基于最优变换关联因子和优化平移分裂法的窃电检测  

Electricity Theft Detection Based on Optimal Transform Correlation Factor and Optimized Translation Splitting Method

作　　者：杨挺[1] 徐嘉成 叶芷杉 王军 卢利军 YANG Ting;XU Jiacheng;YE Zhishan;WANG Jun;LU Lijun(Key Laboratory of Smart Grid(Tianjin University),Ministry of Education,Nankai District,Tianjin 300072,China;Henan Xuji Instrument Co.,Xuchang 461000,Henan Province,China)

机构地区：[1]智能电网教育部重点实验室(天津大学),天津市南开区300072 [2]河南许继仪表有限公司,河南省许昌市461000

出　　处：《电网技术》2023年第9期3913-3923,共11页Power System Technology

基　　金：国家重点研发计划项目(2022YFB2403800);国家自然科学基金项目(61971305)。

摘　　要：传统的窃电检测算法应用于用户众多的复杂配用电台区中时,随着计算模型参数增多,存在由于个别参数估计失准导致的最终辨识结果不准确的问题,即表现为抗扰动性差。针对该问题,该文研究并提出基于最优变换关联因子和优化平移分裂法的窃电检测算法。首先通过数值分析和窃电机理验证,发现台区线损率与窃电用户用电占比间存在的正相关关系中含有非线性度。进而根据最优变换关联因子法实现窃电用户初筛,再根据广义电量守恒建立低维模型,采用优化平移分裂法求解,以克服模型的病态性,实现窃电用户二次准确辨识。算法通过两步筛选计算,有效解决了大台区高维参数的误差扰动致使窃电用户辨识准确度下降的难题,具有更强的抗扰动性。采用实际多种复杂场景下不同规模台区的用电信息采集系统的日用电量数据进行窃电检测,并与现场稽查结果比对,验证了所提方法的有效性和优越性。The traditional algorithms for electricity theft detection,when applied in a complex environment with many users,may have inaccurate final identification results,showing poor disturbance resistance.This results from the increase of the model parameters and the inaccurate estimation of some individual parameters.This paper proposes an algorithm for electricity theft detection based on the optimal transformation association factor and the optimized translation splitting method.Firstly,through the mathematical analysis and the verification of the electricity stealing mechanism,it is found that there is a nonlinear correlation between the line loss rate of the station area and the proportion of the electricity users who steal electricity.Consequently,this paper can realize initial screening of users stealing electricity through optimal transform correlation factor method.Then,in second identification of users stealing electricity,a low-dimensional model is established according to the generalized conservation of energy.The optimized translation splitting method is used to solve the ill-conditioned model,realizing the accurate identification of the electricity stealing users.Through the two-step screening and calculation,the algorithm solves the problem that the error disturbance of the high-dimensional parameters leads to the decrease of the identification accuracy of the power theft users,which shows a stronger resistance to disturbance.To analyze and verify the effectiveness and superiority of the proposed method,electricity theft detection is performed in distribution areas with different scale and complex scenes.The power data is collected by the smart power information collection equipment and the analysis result is verified on-site.

关 键 词：窃电检测 用电量估计 关联关系 优化平移分裂法 

分 类 号：TM721[电气工程—电力系统及自动化]