Hybrid Centralized-Decentralized (HCD) Charging Control of Electric Vehicles

Abstract

Integrating massive electric vehicles (EVs) into the power grid requires the charging of EVs to be coordinated to reduce the cost and guarantee system stability. The coordination becomes more challenging when the EV owners have different charging preferences. To tackle this problem, a hybrid centralized–decentralized charging control scheme is designed in this paper which mainly includes three parts, as follows. On the centralized charging side, an offline optimal scheduling approach is first presented, aiming at minimizing the energy cost while satisfying the charging requirements of EVs. To deal with the system dynamics and uncertainties, a model predictive control-based adaptive scheduling strategy is then developed to determine the near optimal EV charging profiles in real time. On the decentralized charging side, the interactions between EVs and the charging system controller is modeled as a leader–follower noncooperative Stackelberg game in which the system controller acts as the leader and the EVs act as followers. The existence of the equilibrium state and its optimality are proved and analyzed. It is shown that by adopting the proposed decentralized charging algorithm, the communication burden between EVs and the system controller is low and that the charging scheme is robust to poor communication channels. Last, we investigate how the size of these two charging groups impacts the system utility and propose an algorithm maximizing the total revenues of the whole system. Simulation results evaluate the performances of the scheme and investigate the parameters’ impacts on the system utilities. The proposed approach and obtained results may provide guidelines for improving the efficiency of the charging park operation and provide useful insights, helping the system operator develop rational investment strategies.