On the determination of grid size/smoothing distance in un−/semi-resolved CFD-DEM simulation of particulate flows

Abstract

Unresolved and semi-resolved CFD-DEM coupling are frequently used for modeling particulate flows due to their high efficiency and satisfying accuracy. The grid size (in unresolved CFD-DEM) and smoothing distance (in semi-resolved CFD-DEM) are usually around three times of the particle diameter so as to properly reproduce the background flow field. An over-estimation of the grid size or smoothing distance can lead to low-fidelity flow fields, while under-estimations of these parameters can produce an inaccurate drag force. Therefore in practices, the size ratio is usually empirical, and how does it vary with cases remains unclear. In this paper, we provide an approach to identify the most suitable grid size or smoothing distance in coupled CFD-DEM. Firstly, resolved simulations based on immersed boundary method (IBM) are conducted to compute particles' drag forces. The obtained forces are then compared with those calculated by force models with certain smoothing lengths to identify an optimal smoothing length or effective grid size. During this process, integral scale of each case is also computed. Finally, linearity between integral length and the optimal smoothing distance or grid size is found. This correlation can offer guidance to CFD-DEM modeling in future and provide insights for novel drag force models.