LES-DEM Investigation of Dense Flow in Circulating Fluidized Beds

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 102 (2015) 1446 – 1455

The 7th World Congress on Particle Technology (WCPT7)

LES-DEM Investigation of Dense Flow in Circulating Fluidized Beds Kun Luo*, Shiliang Yang, Junhua Tan, Jianren Fan State Key Laboratory of Clean Energy Utilization, Zhejiang University, Hangzhou 310027, P.R. China

Abstract

In the Eulerian-Lagrangian framework, the dense gas-solid motions in the conceptually-similar gas-solid fluidizing system — internally circulating fluidized bed (ICFB) and circulating fluidized bed (CFB) are modeled using discrete element method coupled with large eddy simulation. In the coupling procedure, the gas motion is resolved at the computational cell level while solid motion is tracked each individually. After validating the proposed model, the intrinsic mechanisms governing the internal movement of gas-solid phases are explored. In the ICFB, the vertical component of gas or solid flux is extremely larger than its lateral one. Large flux of gas phase mainly concentrates in the right region of each chamber. Vigorous transportation of solid phase locates in the two sides of the reactor chamber. The falling intensity of solid motion near the wall of heat exchange chamber enlarges along the bed height. In the CFB, typical internal circulation property of solid phase can be observed in the riser. Moreover, non-uniform distributions of gas-solid properties appear in the riser especially the bottom region of this component. Strong backmixing of solid phase can be observed in the vicinity of wall. © Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ©2015 2014The The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection and peer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese Selection and peer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese Academy Academy Sciences (CAS). of Sciences of (CAS)

Keywords: Circulating fluidized bed; Discrete element method; Particulate flow; Computational fluid dynamics; Multiphase reactor

Nomenclature Fc *

Contact force, N

Corresponding author. E-mail address: [email protected]

1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection and peer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese Academy of Sciences (CAS)

doi:10.1016/j.proeng.2015.01.278

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

Fd Fp g Gg_x Gg_y Gg_z Gs_x Gs_y Gs_z I k M m n pf Sp Δt t Umf uf ui, uj Usx Usy Usz vi Vp ΔV

1447

Drag force, N Far field pressure force, N Gravitational acceleration, m/s2 Gas flux in X direction, kg/(m2s) Gas flux in Y direction, kg/(m2s) Gas flux in Z direction, kg/(m2s) Solid flux in X direction, kg/(m2s) Solid flux in Y direction, kg/(m2s) Solid flux in Z direction, kg/(m2s) Particle moment of inertia, kg·m2 Total number of particles in contact with the current one Torque exerted on particle, N·m Particle mass, kg Number of particle locating in the current cell Pressure, Pa Momentum sinking item, kg/(m2s2) Time step, s Time, s Minimum fluidizing velocity, m/s Fluid velocity in the current cell, m/s Components of gas velocity, m/s Velocity component of solid phase in X direction, m/s Velocity component of solid phase in Y direction, m/s Velocity component of solid phase in Z direction, m/s Particle velocity, m/s Particle volume, m3 Volume of the current cell, m3

1. Introduction Circulating fluidized beds (CFB) have been extensively applied in many chemical processes, such as the catalytic cracks, the power generation and the biomass gasification/combustion [1-3]. The circulating property of solid phase in the system leads to substantial advantages over conventional gas-solid fluidizing reactors, such as the bubbling fluidized bed and the spouted bed. To achieve an optimized system performance, the internally circulating behavior of gas-solid phase in the CFB should be well understood. In all the kinds of CFB, the internally circulating fluidized bed (ICFB) is an apparatus with a partition plate centrally located in the system. The bed is vertically divided into several chambers for different usages, and the solid circulation is established between these chambers of the system. It has been widely used in the coal gasification/combustion [4] and solid waste disposal [5]. With the developments of computational algorithm and capacity, numerical simulation of the gas-solid motion in the fluidizing apparatus has been a powerful tool in exploring the solid transportation mechanism in the fluidizing apparatus [6, 7]. In general, two approaches can be used to model the gas-solid motion in the dense two-phase flow, namely, the Eulerian-Eulerian method and Eulerian-Lagrangian method. In the former approach, both the fluid and solid phases are treated as the continuous mediums, while the solid motion in the later approach is tracked at the particle-scale level [8]. Marschall et al. [9] modeled the gas-solid motion in an ICFB with two-fluid model in the Eulerian framework. The results demonstrated that the height of the annulus is a key factor for the control of the solid circulation. Recently, Feng et al. [10] modeled the gas-solid flow in an ICFB using the Eulerian-Eulerian model with kinetic theory of granular flow. They claimed that the baffle is the key to introduce and control the solid circulation in the system.

