基于VMD模糊熵与GG聚类的直流配电网故障检测方法  

Fault detection method for DC distribution network based on VMD fuzzy entropy and GG clustering

作　　者：韦延方[1,2] 王志杰 王鹏 曾志辉[1] 王晓卫 WEI Yanfang;WANG Zhijie;WANG Peng;ZENG Zhihui;WANG Xiaowei(School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454003,China;Henan Key Laboratory of Intelligent Detection and Control of Coal Mine Equipment,Henan Polytechnic University,Jiaozuo 454003,China;State Grid Henan Electric Power Company Scientific Research Institute,Zhengzhou 450052,China;School of Electrical Engineering,Xi’an University of Technology,Xi’an 710000,China)

机构地区：[1]河南理工大学电气工程与自动化学院,河南焦作454003 [2]河南理工大学河南省煤矿装备智能检测与控制重点实验室,河南焦作454003 [3]国网河南省电力公司科学研究院,河南郑州450052 [4]西安理工大学电气工程学院,陕西西安710000

出　　处：《电机与控制学报》2024年第2期129-141,共13页Electric Machines and Control

基　　金：国家自然科学基金(61703144);河南省矿山电力电子装置与控制创新型科技团队项目(CXTD2017085)。

摘　　要：针对直流配电网存在的故障信号难以提取、不易对各类故障进行诊断等问题,提出一种基于变分模态分解(VMD)模糊熵与Gath-Geva(GG)聚类的故障检测方法。首先,提取出暂态电流,采用VMD算法将故障暂态电流分解成若干个固有模态分量(IMF)。然后,分别计算分解得到的若干个IMF的模糊熵,将其作为特征向量。最后,采用GG聚类算法对故障特征的特征向量进行聚类识别。GG聚类的主要算法为将聚类样本划分为c类,设出隶属度矩阵,通过设定迭代来计算聚类中心与最大似然估计距离,更新隶属度矩阵,当隶属度矩阵满足条件矩阵时终止迭代,从而实现对单极故障、极间故障以及区外交流侧接地故障的聚类识别。仿真结果表明,所提保护方案可靠性强、准确率高,在不同故障类型、故障位置和过渡电阻等工况下均能可靠检测直流线路故障并准确识别故障类型,且具备一定的抗干扰能力。In response to the difficulties in extracting fault signals and diagnosing various types of faults in DC distribution networks,a fault detection method for DC distribution network based on variational mode decomposition(VMD)fuzzy entropy and Gath-Geva(GG)clustering was proposed.First,the transient current was extracted.The VMD was used to decompose the current into several intrinsic mode functions(IMF).Then,the fuzzy entropy of several IMFs obtained by decomposition was calculated and used as the feature vector.Finally,the feature vector was input into the GG clustering analyzer for cluster recognition.In the specific algorithm of GG clustering,the cluster samples were divided into c categories.The membership matrix was set,the estimated distance was calculated between the cluster center and the maximum likelihood by setting iteration.Then the membership matrix was updated,and the iteration was terminated when the condition matrix is met.Then the cluster identification of singlepole fault,inter-pole fault and ground fault on the AC side outside the zone can be realized.The simulation results show that the proposed protection scheme has strong reliability and high accuracy,which can reliably detect DC line faults and accurately identify fault types under different fault types,fault locations and transition resistance conditions,and has a certain anti-interference ability.

关 键 词：直流配电网 故障暂态电流 变分模态分解 模糊熵 Gath-Geva聚类 故障检测 

分 类 号：TM721[电气工程—电力系统及自动化]