Abstract

Electric Vehicles (EVs) are emerging among the Distributed Energy Resources (DERs) as a promising flexibility source. The small-scale of DERs still constitutes a barrier to their direct participation in electricity markets. Aggregators can exploit EV smart charging to lower the energy procurement cost in the day-ahead (DA) scheduling phase. This cost minimization should consider the uncertainty stemming from the availability of EVs, the DA and the imbalance market prices. To catch all these uncertainties, a novel approach combining robust optimization and stochastic programming is proposed to define the DA charging schedule for an EV fleet, considering the DA market prices and the possible imbalance realizations in real-time. This method is compared with a two-stage stochastic (TSS) programming approach and with the EV uncontrolled charging. Two weeks of data from the DA spot market in the Netherlands have been used for comparing the methods with different fleet sizes: 100, 200 and 400 EVs. The results show that the hybrid robust-stochastic method, while keeping approximately the same average DA energy cost, can estimate better than the TSS method the actual daily energy cost with reduced computational time.