Elsevier

Powder Technology

Volume 369, 1 June 2020, Pages 184-201
Powder Technology

Analysis of flexible ribbon particle residence time distribution in a fluidised bed riser using three-dimensional CFD-DEM simulation

https://doi.org/10.1016/j.powtec.2020.05.025Get rights and content

Highlights

  • Spring constants of the double chain model is calibrated.

  • Residence time distribution of flexible ribbon particles is validated against experimental datum.

  • Flexible ribbon particle in the riser is shown vividly.

  • Axial and radial residence time distribution of this special particle is discussed.

Abstract

Fluidised bed technology is a very common industrial process and has a wide range of applications in process engineering. Though indirectly, a common way to infer the end product uniformity is through investigating the solid residence time distribution (RTD) in the system. A good understanding of the particle RTD is crucial for the proper design, optimisation, and scale-up of a fluidisation process. Flexible ribbon particles have a very different flow behaviour in a fluidised bed compared to that of spheres which in turn influences the RTD of the particles. However, this issue is still not very well addressed to date. In this paper, a double chain model is adopted to model cut tobacco strips in a fluidised bed riser. The spring constants of this chain model are calibrated, and the model is validated against experimental datum. Finally, the particle RTD as a function of different process parameters is investigated.

Introduction

Due to the high heat and mass transfer efficiency, the fluidised bed is widely used in various fields, such as chemical engineering, bioengineering, mineral industry and environmental science. In such a system, particles are first fed into the riser from the bottom inlet and subsequently get fluidised and carried towards the outlet, during which the solids exchange mass and energy with the gas. The transfer efficiency of mass and energy in a fluidised bed can be inferred by the particle Residence Time Distribution (RTD), which is defined as the probability distribution of time that solid materials stay inside an operation in a continuous flow system [1,2]. The particle RTD is a function of particle properties and operating conditions such as superficial gas velocity and particle flow rate [3]. A narrow RTD indicates a more consistent product quality attributes (drying and mixing uniformity, coating consistency, and etc.), while a wide RTD suggests high variability. A good understanding of the particle RTD is therefore crucial for the proper design, optimisation, and scale-up of a fluidisation process.

A lot of experimental studies on the particle RTD characteristics of the fluidised bed have been reported in the past few decades [[4], [5], [6], [7], [8], [9]]. Zhang et al. [10] investigated the mean residence time (MRT) of the particles with wide size distribution in a lab-scale fluidised bed. It was found that the MRT difference for the different size particles increases by adding more baffles. Ambler et al. [11] proposed a mathematical model based on the core-annular flow structure to characterise the RTD of particles within the riser of circulating fluidised beds. The model showed a good prediction of the bimodal distributions of particle residence time in the system. Chan et al. [12] found that the particle residence time of a circulating fluidised bed (CFB) riser can be determined by tracking the particle velocity using particle emission positron tracking method (PEPT). Guío-Pérez et al. [13] investigated the effect of the ring-type internals on the RTD of particles in the fuel reactor of a dual circulating fluidised bed system based on inductance measurement of the detection of ferromagnetic tracer particles.

Numerical studies on the particle RTD characteristics of the fluidised bed are also widely reported in the literature [[14], [15], [16], [17], [18], [19], [20], [21]]. Li et al. [22] discovered a relationship between the particle flow characteristics and the particle RTD in an opposed multi-burner gasifier through the Direct Simulation Monte Carlo (DSMC) method and hard-sphere model. Geng et al. [23] proposed a semi-empirical approach for predicting particle RTD through computational fluid dynamics (CFD) model. Yoo et al. [15] used the Eulerian-Eulerian model with the kinetic theory of granular flow to simulate the motion of tracer particles in a CFB riser. The effect of the gas jet direction on the solid residence time was investigated using the tracer technique. Hua et al. [14] established a CFD model to predict particle RTD of gas-solid flow and to account for the effect of the presence of particle clusters on particle RTD. The gas-solid drag force was modified by using the energy minimisation multi-scale drag model, and the influence of particle clusters was emphasised. The established CFD model showed a good agreement with the available experimental data. Harris et al. [24] developed a set of stochastic mathematical models based on a Markov chain to simulate the particle RTD in a circulating fluidised bed riser. These discrete-time four-zone model outputs were validated against a variety of experimental data reported and found to provide a good agreement.

