No abstract

No abstract

We present a novel approach for analyzing the experimental voltage-current curves of a polymer electrolyte membrane ͑PEM͒ fuel cell. State-of-the-art numerical models involve many poorly known parameters. This makes a comparison of numerical and experimental polarization curves unreliable. We suggest characterizing the cell by first using a simplified analytical model, which contains a minimal number of parameters and ignores three-dimensional ͑3D͒ effects. The resulting physical parameters are then used as input data for a 3D numerical simulation of the PEM fuel cell. Comparison of experimental, analytical, and numerical polarization curves enables us to estimate the contribution of 3D effects to the voltage loss. This procedure is performed using specially designed experiments, our recent analytical model, and the newest version of a numerical quasi-3D model of a cell. The results show that this approach may serve as a tool for the optimization of the flow field design.The performance of a polymer electrolyte fuel cell ͑PEFC͒ is determined by several tens of parameters, which describe fundamental electrochemical and physical properties of the membrane electrode assembly ͑MEA͒, operational conditions, geometry of the MEA, and the structure of the flow field. Many of these parameters are strongly coupled. For instance, an increase in temperature improves the kinetics of the electrochemical reactions but decreases water content and conductivity of a polymer electrolyte membrane ͑PEM͒. The overall effect of temperature variation hence depends on the humidification conditions, which in turn depend on the geometry of the flow field. This chain of dependencies is typical for PEM fuel cells. Obviously, experimental investigations of these dependencies are time consuming and expensive, thus the use of modeling activities is desirable.The basic features of fuel cells can be analyzed with onedimensional 1D models that take into account transport across the cell and ignore any variations along the cell surface. 1-6 Twodimensional ͑2D͒ models 7-22 give more detailed information, generating a map of parameters in a cross section of the MEA in one of the two planes: across-the-channel ͑x-y plane, Fig. 1͒ or along-thechannel ͑x-z plane, Fig. 1͒. In essence, either model disregards the distribution of the parameters in the other plane.The most detailed information is provided by fully threedimensional ͑F3D͒ models. 23-28 However, these models are very time consuming. To reduce the run times, usually just a small fragment of the fuel cell is simulated ͑3D element, Fig. 1͒, which typically covers a 10 cm distance along the channel ͑in Ref. 25 and 28 small cells with a meander-like flow field are simulated͒. The effects specific to large cells with long meander channels are beyond the scope of F3D models. Probably for efficiency, the catalyst layers in Ref. 23-27 are replaced by infinitely thin interfaces. Our results show that the distribution of the reaction rate over the catalyst layer volume can be strongly nonuniform. This no...

A new in situ measurement method of mapping the current density distribution in polymer electrolyte fuel cells (PEFC) is used to analyze the performance of a fuel cell under different operating conditions. The present method is useful in investigating the current density distribution in a single cell as well as a stack, which carries the information about the local reactant activity over the electrode area. It was found that the current density close to the gas inlets is strongly influenced by the reactants' relative humidity. The performance close to the gas outlets is greatly influenced by the inlet gas pressures and the stoichiometry factors of the reactant gases, mainly on the cathode side. It was also observed that the performance of the fuel cell drops with the increase in operating temperature if the reactant gases are not sufficiently humidified.