1. Introduction
The destabilization of power systems due to the uncertainties in variable renewable energy (VRE) impedes the further deployment of VRE and the realization of a more sustainable society. To address this issue, it is necessary to enhance power system flexibility by integrating both supply- and demand-side resources [
1]. A technology that allows bidirectional power flow from electric vehicles (EVs) to the grid by controlling the charging and discharging of EVs offers demand-side flexibility. This technology is known as vehicle-to-grid (V2G) technology [
2]. Aggregated EV fleets are expected to provide energy and ancillary services for the system-wide (transmission level) or local network (low-voltage and medium-voltage distribution grids) level by taking advantage of their high responsiveness [
3]. However, stakeholders need to better understand and appreciate the benefits of V2G when promoting and supporting its adoption.
Economic analysis has shown that revenue from V2G ancillary services to system-wide level (e.g., frequency regulation and spinning reserves) can offset the high capital cost of EVs, and previous findings have shown that this is a major motivation for introducing V2G [
4]. Furthermore, a previous study was conducted to analyze the economics of V2G ancillary services to spinning reserves with an EV fleet that could connect to the Electricity Reliability Council of Texas electric power system for 2 h, using a unit commitment model [
5]. Researchers have also proposed the dual use of V2G for frequency regulation and peak-load reduction, finding that the revenue from frequency regulation was much larger, but that the two functionalities could be provided together, yielding higher combined revenue [
6]. The simulation results estimate annual benefits ranging between EUR 100 and EUR 1100 per vehicle, which demonstrates that EV owners can obtain significant benefits by providing frequency-controlled normal operation reserves in the Nord Pool market [
7]. A recent review [
8] explains that ancillary services such as spinning reserves or frequency regulation are argued to be the best cases for V2G and frequency regulation is especially considered the most valuable service that V2G can participate in. The generally accepted practice of participating in the ancillary services market is the assumption that aggregators combine EV fleet clusters to ensure a minimum bid volume [
9]. A previous study employed stochastic travel models to consider the availability of EV fleets and determined that the V2G aggregative communication architecture can improve the reliability of ancillary services [
10]. Because the efficient aggregation of a highly available fleet is important for increasing revenues, statistical estimates have been generated regarding the availability of ancillary services for private EVs for daily home–work commuting [
11]. The results showed that bidding at work was more reliable than bidding at home because of the lower uncertainty. To demonstrate commercial viability of V2G ancillary services, many V2G demonstration projects have focused on frequency regulation [
12]. V2G ancillary service demonstrations have participated in the actual regulation market in the Pennsylvania–New Jersey–Maryland Interconnection in the eastern part of the United States of America [
13] and in Denmark [
14].
Motivating V2G penetration is currently limited due to concerns that the market is a niche one and that its value will decrease due to saturation [
2,
4,
5,
6,
8,
14]. However, these studies used prices and marginal costs based on the market at the time and did not account for future changes in the power system structure. In the future, the value of flexibility and the size of the power market may differ due to the historical structure, VRE penetration, network evolution, and the availability of technology options. For example, the flexibility demand of a power system was found to increase with VRE penetration in Europe and Germany [
15]. Furthermore, investigators have reviewed several technology options and market designs that provide flexibility for power system management, and V2G represents only a single demand-side option [
16,
17], although it is one of the most cost-effective options [
18]. Therefore, to assess the value of V2G in future power systems, a supply and demand analysis model that explicitly analyzes power system operations specific to each country is needed. The availability of the EV fleet should also be considered.
A Spanish case study [
19] was conducted to investigate various scenarios of combined VRE and EV penetration levels in the Spanish electricity system using a medium-term operational model that calculated the power generation dispatch and optimized the charging/discharging of EVs, based on five different driving patterns and hourly power system operations. The authors of that study highlighted the benefits of combining high levels of VRE and EV penetration to reduce power system operating costs. Another study [
20] was performed to investigate the value of V2G focused on spinning reserves in future power scenarios in the UK, by modeling power system operations that considered primary, secondary, and tertiary reserve services using mixed-integer linear programming (MILP). The authors found that when wind power became prevalent, the value of secondary reserve provision by V2G was high with inflexible systems, but low with flexible systems where pumped storage hydroelectricity plants (PSHPs) were utilized. Furthermore, a case study of the German power system [
21], including interconnection lines to neighboring countries, employed a unit commitment model to analyze the impact of V2G flexibility on the decarbonized power system in 2030.
Although the value of future V2G has been assessed in accordance with national power system scenarios, there is currently a lack of discussion regarding the trends of saturated value relative to the number of EVs participating in the market. Motivating market participation with an aggregated EV fleet requires analysis of the marginal value (revenue in relation to the number of incrementally added EVs), which is the value derived from the niche market. Strbac et al. [
22] focused on distributed-energy storage systems (ESSs) and analyzed the marginal values of ESSs in terms of ancillary services and market saturation, using an assumed power system model for the UK in 2030. However, evaluation of the marginal value of V2G was not considered. Because EVs are used for mobility purposes, the time period during which they can be connected to the power system to provide flexibility is more limited than that of ESSs. Zhou et al. [
23] analyzed the value of flexibility of demand-side resources (e.g., EV charging, heat pumps, etc.) and suggested that the marginal value tends to saturate as more resources are added. If the value of flexibility depends on the time of day during which VRE penetration increases, then it is important for stakeholders who utilize V2G to consider when to aggregate vehicles and provide flexibility through V2G.
