Application of the Ward´s method with TABU-search in the classification of load characteristics

Ward´s method is an effective method for classification of power system load characteristics. In view of the limitations of the traditional Ward´s method which can only reduce the number of clusters through amalgamating, an improved measure to optimize the cluster result based on TABU search is put forward along with a statistical method to guide the clustering process and test results. In this paper, the daily load curve is used as the feature vector, while the traditional Ward´s method is used to find a suitable number of clusters. The clustering results are optimized by means of the method of TABU search. A classification case of the load curves of 744 days in Guangzhou shows the advantage of those methods.