Forecasting electric vehicle charging with uncertainty across changing infrastructure
Probabilistic Forecasting of Electric Vehicle Charging in a Time-varying Hierarchical Infrastructure
Machine Learning
Summary
Predicting how many electric vehicles will charge at public stations is difficult because new stations are often added or removed over time. The authors studied a large dataset from Scotland and developed a method that can make accurate predictions while acknowledging uncertainty at different levels, from single chargers up to the entire country. Their method adapts as the charging network changes, providing reliable estimates that can help balance power supply and demand better.
What this means in practice
- •For power grid operators: Improve grid management by integrating probabilistic forecasts of EV charging demand that adjust for network expansion and changes.
- •For energy infrastructure planners: Support planning of new charging stations by understanding uncertain demand patterns across multiple regions and time periods.
Tested on one dataset.
Authors
Guillaume Principato, Yvenn Amara-Ouali, Yannig Goude, Bachir Hamrouche, Jean-Michel Poggi
Abstract
Uncertainty quantification for electric vehicle (EV) charging demand is essential to turn the increasing pressure of EV adoption on power systems into a source of flexibility. We study a large-scale dataset of public charging stations in Scotland, where demand exhibits a hierarchical structure, organized across multiple aggregation levels, from individual charging points to regional and national scales. The charging infrastructure is continuously evolving, with new stations deployed and existing ones occasionally removed, which calls for forecasting and uncertainty quantification methods tailored to this evolving structure. To address these challenges, we combine hierarchical forecasting with recent advances in conformal prediction with missing data. The conformal procedures considered produce valid prediction intervals at every node of the hierarchy and the algorithm CP-MNR-Nested successfully accounts for the time-varying hierarchical structure of the data.