Active Adaptive Fault-Tolerant Control Design for PEM Fuel Cells

A novel control design is introduced for pressure control in proton exchange membrane (PEM) fuel cells. The proposed control design is able to control the system under faults in actuators. In this design, an artificial neural network (ANN) observer is developed for online fault detection and isolation (FDI) in the pressure valves of the PEM fuel cell (PEMFC). Extended Kalman Filter (EKF) is used to update ANN weights which resulted in real-time and more accurate FDI. Based on the feedback data received from the designed FDI system, a nonlinear controller is designed to compensate for the fault effects in real time. The simulation results and comparison with state of art control teckniques demonstrate the advantages of the proposed active fault tolerant controller (FTC) in presence of a fault in the actuators and its ability to detect faults and compensate for their effect in the system performance.