Research On Mechanical Fault Diagnosis Method Of High Voltage Circuit Breaker

High voltage circuit breaker is one of the most important switch devices inpower system, which bears the protection and control function in the system （playsthe role of protecting and controlling in the systems）. In order to ensure its reliableoperation, monitoring of state information and diagnosing the fault based on thecollected signal is very necessary. According to the statistics, the failures of highvoltage circuit breaker are mainly mechanical failures, this paper studied themethod based on vibration acoustic joint mechanical failure of high voltage circuitbreaker fault diagnosis.Firstly, collect a variety of signals of high-voltage circuit breakers, and then theoverall program of a high voltage circuit breaker condition monitoring and faultdiagnosis is designed. Secondly, based on the characteristics of vibration andacoustic signal of the high voltage circuit breaker points in the process of closing,this paper uses the empirical mode decomposition (EMD) to extract the featurevector, and treat this feature vector as SVM input circuit. The typical mechanicalfailures of high voltage circuit breaker are recognized, which is verified byexperiments. Thirdly, a new method of high voltage circuit breaker mechanical faultdiagnosis based on the joint of vibration signal and acoustic signal is proposed. Thismethod apply fast kernel independent component analysis (Fast KICA) to makeblind source separation processing for acoustic signals collected, through theprocessing of acoustic signal and the original vibration signal to improve thecollection of empirical mode decomposition (EEMD), to calculate each intrinsicmode function (IMF) of two-dimension spectrum entropy, and the matrixtransformation matrix of two-dimension spectrum entropy production, as the inputfeature vectors of support vector machine (SVM) to identify the mechanical state ofthe circuit breaker. The last, this paper proposed a new method for fault diagnosisbased on three-dimensional diagram of the vibration-acoustic and time. By usingthe method of hierarchical multiple time-scale signals aimed at buildingthree-dimensional diagram of the vibration, acoustic and time, and then extractthree-dimensional diagram envelope’s six sides and eight angle characteristicvectors as the firstly level of feature, extract object region chord distribution oftwo-dimensional projection of three-dimensional diagram as the secondary feature,based on shape feature classification built circuit breaker fault diagnosis SVM inputvector, and gives practical diagnostic procedures and implementation process.