1448

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

Besides the ICFB, the full-loop gas-solid motion in the CFB is achieved through the riser, the cyclone, the standpipe and the L-Valve. There has been a continuous numerical interest focusing on the hydrodynamics in this kind of CFB. However, nearly all of them are carried out by means of two-fluid model [11, 12]. Rare report on the flow characteristics of CFB with solid motion tracked each individually is available. Since the particle-scale modeling of solid motion can present more meaningful information for understanding the solid transportation behavior, there is a need to conduct this kind of simulation. Recently, Chu and Yu [13] studied the general gas-solid motion and the distribution of particle-particle or particle-wall interaction force in the CFB with the CFD-DEM coupling approach. However, only qualitative descriptions of gas-solid motions in the riser have been presented. Towards the final goal to develop a high-fidelity numerical platform for the prediction of coal/biomass gasification and combustion in the CFB at the particle-scale level, a 3-D large eddy simulation combined with the discrete element method (LES-DEM) [8] code with capability of high performance parallel computation is developed. In the coupling approach, the gas motion is solved in the Eulerian framework while the solid motion is resolved in the Lagrangian framework. Based on the proposed model, the flow characteristics of gas-solid phase in the ICFB and the CFB have been explored. The obtained results provide information that cannot be easily achieved by measurement and could be used for process optimization and reactor design. 2. Mathematical model 2.1. Governing equations for fluid motion In the LES-DEM coupling approach, the fluid motion is resolved at the computational grid level using the large eddy simulation. The governing equations for the fluid motion are the Navier-Stokes equations taking into account of the presence of particles. It can be formulated as

w( U f H f )



wt

w ( U f H f u f ,i )

w ( U f H f u f ,i ) wt

0

wxi



w ( U f H f u f ,i u f , j ) wx j

(1)

UfH f g H f

wp f wxi



w(H f (W f ,ij  W f ,ij SGS )) wx j

 Sp

(2)

In the above equations, εf is the volume fraction occupied by the fluid phase. t is the time. The symbol ‘~’ stands for the filtering procedure. ρf and pf stand for the density and the pressure of fluid phase, respectively. ui and uj are the components of gas velocity. g is the gravitational acceleration. τf and τfSGS represent the resolved stress tensor and the sub-grid stress tensor of the computational cell, respectively. For short, the model details can be observed in the previous work [8]. The Sp item in the momentum equation stands for the momentum exchange between the fluid and solid phase, and can be expressed as

Sp

¦

n

Fd ,i / ('V )

(3) where Fd,i is the drag force exerted on the particle i. ΔV is the volume of current cell. n is the total number of particles locating in the current cell. i 1

2.2. Governing equations for solid motion The translational and rotational motions of a particle can be governed by the Newton’s second law as

dvi dt d Zi Ii dt

mi

Fd ,i  Fp ,i  ¦ j 1 Fc ,ij  mi g

(4)

