Computational characterization of nonwoven fibrous media: I. Pore-network extraction and morphological analysis

Nonwoven fibrous materials are a unique class of porous materials that are widely used for filtration and separation applications. The complex pore structure in fibrous media, with high connectivity and low tortuosity, enhances filtration efficiency while preventing filter fouling. The pore size distribution is a key to predicting and improving transport properties by tailoring the microstructure of fibrous media. However, the statistical properties of the pore space are not easily accessible in experiments, and often only effective geometric properties such as hydraulic radius and specific surface area are reported. In this study, we employ a computational framework and generate realistic nonwoven fibrous media with a wide range of porosities. We obtain the pore network by extracting a subnetwork of the watershed segmentation of pore space and analyze the statistical distribution of pore and throat sizes and their connectivity. We find that a random fiber orientation has a pronounced effect on the effective geometric properties and their dependence on porosity. We investigate the influence of processing parameters such as fiber size, orientation, and overlaps between fibers as a prerequisite to predicting transport properties of nonwoven fibrous media. The computational approach can be combined with experimental imaging techniques, which enables rapid characterization of porous microstructures in terms of the statistical pore space properties resulting from specific manufacturing processes.