In the last decades it has become clear that the transport of gas and water inside the mixed-wettable gas diffusion medium (GDL) plays a significant role for the improved understanding and optimization of the polymer electrolyte membrane fuel cells (PEMFC). In the present thesis the influence of liquid water and gas on the transport properties of gas diffusion media of polymer electrolyte membrane fuel cells (PEMFC) is examined numerically and experimentally. The different arising transport mechanisms and their influence as well as their representation in theoretical models (especially REV-based Darcy models) are presented. Moreover, an approach for modelling and simulation of the water distribution inside mixed-wettable porous media, especially gas diffusion layers, is discussed. To this end, a thermodynamical-based approach is chosen – the interactions between gaseous, liquid, and solid (carbon and PTFE) phases are treated with the help of a stationary scheme based on the interfacial energies which have to be minimized. For the optimization task itself the (parallel) simulated annealing approach is chosen and discussed. In addition algorithms for the generation and discretization of the virtual porous structures are described.
Based on the results the modelling of constitutive relationships and transport parameters depending on water and PTFE content is performed. Besides that experiments for the measurement of those relationships are developed. A special focus is on the precise compression of the GDL sample and the influence on capillary pressure-saturation relationship, relative permeabilities, and effective diffusivities. Different apparatus for in- and through-plane measurements are developed.
At the end the derived transport parameters and relationships are applied to a REV-based Darcy model which is compared with an integral experiment. The experimental setup is motivated by the counter-current flow regime of liquid water and gas at the cathode side of the PEMFC. It has been demonstrated that Darcy-flow based models for porous media are also applicable to thin technical porous layers.