In this study, we evaluated the marginal value of V2G ancillary service on a load-frequency control (LFC) timescale. The LFC timescale is an ancillary service that corresponds to a secondary frequency-control reserve [
24]. We analyzed the electricity supply and demand using a production-cost model [
25] that endogenously optimizes the operation of the power system, including generators, hydropumped storage, interconnection lines, and the balancing capacity provided by V2G power control on the LFC timescale (V2G LFC). The operational cost savings of the power system attributable to the V2G LFC were determined considering VRE penetration and EV fleet availability in each interconnected area. As a case study, the Japanese power system in 2030 (with increased VRE penetration and grid-side flexibility) was assumed. The Japanese power supply system is divided into 10 areas joined by grid-interconnector lines in the longitudinal direction and is classified as a “longitudinal transmission system”. The use of interconnected lines has been uncommon to date; however, this is expected to change with the indirect auctioning of power transactions, which started in 2020. Furthermore, the Japanese power system is equipped with several PSHPs [
26]; hence, our study is of interest to other power systems aimed at introducing interconnections and energy storage systems (e.g., batteries and PSHPs) as flexibility measures. We assume that the aggregated EV fleet for V2G consists of private EVs on daily work–home commutes. An analysis of Japanese vehicle driving patterns [
25] showed that private commuter cars had high switching potential from conventional vehicles to EVs, due to small driving distances, and the fleet behavior was predictable because most activities were between the home and office before and after working hours.
The contributions of this paper are threefold: (a) we evaluated the dependence of the marginal value of LFC timescale flexibility on the hours of the day under various VRE penetration scenarios; (b) the reasonable maximum capacity of the V2G LFC to be provided for the power system was determined by considering the limited availability time based on the duty cycle of commuter EVs; (c) studying changes in the operation of the power system due to V2G LFC revealed that improving the value of the V2G LFC required coordination with storage and excess VRE generation.
This paper is structured as follows:
Section 2 describes the production cost model including VRE operations, PSHPs, interconnection power flows, LFC capacity provided by V2G, and a definition of the marginal value of V2G;
Section 3 describes scenarios for the Japanese power system in 2030;
Section 4 presents the marginal value in power system operations as a result of the V2G LFC;
Section 5 discusses our results in relation to previous studies; and our conclusions are presented in
Section 6.
6. Conclusions
In this study, we examined the marginal value of the balancing capacity of V2G on the LFC timescale by considering the market size and saturation trends of ancillary services for future structural changes in the Japanese power system. Production cost simulations were performed under various VRE penetration scenarios with grid-side flexibility, including PSHPs, interconnected lines, and LFC, via the partial-load operation of coal-fired power plants. The aggregated EV fleet used for providing the V2G LFC capacity was assumed to be the private commuter duty cycle, which demonstrated predictable behaviors and was able to connect at both the workplace and home. The marginal value of the V2G LFC capacity was obtained by differentiating the annual power generation cost savings. Based on the marginal value of V2G LFC capacity, the reasonable maximum size of the EV fleet that could be deployed in the ancillary service market was evaluated. Our results quantitatively support the findings of several literatures. The results are as follows.
The marginal value of the V2G LFC capacity increased with higher VRE penetration. Comparing the low- and high-VRE scenarios showed that the maximum marginal value increased by approximately 3.2-fold during the daytime (from USD 125 to USD 400/kW/year) and by approximately 2.1-fold overnight (from USD 85 to USD 175/kW/year). The reasonable EV fleet size for the power system increased from approximately 292,000 to 1,483,000 vehicles during the day and from 0 to approximately 542,000 vehicles overnight. The maximum cost saving (USD 705.6/EV/year) occurred during the daytime under the high-VRE scenario. Note that this value might be underestimated because it does not reflect the investment and maintenance costs of generators.
Daytime V2G LFC not only increased daytime VRE generation, but it also reduced thermal power generation by causing changes in the PSHP operations before and after V2G. Under higher VRE penetration scenarios, the VRE that avoided curtailment during the day due to V2G was temporarily stored in the PSHPs, and the discharging of PSHPs in the evening further reduced LNG and coal generation with higher fuel costs. Overnight V2G LFC resulted in cost savings by substituting coal for LNG. Under higher VRE scenarios, VRE that avoided curtailment overnight was stored in PSHPs, and PSHP discharge reduced the evening thermal-generation peak. Improving the value of V2G LFC required coordination with storage and excess VRE generation.
The transactions of LFC capacity through the interconnected lines decreased because the daytime V2G LFC increased the power system flexibility in each area. Conversely, energy transactions increased from areas with abundant VRE and low demand to areas with high demand. Overnight V2G reduced both the LFC capacity and energy transmission, while mitigating transmission losses.
Our model has limitations that merit future research. First, the impact of the EV battery C-rate and SOC on the available LFC capacity was not considered. In addition, detailed modeling is required based on the distinction between the upward and downward ancillary services. Second, we did not factor in the decision-making of individual EV users. An analytical model that reflects the benefits of demand-side energy management is desirable. Finally, there was a lack of assumptions regarding integration with other newly distributed resources. Future work will incorporate market-saturation trends due to the deployment of stationary battery-storage systems, which serve as a decentralized power source capable of a fast response. It includes some sensitivity analysis and the more specific setting of LFC capacity required for upward and downward reserves. In evaluating multi-use in combination with other timescale electricity markets, it is important to consider demand-response technologies such as heat pumps.