| Polymer Electrolyte Membrane (PEM) fuel cells present an opportunity to transition to cleaner alternative energy sources such as hydrogen. The use of fuel cells in energy applications requires developing a deeper understanding of how they respond and interact in an integrated system. In this thesis, electrochemical analysis techniques are combined with systems analysis techniques, e.g. frequency and transient response, to investigate the dynamics of PEM fuel cell stacks. Electrochemical Impedance Spectroscopy (EIS), along with an electrochemical model based on constant phase elements (CPE) originally presented by Orazem is used in the analysis. This model is enhanced by providing better estimates of the parameters alpha (a measure of the distributed nature of the fuel cell capacitance), Ce (the effective capacitance in the distributed system), and Rs (the electrolyte resistance), in addition to developing techniques for estimating L (the inductance) and R f (the faradaic resistance). This research evaluates the sensitivity of Orazem's model estimation technique for alpha (and in turn C e) to the inductance in the system. Here, an alternative data filter based on Kramers-Kronig relations, which allows the incorporation of selective high and low frequency data into the model, is also proposed. Voltage-dependent regression models of the fuel cell stack parameters are derived, noting the ability to reduce the order of the regression models if modeled as a function of voltage instead of current. Transient response parameters (step input) are related to frequency response parameters. Further, Chf (high frequency capacitance), as determined by the transient response analysis, is shown to be orders of magnitude smaller than Ce, as determined by the frequency response. |
