Design of a Fault Diagnosis System for PEM Fuel Cells

This paper develops a fuel cell fault diagnosis system for Proton Exchange Membrane Fuel Cells (PEMFCs). First, it uses a set of ZigBee wireless sensors with a Modbus interface to accomplish PEMFC condition monitoring. The sensor modules are used to construct the data for PEMFC voltage, current, temperature and supplied gas pressure. The fault diagnosis system divides 7 fault types and uses 13 output characteristics to construct the matter-element models. Then, using an extension neural network (ENN) based method, builds the fault diagnosis mathematical platform. The hardware for the fault forecasting system is constructed by a Lab VIEW man-machine interface. To demonstrate the effectiveness of the proposed method, comparative studies using a multi layer neural network (MNN) and k-means algorithm are conducted on 400 sets of PEMFC field test patterns with rather encouraging results.