Research on Classification Performance of Circuit Breaker Vibration Signal Based on Fuzzy C-Means Clustering Analysis

Abstract

With permanent magnetic actuator monostable circuit breaker as the research object, it gets three different running status signal data, normal circuit breaker closing state , break-brake spring single loss and Institutions unsmooth Through constructing data acquisition platform. The performance of the three methods of feature extraction is compared, short-time Fourier transform, wavelet packet energy entropy and Hilbert huang transform, through fuzzy c-means clustering analysis algorithm. Through the comparative study, it is concluded that adopt wavelet packet transform method to get the best classify performance of time-frequency entropy vector .