A CFD-DEM sensitivity analysis: The case of a pseudo-2D spouted bed

Abstract

The coupling of computational fluid dynamics (CFD) and the discrete element method (DEM) permits to simulate complex multiphase flows, and has been widely used for research and industrial purposes. However, this approach relies on many sub-models and parameters, whose influence is not always clear. In this work, we performed an extensive sensitivity analysis, modifying one sub-model or parameter at a time and assessing the effect on the results. We run our simulations through the commercial program Ansys Fluent 19.1. We considered experimental data reported in literature for a pseudo bi-dimensional spouted bed, containing 2.033 mm glass particles. Spouted beds are suitable to process coarse, poli-dispersed and irregular particles, and are hence a popular research subject, often simulated through the CFD-DEM approach. Our results slightly overestimate the highest position reached by the particles, possibly because of the limits in the mesh refinement. The sensitivity analysis shows that the drag model, restitution coefficient, friction coefficient and Magnus lift model have the strongest influence on the particles' trajectories. It is important to include turbulence, the Magnus lift effect and to consider the rotation of particles. These results clarify the influence of the considered parameters and models in CFD-DEM, and can be considered for the optimisation of CFD-DEM simulations. Thanks to the conclusions of the sensitivity campaign, we enhanced the accuracy of our model.