¦

(5)

k

k j 1

M ij

1449

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

Here, mi and Ii are the mass and the moment of inertia of particle i, respectively. ωi and vi represent the angular velocity and the translational velocity of particle i, respectively. Mij stands for the torque exerted by the other colliding particles or walls on particle i. k is the total number of particles colliding with the current one. Fc,ij represents the force exerted by the colliding particles or walls on the current one, and it is estimated with the softsphere contact model. For short, the calculation details can be found in the previous work [8]. Fd,i is the drag force exerted on the current particle and can be calculated as

Fd ,i

Vp ,i E gs

H p ,i

( u f  vi )

(6)

Here, uf is the velocity vector of fluid phase in the current cell. vi is the particle volume. βgs is the momentum exchanging coefficient [8]. Fp,i is the pressure gradient force exerted on the particle i. 3. Model validation of LES-DEM coupling Model validation of the LES-DEM coupling approach is carried out to verify the particularity of solid properties referring to the combined experimental and numerical work of Müller et al. [14]. The bed has a dimension of 44×10×30 mm3 and the gas-solid properties are exactly the same as those adopted in the numerical work of Müller et al. [32]. Since the minimum fluidization velocity and the time-averaged values of gas-solid hydrodynamics are the important properties of the bubbling fluidized bed, both of them have been investigated and the detailed comparison between the simulated results and the experimental data can be found in our previous work [8]. The simulated results agree well with experimental results both qualitatively and quantitatively, indicating that the proposed model captures the important dynamical properties of gas-solid phases in the fluidizing apparatus. 4. Results and discussion 4.1. Hydrodynamics of the ICFB 4.1.1

Simulation setup

Fig. 1 Geometrical configuration of the investigated ICFB (unit: mm).

Gas-solid motion in an ICFB has been numerically modeled with the LES-DEM coupling approach. In the simulation, the drag force exerted on a particle is calculated from the correlation proposed by Koch and Hill [34]. As shown in Fig. 1, the geometrical configuration of the investigated ICFB is a cubic column with size of 132×10×600mm3 in the X, Y and Z directions, respectively. A partition plate is vertically inserted into the system with a height of 12mm. Through this plate, the bed is divided into two chambers, namely the reaction chamber (RC)

1450

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

and the heat exchange chamber (HEC). Details of the parameters used in the simulation are listed in Table 1. Table 1. Physical and numerical parameters used in the ICFB simulation Gas phase Density, kg/m3 Viscosity, kg/(m·s) Outlet pressure, Pa Gas velocity of the RC, m/s Gas velocity of the HEC, m/s Minimum fluidizing velocity, m/s Solid phase

1.225 1.8 × 10-5 1.013 × 105 1.2 0.6 0.3

Particle number Initial bed height, mm Density, kg/m3 Diameter, mm Young modulus, Pa Poisson’s ratio Restitution coefficient Friction coefficient Geometrical configuration

