Data-driven multi-fault approach for H2/O2 PEM Fuel Cell diagnosis

Proton Exchange Membrane Fuel Cells (PEMFCs) are promising devices in energy conversion domain. Improper operating conditions can severely affect their performance. Particularly, improper water management, fuel quality and reactants starvation conditions, if recurrent or continued for a long period can have critical effects, introducing degradations phenomena and reducing the FC lifespan. This work aims to investigate the impact of improper water managements and reactants starvation conditions on stack voltage and Electrochemical Impedance Spectroscopy (EIS) measurements for multi-fault detection purposes. In-house tests performed on H-2/O-2 PEMFC short-stacks performance are presented. Data are then analyzed and processed through a double-fuzzy multi-fault detection approach. The procedure can be also used for H-2/Air PEMFC diagnosis. Experimental data and data-driven multi-fault detection procedure improvement are developed in the framework of the European Project Health-code.