Percolation-based model for straining-dominant deep bed filtration

• An algorithm is developed for filtration infinite cluster backbone identification.
• The relationship between Ce/C0 and infinite cluster flow property is established.
• The Ce/C0 from cluster analysis agrees with the experimental and simulated results.

A straining-dominant filtration model based on directed bond percolation is proposed. The infinite cluster structure of this model is relevant to filtration properties. Thus, a direct electrifying algorithm is modified for infinite cluster backbone identification, and the geometrical characteristic of the infinite cluster is analyzed. In addition, the relationship between the normalized effluent concentration and infinite cluster is established. The model is validated by comparing the particle concentrations estimated from analyzing the property of filtration cluster of network model with the normalized effluent concentration obtained from filtration process simulation and the experimental data. The results show that normalized effluent concentrations estimated from cluster analysis are consistent with the experimental data and filtration simulated result. Finally, the sensitivity of pore size distribution (PSD) parameters to normalized effluent concentrations estimation is analyzed. Assuming that the PSD is lognormal, the normalized effluent concentrations increase as the location parameter μ increases and scale parameter σ decreases.