Fault detection method for buck converter based on inverse kalman filter

Abstract

The disclosure relates to a fault detection method for a Buck converter based on an inverse Kalman filter, which belongs to the field of fault detection for power converters. The method includes: establishing a hybrid discrete model of a non-ideal Buck converter in Continuous Conduction Mode (CCM); acquiring a state equation and an observation equation of the Buck converter; acquiring the input voltage, the switching state, the inductive current and the output voltage in a working state, and determining the state matrix of the Buck converter; estimating a parameter matrix at time k according to the state matrix, the state equation and the observation equation; determining an estimated value of each component according to the estimated parameter matrix of the Buck converter at time k; and determining whether the Buck converter is faulty or not according to the estimated value of each component. The problems of poor real-time performance and low accuracy of fault diagnosis for the Buck converter at present are solved, and the real-time performance and accuracy of fault detection for the Buck converter are improved.