Improving WFA k-means technique for demand response programs applications

There are several pattern-based clustering methods which are used for different applications such as pattern recognition, data mining, etc. In recent years, some of these methods are implemented in power system studies, especially for clustering load curves for designing suitable tariffs, demand response programs selection, etc. Choice of the best clustering method for certain application is one of the most important issues which is case dependent and should be considered in using of clustering load curves. Demand response programs are widely used in power system for different applications such as peak clipping, demand reduction, etc. since demand response programs are featured with different characteristics. Therefore, selection of suitable programs for different customer classes is of great importance. In this paper, an improved weighted fuzzy average (WFA) K-means for the purpose of demand response programs applications is developed. This method is implemented on 316 load curves of Tehran distribution network and the results are investigated.