Fault Detection for Proton Exchange Membrane Fuel Cell using Adaptive Extended Kalman Filter

One promising technique for producing electricity that is both clean and high-performing is proton-exchange membrane fuel cells or PEMFCs. It is important to diagnose the fault in the PEMFCs system to ensure its performance and safety in operation. Drying is one of the faults in PEMFCs that occurs when relative humidity drops and temperature uplift takes place simultaneously. Here, we proposed the fault detection (FD) model utilizing an adaptive extended Kalman filter (AEKF) to detect drying that affects the membrane and catalyst's proton conductivity by implementing three drying scenarios based on severity. PEMFCs were used to realize the electrochemical properties and accommodate the characteristics of drying. An AEKF was used as a residual generator and fixed threshold for the residual evaluation. The findings show that an increase in temperatures and a decrease in relative humidity representing the drying condition caused the membrane’s water uptake and catalyst to drop. It further decreased the PEMFC’s proton conductivity and performance. The results of the proposed FD scheme using AEKF showed that the fault detection succeeded in detecting all drying fault scenarios within less than one second.