A fast fault diagnosis method based on reset augmented observer

This paper presents a novel fault diagnosis approach based on the reset observer to improve the rapidity as well as the accuracy of fault diagnosis. By introducing an integral reset mechanism into a traditional observer, the reset observer overcomes the linear fundamental limitations and can decrease the overshoot and settling time simultaneously in the estimation process. In this paper, an augmented system is firstly constructed by considering the actuator fault as an auxiliary state vector. Secondly, the structure of the proposed reset observer for the augmented system is introduced. Based on the Lyapunov stability theory, we give the sufficient and necessary condition of stability analysis and formulate it as an LMI problem which can be solved easily. In order to verify the superiority of this method, it is compared with the proportion integral observer without reset mechanism. The simulation results demonstrate that the reset observer method can achieve fault detection and diagnosis accurately and rapidly.