Diagnosis of a hydrogen/air fuel cell by a statistical model-based method

This paper proposes a diagnosis method of a hydrogen/air fuel cell using a quasi-static model coupled to a parameter identification method that are both described. An original statistical approach is proposed in an effort to obtain a certain guaranty on the validity of the identified parameters and raise in this way the associated "confidence index". The analysis of the degradation of a fuel cell is afterwards achieved by comparing parameters identified before and after the degradation. A diagnosis is then presented based on the analysis of the different losses occurring within the fuel cell in an effort to monitor and control on-board systems.