Elsevier

Chemical Engineering Science

Volume 92, 5 April 2013, Pages 146-156
Chemical Engineering Science

Computational and experimental study of electrostatics in gas–solid polymerization fluidized beds

https://doi.org/10.1016/j.ces.2013.01.023Get rights and content

Abstract

Gas–particle flows are present in many industrial applications such as polymerization, fluid catalytic cracking, chemical vapor deposition, combustion and drying. Particle–particle, particle–wall and gas–particle interactions cause electrostatic charge to form on particles. The motion of charged particles creates an electric field, affecting the hydrodynamics in reactors such as polymerization fluidized beds and fluid catalytic crackers (Hendrickson, 2006). In this work, a combined multi-fluid and electrostatic model previously developed in Rokkam et al. (2010) is used to simulate laboratory-scale experiments on electrostatics in gas–solid fluidized beds conducted by Sowinski et al. (2010). The fluidized-bed experiments were operated in two flow regimes, bubbling and slug flow. Charge-to-mass ratio (q/m) measured in the experiments was used as an input to the computational fluid dynamic (CFD) electrostatic model. Particle-phase segregation from CFD simulations with electrostatic forces compared well with experimental measurements and observations.

Graphical abstract

Mean volume fraction of wall particles in bubbling-bed simulations with refined grid (a) charged, (b) left wall near the distributor, (c) left wall near the distributor with different ranges.

  1. Download : Download high-res image (199KB)
  2. Download : Download full-size image

Highlights

► A Faraday cup technique is used to study the electrostatics in fluidized-bed. ► The measured particle charge is used as an input to the multiphase CFD model. ► The CFD multiphase model predicted the segregation observed in experiments.

Introduction

Polyethylene is one of the most widely used plastics with over 60 million tons produced worldwide every year. A major portion of this polymer is produced by catalytic polymerization of ethylene in gas- and liquid-phase reactors. The gas-phase processes are more advantageous and use fluidized-bed reactors for commercial production of polyethylene. Although gas-phase processes have advantages, there are some disadvantages like particle agglomeration and sheeting. These disadvantages are due to electrostatics and hot spot formation (Hendrickson, 2006). Particle–particle, particle–wall and particle–gas contacts cause electrostatic charge to form. The motion of charged particles in reactors then results in non-uniform electric fields affecting the hydrodynamics of fluidized beds. Electrostatic charge generation is known to affect the bubble dynamics such as size, shape and also particle elutriation rates (Rokkam et al., 2010). With charge accumulation, process upsets such as defluidization can occur (Hendrickson, 2006, Rokkam et al., 2009, Moughrabiah et al., 2012, Sowinski et al., 2012). Understanding the dynamics of the bed aids in the design and process optimization of such reactors. In this work the electrostatic model previously developed in Rokkam et al., 2009, Rokkam et al., 2010 is compared with results obtained from experiments performed by Sowinski et al. (2010).

Gas–particle flows can be modeled using two different approaches, Euler–Lagrange and Euler–Euler. In the Euler–Lagrange models a particle's position and velocity are obtained by Newton's law and the fluid-phase density and velocity from the modified Navier–Stokes (Hoomans et al., 1996, Tsuiji et al., 1993, Kaneko et al., 1999). The particle–particle and particle–wall interactions are accounted for a realistic manner. A closure is required to model the momentum exchange between the gas and particle phases (Ergun, 1952). This method is computationally expensive when the number of particles to be tracked is high and thus it is unsuitable for simulating large-scale reactors. The Euler–Euler two-fluid model treats the particle phase as a pseudofluid, and modified Navier–Stokes equations are used to solve the mass and momentum balance for the fluid and particle phases. Kinetic theory of granular flow (Gidaspow, 1994) and frictional theory (Schaeffer, 1987) are used to describe the rheology of the particle phase and a closure is required for momentum exchange between the phases. Euler–Euler models have been used to simulate pilot and commercial-scale reactors (Rokkam et al., 2010, Gobin et al., 2003, Fede et al., 2010).

Sowinski et al., 2009, Sowinski et al., 2010 used an online Faraday cup technique to measure the electrostatic charge of the particles in their entirety using a lab-scale gas–solid fluidized bed. The system was operated with dry air and with a wide size distribution of commercially produced polyethylene particles. The experiments were conducted at two different fluidization regimes, bubbling and slug flow. Bipolar charging was observed with fine entrained particles being predominantly positively charged and the bulk bed particles being predominantly negatively charged. The charge-to-mass ratio (q/m) of the particles was measured and used as an input to the electrostatic model. The solid-phase distribution in the CFD simulations was compared with the experimental observations. The remainder of this paper is organized as follows. The description of the experiments is presented where the experimental setup, procedure conditions, and observations is given in Section 2. Section 3 discusses the CFD model, simulation parameters and boundary conditions. Section 4 describes the simulation results and comparison with experiments. The final section summarizes the conclusions of this work.

Section snippets

Experimental section

The following section gives a brief summary of the experimental setup, procedure, overview of conditions and observations. For more details, refer to Sowinski et al. (2010).

CFD model

The following section describes the governing equations, simulation parameters and boundary conditions. An Euler–Euler multi-fluid CFD model in ANSYS FLUENT 12.1 was used to describe the hydrodynamics of gas–particle fluidized-bed. The electrostatic model (Rokkam et al., 2010) was coupled to the multi-fluid model as a user-defined function. The multi-fluid model (Gidaspow, 1994, Fan and Fox, 2008) solves for one gas and three particle phases (FINES, DROPPED and WALL). The model assumes the

Simulation results

In this section, the bubbling-bed and slug-flow simulation results and comparison with experiments will be discussed. The instantaneous contours of gas volume fraction for the bubbling bed and slug flow are shown in Fig. 3, Fig. 4. It is important to note that the same fluidized-bed (diameter: 0.0859 m, height: 1.27 m) column is used for both flow regimes, even though the simulation contour plots show different column widths. Second-order discretization schemes are necessary to resolve the bubble

Conclusions

The predictions of the electrostatic model (Rokkam et al., 2010) were compared with experimental observations. The model predicted the hydrodynamic and electrostatic segregation observed in the experiments. The height of the wall coatings in both flow regimes was predicted fairly accurately. The model was able to predict mean properties such as bed height and solid-phase distribution inside the bed, even though a mechanistic model for the adhesion of particles to the column wall and charge

References (27)

Cited by (0)

View full text