利用凝聚层次聚类辨识电网中脆弱输电断面

Identifying vulnerable transmission sections in power grids using agglomerative hierarchical clustering

针对传统的输电断面识别主要依靠电网运行调度人员经验划分电网、寻找断面存在不准确问题,采用改进的凝聚层次聚类算法对电网薄弱断面进行辨识;建立基于改进凝聚层次聚类算法的薄弱辨识模型,首先以潮流介数作为线路权重来构建节点相似矩阵,其次通过聚类算法确定电网最优分区数,完成电网分区;根据电网分区结果结合图论,搜索初始输电断面,对所有初始断面搜索结果进行电气连接检验,最终确定脆弱输电断面;通过聚类算法辨识脆弱断面,并应用机器学习方法提高识别脆弱断面的效率。结果表明,基于聚类的脆弱断面辨识方法在识别电网系统中的脆弱输电断面具有可行性,可以有效提升辨识效率。

For the traditional transmission section identification mainly relies on the grid operation dispatchers to divide the grid based on their experience and find the problems existing in the section,an improved agglomerative hierarchical clustering algorithm was used to recognize the vulnerable sections of the power grid.To establish the vulnerability identification model based on the improved agglomerative hierarchical clustering algorithm,the node similarity matrix was firstly constructed by using the power flow betweenness as the line weights,and the optimal number of grid partitions was determined by the clustering algorithm to complete the grid partitioning.Based on the grid partitioning results combined with the graph theory,the initial transmission section search was carried out,and the electrical connection inspection was performed on all the initial section search results to finalize the vulnerable transmission sections.The identification of vulnerable sections by clustering algorithm,and applying machine learning methods to improve the efficiency of identifying vulnerable sections.Simulation results show the feasibility of the clustering-based weak section identification method in identifying vulnerable transmission sections in the grid system.