Research On Circuit Breaker Mechanical Fault Diagnosis Method Based On Multi-information Fusion

With the continuous expansion of power system construction,the reliability requirements for the protection and control of circuit breakers are also increasing.The probability of mechanical failure in the circuit breaker of the spring-operated mechanism is high,and a variety of signals will accompany the operation.Through the fusion of various characteristic information extracted from these signals,the type of typical mechanical failure of the circuit breaker can be distinguished.Therefore,it is of great significance to study the circuit breaker mechanical fault diagnosis technology for the circuit breaker maintenance and power system operation.Based on the analysis of the basic structure and working principle of ZN65-12 circuit breaker spring operating mechanism,this paper selects appropriate sensors to collect sound signals,vibration signals and closing spring force signals during circuit breaker operation.This kind of signal extracts the effective information to fuse,researches the circuit breaker mechanical fault diagnosis method deeply.Combined with the characteristics of various externally expressed signals during the operation of the circuit breaker,the characteristics of the externally expressed signals during the operation of the spring operating mechanism were first analyzed.Sound,vibration and mechanical signal sensors were selected.Three characterization signals.Secondly,a variety of signal feature extraction methods are listed.Wavelet spectral subtraction is used to denoise the sound signals.Then,the features of sound signals and vibration signals are extracted using local mean decomposition(LMD)and multi-scale entropy algorithms.The time domain method is used.Extract the characteristics of the closing spring force signal.Finally,a multi-information fusion method based on principal component analysis is studied,and the fused features are input into a fireworks algorithm to improve the support vector machine(SVM)model for classification.The example analysis shows that the accuracy and credibility of the circuit breaker accident identification can be effectively improved by applying the method in this paper.The more information is fused at the same time,the higher the accuracy of fault diagnosis is.The feasibility of fault diagnosis methods.