A New Classification Pattern Recognition Methodology for Power System Typical Load Profiles

In this paper a new pattern recognition methodology is described for the classification of the daily chronological load curves of power systems, in order to estimate their respective representative daily load profiles, which can be mainly used for load forecasting and feasibility studies of demand side management programs. It is based on pattern recognition methods, such as k-means, adaptive vector quantization, selforganized maps (SOM), fuzzy k-means and hierarchical clustering, which are theoretically described and properly adapted. The parameters of each clustering method are properly selected by an optimization process, which is separately applied for each one of six adequacy measures: the error function, the mean index adequacy, the clustering dispersion indicator, the similarity matrix indicator, the Davies-Bouldin indicator and the ratio of within cluster sum of squares to between cluster variation. This methodology is applied for the Greek power system, from which is proved that the separation between work days and non-work days for each season is not descriptive enough. Key-Words: Load profiles, clustering algorithms, adaptive vector quantization, fuzzy k-means, hierarchical clustering, k-means, self-organized maps, pattern recognition, adequacy measures