A dual-geometry pore-size-resolved model to predict deep-bed loading in a wall-flow filter

A gasoline particulate filter (GPF) is an effective wall-flow technology for gasoline engine particulate matter (PM) emission control via deep-bed filtration through porous ceramic walls within a honeycomb channel structure. An analytical filter model is desired to understand its efficiency and pressure drop under engine operating conditions and, thereby, reduce the development costs for meeting emission regulations and engine performance targets. Most existing models either oversimplify the filter geometry or do not offer good predictions in the soot loading regime. The present work treats the deep bed regime as two co-existing filters with different geometries based on their natural morphologies, one that describes the porous wall and the other for the fiber structure created by soot deposition. Both models are dynamic to account for changes in pore size distribution and soot dendrite growth during loading. Model validation by 12 sets of experimental filter loading data shows that it provides reasonable predictions of both filter efficiency and pressure drop during deep-bed loading for different GPF wall-flow filters under various operating conditions. Scanning electron microscopy (SEM) confirms model predictions for the unit collector size of the dendritic filter formed by soot deposition. Numerical simulation using GeoDict was also carried out to validate the model predictions on particle deposition location. The model is used to investigate the evolution of pore-size-resolved aerosol flow, filtration efficiency, mass loading, and particle packing density during loading. The impact of porosity, mean pore size, and pore size distribution on filter loading performance is studied. The results suggest that increasing porosity, narrowing pore size distribution, and increasing mean pore size can improve filter performance, offering pathways for future filter optimization.