Elsevier

Powder Technology

Volume 260, July 2014, Pages 52-58
Powder Technology

A computational fluid dynamics study on the flow field in a liquid–solid circulating fluidized bed riser

https://doi.org/10.1016/j.powtec.2014.03.030Get rights and content

Highlights

  • The CFD model for turbulent two-phase flows in a LSCFB riser is proposed in this work.

  • k–ε dispersed turbulence model is more efficient than other k-ε multiphase models.

  • A model to predict the solid and liquids residence time distributions (RTDS) is developed.

Abstract

A detailed study on the hydrodynamics of liquid–solid circulating fluidized bed (LSCFB) reactors is crucial in the efficient design and scale-up of these reactors. In this paper, an axisymmetric CFD model is developed to simulate the flow field in a LSCFB riser. The model is based on Eulerian–Eulerian approach incorporating the kinetic theory of granular flow. The predicted results agree well with our earlier experimental data. Furthermore, it is found that the dispersed k–ε multiphase turbulence model is more accurate and computationally efficient than other k–ε multiphase turbulence models. Also, the model predicts the residence time of both liquid and solid phases in the riser by using a Pulse technique. Finally, the proposed CFD model is used to investigate the effects of the liquid stream velocity and the solids circulation rate on the performance of the LSCFB. It is demonstrated that the proposed CFD model can be a robust tool for the scale-up and design of industrial LSCFB reactors.

Graphical abstract

  • An axisymmetric CFD model is developed to simulate the turbulent flow field in a LSCFB riser.

  • The model is based on Eulerian–Eulerian approach incorporating the kinetic theory of granular flow.

  • The CFD model is able to predict the residence time of both the liquid stream and the solid particles in the riser with a Pulse technique.

  1. Download : Download full-size image

Introduction

Liquid–solid circulating fluidized bed (LSCFB) reactors are obtaining extensive attraction in diverse fields of industrial processes, such as many new processes in biochemical technology, wastewater treatment, petroleum and metallurgical industries (Atta et al. [1]). This is because this new type of liquid–solid contacting equipment has a large number of unique features, such as effective liquid–solid contacts, short and narrow residence time for both phases and independent control of solids holdup by varying the mass flow rate of particles (Zheng et al. [2], [3]).

A typical LSCFB is comprised of a riser, a downcomer, a liquid–solid separator, a top solid-return pipe and a bottom solid-return pipe. Particles are entrained up by the liquid stream along the riser under a co-current pattern, then separated at the riser top (separator), and finally recirculated back through a particle storage vessel or downcomer to the bottom of the riser (Zheng et al. [2], Razzak et al. [4]). A proper selection of the reactor is crucial to minimize the costs of the plant and also the negative impact of the reaction products on the environment. The reactor modeling approach that illustrates the key features of the multiphase flow pattern and predicts the relevant physical quantities can be a reliable technique to gain the aim. However, a kinetics model describing the reaction chemistry can predict the reactor performance meaningfully, only when the comprehensive flow field information in the reactor is known (Roy et al. [5]).

In the recent decades, computational fluid dynamics (CFD) techniques have received many attentions in simulating the flow field in two phase flow. Generally, two different types of the CFD models can be used to simulate two phase flow: the Eulerian–Lagrangian (E–L) approach and Eulerian–Eulerian (E–E) approach. In the E–L approach, the carrier phase is considered as a continuous phase and the solid phase is treated as a discrete phase and each solid particle is tracked by solving the Lagrangian force balance equation. In the E–E approach, also well known as the two-fluid model, each phase is treated as an interpenetrating continuum. In order to estimate the solids viscosity and solids stresses, the kinetic theory of granular phase (KTGP) is incorporated into the two-fluid model (Sinclair and Jackson [6]; Gidaspow [7] and Ding et al. [8]).

In the case of two-phase flow in the fluidized beds where the number of solid particles is huge, the E–E approach is the more attractive and practical method. The prior CFD studies on this area mostly focused on the gas–solid fluidized bed; and less attention has been dedicated to the CFD modeling of the liquid–solid fluidized bed. Roy et al. [5] presented a two-fluid model based on the KTGP to simulate the LSCFB riser for alkylation process. The liquid-phase turbulence was modeled by using the standard k–ε model. They found out that the predicted flow field was not very sensitive to the restitution coefficient in the vicinity of e = 1.0. Doroodchi et al. [9] applied the E–E approach to investigate the influence of the inclined plates on the expansion behavior of the liquid–solid fluidized bed. The viscous stress of the solids was neglected in the simulation. The CFD models successfully predicted the general trends in the experimental data.