Most of the aforementioned work revolves around the studies of RTD of spheres in the fluidised bed. However, in reality, the feeding materials could exhibit a wide range of particle size, shape and physical properties [25]. For instance, flexible ribbons are typically encountered in the tobacco industry, where cut-tobacco strips are dried through fluidisation [26]. These tobacco ribbons can easily hold together due to their flexibility and non-uniformity in a ribbon shape, which in turn influences the RTD of the particle in a fluidised bed. Also, investigation and analysis of the RTD of this particular particle in fluidised bed riser by Lagrangian particle tracking method are rarely reported. Therefore, this issue is still not very well addressed to date. In this paper, a double chain cut tobacco model is adopted to model cut tobacco strips in a fluidised bed riser using a coupled CFD-DEM simulation. The spring constants of this double chain model are calibrated by the experiments. Compared to the spherical particle model, the cut tobacco particle model produces a better agreement in the residence time distribution (RTD) when compared to the experimentally obtained data of cut tobacco strips. Finally, the residence time of the flexible ribbons as a function of different process parameters such as the superficial gas velocity, particle flow rate and riser height is investigated.

Section snippets

Double chain model

Multi-sphere model is one of the most efficient composite approaches for modelling flexible particle, which is first reported by Satoru Yamamoto et al. [27,28]. In this model, the flexible particle is made up of n spheres, as shown in Fig. 1(a). The double arrow line in Fig. 1 indicates the spring force between the two spheres. The spring force (FS) between the two spheres (sphere i and j) is calculated in Eq. (1).FS=ksrijr0where ks is the spring constant. r0 is the equilibrium distance of

Experimental setup

The fluidised bed unit (Fig. 4(a), 4(b)) mainly consists of a riser, a fan, an air pipe, a separator, a feeding device and two high-speed CCD cameras. In this study, only the particle flow behaviour in the riser (0.21 m × 0.11 m × 5.20 m), which is made of transparent tempered glass is studied. Air is introduced from the bottom of the riser to the bed by the fan. Flue-cured cut tobacco strips (produced from Yunnan, China) are fed into the riser from the hopper, which subsequently got carried upwards

Simulation setup

A sensitivity analysis was performed to determine the optimal grid size for the fluid domain of the riser. Three different grid sizes were tested, and it was found that a computational domain containing 967,680 CFD cells (24 × 42 × 760 cells in the x, y and z-direction, respectively) provides the best balance between the accuracy and computational cost of the numerical calculation (Fig. 4(c)). The size of the CFD cell is 0.005 m, and it is 10 times of sphere diameter. The air is supplied in a

Calibration of interparticle spring constants

To obtain a more realistic bending and flexing behaviour of the ribbon particle, a test with the straight cut tobacco strip is proposed to calibrate the interparticle spring constants against experimentally obtain data. The detailed procedure is as follow:

A transparent glass tube (diameter = 0.050 m) is horizontally placed, and a slit is made on the high or low part (at the mid-length) of the tube. A strip of straight cut tobacco is first inserted and fixed vertically into the tube through the

Conclusion

The uniformity of the end product of the fluidised bed process can be inferred by the particle RTD. A good understanding of the particle RTD is therefore crucial for the proper design, optimisation, and scale-up of a fluidisation process. In this paper, a double chain cut tobacco particle model is carefully verified. The effect of the riser height, superficial gas velocity and particle flow rate on RTD of flexible ribbon particle is discussed in detail. The main conclusions are as follows:

Nomenclature

    Cd

    Drag force coefficient on spherical particle

    dp

    Diameter of the spherical particle

    E

    Probability density function of residence time

    Emax

    Maximum probability density value of residence time

    FS

    The force exerting on a pair of bonded spheres

    Fi

    The total force acting on the particle i

    Fd,i

    The drag force on the particle i

    Fc,i

    The contact force on the particle i

    Fn,ij

    The normal contact forces on the particle i

    Ft,ij

    The tangential contact forces on the particle i

    g

    The acceleration of gravity

    H

    Riser height, (m)

    Ii

Disclosures

The double chain model and the software adopted in this paper were developed by the Institute of Process Engineering of the Chinese Academy of Sciences and Zhengzhou Tobacco Research Institute.

Declaration of Competing Interest

None.

Acknowledgement

Financial support from the National Natural Science Foundation of China (No. 51576046, No. 51306213), Science and Technology Plan Projects of Jiangsu Province (No. BY2015070-15) and Key Lab of Tobacco Processing Technology of Zhengzhou Tobacco Research Institute (No. 212016AA0300) are gratefully acknowledged. We are thankful to Dr. Wei Pin Goh from the University of Leeds for proofreading.

