Study on PD Pattern Recognition of Power Transformer Considering External Corona Interference Signal

Transformer plays an important role in the stable operation of the power system. In the process of on-site monitoring, partial discharge often occurs, but all kinds of electromagnetic interference and vibration interference are very serious. In order to check and identify the discharge fault of transformer more efficiently, it is necessary to carry out the research on the discharge identification and classification under the interference in the partial discharge detection of transformer. In this experiment, the partial discharge pattern recognition of power transformer considering corona interference signal was studied. The partial discharge experiment model of transformer was set up, and six discharge types were set up in the transformer. The external corona interference signal was set, and six kinds of discharge signals under corona interference were collected. Based on the BP(Back Propagation) neural network pattern recognition classifier, six kinds of discharge characteristic parameters were classified. The results showed that the accuracy of training confusion matrix was 87.1%, verifying confusion matrix was 86.7%, testing confusion matrix was 91.1%, and overall confusion matrix was 87.7%. In the overall ROC(Receiver Operating Characteristic) curve, the overall classification accuracy was high, and the network output error was gradually close to the best output, which can meet the requirements of pattern recognition accuracy. The results could provide theoretical basis and technical support for online monitoring and fault diagnosis of primary equipment in substation.