High Voltage Circuit Breaker Mechanical Fault Diagnosis Based On Combined Sound And Vibration Analysis

High-voltage circuit breakers are important protection devices for power grids,and their operation status directly affects the reliability and stability of power grids,and once a fault occurs,it will bring huge economic losses or even cause personal injuries.In order to improve the accuracy of mechanical fault diagnosis of high-voltage circuit breakers,a high-voltage circuit breaker fault diagnosis system based on the joint diagnosis of sound and vibration signals is proposed.Firstly,the sound and vibration signals were collected using the laptop’s own sound card and vibration sensor,and the data acquisition card and Lab VIEW formed an acquisition system,which realized the collection of sound and vibration signals in five states: normal tripping,tripping with energy storage spring off,tripping with worn solenoid contact,tripping with loose base screw and tripping with cocked lever paddle.Secondly,the sound signal is noise-reduced by wavelet transform method and EEMD decomposition is performed to find the corresponding energy entropy as feature values,which are input to ANFIS network and WOA-SVM classifier for state identification respectively.The method of dimensionality reduction of correlation data is introduced to remove redundant information in the signal and further improve the accuracy of fault diagnosis.The vibration signal is then denoised by the wavelet packet transform,and the Hilbert yellow transform is used to find the envelope value of the signal and the corresponding energy entropy as the feature value to input into the WOA-SVM classifier for state identification.The PCA data dimensionality reduction method is introduced to further improve the accuracy by effectively compressing the data while retaining a large amount of useful information.Finally,the BPAs of the sound and vibration signals are derived from the diagnostic results of the support vector machine,and the BPAs of the sound and vibration signals are weighted by the accuracy of each state of the single signal,respectively,to perform DS evidence theory-based high-voltage circuit breaker fault diagnosis on the sound and vibration signals at the decision level.The study shows that the joint diagnosis method further improves the accuracy of fault diagnosis on the basis of single signal diagnosis results compared with traditional single signal fault diagnosis results,and provides a strong theoretical support and experimental basis for online monitoring of circuit breakers.