Determination of Typical Electricity Load Profile by Using Double Clustering of Fuzzy C-Means and Hierarchical Method

Abstract

The Fuzzy C-means (FCM) and Hierarchical clustering method are widely used by many researchers in clustering data sets of electricity consumption to determine the typical consumers’ electricity load profile. FCM method clustered the data sets into several clusters by assigning the membership degree for each data while Hierarchical clustering method clusters the data sets by finding the distance between each data to find the similarity and dissimilarity of each data. This study presents the determination of typical electricity load profile by using double clustering of FCM and Hierarchical methods. Hierarchical clustering was performed as the second method after the data sets had been clustered by FCM. Cluster validation is completed by using Davies-Bouldin Index, Calinski-Harabasz Index and Silhouette Index to determine the compactness of the resulting clusters and to find the optimal number of clusters for a data collection. As a result, the number of clusters 3 is chosen as the optimal number of the cluster by using double clustering method.