基于最小二乘双支持向量机的配电网短路故障辨识方法研究  

Research on identification method of short-circuit fault in distribution network based on least squares twin support vector machine

作　　者：卢晓强 李刚 周万竣 张逸凡 李勇汇 石勇[3] 王耀[3] LU Xiaoqiang;LI Gang;ZHOU Wanjun;ZHANG Yifan;LI Yonghui;SHI Yong;WANG Yao(Western Maintenance Test Branch,China Southern Power Grid Peak Load Regulation and Frequency Modulation Power Generation Co.,Ltd.,Xingyi 562400,China;School of Electrical Engineering and Automation,Wuhan University,Wuhan 430072,China;Nanjing Nanrui Relay Electric Co.,Ltd.,Nanjing 211102,China)

机构地区：[1]南方电网调峰调频发电有限公司西部检修试验分公司,贵州兴义562400 [2]武汉大学电气与自动化学院,湖北武汉430072 [3]南京南瑞继保电气有限公司,江苏南京211102

出　　处：《武汉大学学报（工学版）》2022年第4期401-408,共8页Engineering Journal of Wuhan University

基　　金：南方电网调峰调频发电有限公司科技项目(编号:021900KK52190002)。

摘　　要：配电网短路故障时有发生,若无法及时辨识和清除故障,对电网的稳定运行和用户的生产生活都将造成严重影响。为了对配电网短路故障类型进行准确辨识,对传统支持向量机算法进行了改进,搭建了基于决策树和最小二乘双支持向量机的多分类辨识模型,并将其应用于配电网短路故障辨识中。首先在短路故障发生后对采样后的电流暂态信号作希尔伯特-黄变换,其次分别提取变换后各电流分量的能量特征构成特征向量,最终输入多分类辨识模型,实现对配电网短路故障类型的快速准确识别,并通过仿真验证了该方法的可靠性和有效性。Short-circuit faults in distribution networks occur from time to time.If the faults cannot be identified and cleared in time,there will be a serious impact on the stable operation of the power grid and the production and life of users.In order to identify the types of short-circuit faults in the distribution network accurately,this paper improved the traditional support vector machine algorithm,built a multi-class identification model based on decision trees and least squares twin support vector machine,and applied it to the short-circuit fault identification in distribution networks.First,Hilbert-Huang transformation was performed on the sampled transient current signal after the occurrence of a short-circuit fault.Secondly,the energy characteristics of each current component after the transformation were extracted to form a feature vector,which was finally input to the multi-class identification model to realize the rapid and accurate identification of the types of short-circuit faults in the distribution network.The reliability and effectiveness of the method were verified by simulation.

关 键 词：短路故障 类型辨识 希尔伯特-黄变换 最小二乘双支持向量机 

分 类 号：TM743[电气工程—电力系统及自动化]