Adaptive unknown input observer approach for multi-fault diagnosis of PEM fuel cell system with time-delays

An adaptive unknown input observer (AUIO) for polymer electrolyte membrane fuel cell system with time-delays is developed and applied to formulate a fault diagnosis algorithm for multiple actuator fault detection and isolation (MAFDI). In a scheme based on multiple-model, a group of observers are developed based on a model that describes the system in the presence of a specific actuator fault. The observers are used to generate fault-dependant residual signals. When a model matches the system, the residual signal is zero and becomes non-zero in case of a faulty scenario. Two types of actuator faults are considered namely stuck fault, where the inputs delivered by the faulty actuators are blocked to constant value, and loss of effectiveness fault. The effectiveness of the proposed AUIO and MAFDI algorithm is shown with simulation results. © 2020 Northeastern University, China.