Load curves partitioning with the application of soft clustering algorithms

Load profiling refers to a procedure which leads to the formulation of daily load curve and consumer categories regarding the similarity of their curves shapes. This procedure incorporates a set of pattern recognition algorithms. While many crisp clustering algorithms have been purposed for grouping load curves into classes, only one soft clustering algorithm is utilized for the aforementioned purpose, namely the Fuzzy C-Means (FCM). Since the benefits of the soft clustering is demonstrated in a variety of applications, we examine the potential of introducing soft clustering algorithms in the electricity demand patterns segmentation. This paper introduces in the load profiling studies, two soft clustering algorithms which have been already used in other clustering problems, namely the Possibilistic C-Means (PCM) and the Gustafson-Kessel Fuzzy C-Means (GKFCM). A detailed comparison takes places between the algorithms and their performance is checked by a set of adequacy measures that have been proposed in the load profiling related literature and by a set of traditional fuzzy clustering validity measures.