基于分区和滑动时间窗策略的配电网线路参数辨识  

Line Parameter Identification in Distribution Networks Based on Partition and Sliding Time Window Strategies

作　　者：李晨涛 欧颖雅 季天瑶 张禄亮[1] LI Chentao;OU Yingya;JI Tianyao;ZHANG Luliang(School of Electric Power,South China University of Technology,Guangzhou 510006,China)

机构地区：[1]华南理工大学电力学院,广东广州510006

出　　处：《智慧电力》2024年第3期117-124,共8页Smart Power

基　　金：国家自然科学基金资助项目(52077081);广东省基础与应用基础研究基金项目(2022A1515011587)。

摘　　要：针对配电网线路参数辨识困难、不精确的问题,提出了一种基于分区和滑动时间窗策略的配电网线路参数辨识的方法。首先,使用线性解耦潮流模型最小二乘回归得到线路初始辨识参数。然后,将配电网分成若干个区域,并采用滑动时间窗策略,在每个时间窗中对每个区域内的线路参数通过高斯-牛顿法进行精确辨识。最后,对所有时间窗内的辨识结果进行离群值检测,综合各个时间窗的辨识结果得到最终的辨识值。所提方法提高了线路参数的辨识精度,并且可以避免精确辨识时选取到坏数据导致辨识结果偏离真实值。Targeting the problems of low accuracy and difficulty in line parameter identification in distribution networks,the paper proposes a method for identifying the line parameters of the distribution network based on the partition and sliding time window strategies.Firstly,the initial identifying parameters of the line are obtained by using the linear decoupling power flow model and least squares regression.Then the network is divided into several regions,and the line parameters in each region are accurately identified using the Gauss-Newton method within a sliding time window.Finally,outlier detection is conducted on the results obtained from all time windows,and the final identifying value is synthesized from the identifying results of each time window.This method improves the precision of the line parameter identification and can avoid the deviation of identifying results from the real value caused by the selection of bad data during the identifying process.

关 键 词：配电网 线路参数辨识 网络分区 滑动时间窗 数据驱动 

分 类 号：TM726[电气工程—电力系统及自动化]