Fuel Cell Electric Vehicle-to-grid feasibility

Abstract

Fuel Cell Electric Vehicles (FCEVs) in combination with green hydrogen (obtained from renewable sources), could make a significant contribution in decarbonizing the European transport sector, and thus help achieve the ambitious climate goals. However, most vehicles are parked for about 95% of their life time. This work proposes the more efficient use of these vehicles by providing vehicle-to-grid (V2G) services achieving the integration of the transport and energy systems. The aim of this work is to determine the technical and financial potential value that FCEVs could have by providing frequency reserves. Experiments were carried out with a Hyundai ix35 FCEV that was adapted with a power output socket so it can operate in V2G when parked, delivering maximum 10 kW direct current power. Results show that both power sources in the fuel cell electric vehicle, which are the fuel cell stack and the battery, can react in the order of milliseconds and thus are suitable to offer fast frequency reserves. The challenge lays in the communication between the car and the party that sends the signal for the activation of the frequency reserves. As one unit does not provide enough power to be able to participate in the electricity market, a car park acting as aggregator of FCEVs was designed taking into account current technology developments. A carpark with a direct current microgrid, a hydrogen local network and only occupied by FCEVs was designed. A financial model was developed to evaluate the economic potential of the car park to participate in the electricity market providing frequency reserves. Results show that by using the fuel cells in the FCEVs in V2G, monetary benefits could be obtained when providing automated frequency restoration reserves (aFRR) upwards. Key parameters are found to be the investment costs, amount of vehicles available, hydrogen price and price of aFRR. With a car park of approximately 400 cars all year long available, payback times of 11.8 and 3.5 years were obtained taking into account worst and best case scenarios for a 15 year period analysis, respectively.