Probabilistic charging management of electric vehicles with stochastic optimization and flexibility analysis under consideration of uncertainties

This paper examines a charging management strategy for electric vehicles, incorporating uncertainties to evaluate estimated flexibility potentials. Using real data collected from a German workplace, the analysis identifies typical patterns and uncertainties in arrival and parking times. A probabilistic approach is adopted to model parking time distributions and to develop a stochastic optimisation strategy to minimise the costs for charging demands. Results show that the stochastic optimisation reliably predicts charging behaviour, also achieving significant cost savings. Moreover, the flexibility potential is assessed, showing significant amounts of further flexibility usage. However, forecasting errors indicate an increased risk of overestimation at higher flexibility levels. The methodology demonstrates effectiveness for managing predictable yet uncertain charging behaviours and offers further potential for applications in grid stability, redispatch, and congestion management, while ensuring that risks associated with flexibility estimation errors are taken into account.