Research On Cable Joint Defect Detection Based On Multi-feature Fusion

As the main equipment of urban transmission and distribution network,power cables are of great significance to the normal operation of power systems to ensure their safety and reliability.Compared with the cable body,the cable joint is very easy to cause main insulation defects during the manufacturing and installation process due to its special and complex structure.With the long-term operation of the cable,the existence of defects will gradually lead to the deterioration and failure of the insulation performance,which will lead to the occurrence of power accidents.Therefore,accurate and timely detection and diagnosis of the defects of cable joints is of great significance to the later maintenance of equipment and stable operation of the distribution network.At present,the diagnosis method of cable defective joint is usually to perform feature extraction and pattern recognition on a single discharge signal.Affected by factors such as detection environment and human operation,the diagnosis results have the problem that the recognition rate is not high enough,and the rich insulation state information cannot be mined to the maximum extent.Therefore,a joint defect diagnosis method based on multi-feature fusion to improve the identification rate of defect type diagnosis was proposed.The main research contents are as follows:(1)Aiming at the phenomenon of temperature change and partial discharge when the cable joint has defects,a finite element simulation model was established for the air gap defect and moisture defect of the cable joint by using COMSOL Multiphysics software,and the temperature field and electric field distribution under different defects were analyzed.;(2)On the basis of simulation research,a joint temperature detection platform based on fiber Bragg grating temperature sensor and a joint partial discharge detection platform based on pulse current method were built,and the data acquisition of temperature and partial discharge signals of two types of defects was completed.The temperature variation characteristics of the main insulation outer layer of the two types of defective joints were compared and analyzed by data processing software;meanwhile,the characteristic parameters of the partial discharge signal in the time domain distribution and the phase distribution were extracted respectively,and they were used as the support vector machine algorithm.Input,the results of the joint defect diagnosis with different feature quantities were obtained;(3)In order to improve the recognition rate of joint defect diagnosis results,a multi-feature fusion recognition system based on DS evidence theory was established,which realized the information fusion of temperature features,TRPD features and PRPD features.The average recognition rate after fusion can reach 93.75%.The detection accuracy and reliability of joint defects have been effectively improved.