CFD simulation and experimental validation of oil-in-water emulsion flow on CT images of a sand filter medium: Impact of pore morphology on trapping phenomena

Porous medium clogging is a major challenge for infiltration systems. While the physical mechanisms of clogging have been widely investigated, the mesoscale processes and dynamics in porous structures remain poorly understood. This study examines the trapping phenomena in a quartz sand filter using a combination of experiments and numerical simulations. The numerical approach employs the mixture model in COMSOL Multiphysics, applied to simulation geometries reconstructed from CT images of porous media. The investigation is divided into two parts. First, the effects of injection velocity, column height, and porosity on oil entrapment are analyzed experimentally. Second, the influence of sand-pack heterogeneity and oil concentration is examined using CT-based filter bed geometries. The sensitivity analysis highlights column height as the most influential factor: increasing the column height from 30 to 70 cm at a porosity of 0.38 nearly doubles the trapped oil volume fraction. In contrast, raising the injection velocity from 1.7 × 10⁻⁴ to 3.74 × 10⁻⁴ m/s reduces the trapped oil fraction by 28.5 % at a porosity of 0.46. Overall, the simulation results show satisfactory agreement with experimental data, with a deviation below 15 %. The novelty of this work lies in the quantification of trapped oil and its impact on porosity reduction using realistic CT-based sand filter geometries. These findings reveal previously unknown facts about mesoscale clogging mechanisms and offer practical guidance for the design and optimization of infiltration systems.