基于改进XGBoost算法的XLPE电缆局部放电模式识别研究  

Research on partial discharge pattern recognition of XLPE cable based on improved XGBoost algorithm

作　　者：刘维功 王昊展 时振堂 黎德初[3] 胡学良[3] 李劲松 Liu Weigong;Wang Haozhan;Shi Zhentang;Li Dechu;Hu Xueliang;Li Jinsong(Sinope Dalian Research Institute of Petroleum and Petrochemicals,Dalian 116045,Liaoning,China;School of Electrical Engineering,Dalian University of Technology,Dalian 116024,Liaoning,China;Sinopec Guangzhou Company,Guangzhou 510726,China)

机构地区：[1]中国石油化工股份有限公司大连石油化工研究院,辽宁大连116045 [2]大连理工大学电气工程学院,辽宁大连116024 [3]中国石油化工股份有限公司广州分公司,广州510726

出　　处：《电测与仪表》2022年第4期98-106,共9页Electrical Measurement & Instrumentation

基　　金：国家自然科学基金青年科学基金(51807106);济宁市重点研发计划(2020PZJY006);中央高校基本科研业务费(DUT20RC(3)018);中国石油化工股份有限公司科技项目(CLY21062)。

摘　　要：局部放电模式识别对交联聚乙烯(XLPE)电缆绝缘性能的判定具有重要意义。在XLPE电缆的局部放电模式识别的研究中,传统机器学习算法存在收敛速度慢、易过拟合、识别准确率低等问题。文章采用一种基于改进XGBoost算法的XLPE电缆局部放电模式识别方法。通过搭建电缆局部放电试验平台人为构造四种35 kV XLPE电缆局部放电缺陷模型进而获取原始数据,利用MATLAB软件完成统计特征参数的计算,以特征参数为输入量,放电类型预测结果为输出量,通过交叉验证、学习曲线确定最优参数进而得到有效的模式识别模型。实验分析结果表明,与决策树、随机森林、BP神经网络和SVM等局部放电模式识别方法相比,文中方法可进一步提升识别准确率,总体识别准确率为96.93%。Partial discharge pattern recognition is of great significance to the determination of the insulation performance of cross-linked polyethylene(XLPE)cables.In the study of partial discharge pattern recognition of XLPE cables,traditional machine learning algorithms have problems such as slow convergence,easy over-fitting,and low recognition accuracy.In response to this problem,this paper adopts an improved XGBoost algorithm-based partial discharge pattern recognition method for XLPE cables.By building a cable partial discharge test platform,four types of 35 kV XLPE cable partial discharge defect models were artificially constructed to obtain the original data,and the MATLAB software was used to complete the calculation of statistical characteristic parameters.The characteristic parameters were used as the input and the discharge type prediction result was the output.An effective pattern recognition model is obtained by cross validation and the learning curve to determine the optimal parameters.The experimental analysis results show that compared with the partial discharge pattern recognition methods such as decision tree,random forest,BP neural network and SVM,the method proposed in this paper can further improve the recognition accuracy rate,and the overall recognition accuracy rate is 96.93%.

关 键 词：XLPE电缆 模式识别 XGBoost 局部放电 学习曲线 

分 类 号：TM855[电气工程—高电压与绝缘技术]