Particle-scale modelling of rapid granular filtration in a dual-media filter

Abstract

Multi-media filters are widely applied in water treatment, yet the understanding of powder transportation mechanisms at particle-scale and clogging prediction is still lacking. In this work, a dual-media (anthracite and sand) rapid granular filtration is numerically studied at particle-scale by a smoothed computational fluid dynamics-discrete element method (CFD-DEM) model. In this model, a multi-sphere model and a modified Johnson-Kendall-Roberts (JKR) model are integrated to describe anthracite shapes and material cohesion, respectively. The model is validated in terms of pressure drop across a packed bed and powders migration through another packed bed. The powder transportation in the filtration through a dual-media filter is studied. The results show that the motion of powders in the anthracite layer are governed by deposition, penetration, and connection, while the motion of powders in the sand layer experience core formation, core growth, and cluster connection. The evolution of solid volume fraction Φ indicates a saturation value (0.67) exists in the entries of layers. The small powders tend to be clogged in the anthracite layer and the large powders prefer to be clogged in the sand layer. The density evolution of the clogged powders nv shows that the sand layer has a saturation nvs=0.21 in the entry, and that of the anthracite layer nva=0.252 is obtained by extrapolation. Finally, a probabilistic model is originally derived for quantifying the clogging by clogging probability. After verification, the sand layer has the clogging probability of 0.47 while the anthracite layer has the clogging probability of 0.10.