A signal-based method for fast PEMFC diagnosis

•A novel signal-based approach is proposed for the PEMFC fault diagnosis.•The state-of-health is estimated by using the wavelet transform approach.•The results are coming from the analysis of the energy and the entropy of a signal.•The robustness of the method is carried out by using a large database.•Experimental data verify the efficiency of the method for an air supplying fault. This paper deals with a novel signal-based method for fault diagnosis of a proton exchange membrane fuel cell (PEMFC). Thanks to an in-lab test bench used for the experimental tests, various parameters can be recorded as electrical or fluidic measurements. The chosen input signal for the diagnosis uses no additional expensive and no intrusive sensors specifically dedicated for the diagnosis task. It uses insofar only the already existing sensors on the system. This paper focuses on the detection and identification of a high air stoichiometry (HAS) fault. The wavelet transform (WT) and more precisely the energy contained in each detail of the wavelet decomposition is used to diagnose quickly an oversupply of air to the fuel cell system. Finally, some experimental results are presented according to different input signals, in order to prove the efficiency of the patented method.