On-line Fault Diagnosis of Proton Exchange Membrane Membrane Fuel Cell Using Polarization Curve

The study investigates performance of using polarization curve for fuel cell fault detection and isolation. As polarization curve can be obtained at practical proton exchange membrane fuel cell (PEMFC) system efficiently, its diagnostic procedure can be used for on-line purpose. The paper starts with background knowledge about polarization curve and its use in fault diagnosis. Polarization curves obtained at a PEMFC system will then be employed to detect abnormal behavior. By studying variation in the gradients of collected curves, PEMFC fault is identified. Furthermore, two classification techniques are applied to isolate different PEMFC conditions, including healthy and faulty conditions. Results demonstrate that by monitoring and analyzing polarization curves, PEMFC fault can be identified accurately, and processing of such information can meet on-line diagnosis requirements.