Classification and identification of temperature status of gas insulated switch gear isolation switch contacts based on a learning vector quantization neural network

In the high-voltage power transmission and transformation system, the safety and reliability of gas-insulated switchgear (GIS) has become the key to maintaining safe operation of high-voltage power grids. The problem of local temperature rise of GIS is one of the important factors affecting the safe and stable operation of GIS. In this article, the temperature change of the GIS isolation switch contact is detected by changing the contact resistance and using fiber Bragg grating (FBG) sensors. Using experience to judge the thermal state of the isolation switch contacts, the traditional identification method involves significant subjectivity. For this reason, a learning vector quantization (LVQ) identification mode is proposed. The LVQ neural network has the advantages of simple mode, self-learning, and self-organization. It is very suitable for constructing the nonlinear mapping relationship between GIS isolation switch contacts temperature and working state to identify the thermal state of GIS isolation switch contacts.