Demand side energy management of EV charging stations by approximate dynamic programming

• Published in: Energy conversion and management Vol. 196; pp. 878 - 890
• Main Authors: Wu, Yu, Ravey, Alexandre, Chrenko, Daniela, Miraoui, Abdellatif
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 15.09.2019; Elsevier Science Ltd; Elsevier
• Subjects: Approximate dynamic programming; Automatic; Autonomy; Batteries; Demand side energy management; Dynamic programming; Electric power; Electric vehicle charging; Electric vehicle charging stations; Electric vehicles; Electricity consumption; Electricity pricing; Energy management; Engineering Sciences; EV charging station; Evolutionary algorithms; Flexible charging pattern; Fluid mechanics; Fuzzy logic; Fuzzy systems; Mechanics; Multiple types of EV charger; Physics; Thermics; Approximate dynamic programming; Flexible charging pattern; EV charging station; Demand side energy management; Multiple types of EV charger
• ISSN: 0196-8904; 1879-2227
• DOI: 10.1016/j.enconman.2019.06.058

•Demand-side energy management is studied for EVCS with multiple types of chargers.•An approximate dynamic programming based energy management system is proposed.•Optimal charging start time of each EV is determined by the proposed EMS.•Maximum charging autonomy can be guaranteed at the device level.•Operation costs can be reduced by over 50% in the numerical studies. The high operation cost of the EV charging station (EVCS) is a severe challenge for the development of electric vehicles, which lead to the general shortage of the EVCS. In order to reduce the operation costs of the EVCS, an approximate dynamic programming (ADP) based energy management system (EMS) is proposed for the EVCS equipped with multiple types of chargers (EVCS-MTC). A fuzzy logic guiding system has been designed to allocate each vehicle an appropriate charging spot based on its charging urgency. Multiple EVs can acquire the charging service through a common charger in the EVCS-MTC. In the proposed EMS, the approximate dynamic programming (ADP) and the evolution algorithm (EA) are combined to determine the optimal charging start time for each EV. This characteristic provides the charging device with the maximum autonomy to select the preferred flexible charging pattern, which can prolong the battery lifetime and reduce the communication requirements of the control system. With taking the dynamic electricity price and uncertain future charging demand into account, the proposed EMS can achieve a total cost reduction of over 50% compared with the conventional charging scheme in the numerical studies.