Gas-insulated Switchgear Discharge Pattern Recognition Based on KPCA and Multilayer Neural Network

Identifying the discharge type based on partial discharge signals in Gas-insulated Switchgear (GIS) equipment is very important for fault diagnosis and early warning of GIS equipment. This paper proposes a method to recognize different typical discharge types of GIS equipment through experiment. Firstly, different discharge defect models of GIS are built, and UHF sensors are used to detect and obtain partial discharge signals of discharge defects. Then, 3-dimentional PRPS patterns of different defects are drawn based on discharge signals. After that, 2-dimentional Gabor filter is applied to these patterns and texture and shape features are obtained. The dimension of the characteristics of discharge patterns is reduced by Kernel Principal Component Analysis (KPCA), and a multi-layer neural network is built to train and test the samples of discharge data. The results show that the recognition accuracy of three discharge types by using the multi-layer neural network can reach 83.3%, which is better than other algorithms, and can accurately diagnose the defect types of GIS.