Parameterization and prediction of temporal fuel cell voltage behavior during flooding and drying conditions

This paper describes a simple isothermal two-phase flow dynamic model that predicts the experimentally observed temporal behavior of a proton exchange membrane fuel cell stack. This model is intended for use in embedded real time control where computational simplicity is of critical importance. A reproducible methodology is presented to experimentally identify six (6) tunable physical parameters based on the estimation of the cell voltage, the water vapor transport through the membrane and the accumulation of liquid water in the gas channels. The model equations allow temporal calculation of the species concentrations across the gas diffusion layers, the vapor transport across the membrane, and the degree of flooding within the cell structure. The notion of apparent current density then relates this flooding phenomena to cell performance through a reduction in the cell active area as liquid water accumulates. Despite the oversimplification of many complex phenomena, this model provides a useful tool for predicting the resulting decay in cell voltage over time only after it has been tuned with experimental data. The calibrated model and tuning procedure is demonstrated with a 1.4 kW (24 cell, 300 cm2) stack, using pressure regulated pure hydrogen supplied to a dead-ended anode, under a range of operating conditions typical for multi-cell stacks.