Detailed modeling of solids separation by microsieving in a rotating belt filter: Explicit effect of particle size, mesh size, and polymer dose

Microsieving by rotating belt filters is an engineered fractionation method for particle separation from wastewater that can be used as an alternative to primary clarification due to its smaller footprint and lower total cost of ownership. Because of the multiscale nature of the filtration processes involved, and the interdependent variables controlling performance and the system design, the availability of verified and comprehensively validated models that, based on fundamental principles, can explicitly capture the effect of particle size, mesh size and polymer dose are essential for performance optimization and scale-up. In this paper, a detailed microsieving filtration model was derived by extending Darcy's law to a dual-layer dynamic framework whose predictions were validated against batch (column) and continuous flow (RBF pilot) experimental data. The ability of the microsieving model to accurately capture observed trends confirmed that the model formulation, derived from column experiments, were sufficiently accurate and flexible to capture the relevant physics involved, with a relative error in performance scale up (from column to pilot) of 33% and 14% for filter capacity and effluent total suspended solids, respectively. Upon further calibration against pilot data, the relative error was reduced to 9% and 5%, respectively, indicating the suitability of the model structure in further adapting to filtration process conditions occurring at pilot scale. Finally, the calibrated model was used to derive guidance for future pilot studies. It was determined that, under naturally varying influent conditions, at least 8 h of continuous pilot data, with sampling frequency of at least 15 min for filter capacity and suspended solids concentration, were necessary for a satisfactory estimate of the model parameters.

Automated summary

The study presented a detailed microsieving filtration model and demonstrated its effectiveness in capturing observed trends in filtration processes, with the relative error in performance scale up reduced through further calibration against pilot data. The calibrated model provided guidance for future pilot studies, emphasizing the importance of continuous pilot data collection and sampling frequency for accurate estimation of model parameters.