References (54)

  • P.A. Ambler et al.

    Residence time distribution of solids in a circulating fluidized bed: experimental and modelling studies

    Chem. Eng. Sci.

    (1990)
  • D.C. Guío-Pérez et al.

    Influence of ring-type internals on the solids residence time distribution in the fuel reactor of a dual circulating fluidized bed system for chemical looping combustion

    Chem. Eng. Res. Des.

    (2014)
  • L. Hua et al.

    CFD simulation of solids residence time distribution in a CFB riser

    Chem. Eng. Sci.

    (2014)
  • H. Yoo et al.

    Effect of the jet direction of gas nozzle on the residence time distribution of solids in circulating fluidized bed risers

    J. Taiwan Inst. Chem. Eng.

    (2017)
  • L. Wei et al.

    Hydrogen production by supercritical water gasification of biomass: particle and residence time distribution in fluidized bed reactor

    Int. J. Hydrog. Energy

    (2013)
  • Z. Zou et al.

    Hydrodynamic and solids residence time distribution in a binary bubbling fluidized bed: 3D computational study coupled with the structure-based drag model

    Chem. Eng. J.

    (2017)
  • L. Li et al.

    Residence time distributions of different size particles in the spray zone of a Wurster fluid bed studied using DEM-CFD

    Powder Technol.

    (2015)
  • H.-J. Warnecke et al.

    On a mathematical model for loop reactors—I. Residence time distribution, moments and eigenvalues

    Chem. Eng. Sci.

    (1985)
  • A.T. Harris et al.

    Particle residence time distributions in circulating fluidised beds

    Chem. Eng. Sci.

    (2003)
  • M. Börner et al.

    DEM-CFD investigation of particle residence time distribution in top-spray fluidised bed granulation

    Chem. Eng. Sci.

    (2017)
  • C. Li et al.

    Numerical study of the particle residence time and flow characters in an opposed multi-burner gasifier

    Powder Technol.

    (2015)
  • S. Geng et al.

    Prediction of solids residence time distribution in cross-flow bubbling fluidized bed

    Powder Technol.

    (2017)
  • A.T. Harris et al.

    Stochastic modelling of the particle residence time distribution in circulating fluidised bed risers

    Chem. Eng. Sci.

    (2002)
  • H. Zhao et al.

    Modulating the mean residence time difference of wide-size particles in a fluidized bed

    Chin. J. Chem. Eng.

    (2018)
  • F. Geng et al.

    Simulation of dynamic transport of flexible ribbon particles in a rotary dryer

    Powder Technol.

    (2016)
  • T. Sasayama et al.

    Simplified bead-chain model for direct fiber simulation in viscous flow

    J. Non-Newtonian Fluid Mech.

    (2017)
  • Z. Pei et al.

    A model for the particle-level simulation of multiple flexible fibers moving in a wall-bounded fluid flow

    J. Fluids Struct.

    (2018)
  • Cited by (7)

    • A review of recent development for the CFD-DEM investigations of non-spherical particles

      2022, Powder Technology
      Citation Excerpt :

      The influences of initial bed height, fiber flexibility and degree of cohesion on fluidization characteristics had been comprehensively discussed. Wu et al. [269,270] also investigated the flow characteristics of flexible ribbon particles in a bubbling fluidized bed by bonded-sphere CFD-DEM. Subsequently, the influence of process parameters (e.g., the gas superficial velocity and bed height) on the dynamic behaviors of flexible ribbon particles (e.g., the particle residence time distribution) was explored.

    • Fluid dynamic analysis and residence time distribution determination for rectangular based spouted beds

      2021, Powder Technology
      Citation Excerpt :

      Several tracers are mentioned in the literature, standing out among them the use of salts [7–9], radioactive particles [10,11] and pigments [12,13]. The distributions are used in various applications that operate with continuous feed of solids such as rotary kiln [16], extruders [17], continuous blenders [18], fluidized bed [1,19] and spouted bed dryers [4,20], being applied for equipment design and evaluation of operational parameters such as: superficial air velocity [1,19], solids feed flow [1,18] and their properties [19]. Spouted beds have been used in the drying of seeds [21,22], small fruits [23], pastes or suspensions [24,25] and biomass [26], evaluating the degree of moisture removal in comparison to other available technologies and/or the effect of the variation of operational parameters such as air temperature, air flow and solids feed flow.

    View all citing articles on Scopus
    View full text