Stochastic Management of Bidirectional Electric Vehicles: The Case of an Electric Vehicles Aggregator

Abstract

This paper proposes a development of a support management system (SMS) for an aggregator managing electric vehicles (EVs), in order to present optimal and profitable energy bids in a day-ahead market (DAM). The EVs allow bidirectional energy flow, i.e., grid-to-vehicle (G2V) and vehicle-to-grid (V2G) charging. The development of the support management system takes into account an augmented level of uncertainty regarding DAM prices, driving requirements and availability periods of EVs from driving patterns of European drivers. The uncertainty is modeled by means of a set of convenient scenarios simulated by a Kernel density estimation (KDE) subsequently reduced by a K-means algorithm. The development of the SMS is based in a stochastic optimization problem (SOP) rewritten as a mixed-integer linear programming (MILP) approach. An advanced interaction between the owners of EVs and the EVs aggregator (EVA) is assumed, being the EVA in charge of decisions of charging and discharging only during the periods of idling of EVs. The results show that the aggregation of EVs is profitable for the aggregator achieving a positive value, covering battery degradation (BD) cost and without any changes in the usual driving patterns of the EV owners. Also, the results show that better-performing batteries can be crucial in the increment of expected profit of the EVA.