(756d) Eulerian-Eulerian Modeling of Segregation and Dispersion in Multi-Component Fluidized Beds: Verification, Validation and Limitations

In many industrial-scale fluidized-bed reactors, particle mixing and segregation play a vital role in determining the reactor performance. Detailed information about the phase voidage distribution throughout the bed at different operating conditions is essential for design, scale-up, and optimization. Several experimental and theoretical studies are published in literature focusing on solid-liquid fluidized beds (SLFB) consisting of mono and bi-dispersed particle systems. However, very few studies in the literature emphasize the fluidization of systems involving more than two types of particles. Such kinds of systems are of high industrial importance in mineral and chemical process industries. Therefore, it is desirable to study the coupling between superficial liquid velocity and multi-size (and multi-density) particle systems consisting of more than two types of particles.

Multi-component SLFB cases exhibit several possibilities: complete segregation, partial mixing, segregation with an interface, and segregation without interface. A few experimental investigations have also shown that the layer inversion phenomena exhibited by SLFB at specific conditions also exist for ternary systems. In the present work, Eulerian-Eulerian simulations have been carried out to investigate particle density and diameter on the segregation and dispersion behavior in SLFB comprising of two to six different sizes and densities of solid phases.

In the initial part of this study, many bidisperse systems are simulated to evaluate the advantages and limitations of the commonly used boundary conditions, geometric simplifications, and empirical models used for modeling of SLFB. The limitations identified are (i) inadequacy of the fluid-solid drag models in accounting for the hindrance effect, (ii) unsuitability of solid-solid interaction force closures for the flow physics exhibited by the SLFB, and (iii) the broad spectra of time range at which different systems achieve a steady state. The identified limitations are addressed by introducing physically appropriate empirical models, boundary conditions, and running the simulations for sufficiently long-time steps.

The improved modeling approach was then used to simulate multi-component fluidized beds. The solid fractions, pressure drop, and voidage profiles obtained from the Eulerian-Eulerian simulations of multi-component systems were compared with the experimental data and theoretical model predictions. The simulations show that although Eulerian-Eulerian simulation may not be as detailed and accurate as the Discrete Element Model (DEM), they still provide precise predictions of solid axial profiles, voidage, and overall segregation, and dispersion behavior. However, the accurate predictions of the SLFB hydrodynamics completely depend upon the appropriate selection of the interfacial closure models, boundary conditions, and time range over which the simulations are run. Moreover, the computational time required to simulate a multi-component SLFB using the Eulerian-Eulerian model is significantly less (~10 times) than that required by DEM to simulate the same system. Thus, the present investigation proposes an economically viable and rational approach for CFD and process engineers to get quick estimates of the SLFB hydrodynamics.