Alexander Scheuermann / The University of Queensland
Filtration is the process of the migration of finer base particles through the pore space formed by coarser filter particles. Filtration is a complex natural process with many, often interacting, factors affecting observed filtration characteristics. Filtration is often studied at the laboratory scale with experimental apparatus that predominantly focus on macro-scale filtration characteristics. However, investigation at this scale does not consider potential variability caused by particle scale mechanisms underlying the filtration process. Therefore, this study considers the application of the Discrete Element Method (DEM) to enable the assessment of the variability of the filtration process. One factor seldom considered is that of initial particle position. This study employs DEM to simulate gravity-driven filtration in base-filter combinations that can be regarded as borderline incompatible, where incompatibility implies that the filter cannot retain the base particles. For each case, multiple filters and bases are generated independently by compressing a non-interacting cloud of particles with near identical log-linear particle size distributions. Filtration simulations are then carried out using all possible filter-base combinations. Using this extensive simulation data, this study utilised a statistical approach based on repeated measures Analysis of Variance (ANOVA) to quantify the significance of both filter and base particle position on multiple variables used to characterise filter performance, including the percentage of base particles remaining, the timescale of filtration and the filling degree. Additionally, the relationship between domain size and variability in filter performance was investigated. The repeated measures ANOVA demonstrated that the initial particle position of base and filter particles significantly influenced the filtration characteristics, and this variability in the filtration process is yet to be properly considered when assessing the compatibility for base-filter combinations.