Load Consumption Characterization and Tariff Design Based on Data Mining Techniques

In the context of intelligent management of electricity consumption, the knowledge of the customers’ representative load consumption, and, consequently, of how and when this consumption occurs, constitutes a competitive advantage, in terms of positioning in the electricity market. This work presents the establishment of a methodology to characterize the electricity consumption of industrial consumers based on data mining techniques. A classification model was also implemented to classify new consumers in one of the obtained classes. Clustering validity indices were used to evaluate the clustering partition and also to support the decision of the best number of classes. Based on the results obtained from the load characterization approach, a methodology for defining electricity tariff structures is implemented taking into account the typical consumption profile and the evolution of electricity prices formed in the market. The results point to a clear distinction between clusters as well as tariff structures adapted to the identified set of industrial customers.