Stochastic Model Predictive Control of Charging Energy Hubs with Probabilistic Forecasting

We introduce an energy management system for an energy hub in an electric vehicle charging station with photovoltaic generation and a battery energy storage system. First, we design a scenario-based stochastic model predictive control that uses probabilistic day-ahead forecasts of charging load and solar generation to achieve cost-optimal operation of the energy hub. Second, the probabilistic forecasts leverage conformal prediction providing calibrated distribution-free confidence intervals starting from machine learning models that generate no uncertainty quantification. In preliminary results, we run a 10-day evaluation in a closed-loop simulated environment to compare the observed cost of the scenario-based stochastic control (100.21%) with two deterministic alternatives: a version with point forecast (100.32%) and a version with perfect forecast (100%).