基于SVM的输变电线路故障原因分析  

Analysis of Line Fault Reasons of Power Transmission and Transformation Based on SVM

作　　者：李维[1] 张利强[1] 魏娇龙 LI Wei;ZHANG Liqiang;WEI Jiaolong(Beijing Sifang Relaying Automation Co.,Ltd.,Beijing 100080,China)

机构地区：[1]北京四方继保自动化股份有限公司,北京100080

出　　处：《东北电力技术》2023年第10期31-35,共5页Northeast Electric Power Technology

摘　　要：输变电线路由于长年运行在恶劣的自然环境中,故障跳闸现象时有发生,为快速找到故障点和故障原因,提出基于支持向量机(SVM)的输变电线路故障原因分析模型,通过对154条样本数据进行探索分析,提取故障不同时期的谐波分布特征,构建专家样本集;采用一对一模式的支持向量机进行模型学习;利用混淆矩阵和准确度指标对已建立的故障原因识别和预测模型进行评价,所建SVM预测准确率可达81.7%,为故障原因识别分析提供一定参考。Due to power transmission and transformation lines have been operating in harsh natural environments for many years,fault tripping often occurs sometimes.In order to quickly identify the fault point and cause,a support vector machine(SVM)based fault cause analysis model for power transmission and transformation lines is proposed.Through the exploration and analysis of 154 sample data,it extracts the harmonic distribution characteristics of different fault period,and constructs an expert sample set.It uses one-to-one support vector machine for model learning.The confusion matrix and accuracy index are used to evaluate the established fault cause identification and prediction model.The prediction accuracy of SVM built reaches 81.7%,which provides reference for fault cause identification and analysis.

关 键 词：输变电线路故障 支持向量机 故障录波 混淆矩阵 

分 类 号：TM75[电气工程—电力系统及自动化]