Abstract:

A partial discharge diagnosis algorithm for cable accessories based on multi-sensor information fusion is proposed to solve the relative questions.An experimental platform is built.The transient earth voltage sensors(TEV),high-frequency current sensors(HFCT),and ultrasonic sensors(AA)are used to collect multimodal data of partial discharge in cable accessories.A partial discharge phase distribution(PRPD)feature mapis draw.A dataset for partial discharge is constructed.The convolutional neural network(CNN),backpropagation neural network(BPNN),and support vector machine(SVM)algorithms are used for classification and recognition to obtain the recognition credibility of the three algorithms.The intelligent fusion algorithm is obtained through the credibility fusion,and the recognition results show that the intelligent fusion algorithm exhibite high recognition accuracy in the different scenarios to improve the accuracy and reliability of partial discharge diagnosis of cable accessories.

Key words: partial discharge, convolutional neural network(CNN), backpropagation neural network(BPNN), support vector machine(SVM), multimodal data