Consumer-branch Connectivity Identification of Low Voltage Distribution Networks Based on Data-driven Approach  

作　　者：Yongjun Zhang Yingqi Yi Wenyang Deng Siliang Liu Lai Zhou Kaidong Lin Yongzhi Cai 

机构地区：[1]the School of Electric Power Engineering and Guangdong Key Laboratory of Clean Energy Technology,South China Uni-versity of Technology,Guangzhou 510641,China [2]the School of Elec-tric Power Engineering,South China University of Technology,Guangzhou 510641,China [3]the Guangzhou Panyu vocational and technical college,Guangzhou 511483,China [4]the Metrology Center Guangdong Power Company Grid,Qingyuan 511500,China

出　　处：《Protection and Control of Modern Power Systems》2024年第4期69-82,共14页现代电力系统保护与控制（英文）

基　　金：supported by the National Natural Sci-ence Foundation of China(No.52177085);Science and Technology Planning Project of Guangzhou(No.202102021208).

摘　　要：Accurate topological information is crucial in supporting the coordinated operational requirements of source-load-storage in low-voltage distribution networks.Comprehensive coverage of smart meters provides a database for low-voltage topology identification(LVTI).However,because of electricity theft,power line commu-nication crosstalk,and interruption of communication,the measurement data may be distorted.This can seriously affect the performance of LVTI methods.Thus,this paper defines hidden errors and proposes an LVTI method based on layer-by-layer stepwise regression.In the first step,a multi-linear regression model is developed for consumer-branch connectivity identification based on the energy conservation principle.In the second step,a significance factor based on the t-test is proposed to modify the identification results by considering the hidden errors.In the third step,the regression model and significance threshold parameters are iteratively updated layer by layer to improve the recall rate of the final identification results.Finally,simulations of a test system with 63 users are carried out,and the practical application results show that the proposed method can guarantee over 90%precision under the influence of hidden errors.

关 键 词：Data driven hidden error linear re-gression low voltage distribution network topology identification 

分 类 号：TM73[电气工程—电力系统及自动化]