Patterns discovery of load curves characteristics using clustering based data mining

The paper presents a comprehensive approach using clustering based data mining for load curves characterization in real distribution networks. The load curves characterized by their main indicators is made using information provided by Smart Meters. The proposed method was tested using a real database with 60 rural substations. The results demonstrate the ability of the methodology to be efficiently used by distribution operators in patterns discovery of the load curves characteristics with the help of which an energy losses estimation in transformers from the distribution substations could be made.