MILP-based Model Predictive Control of Electric Vehicle Fleet Charging in the presence of Electricity Production from Renewable Energy Sources

Electric vehicles (EV) play a crucial role in transforming the transportation sector to become more sustainable, cleaner, and energy efficient. Given that EVs are generally parked and ready for charging for a great majority of time, advanced charging management techniques can take advantage of this opportunity to lower charging costs and more effectively harness the potential of renewable energy sources (RES). To address this objective, the paper proposes an optimal charging management strategy based on model predictive control (MPC), which relies on a mixed-integer linear programming (MILP) algorithm for online optimization of charging power of individual EVs within a fleet. The proposed strategy is formulated in two forms: (i) direct optimization of individual EV charging power and (ii) optimization of aggregate charging power distributed to individual EVs through a heuristic algorithm (hierarchical approach). The two strategies are evaluated against an offline optimization benchmark in terms of charging cost, RES energy utilization, and related optimization time execution for the case of one-week period, virtually electrified delivery vehicle fleet of a local retail company, and a two-tariff electricity price model. A dumb charging strategy, which simply performs charging under maximal power immediately upon EVs connection, is also used as an evaluation baseline.