Impact of Uncertainties and Price of Robustness in receding-horizon EV Smart-Charging

Abstract

The large Electric Vehicle (EV) fleet penetrations can provoke several grid impact issues if no EV smart-charging is implemented. However, many EV smart-charging works assume an accurate prediction of input data, such as the EV driving patterns, which are highly uncertain. This paper addresses the impact and potential management of several uncertainties related to EV smart charging, such as photovoltaic (PV) generation, load demand, arrival state-of-charge (SOC), requested energy, and arrival and departure time of the EVs. The application of different levels of uncertainty budgets is proposed to account for the gradual impact of every uncertainty on smart charging performance. Moreover, potential uncertainty management is investigated with the use of robust optimization (RO) in predictive receding-horizon EV smart charging under the worst-case uncertainty level, and the ''price of robustness"is calculated. The results show that the EV driving uncertainties are more hazardous for the provided charging energy. In contrast, PV generation and load demand uncertainties have a significant impact mostly on the charging cost. Moreover, the price of robustness is very low for EV charging under every uncertainty case.