套索回归在配电网谐波源定位的应用  

Application of LASSO Regression to Harmonic Source Location in Distribution Network

作　　者：程宏波 万紫彤 李宗伟 蔡木良 辛建波 CHENG Hongbo;WAN Zitong;LI Zongwei;CAI Muliang;XIN Jianbo(School of Electrical and Automation Engineering,East China Jiaotong University,Nanchang 330013,China;Electric Power Research Institute,State Grid Jiangxi Electric Power Co.,Ltd.,Nanchang 330096,China)

机构地区：[1]华东交通大学电气与自动化工程学院,南昌330013 [2]国网江西省电力有限公司电力科学研究院,南昌330096

出　　处：《电力系统及其自动化学报》2025年第3期59-65,共7页Proceedings of the CSU-EPSA

基　　金：江西省主要学科学术和技术带头人培养项目(20232BCJ22004);江西省重点研发计划项目(20223BBE51013)。

摘　　要：为实现配电网谐波源定位欠定方程组的准确求解,提出用套索回归实现配电网谐波源定位。套索回归通过引入惩罚项,对无谐波源的节点电流进行压缩,以降低方程组的欠定程度。以残差平方和最小为目标对节点的谐波电流进行估计,得到配电网谐波源定位稀疏方程的最优解,根据求解的谐波电流判断谐波源的位置,并利用IEEE33节点系统进行仿真验证。结果表明,本文方法可准确确定谐波源位置,与最小二乘法和岭回归及正交匹配相比,本文方法求解结果的误差更小、精度更高,当量测点数量越少时,本文方法的优势越明显。因此,本文方法估计结果准确,抗干扰能力强。To accurately solve the underdetermined equations of harmonic source location in a distribution network,least absolute shrinkage and selection operator(LASSO)regression is proposed to realize the harmonic source location in the distribution network.Through the introduction of a penalty term,the LASSO regression compresses the node current of the non-harmonic source to reduce the degree of under determination of equations.The harmonic current at the node is estimated with the minimum sum of residual squares as a goal,the optimal solution of the sparse equation of har-monic source location in the distribution network is obtained,and the harmonic source location is judged according to the solved harmonic current.The simulation results of an IEEE33-node system show that the proposed method can accurately determine the harmonic source location.Compared with the least squares method,ridge regression and orthogonal matching,it has a smaller error and a higher accuracy,and its advantages are more obvious when the number of equivalent measuring points is less.Therefore,the estimation results of this method are accurate and its anti-interference capability is strong.

关 键 词：谐波源 欠定方程 套索回归 谐波状态估计 惩罚系数 

分 类 号：TM72[电气工程—电力系统及自动化]