Fault Diagnosis in DC-DC Power Converters Based on Parity Equations

Abstract

This paper presents an approach based on parity equations for the problem of fault diagnosis in Buck and Boost converters. The basic idea of this method consists of to formulate a set of residues that compare the difference between the signals measured from the power circuit with an equation model describing the nominal behavior of the power converter. The fault detection algorithm is based on checking the residuals of the inductor current and output voltage. The fault signatures matrix was obtained using a theoretical analysis of residuals in order to establish an appropriate and effective fault isolation algorithm for converters running in continuous mode. Simulation results demonstrate the effectiveness of the method to isolate the main faults in the Buck and Boost power converters.