Vibration Detection and Analysis of Partial Discharge Defects in GIS

Partial discharge is one of the main reasons for the performance decline and damage of GIS equipment. Moreover, different types of discharge defects cause different kinds of hazards to the equipment, and the corresponding treatment measures also have certain differences. The vibration signals of GIS equipment under normal working conditions and three types of partial discharge defects, namely surface, suspension and tip, were simulated by using the GIS partial discharge fault defect testing platform. The vibration signal data were collected and sorted out. Through the comparison and analysis of the results in the time-frequency domain, the characteristic differences of different defects were analyzed. Two time-domain and five frequency-domain characteristic quantities were mainly extracted to explore the distribution laws of different defect characteristic quantities. Finally, the Support Vector Machine model based on Bayesian optimization algorithm (BO-SVM) is obtained, and the partial discharge defects of GIS equipment are diagnosed. The results show that the diagnostic accuracy rate of the BO-SVM model for the four test states is 92.3%, providing a certain reference for the fault diagnosis of partial discharge defects in GIS equipment.