Statistical capillary tube model for porous filter media: An application in modeling of gasoline particulate filter

Abstract

A new statistical capillary tube model is presented to predict both filtration efficiency and permeability of particulate filters. The model considers the porous structure of the filter media as an assembly of curved capillary tubes with a statistical size distribution. The key model descriptors of pore microstructure are porosity, pore size, pore size distribution, and tortuosity. The model was validated by permeability and size dependent filtration efficiency measurements of bare and washcoated gasoline particulate filters (GPFs), using porosimetry data of the filter media as model inputs. Comparing with a more widely used spherical model for particulate filters with typical porosity range of 30–60%, the capillary tube model has similar input parameters but provides additional permeability prediction as well as better filtration predictions over a wider porosity range including the lower porosity of washcoated filters. Furthermore, it offers a more visually straightforward and intuitive correlation between filtration paths and filter performance. This allows convenient adoption of the capillary tube model in numerous applications that are currently based on spherical model assumption and provides improved guidance for filter optimization. Beyond ceramic automotive exhaust filters, this geometric model offers a new framework for examining porous filter media in general.