Analyzing the medium-term impact of the aggregated peak load from electric vehicles using a clustering-based approach

The rapid increase of electric vehicles (EVs) has a large impact on distribution networks in residential areas. To prevent congestion issues, an efficient reinforcement strategy is crucial. Until now, medium-term studies focus on model-based approaches to calculate charging loads of EVs starting bottom-up at the individual level of the users, EV properties, and charging power. However, analysis of a large number of EVs limits the applicability of these studies for distribution system operators in practice. Therefore, this study proposes an alternative top-down approach to calculate the weekly aggregated peak load for a large number of EVs based on individual charging powers and battery capacities. This flexibly adaptable calculation can be used to study the annual duration and rate of congestion of distribution network assets due to additional EV charging loads in residential areas. First, charging loads of 235 EVs were clustered based on their normalized load duration curves. Second, these clusters were used to propose a methodology to calculate the weekly aggregated peak load for any group of EVs based on three representation functions, the battery capacity, and the charging power. Finally, the weekly aggregated peak load and the related impact on a medium to low voltage transformer were studied for three use cases.