Clustering Methods for the Profiling of Electricity Consumers Owning Energy Storage System

Active consumers, as new players that can contribute for the market transactions, will provide flexibility through demand response events to deal with the volatile behavior from the distributed generation resources such as solar and wind. However, their lack of knowledge and uncertain behavior will add a new level of complexity to the community management. In this way, the authors propose a methodology capable of dealing with these resources and promoting fairness between all players. In present study, to add more features and enrich the several methodology steps from previous works, several clustering methods were compared to find patterns between the participants in the demand response events triggered. Since some clustering methods are sensitive to the initial number of clusters, an optimal must be found for the dataset. The authors also made a comparison between algorithms and found out that a low number of groups can be reductive since there are different consumer behaviors throughout the different contexts in each demand response event.