Study On The Algorithm For Predicting The Reliability And Lifetime Of XPLE Insulated Cable

Since the1970s, the production and application of XLPE power cable has been rapiddevelopment in China. With the increase in the number of cable and the extension of runningtime, early aging of XLPE cable laid frequent failures, resulting in insulation breakdownaccident and power outage. This inconvenience to the people’s lives, and causing significanteconomic losses and security risks. Therefore, to study life prediction of XLPE power cableshas important security implications and economic value.The survey found that the cable failure is often caused by the ageing of the insulation,which will be affected by electricity, thermal, mechanical, environmental factors.Thus, thispaper first comprehensive analysis of the cable insulation aging and aging mechanisms.Then,The single factor aging models and the principle of accelerated aging test was studied. On thisbasis, analysis of the synergy of multiple factors aging, and comparative study several modelsto describe combined thermal-electrical aging.Through the statistics of the cable fault causes analysis found that, in addition toinsulation aging, the external damage and improper installation also account for a large part,which makes cables fault before reaching the insulation life. Therefore, in order to correctlypredict the cable running life, besides considering the cable insulation life, but also need topredict the reliability of the power cables in service.Based on the study of the XLPE cable insulation life, multi-factor aging models, the cablefailure analysis and reliability, this paper proposes a new life prediction method based on thecable reliability, and gives the XLPE cable life prediction index system and system model.The cable reliability model based on Weibull distribution is established. For small samples,the LS-SVM regression method is proposed for parameter estimation. Finally, through the lifeprediction and reliability analysis of8.7/10KV XLPE cable, the effectiveness of the modelwas verified,and accumulated some experience.