An efficient method for numerical predictions of the performance characteristics of fuel cells. I. Model development and validation

This study presents the model development and validation of an efficient method for numerical predictions of the performance characteristics of proton exchange membrane (PEM) fuel cells. To efficiently execute extensive modeling in a short time period, the computational model uses a half-cell at the cathode with only one oxidant channel as the modeling domain. The original three-dimensional (3D) model is replaced with a reduced two-dimensional model. In terms of these computational models and reduced dimensions, mathematical formulations describing physical phenomena within fuel cells are developed. In addition, two-phase Darcy's laws, which are originally only applied to porous media, are further extended to describe the transport and formation of liquid water within channels. In this study, the Fortran programming language is used as a numerical program to implement iterative calculations and predict the overall performance characteristics for three different cases. By comparing the modeling and experimental polarization curves for case 1 and the modeling of points of current density vs. cell voltage (IV) of different computational models for cases 2 and 3, this computational model is found to possess a certain degree of accuracy and reliability. In addition, the complete extensive modeling also works more efficiently and with fewer computational resources.

► This study can well predict cell performances under different operating conditions.
► This study can well predict cell performances under different design parameters.
► Reactant flow rates, pressure drops and cell-voltage components can be evaluated.
► Modeling works can be efficiently completed with less computational resources.