Polymer Electrolyte Membrane (PEM) fuel cell performance can be optimized and improved by modeling the complex processes that take place in the various components of a fuel cell. Operability over a range of conditions can be assessed using a robust design methodology. Sensitivity analysis can identify critical characteristics in order to guide hardware and softgoods development. A computational model is necessary which captures the critical physical processes taking place within the cell. Such a model must be validated against experimental data before it can be used for product development. A computational model of an experimental PEM fuel cell has been developed. The model is based on the FLUENT CFD solver with the addition of user-defined functions supplied by FLUENT. These functions account for local electrochemical reactions, electrical conduction within diffusion layers and current collectors, mass and heat transfer in the diffusion layers and the flow channels along with binary gas diffusion. The results of this model are compared to experimental data. A PEM fuel cell consists of an ion conducting membrane, anode and cathode catalyst layers, anode and cathode gas diffusion layers, flow channels, and two bipolar plates. Hydrogen and oxygen are supplied to the anode and cathode respectively. As a result of hydrogen oxidation at the anode catalyst layer, hydrogen ions and electrons are produced. The hydrogen ions are conducted through the membrane to the cathode catalyst layer where they combine with oxygen and electrons to produce water and heat. Therefore, a PEM fuel cell model has to take into account: • Fluid flow, heat transfer, and mass transfer in porous anode and cathode diffusion layers; • Electrochemical reactions; • Current transport and potential field in porous anode, cathode, and solid conducting regions. FLUENT Inc. has developed such a model based on their commercially available FLUENT CFD code. This model was exercised on an experimental Plug Power fuel cell. The voltage characteristic of the model was compared to the experimentally measured values. The preliminary comparison between the predicted polarization curve and the experimental results are very favorable.

This content is only available via PDF.
You do not currently have access to this content.