92 000 90 1 000 1.20 1.20 × 105 0.33 0.97 0.1

Width × depth × height, mm3 Gap height, mm Baffle height, mm Grid number

132 × 10 × 600 12 78 44 × 3 × 200

4.1.2

Gas-solid motion in the ICFB

Due to the unequal gas flow rate introduced into the system, the internal circulation of solid phase is established in the two chambers of the system. The interaction between these two chambers combined with the influence of operating parameters and geometrical configuration has been presented in our previous work [8]. Due to the special circulation of solid phase in the system, the flux distributions of gas-solid phases in the system are extremely important for the design and operation of the system since identifying the regions with strong transportation intensity of gas-solid motion gives a direction of performance optimization in some condition by inserting the internal device. Fig. 2 illustrates the contour plots of the time-averaged fluxes of gas and solid phases in the central slice of the ICFB. Due to the large gas flow rate introduced into the RC, larger flux of gas phase can be obtained in this chamber as compared with that in the HEC. In the RC, gas flow mainly passes through the right part of this chamber. Limited by the wall effect, the gas flux in the vicinity of wall is small. In the HEC, the small gas flux in the vicinity of baffle is mainly due to the dense flow of solid motion in this region [8]. For the solid motion, as illustrated in Fig. 2(b), several regions of vigorous solid transportation can be observed. One locates in the near wall region of RC, which is mainly due to the solid downward flow. Another region in the vicinity of baffle of RC is due to the vigorous motion of gas phase. Below the baffle, the lateral motion of solid phase below the baffle is mainly due to the pressure difference between the two chambers [8]. Near the wall of HEC, large intensity of solid motion in the HEC results from the vigorous scatter of particles after the eruption of bubble phase. Time-averaged distributions of gas flux in the central slice of the ICFB are quantitatively illustrated in Fig. 3. The three investigated heights locate in the bottom region, the central region and the top region of the ICFB, respectively. The lateral and vertical components of gas flux in the two chambers distribute differently. In the HEC, gas phase moves towards the RC. However, its intensity is small, mainly resulting from the falling particles in this chamber. Comparatively, the vertical flux of gas phase is more regular that the lateral one. The vertical flux of gas phase mainly concentrates in the right region of each chamber.

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

1451

Fig. 2 Contour plots of the time-averaged transportation intensity of gas-solid phase in the central slice of ICFB. (a)gas flux; (b)solid flux.

Fig. 3 Distribution profiles of gas flux in the central slice of ICFB. (a)lateral flux; (b)vertical flux.

Fig. 4 Distribution profiles of solid flux in the central slice of ICFB. (a)lateral flux; (b)vertical flux.

Fig. 4 illustrates the transverse distributions of time-averaged flux of solid phase in the central slice of ICFB. In general, the vertical transportation intensity of solid phase is extremely larger than the vertical one in these two chambers. Larger transportation intensity of solid phase can be spotted in the RC as compared with that of the HEC. Limited by the wall and baffle, vigorous lateral transportation of solid phase appears in the central region of each chamber. In the RC, influenced by the gas motion, vigorous upward transportation of solid phase appears near the baffle of this chamber while the strong back-mixing of solid phase can be spotted near the wall region. The rising intensity of solid phase enlarges with increasing the bed height in the RC. Along the bed height, the falling intensity of solid phase near the wall of HEC increases due to the relatively dense distribution of solid phase. However, the downward motion of solid phase near the baffle of this chamber is enhanced due to the solid motion from the HEC to RC below the baffle. 4.2. Hydrodynamics of the CFB 4.2.1

Simulation setup

The LES-DEM coupling approach is also used to investigate the gas-solid flow inside a lab-scale circulating fluidized bed. The drag force exerted on the particles in the CFB is calculated using the correlation proposed by Gidaspow [15]. The CFB consists of a cylindrical riser, a cylindrical cyclone and a dipleg. The geometrical configuration and the grid representation of the investigated apparatus are shown in Fig. 5. Details of the gas and solid properties used in the simulation are given in Table 2. A total number of 81500 grids are used to represent the calculation domain. Before the calculation, a half number of the total particles are randomly generated in the calculation domain and then accumulate in the bottom of the riser and the L-Valve, respectively

1452

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455 Table 2. Details of the gas-solid properties in the CFB simulation.

Gas property (Air) Gas density, kg/m3 Viscosity, kg/(m·s) Superficial velocity, m/s Revert aeration, m/s Outlet pressure, Pa Time step for gas motion, s Solid property

1.225 1.8 × 10-5 5.5 1.0 1.013 × 105 1.0 × 10-5

Particle number Density, kg/m3 Diameter, mm Young modulus, Pa Poisson rate Restitution coefficient Friction coefficient Time step for solid motion, s

200000 1500 1.6 5.0 × 107 0.33 0.90 0.10 1.0 × 10-6

(a)

(b)

Fig. 5. (a)Geometrical configuration of the investigated CFB; (b)grid representation of the calculation domain

