A Study on PEMFC Faults Diagnosis Based on Wavelet Analysis

The study of Proton Exchange Membrane Fuel Cell (PEMFC) faults mechanism and diagnosis can afford theoretical and technical supports for the future mass production and applications of PEMFC stacks and systems, which need modularization and high reliability. Most of the existing fault diagnosis methodologies, such as Cell Voltage Monitoring (CVM) method, require the knowledge of numerous parameters which may lead to a special inner parameter monitoring setup. The corresponding devices increase the cost and are not suitable for stack modularization and system application. In this paper, a simple and low-cost PEMFC faults diagnosis method using discrete wavelet transform is described. The method only uses the stack information for signal feature extraction. After that, by faults identification and classification, several typical faults including temperature fault, inlet flow inefficiently supplying fault and membrane dehydration fault have been distinguished. A semi-empirical distributed parameter stack model is employed for simulating typical faults that may occur. The simulation results demonstrate that the application of wavelet theory to PEMFC diagnosis is feasible, reliable and promising.