Impact of disperse populations of non-equant crystals on filter cake resistance: Insights from filtration experiments and simulations

Particle size and shape distributions (PSSDs) critically influence filter cake resistance in industrial filtration, yet quantitative prediction remains challenging, hindering design and optimization of crystallization and filtration processes. In this study, Monte Carlo packing simulations are combined with Lattice Boltzmann fluid dynamics simulations on the resulting packing structure to estimate the resistance of disperse populations of non-equant particles, common in industrial crystallization. Filtration experiments conducted with diverse crystal populations are compared to simulations in terms of cake porosity and resistance, yielding qualitative agreement. A sensitivity analysis reveals that even the small, currently unavoidable uncertainties in simulated porosities ( ≈ 0.05 in absolute porosity units), can lead to over 150 % uncertainty in resistance, indicating that sequential coupling of packing models and LB yields unreliable quantitative predictions. Systematic variation of particle elongation, flatness, and dispersity, using simulations, reveal how four packing-structure attributes, porosity, specific surface area, flow path tortuosity, and bottleneck pores, mediate the relationship between PSSD and cake resistance. Varying particle elongation produces a resistance maximum, driven by a porosity minimum, whereas increasing flatness raises resistance monotonically, primarily via increased tortuosity. Increasing population dispersity reduces resistance for any non-equant particles through decreased specific surface area, and for plate-like particles, also due to increased porosity while increasing pore bottlenecks lessen the decrease. Finally, comparisons with the Kozeny–Carman estimate show that it systematically underestimates true resistance, by up to a factor of 2.6 within the examined range of PSSDs, attributable to increased tortuosity and bottelneck pores. These findings offer valuable qualitative insights into the links between PSSD, packing structure, and resistance, and highlight the current limitations of predicting resistance through packing models for non-equant particles. • Monte Carlo model reproduces porosity trends of filter cakes of disperse, non-equant particle crystals. • Particle size distribution affects resistance through porosity, surface area, tortuosity, and bottleneck pores. • Combined effects of elongation, flatness, and dispersity on resistance are non-additive.