Sensor fault detection and estimation for switched power electronics systems based on sliding mode observer

Abstract

In this paper, the problem of sensor fault detection and estimation (FDE) for power electronics systems such as DC-DC buck-boost converter is investigated. Firstly, the power electronics system is modeled, and the fault model is given. Then based on the switched Lyapunov function technique, a sliding mode observer (SMO) is designed for the switched error systems, which can guarantee that the error systems are asymptotically stable with a prescribed H∞ performance index. Next, comparing the set threshold with the residual which is produced by the sliding mode observer, the sensor fault detection can be achieved. Meanwhile, based on the method of fault reconfiguration, fault estimation is realized. Finally, in order to show the feasibility of this sensor fault detection and estimation, the simulation of three kinds of sensor faults including open-circuit, gain deviation and noise abnormity is given in this article.