Adaptive Model Predictive Control for Large-scale Coordinated PEV Recharging

This paper presents a centralised model predictive control (MPC) approach for the control of the recharging process of a fleet of plug-in electric vehicles (PEVs). The first control goal is to recharge the PEVs so that the user preferences on the final desired state of charge, and on the time available for recharging are respected. A second goal is from the perspective of the grid operator. The charging processes should be coordinated in such a way that the aggregated power profile tracks a proper power reference established by the grid operator to ensure the safe operation of the grid. We propose an adaptive MPC formulation, in which the weights of the objective function are changed based on the feedback on the current progress of the charging sessions. This allows to better adapt the control effort to the current PEVs' status, and to reduce the computation burden by allowing to decrease the MPC prediction horizon. We validate the proposed approach via simulations on a large scale realistic scenario randomly generated.