Research On Defect Analysis And Identification Method Of Power Cable And Accessories

As a widely used cable line category,the safe and stable operation of XLPE and its accessories is one of the important factors influencing the safety of power grid.Partial discharge of XLPE can lead to local destruction of dielectric,especially for organic dielectric.Because the partial discharge is closely related to the insulation state,if it exists for a long time,it can destroy the electrical strength of the insulation device under certain conditions.Therefore,it is of great practical significance to ensure the reliability of power grid and improve the quality of power supply to accurately identify cable defects and timely diagnose faults caused by defects through online monitoring and identification of cable insulation performance.Therefore,in this paper,XLPE insulation partial discharge characteristics,XLPE and its accessories defects identification research,specific research includes the following three points:（1）The typical defects of XLPE power cable and cable accessories were simulated and analyzed.Firstly,seven typical defects of XLPE power cables and accessories are introduced and constructed,and the types and sizes of defects are recorded.Then,the reasons of defects are analyzed,and the influence of defects on cable operation safety is expounded.Finally,the forming conditions of partial discharge and the mechanism of partial discharge caused by defects are analyzed,and the discharge detection of defective cables is carried out.The influence of different defects on cable operation is explained.（2）A graph-based semi-supervised learning method is proposed to optimize the sample space of the partial discharge signals of XLPE cables and accessories.Firstly,the waveform characteristics and equivalent time-frequency characteristics were analyzed,and the characteristic distributions of four types of discharge pulses in timefrequency diagrams were sorted out and analyzed.Furthermore,the multi-scale line modulated characteristic discharge pulses were extracted based on time domain analysis,and the sample space was constructed.Finally,the partial discharge identification method based on the graph semi-supervised learning method is carried out,and the similarity of the samples is judged.The simulation results show that the proposed method is effective.（3）Combined with the actual defects in cable operation,the proposed defect identification method was tested and analyzed on site.Firstly,the cable testing process and wiring method in the field test are introduced.Then,according to the cable defects mentioned above,the cable is tested for the local,internal and body defects of the cable in turn,and the proposed defect identification method is used to identify the test results.Finally,the correctness and reliability of the proposed identification method is verified by comparing with the test cable dissections.