Lettieri et al. [10] used the E–E approach to simulate a liquid fluidized bed of lead shot in slugging mode. The granular kinetic theory was applied to describe the solids pressure and the solid phase stress tensor. The CFD results showed an agreement with the experimental data at low liquid velocities. However, modeling was not able to accurately predict the flow structure at high liquid velocity. Cheng and Zhu [11] developed a two-fluid model to simulate the turbulent liquid–solid flow in an LSCFB riser. KTGP was incorporated into the model. The model predictions agreed well with the experimental data in the literature and it was also found that increase in particles size and bed diameter result in more non-uniform distributions of hydrodynamic parameters in the radial direction. Shi et al. [12] developed a three-dimensional E–E model to describe the liquid–solid flow in a tubular loop propylene polymerization reactor. The predicted pressure gradients showed a good agreement with the classical calculated data.

In this work, an axisymmetric CFD model is proposed to describe the flow field in LSCFB risers. The model is based on Eulerian–Eulerian approach incorporating the kinetic theory of granular flow. The CFD model is applied to capture the detailed information about the local values of volume fraction and velocity, as well as the residence time of both the liquid stream and the solid particles in the riser with a Pulse technique.

Section snippets

Experimental setup of the LSCFB system

In our earlier works, Zheng et al. [2], [3], [13] conducted an experimental study on the structure of the solids and liquid flows in an LSCFB which was designed and manufactured in a lab scale. A schematic diagram of the experimental apparatus is shown in Fig. 1. The main components of the system are the vertical Plexiglass riser column of 76 mm I.D. and 3.0 m in height, a liquid–solid separator, a liquid stream distributor, a dual flipping valve for measuring solids circulation rate and a solids

Mathematical modeling

Since the liquid phase velocity in the riser is higher than the terminal velocity (Ut) of the solid particles, the riser is operated under the circulating fluidization regime (Liang et al. [14]). A CFD axisymmetric model is used to simulate the turbulent flow field in the LSCFB riser. The CFD model is based on Eulerian–Eulerian approach. Thus, the governing equations for the solid phase have similar structure to those for the liquid phase. Furthermore, in order to close the conservation

Numerical methodology

The commercial software, ICEM CFD, Ansys 13.0, is used to create the riser geometry and then generate the mesh. The governing equations are then solved by the commercial CFD code FLUENT, Ansys 13.0. The convection terms and gradients in all transport equations are descritized by the second order upwind method and Green-Gauss cell based method, respectively. The SIMPLE algorithm using a segregated solution technique is used to solve the pressure field and velocity field. The mesh independence is

Results and discussion

The numerical model presented in this study is used to predict the flow field in the LSCFB riser. The effect of turbulence models on the numerical simulation of the liquid–solid turbulent flow is examined by comparing the numerical results with available experimental data (Zheng et al. [2], [3]). Also, the effects of the liquid superficial velocity (Ul) and solids circulation rate (Gs) on the hydrodynamic characteristics of the LSCFB riser are investigated. In addition, the residence time

Conclusions

The CFD model was proposed to provide qualitative and quantitative pictures of the turbulent two phase flows in an LSCFB riser. Three different types of k–ε multiphase turbulence models were examined in this work and it was found that the dispersed k–ε turbulence model is more efficient than the other ones because of the lower computational time and higher accuracy. Comparisons of the predicted liquid velocity profiles and solids holdup profiles are in good agreement with the experimental data.

Nomenclature

    Notation

    C1ε

    Constants in Eqs. (10), (11)

    C2ε

    Constants in Eqs. (10), (11)

    CD

    Drag coefficient, dimensionless

    dp

    Mean particle diameter, m

    D

    Riser diameter, m

    e

    Restitution coefficient for interparticle collisions, dimensionless

    ew

    Restitution coefficient for particle–wall collisions, dimensionless

    g

    Acceleration due to gravity, m/s2

    g0

    Radial distribution function, dimensionless

    Gs

    Solids circulation rate, kg/m2 s

    k

    Turbulent kinetic energy, m2/s2

    Ksl

    Interphase momentum exchange coefficient, kg/m3 s

    kΘs

    Granular conductivity,

References (20)

There are more references available in the full text version of this article.

Cited by (0)

View full text