GIS machinery fault diagnosis based on VMD information entropy and CNN-SVM

Aiming at the problems of low accuracy of mechanical fault identification and difficult fault feature extraction of high-voltage circuit breakers in GIS, a mechanical fault diagnosis method combining information entropy and VMD-CNN-SVM is proposed. Firstly, the GIS experimental platform is built, and the breaking vibration signals under three working conditions are collected; secondly, in order to effectively extract the fault information of the high-voltage circuit breaker, the components containing richer fault information are identified by taking the minimum information entropy as the constraint function and optimizing the VMD parameters using the SABO algorithm; after that, the feature values are inputted into the CNN network to carry out the feature extraction; finally, the feature values extracted by CNN are inputted into the SVM classification method; and finally, the feature values extracted by CNN are inputted into the SVM classification method to carry out the feature extraction. Finally, the CNN-extracted feature values are input into the SVM classifier for classification prediction. The results show that the accuracy of the proposed method is improved compared with other diagnostic methods.