Novel Feature Set Based on Bayesian Approach for Circuit Breaker Fault Detection

Abstract

Smart grid is developing rapidly with the promotion of artificial intelligence (AI) technology, and online condition assessment of circuit breaker (CB), as one of the most important facility in power system, is increasingly requested. Owing to most of CB faults can be reflected on its coil current (CC), lots of papers research on the relationship between CC data and CB fault. This paper proposed a new method to get distinctive CC feature set, which achieves better data distinction and has more practical physical meaning. The classification experiments based on the new feature set show that the proposed method can quickly and accurately recognize fault, meanwhile point out the malfunctioning device.