基于蝠鲼觅食算法优化支持向量机的接地线定位方法  

Grounding line localization method based on manta ray foraging algorithm optimization support vector machine

作　　者：李阎君 张斌 黄翰 王雪蒙 董子健 张占龙[2] 邓军[2] LI Yan-jun;ZHANG Bin;HUANG Han;WANG Xue-meng;DONG Zi-jian;ZHANG Zhan-long;DENG Jun(Bishan Power Supply Branch of State Grid Chongqing Electric Power Company,Chongqing 402760,China;State Key Laboratory of Power Transmission Equipment&System Security and New Technology(Chongqing University),Chongqing 400030,China)

机构地区：[1]国网重庆市电力公司璧山供电分公司,重庆402760 [2]输配电装备及系统安全与新技术国家重点实验室(重庆大学),重庆400030

出　　处：《电工电能新技术》2022年第12期35-43,共9页Advanced Technology of Electrical Engineering and Energy

基　　金：国家自然科学基金青年基金项目(52007011)。

摘　　要：10 kV配网线路检修因漏拆临时接地线引起的带地线合闸恶性事故频发,现有临时接地线漏拆诊断及定位方法诊断效果差、定位精度低,为此本文提出了一种基于蝠鲼觅食算法优化支持向量机的接地线定位方法。搭建有漏拆接地线时的配网线路π形等值电路计算模型,建立回路阻抗与注入信号频率和漏拆接地线位置的关系,将线路最大阻抗值对应的特殊频率点作为漏拆接地线位置的判据,采用PSCAD/EMTDC搭建仿真模型验证计算模型的准确性;将蝠鲼觅食算法应用于支持向量机的核心参数优化,提高模型预测的准确性,建立了特殊频率点与漏拆接地线位置的关系模型,并与另外3种常用算法进行对比分析,结果表明本文所提方法能够实现漏拆接地线的诊断和定位,定位精度达到98%以上。For 10 kV distribution network maintenance the missing of removing the temporary grounding wire will cause the occurrence of grounded short circuit accidents.The existing diagnosis and positioning methods have poor results and low positioning accuracy.The paper proposes a new method based on manta ray foraging algorithm optimization support vector machine(MRFO-SVM)for ground line localization.Theπ-shaped equivalent circuit calculation model of the distribution network is built when there is a missing ground wire,the relationship between the loop impedance and the frequency of the injected signal and the position of the missing ground wire is established,and the special frequency corresponding to the maximum impedance of the network is taken as the criterion of the location of the missing ground wire.Based on the location criterion,PSCAD/EMTDC is used to build a simulation model to verify the accuracy of the calculation model.The manta ray foraging algorithm(MRFO)is applied to the core parameter optimization of the support vector machine(SVM)to improve the accuracy of the model prediction.The relationship model between the special frequency point and the position of the missing ground wire is compared and analyzed with other three commonly used algorithms.The results show that the method proposed in this paper can realize the diagnosis and positioning of the missing ground wire,and the positioning accuracy reaches more than 98%.

关 键 词：漏拆接地线 蝠鲼觅食优化算法 支持向量机 定位 

分 类 号：TM92[电气工程—电力电子与电力传动]