Large-scale simulation of polymer electrolyte fuel cells by parallel computing

A three-dimensional, electrochemical-transport fully coupled numerical model of polymer electrolyte fuel cells (PEFC) is introduced. A complete set of conservation equations of mass, momentum, species, and charge are numerically solved with proper account of electrochemical kinetics and water management. Such a multi-physics model combined with the need for a large numerical mesh results in very intense computations that require parallel computing in order to reduce simulation time. In this study, we explore a massively parallel computational methodology for PEFC modeling, for the first time. The physical model is validated against experimental data under both fully and low-humidified feed conditions. Detailed results of hydrogen, oxygen, water, and current distributions in a PEFC of 5-channel serpentine flow-field are discussed. Under the fully humidified condition, current distribution is determined by the oxygen concentration distribution. Cell performance decreases in low-humidity inlet conditions, but good cell performance can still be achieved with proper water management. Under low-humidity conditions, current distribution is dominated by the water distribution at high cell voltages. When the cell voltage is low, the local current density initially increases along the flow path as the water concentration rises, but then starts to decrease due to oxygen consumption. Under both fully and low-humidified conditions, numerical results reveal that the ohmic losses due to proton transport in anode and cathode catalyst layers are comparable to that in the membrane, indicating that the catalyst layers cannot be neglected in PEFC modeling. (C) 2004 Elsevier Ltd. All rights reserved.