Proton exchange membrane fuel cell(PEMFC) technology has been greatly promoted in recent years, but the fault diagnosis and predictive maintenance are unneglectable issues in practical work. According to the safety and reliability requirement of 60kW automotive fuel cell engine designed by our group, a fault diagnosis method based on T-S fuzzy model which is tuned and optimized thanks to particle swarm optimization is put forward in this paper. Its inputs include voltage, the lowest single cell voltage, current, temperature and air pressure, by setting the output threshold of T-S fuzzy model at 0.85,when the healthy degree and its variety rate are below 0.85 and 0.05 respectively, the flooding fault is distinguished, if the healthy degree is below 0.85 but its variety rate is above 0.05,drying of the proton membrane is on-line diagnosed successfully, which can provide a guidance to its real-time monitoring and optimized control in future.
Data-based flooding fault diagnosis of proton exchange membrane fuel cell systems using LSTM networks