In this paper we combine a stochastic 3D microstructure model of a fiber based gas diffusion layer of polymer electrolyte fuel cells with a Lattice Boltzmann model for fluid transport. We focus on a simple approach of compressing the planar oriented virtual geometry of paper-type gas diffusion layer from Toray. Material parameters -permeability and tortuosity -are calculated from simulation of one phase, one component gas flow in stochastic geometries. We analyze the statistical spread of simulation results on ensembles of the virtual geometry, both uncompressed and compressed. The influence of the compression is discussed with regard to the Kozeny-Carman equation. The effective transport properties calculated from transport simulations in compressed gas diffusion layers agree well with a trend based on the Kozeny-Carman equation.

No abstract

A novel, realistic 3D model is developed describing the microstructure of nonwoven GDL in PEMFC which consists of strongly curved and non-overlapping fibers. The model is constructed by a two-stage procedure. First we introduce a system of random fibers, where the locations of their midpoints are modeled by a 3D Poisson point process and the fibers themselves by random 3D polygonal tracks which represent single fibers in terms of multivariate time series. Secondly, we transform the random fiber system into a system of non-overlapping fibers using an iterative method leaned on the so-called force-biased algorithm. The model is validated by comparing transport-relevant characteristics computed for experimental 3D synchrotron data, and for realizations sampled from the stochastic microstructure model. Finally, we suggest a model for the spatial distribution of PTFE, a wet-proofing agent often used in non-woven GDL, and combine this PTFE model with our new microstructure model for non-woven GDL.

No abstract

No abstract