Bayesian Network based Real-time Charging Scheduling of Electric Vehicles

Abstract

The electricity purchase cost of an electric vehicle (EV) charging station (EVCS) with photovoltaic (PV) facilities can be reduced by matching the EV charging load and PV generation output. However, the uncertainties of the EV charging demand and photovoltaic generation output impose difficulties to real-time charging scheduling. In this paper, an EV charging scheduling model based on the spot pricing of electricity with an objective of minimizing the electricity purchase cost of an EVCS is proposed. The historical data of PV generation output and EV charging demands are employed in forming the daily charging optimization model, and the daily optimal charging scheduling is carried out and a training sample set is then attained. The Bayesian network (BN) based real-time EV charging scheduling, that is carried out in a recursive way with one time period ahead considered, is next addressed, and the BN structure is determined by the hill climbing algorithm with Bayes scoring employed. The EV charging schedule is subsequently determined by the Bayesian inference. A case study based on an EVCS in an industrial park is carried out to demonstrate the proposed method, and comparisons between the proposed method and the deterministic real-time scheduling method are also detailed.