Model-based fault diagnosis of a DC-DC boost converters using hidden Markov model

This paper introduces a hidden Markov model (HMM)-based fault diagnosis technique for DC-DC boost converter. Four HMMs are trained to model parameter variations in the power converters. Each HMM is created based on 14 visible states, which generates probability of each time step matching a signature fault pattern. The proposed method can cover multiple faults that may occur in any element of power electronic circuits. It can also achieve high precision of diagnosing for pre-defined faults in real-time. The simulation results demonstrate an accurate diagnosis performance using HMMs.