Robust fault diagnosis for DC–DC Boost converters via switched systems

Abstract

In this paper, an effective system for fault detection and fault identification is proposed for DC–DC Boost converters, and the effectiveness of the method is proved from theoretical analysis, simulation, and experimental verification respectively. The switched systems are modeled and some of the fault models are analyzed based on the switched characteristics of the DC–DC Boost converter firstly. Then, the core of the method in this paper is to design a robust linear switch fault detection observer and a set of fault recognition observers. The fault detection observer detects a fault based on the residual among physical output and estimated output. Once the system detects faults, it automatically turns on the identification observers. Meanwhile, the system selects appropriate adaptive rules to estimate the value of the faulty component and achieves the adaptive parameter identification. The location of the fault is identified by comparing several residual evaluation functions. Finally, simulation and semi-physical simulation results testify to the effectiveness of the proposed approach.