基于改进K-means聚类算法的配电网区域规划方法研究

Research on distribution network regional planning method based on improved K-means clustering algorithms

针对传统配电网区域规划方法无法满足实际需求,而基于传统聚类算法的配电网区域规划方法存在着初始聚类中心选取不合理和聚类个数不确定的问题。文中对传统K-means聚类算法进行了改进,用于解决配电网区域规划中面临的聚类中心选取和聚类数确定问题。综合考虑密度与距离两个因素,提出了初始聚类中心的选取方法,引入聚类效果指数进行最佳聚类数的确定。性能测试分析的结果表明,改进后的K-means聚类算法能够准确计算配电网规划分区数量、确定聚类中心,聚类效果相比于传统方法有显著提升。

In view of the traditional distribution network regional planning method can not meet the actual needs,the traditional clustering algorithm based distribution network regional planning method has the problems of unreasonable selection of initial clustering center and uncertain number of clusters.In this paper,the traditional K-means clustering algorithm is improved to solve the problem of cluster center selection and cluster number determination in regional planning of distribution network.Considering the two factors of density and distance,the method of selecting initial clustering center is proposed,and the clustering effect index is introduced to determine the best clustering number. The performance test results show that the improved k-means clustering algorithm can accurately calculate the number of distribution network planning partitions and determine the clustering center, and the clustering effect is significantly improved compared with the traditional method.