基于数据驱动的低感知度配电网动态无功优化  

Dynamic Reactive Power Optimization of Low PerceptionDistribution Networks Based on Data-driven Approach

作　　者：徐晓春 卜强生 俞婧雯 赵娜 王涛 窦晓波[3] Xu Xiaochun;Bu Qiangsheng;Yu Jingwen;Zhao Na;Wang Tao;Dou Xiaobo(Huai’an Power Supply Branch.State Grid Jiangsu Electric Power Co.,Ltd.,Huai’an Jiangsu 223001,China;Electric Power Research Institute,State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing Jiangsu 210000,China;School of Electrical Engineering,Southeast University,Nanjing Jiangsu 210096,China)

机构地区：[1]国网江苏省电力有限公司淮安供电分公司,江苏淮安223001 [2]国网江苏省电力公司电力科学研究院,江苏南京210000 [3]东南大学电气工程学院,江苏南京210096

出　　处：《电气自动化》2024年第3期69-72,共4页Electrical Automation

基　　金：国网江苏省电力有限公司科技项目(J2021036)。

摘　　要：由于配电网网络通信基础设施较差,且节点监控覆盖不完全,因此存在无法实时采集数据的节点,导致无法进行传统无功优化。为此,提出了一种数据驱动的低感知度配电网动态无功优化方法。通过K-means算法聚类节点历史负荷,对非实时观测节点依据特征分类;选择最优超参数基于时间卷积网络进行量测数据补全;最终通过改进后的社交网络搜索算法实现动态无功优化,并仿真验证了方法的有效性。Due to poor communication infrastructure in the distribution network and incomplete node monitoring coverage,there are nodes that cannot collect data in real-time,resulting in the inability to perform traditional reactive power optimization.To this end,a data-driven low perception dynamic reactive power optimization method for distribution networks was proposed.Cluster node historical loads using K-means algorithm,and classify non real-time observation nodes based on features;the optimal hyper parameters were selected to complete the measurement data based on the time convolution network;finally,the improved social network search algorithm was used to achieve dynamic reactive power optimization,and the effectiveness of the method was verified through simulation.

关 键 词：时间卷积网络 社交网络搜索算法 K-MEANS算法 动态无功优化 数据驱动 

分 类 号：TM73[电气工程—电力系统及自动化]