4.2.2

Gas-solid motion in the CFB

Fig. 6 shows the 3-D view of the time-averaged gas-solid velocity in the full-loop CFB. The time-averaged gas-solid velocity inside the CFB is complex and heterogeneous. Although the radial velocity of the gas (X- and Ycomponent) is about two orders smaller than the axial velocity, it is proved to be existed in the presumably onedimensional gas flow inside the riser. Meanwhile, it seems that the radial flow direction of gas alters with bed

1453

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

elevation along the riser. This may be due to the coactions of the one-sided inlet/outlet design and the intensive gassolid interactions. The radial flow of gas is benefit to the gas-solid reactions, but it decreases quickly along the riser height. The radial components of solid velocity inside the CFB are a little bit smaller than their counterparts of the gas flow, but they are only about one order smaller than the vertical component. Moreover, it seems that the X- and Ycomponents of solid velocity co-exist at the same height. This reflects the typical internal circulation of solid motion inside the riser. Finally, the Z-component of gas-solid velocity inside the rig demonstrates the typical core-annulus flow structure of fast fluidization flow regime of the CFB.

(a)

(b)

(c)

Fig. 6. 3-D view of the time-averaged gas-solid velocities in the full-loop CFB: (a)X-component;(b)Y-component;(c)Z-component.

The radial distributions of the time-averaged gas-solid velocity in the central slices X=0 and Y=0 of the riser of the CFB are shown in Fig. 7. In the vicinity of the riser bottom, non-uniform distributions of gas-solid velocities can be observed due to the one-sided return of solid phase. With increasing the bed height, this non-uniform distribution property diminishes. At the height of Z=0.3m, the distribution of gas-solid velocity shows an inverted bowl pattern. The velocity difference of gas phase along the radial direction diminishes with bed elevation. In general, the backmixing behavior of solid phase in the wall vicinity can be observed at all the investigated heights. In addition, the upward motion of solid phase in the central region and the downward flow near the wall region reflects the typical core-annulus structure of solid motion in the riser of CFB. 4.2.3

Gas-solid motion in the cyclone

Full-loop simulation of the whole CFB gives advantage of truly capturing the internal movement details of gassolid phases in the cyclone as compared with those carried out only in a single component. Fig. 8(a) presents the vector plot of the gas motion in the cyclone of the CFB. After tangentially introduced into the cyclone, plenty of the gas escapes from the apparatus through the central region while others are accelerated downward due to the reduced flow area. In the lower part of the cyclone, they move towards the central region of the cyclone and flow through the central tube. Besides, both the inward and outward swirling flows of the gas motion in the cyclone have been captured. For the solid motion, the moving particles fall along the surface in a spiral pattern. Vigorous interaction between particles and the apparatus can be obtained in the procedure.

1454

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

Fig. 7 Radial distributions of the time-averaged gas-solid velocities in the central slice X=0 and Y=0 of the riser of the CFB.

(a)

(b)

Fig. 8 Vector plots of the gas-solid motions in the cyclone of the CFB. (a)vector plot of gas motion; (b)vector plot of solid motion.

5. Conclusion Numerical simulation of the gas-solid motion in the CFB is carried out with the LES-DEM coupling approach. Based on the simulated results, the following conclusions can be drawn:

