The major bottleneck in durability of current polymer electrolyte membrane fuel cells (PEMFCs) designs is the occurrence of multiphysical phenomena, especially single/multi-component multiphase flow/diffusion of water in vapor and liquid water, hydrogen & oxygen gasses. The condensation of water vapor to liquid water in microscopic pores of the porous media including gas diffusion layer (GDL) & catalyst layer (CL) block the active reaction sites in the CL as well as decreasing the diffusion of reactant gases, thereby decreasing the overall PEMFC performance. Future designs of better performing and durable PEMFCs require a thorough analysis of the multiphase phenomena. However, continuum level computational fluid dynamics (CFD) analysis of this type of behavior is not sufficient, as the inclusion of the complex microstructure and molecular architecture in the porous media increases the scale of computation time significantly for current methodologies.
To resolve these issues, we develop and utilize a highly computationally efficient mesoscale level methodology based on kinetic theory known as lattice Boltzmann method (LBM). We take advantage of the highly efficient geometry handling techniques in LBM and develop a non-isothermal multiphase LBM schemes descriptive for GDLs and describe how to incorporate molecular porous architecture as an input parameter. The GDL structure used in our simulation is reconstructed from micro/mesoscale x-ray tomography images via the use of stereolithography files. LBM will accurately simulate the multiphenomena occurring at microscale for the porous geometries in the membrane including phase change, Knudsen/molecular diffusion, and micro/nano scale heat transfer. In this work, we investigate the liquid water droplet formation, coalescence and movement in GDLs via developing a novel generalized multiphase LBM scheme for large density ratios (ratio of liquid water density to water vapor density) around 2 000 based on the pseudo potential models, where the non-ideal behavior of fluids arising due the fluid-fluid interactions (occurring because of the difference in interaction strength between molecules in the liquid and the gas phase) is incorporated inside the lattice Boltzmann equation framework via an force term, which then links the multiphase LBM to a non-ideal equation of state (EOS), thereby preserving local thermodynamics consistencies. Depending on the sign and magnitude of the force, we can simulate miscible/immiscible phases for single/multi component systems as well as fluid/solid interactions which result in hydrophobicity/hydrophilicity. We further investigate the effect of temperature changes inside the GDL and in HFC systems on the multiphase condensation and movement & behavior of liquid water. For the first time, we were able to physically understand the thermoosmotic/thermocapillary phenomena that occurs in porous media: we found that depending on the hydrophilicity or hydrophobicity of the porous media, the flow of condensed water changes from hot to cold regions or cold to hot regions, respectively. This phenomenon has only been observed recently, and is not numerically resolved yet [1,2]. Thus, our LBM scheme can play a major role in provide design criteria such as manipulating hydrophobicity of GDL/CL surface to increase the I-V performance of PEMFCs.
- M. Khandelwal, S. Lee, and M.M. Mench, “Model to predict temperature and capillary pressure driven water transport in PEFCs after shutdown,” Journal of The Electrochemical Society, 156 , 6, (2009)
2. M. Tasaka, T. Mizuta, and O. Sekiguchi, “Mass transfer through polymer membranes due to a temperature gradient,” Journal of Membrane Science, 54, 191 (1990).