Research on Fault Diagnosis Method of Circuit Breaker Mechanical Characteristics Based on Relevance Vector Machine

For researching fault diagnosis method of circuit breaker mechanical characteristics, the circuit breaker with permanent magnetic actuator is taken as an example. The fault mechanism is analyzed, and the monitor platform is built in this paper. Aiming at mechanical vibration signal of circuit breaker in closing process, the wavelet packet energy entropy method is adopted in order to extract feature. The diagnosis method based on relevance vector machine theory is researched for diagnosing circuit breaker mechanical characteristic fault. The fault diagnosis model is built. The fault diagnosis experiments which include normal closing, opening spring falling and operating mechanism jam fault are made on circuit breaker prototype. Experimental conclusions indicate that the fault diagnosis results agree with the actual fault types of circuit breaker, which calibrates the accuracy of diagnosis method proposed in this paper.