A decentralized power dispatch strategy in an electric vehicle charging station

Abstract

In this study, a decentralized power dispatch in a charging station serving electric vehicles (EVs) is discussed. The power dispatch problem is solved through a Stackelberg game in real time. In this game, the leader is the EV charging station (EVCS) while the followers are EVs. The preferences of the EVCS are designed as being self-sufficient, providing charging services to the EVs, and maintaining the energy level of the battery energy storage system (BESS), which are described through different utility functions. In addition, the preferences of followers are to maximize their EV charging powers. The learning algorithm utilizes the consensus network to reach the generalized Stackelberg equilibrium as the power dispatch among EVs in an iterative decentralized manner. Both the static and dynamic case studies in the simulation verify the successful implementation of the proposed strategy, the flexibility to uncertainties and the re-configurability to the number of EVs. It also has an excellent performance compared with the centralized benchmark strategy with criteria, that is, the average EV charging time, the number of charge and discharge rate of the BESS and energy exchange to the grid. Finally, a down-scaled experiment implementation is set up to validate the functionality and the effectiveness of the game theory-based strategy.