The Studies On Classified Methods Of Load Data Based On Clustering Techniques

In this thesis, clustering analysis is applied to load modeling in power system, and is used to analyze the statistics data on load components(including five components: industry component, agriculture component, business component, railroad component and others component) in The Northeast of China. In this thesis, the results and clustering course of four clustering methods are analyzed and compared, finally the Between-groups Linkage method in Hierarchical Cluster Analysis is considered as a valid method. In analyzing, the statistics data of 231 substations' load components are divided into four areas(HeiLongJiang province ,JiLin province,LiaoNing province and the North of NeiMengGu municipality )and two seasons(Summer and Winter),thus eight group stat data are gotten. In order to get an optimal cluster number, statistics is applied to this research . The changing of four statistics variables during clustering are compared and on these results basis the optimal cluster number is decided. After elementary analyzing, the method and the result are shown valid.