Online Clustering and Aggregation Method for Industrial Load Considering Differentiated User Preference

High energy consuming industrial loads have high potential for regulation, but have not been fully utilized. The clustering and management of high energy consuming industrial loads is the key to unleashing their potential. Therefore, this paper proposes a fuzzy C-means method that considers user preferences to address high energy consumption. To overcome the curse of dimensionality, comprehensive features are extracted from the basic features to achieve dimensionality reduction, and the analytic hierarchy process (AHP) method is used to weight the features. Using comprehensive features as clustering inputs, an improved weighted fuzzy C-means is used to conduct preliminary clustering and aggregation of industrial loads in a certain area. This refined approach ensures a more accurate and efficient categorization of the loads. Subsequently, based on data updates and existing results, the Canopy algorithm can be used for online real-time clustering. Taking high energy consuming industrial loads in a certain province as a typical case, while considering participation preferences, grouping high energy consuming loads can be achieved.