Kun Luo et al. / Procedia Engineering 102 (2015) 1446 – 1455

1455

Internal circulation of solid is established between the two chambers of the ICFB. Large flux of gas phase mainly concentrates in the right region of each chamber. In these two chambers, the vertical transportation intensity of solid phase is extremely larger than that of the lateral one. Vigorous transportation of solid phase mainly locates at the two sides of RC and the region below the baffle. The falling intensity of solid phase in the near wall region of HEC enlarges along the bed height. In the CFB, the radial flow direction of gas alters with bed elevation along the axial direction. The radial component of solid velocity is nearly one order smaller than the vertical one. It decreases quickly along the riser height. Moreover, the typical core-annulus structure of solid motion can be observed in the riser. Non-uniform distributions of gas-solid motion can be observed in the bottom region of the riser. Strong back-mixing of solid phase appears near the wall. Gas and solid motions fall downward along the surface of cyclone in a spiral pattern. Acknowledgements Financial supports from the Major Program of the Natural Science Foundations of China (Grant Nos. 51390491, 51390493), and the Zhejiang Provincial Natural Science Foundation for Distinguished Young Scholars (Grant No. LR12E06001) are sincerely acknowledged. References [1] K. Matsuoka, S. Hosokai, K. Kuramoto, Y. Suzuki, Enhancement of coal char gasification using a pyrolyzer–gasifier isolated circulating fluidized bed gasification system, Fuel Process. Technol. 109 (2013) 43-48. [2] K. Matsuoka, S. Hosokai, Y. Kato, K. Kuramoto, Y. Suzuki, K. Norinaga, J. Hayashi, Promoting gas production by controlling the interaction of volatiles with char during coal gasification in a circulating fluidized bed gasification reactor, Fuel Process. Technol. 116 (2013) 308-316. [3] X. Hu, T. Xu, C. Li, C. Yang, Catalytic cracking of n-heptane under activation of lattice oxygen in a circulating fluidized bed unit, Chem. Eng. J. 172 (2011) 410-417. [4] J.M. Lee, Y.J. Kim, S.D. Kim, Catalytic coal gasification in an internally circulating fluidized bed reactor with draft tube, Appl. Therm. Eng. 18 (1998) 1013-1024. [5] L. Mukadi, C. Guy, R. Legros, Prediction of gas emissions in an internally circulating fluidized bed combustor for treatment of industrial solid wastes, Fuel 79 (2000) 1125-1136. [6] H.P. Zhu, Z.Y. Zhou, R.Y. Yang, A.B. Yu, Discrete particle simulation of particulate systems: Theoretical developments, Chem. Eng. Sci. 62 (2007) 3378-3396. [7] S. Yang, K. Luo, M. Fang, J. Fan, Influence of tube configuration on the gas–solid hydrodynamics of an internally circulating fluidized bed: A discrete element study, Chem. Eng. J. 239 (2014) 158-170. [8] K. Luo, F. Mingming, Y. Shiliang, F. Jianren, LES-DEM investigation of an internally circulating fluidized bed: Effects of gas and solid properties LES-DEM, Chem. Eng. J. 228(2013) 583-595. [9] K.J. Marschall, L. Mleczko, CFD modeling of an internally circulating fluidized-bed reactor, Chem. Eng. Sci. 54 (1999) 2085-2093. [10] Y. Feng, T. Swenser-Smith, P.J. Witt, C. Doblin, S. Lim, M.P. Schwarz, CFD modeling of gas–solid flow in an internally circulating fluidized bed, Powder Technol. 219 (2012) 78-85. [11] B. Lu, N. Zhang, W. Wang, J. Li, J.H. Chiu, S.G. Kang, 3-D full-loop simulation of an industrial-scale circulating fluidized-bed boiler, AIChE J. 59 (2013) 1108-1117. [12] A. Nikolopoulos, N. Nikolopoulos, A. Charitos, P. Grammelis, E. Kakaras, A.R. Bidwe, G. Varela, High-resolution 3-D full-loop simulation of a CFB carbonator cold model, Chem. Eng. Sci. 90 (2013) 137-150. [13] K.W. Chu, A.B. Yu, Numerical simulation of complex particle–fluid flows, Powder Technol. 179 (2008) 104-114. [14] C.R. Müller, D.J. Holland, A.J. Sederman, S.A. Scott, J.S. Dennis, L.F. Gladden, Granular temperature: Comparison of Magnetic Resonance measurements with Discrete Element Model simulations, Powder Technol. 184 (2008) 241-253. [15] D. Gidaspow, Multiphase flow and fluidization: continuum and kinetic theory descriptions, first ed., Academic press, San Diego, 1994.