Abstract

During long-term operation, cables and their accessories may experience partial discharge,which can easily lead to insulation degradation or insulation faults.This article adopts Convolutional Neural Network (CNN) and introduces genetic algorithm to optimize the CNN network model,achieving pattern recognition for typical discharge types.The research results indicate that the CNN model optimized by genetic algorithm has significantly improved recognition accuracy compared to traditional methods,and this method has significant advantages in improving the accuracy and robustness of cable terminal partial discharge recognition.