A data-driven probabilistic framework for estimating grid impacts of EV charging at scale

The rapid uptake of electric vehicles presents both opportunities and challenges for modern power systems. As charging demand increases, intelligent and data-driven approaches are essential to predict and manage future grid impacts. This paper presents a probabilistic and data-driven approach for generating realistic electric vehicle charging profiles, which can be integrated into probabilistic load flow simulations to assess their effect on grid capacity. In contrast to many existing studies that focus on low-voltage grids, this work targets the sub-transmission grid, analyzing the impact of aggregated charging demand. The approach is applied to a Swedish sub-transmission grid using a combination of real-world data: travel statistics, demographic data, charger usage rates, and two empirical home charging datasets, one representing uncoordinated charging, and another incorporating smart charging technologies. The methodology generates synthetic but realistic charging demand profiles, which are coupled to a full-scale grid model to evaluate the system under 100 % fleet electrification. Results show that substantial congestion may emerge in the primary substation transformers, which may, to some extent, be mitigated by adopting planned or intelligent charging strategies. By combining real-world data with a detailed grid model and a probabilistic modeling approach, the study lays a foundation for analyzing grid impacts under high EV penetration and supports the development of flexibility services to inform more proactive grid planning. • Probabilistic framework for large-scale grid model with EV integration. • Data-driven EV charging demand profiles for fully electrified car fleet. • Quantified risk of grid congestion due to EV charging integration. • Grid congestion at sub-transmission grid level due to aggregated EV charging.