Relative Wavelet Energy as a Diagnosis Tool for PEM Fuel Cells

This paper deals with the relative wavelet energy as a diagnosis tool for PEMFC. The reliability and the durability of the fuel cell system have to be improved and the diagnosis approach is a way to extend the PEMFC lifetime. A signal- based method is used to estimate the state of health of the system. The energetic distribution is given according to different experimental conditions. The energy contents of a voltage signal are selected and analyzed to define the PEMFC's state of health. A clustering step is added to identify a fault. This study is based on a high air stoichiometry fault.