Gunn, Transfer of heat or mass to particles in fixed and fluidized beds, Int. J. Heat Mass Transf. 21 (1978) 467–476. [33] H.M. Li, J.J. McCarthy, Cohesive particle mixing and segregation under shear, Powder Technol. 164 (2006) 58–64. [34] S. Antonyuk, S. Heinrich, I. Smirnova, Discrete element study of aerogel particle dynamics in a spouted bed apparatus, Chem. Eng. Technol. 35 (8) (2012) 1427–1434. [35] L. Fries, S. Antonyuk, S. Heinrich, D. Dopfer, S. Palzer, Collision dynamics in fluidised bed granulators: a DEM-CFD study, Chem. Eng. Sci. 86 (2013) 108– 123. [36] Z. Peng, E. Doroodchi, C. Luo, B. Moghtaderi, Influence of void fraction calculation on fidelity of CFD-DEM simulation of gas-solid bubbling fluidized beds, AIChE J. 60 (6) (2014) 2000–2018. [37] Z. Peng, E. Doroodchi, Y. Alghamdi, B. Moghtaderi, Mixing and segregation of solid mixtures in bubbling fluidized beds under conditions pertinent to the fuel reactor of a chemical looping system, Powder Technol. 235 (2013) 823– 837.
Please cite this article in press as: C.-B. Dinh et al., Numerical study of hydrodynamics with surface heat transfer in a bubbling fluidized-bed reactor applied to fast pyrolysis of rice husk, Advanced Powder Technology (2016), http://dx.doi.org/10.1016/j.apt.2016.10.013