Establishing Accuracy of Watershed-Derived Pore Network Extraction for Characterizing In-Plane Effective Diffusivity in Thin Porous Layers

Abstract

The work presented in this paper reports a versatile and customizable watershed-based pore network extraction tool. It was used to build pore networks from microscale computed tomography images of highly porous carbon-fiber materials. The extracted networks were used to characterize the structural properties (pore and throat diameter, and porosity distributions) and to predict the dry, in-plane, effective diffusivities of two commercial gas diffusion layers under various levels of compression. The relationship between compression and effective diffusivity was used to accurately predict experimental values available in the literature to within 10% of the reported values for compressive strains less than 0.55. Through this work, we validate the accuracy of watershed-derived pore networks for predicting in-plane effective diffusivities, further establishing the reliability of watershed-derived pore network modeling for fuel cell applications.