Clean and efficient conversion of energy to useful forms is an important societal need. This has brought fuel cells into prominence as a potential source for energy conversion. Among the H2-O2 fuel cells, Proton exchange membrane fuel cells (PEMFC) are in an advanced state of development with potential applications in portable power generation. They are also considered the most promising technology for obtaining vehicles with intrinsic possibility of zero emissions. In spite of the advances made in PEMFC, technical advances and breakthroughs are still required to make this technology viable. Cost and reliability are two critical issues in the commercialization of the fuel cell technology.

Achievement of the reliability goals will depend on detailed phenomenological understanding of the various failure modes, and the ability to predict the behavior of fuel cells under these failure modes. Electrochemical impedance analysis (EIS) is an important approach and has been widely used for diagnostics. Several papers exist on the use of EIS for diagnostics and this appears to be the state-of-art in fuel cell diagnostics. Much of the previous work focused individually on a single problem such as flooding, anode poisoning and so on. It is also noted in that with just EIS, there is a risk of confusion when discriminating between flooding and catalyst poisoning.

In this paper, a detailed dynamic model for a PEM fuel cell will be discussed. We will also present comprehensive validation results for this model at different operating conditions. Further, we will outline results on using wavelet analysis of current transients for fault diagnosis in PEM fuel cell systems. We analyze the time-domain properties of the wavelet coefficients at different scales, which represent the frequency information for diagnostics. This provides a combined time-frequency analysis that is missing in pure spectral or pure time domain analysis.

Extended Abstract Status: Not Uploaded