Research On Intelligent Detection Technology For Partial Discharge Of 35kV Cable Terminal

With the rapid development of the electric power industry, the number of XLPE cable and its accessories is increasing day by day. With the running time increased, the possibility of partial discharge is also increased. Since the1970s, the partial discharge detection equipment for fault location, partial discharge magnitude calibrating and aging study was made. In consideration of the characteristics of partial discharge, water tree and electricity tree, system with partial discharge detection and intelligent fault diagnosis is built for35kV cable and terminal, which is based on the high frequency pulse current method.The principles, detection methods, and relative experimental results of partial discharge, water tree and electric tree are analyzed in detail in this article. In order to study the features such as field intensity distribution and potential distribution inside the cable termination, partial discharge defect models with different sizes and shapes were built and simulated by finite element method where partial discharge may occur. This is used to further study on the partial discharge mechanism inside the terminal, and provide theoretical reference to its aging diagnosis. The conclusion of simulation indicates that the fracture of the copper shield is easy to cause the electric field distortion and partial discharge with high field strength. The electric field strength of the incise lesion defect is higher than that of the sag one at the same place where might cause partial discharge. The more acute the defect is, the more serious the electric field distortion becomes and the greater the field strength is. The local field strength is greater than the air breakdown voltage30kV/mm when the defect exists near the copper shield fracture which is more likely to cause partial discharge. The closer to the copper shield fracture, the greater the field intensity becomes.This article applies rough set and neural network to pattern recognition of partial discharge type and diagnostic analysis of ageing at the cable terminal. Based on this method, system for35kV cable and terminal partial discharge detection and intelligent fault diagnosis is built. The system integrates the application of virtual instrument technology, RFCT, and wireless communication technology, and professional data acquisition equipment of partial discharge. It has the ability to display the real time amplitude spectrum, record and analyze data, view the history data, rough set-neural network intelligent diagnosis, and user information management.Comparing the results of rough set-neural network method and BP neural network method, we get that the rough set-neural network method have better partial discharge type recognition rate and fault tolerance, and the training time is lower than the BP neural network.