Diagnosis of PEMFC for automotive application

Fuel Cells (FC) are electrochemical energy converters that have higher efficiency and energy density than conventional power sources. Proton Exchange Membrane Fuel Cells (PEMFC) based on low operating temperature; pressure and solid electrolyte are well suited for the automotive application. However, they suffer from low reliability and short lifetime. To improve these two points without increasing the cost, low-cost diagnosis procedures for PEMFCs are needed. This paper focuses on PEMFC fault diagnosis in automotive application, the diagnosis is achieved first through a 3D sensitive modeling of PEMFC operation, and then by training a model based on artificial neural network (ANN) for fault state of health classification. The 3D sensitive model allows determination of temperature and current distribution and was validated experimentally. The 3D validated model can be used for introducing flooding and drying out faults and study the behaviors of the distributions of voltages and currents in three space directions for the purpose of diagnosis of the FC.