Techno-Economic Feasibility of Fuel Cell Vehicle-to-Grid Fast Frequency Control in Non-Interconnected Islands

Abstract

This paper presents an innovative approach to fast frequency control in electric grids by leveraging parked fuel cell electric vehicles (FCEVs), especially heavy-duty vehicles such as trucks. Equipped with hydrogen storage tanks and fuel cells, these vehicles can be repurposed as dynamic grid-support assets while parked in designated areas. Using an external cable and inverter system, FCEVs inject power into the grid by converting DC from fuel cells into AC, to be compatible with grid requirements. This functionality addresses sudden power imbalances, providing a rapid and efficient solution for frequency stabilization. The system’s external inverter serves as a central control hub, monitoring real-time grid frequency and directing FCEVs to supply virtual inertia and primary reserves through droop control, as required. Simulation results validate that FCEVs could effectively complement thermal generators, preventing unacceptable frequency drops, load shedding, and network blackouts. A techno-economic analysis demonstrates the economic feasibility of the concept, concluding that each FCEV consumes approximately 0.3 kg of hydrogen per day, incurring a daily cost of around EUR 1.5. For an island grid with a nominal power of 100 MW, maintaining frequency stability requires a fleet of 100 FCEVs, resulting in a total daily cost of EUR 150. Compared to a grid-scale battery system offering equivalent frequency response services, the proposed solution is up to three times more cost-effective, highlighting its economic and technical potential for grid stabilization in renewable-rich, non-interconnected power systems.

1. Introduction

Renewable energy sources (RESs) offer a promising solution for non-interconnected islands (NIIs) by reducing electricity costs, decreasing reliance on imported fossil fuels, and lowering CO₂ emissions [1,2]. However, the small to medium size and low inertia of these electrical networks pose significant challenges to frequency stability. To address this, network operators in NIIs impose RES penetration thresholds, which limit the proportion of instantaneous power generation from RESs relative to the total demand [3,4]. For example, in countries such as Greece [4] and France [5], this threshold is set at approximately 30%. As a result, any surplus power generated by RESs beyond this limit is curtailed, restricting their full utilization.

To increase RES penetration in NIIs and address frequency stability challenges, several strategies have been proposed. One approach involves enhancing RES control systems to enable artificial frequency control through techniques such as deloading, inertia emulation, and synthetic inertia response [6,7]. However, these methods often fall short at higher levels of RES penetration, as they struggle to deliver the necessary control performance [8]. Additionally, they underutilize the available renewable energy, as renewable sources are operated below their maximum output, at sub-optimal power levels, to maintain sufficient reserves for covering power deficits. Another strategy focuses on integrating fast energy storage systems (ESSs) [9]. Recent research [10,11,12] has explored various energy storage technologies capable of providing ancillary services to power grids, with particular emphasis on battery energy storage systems (BESSs) and location-specific solutions like pumped hydro and compressed air storage. Despite their potential, current BESS technologies face notable limitations, including accelerated degradation under frequent cycling [10] and high costs for large-scale deployment, which reduce their viability as standalone solutions for ancillary services.

Electric vehicles (EVs), whose adoption has grown rapidly in recent years, present promising solutions for both domestic and industrial energy applications. With EVs typically parked for most of the day (approximately 95%) [13], they offer a significant opportunity to deliver valuable energy services during idle periods. Vehicle-to-grid (V2G) technology allows EVs to act as supplementary power sources by connecting them to the grid, enabling the supply or sale of stored energy [14]. This bidirectional energy flow facilitates power delivery from EVs to homes or utility grids, improving grid flexibility, efficiency, and overall balance [15]. Consequently, V2G technology has gained traction as an innovative solution for meeting energy demands and enhancing grid stability through energy transactions [16].

Research on V2G technology has predominantly focused on battery electric vehicles (BEVs) [17,18], primarily because fuel cell electric vehicles (FCEVs) have historically been considered separately from EVs and have experienced slower market penetration. However, this dynamic is shifting as major automakers, including Hyundai, Toyota, Honda, and Mercedes-Benz, are now commercializing FCEVs such as the Hyundai Nexo [19], Toyota Mirai [20], Honda Clarity [21], and Mercedes-Benz GLC F-CELL [22]. Hydrogen-powered vehicles, spanning both light and heavy models, are gaining momentum in both private and public sectors. For instance, Athens recently introduced its first hydrogen-powered public bus, operating a circular route from Ano Patissia to Zappio and serving hundreds of passengers daily [23]. A recent European Commission report [24] identifies hydrogen technology as a viable solution for modern transportation systems. Compared to BEVs, FCEVs offer notable advantages, including significantly faster refueling times (averaging about 3 min) and extended driving ranges of approximately 500 km, positioning them as a competitive alternative in the evolving transportation landscape.

FCEVs, like other electric vehicles, can be integrated into V2G systems with an appropriate connection interface, serving as a novel source of power generation. However, research on FCEVs in V2G applications remains limited. Early studies have primarily been theoretical, focusing on the potential of FCEVs as distributed power generators and evaluating their economic viability in various electricity markets [25,26,27,28,29]. Other research has explored the role of FCEVs in balancing energy consumption at the building and community levels, yet these investigations have largely relied on theoretical models of FCEV operation under V2G conditions [30,31,32,33,34,35].

Recent studies have highlighted the potential of grid-connected FCEVs to reduce consumer costs and enhance grid efficiency. İnci [36] conducted an energy and profit analysis of a fuel cell electric vehicle-to-grid (FCEV2G) system supported by photovoltaic (PV) panels for continuous 24 h operation. This study analyzed a 100 kW FCEV connected to the grid and residential systems, alongside a 40 kW PV-integrated FCEV2G setup. Qian et al. [37] explored the energy-saving potential of the FCEV2G system using Monte Carlo simulations and optimized its economic benefits through genetic algorithms. Similarly, Li et al. [38] developed a robust economic optimization model using a hybrid algorithm that combined competitive swarm optimization (CSO) and the imperialist competitive algorithm (ICA). Their findings demonstrated significant economic gains with FCEVs integrated into the system, with sensitivity analyses revealing strong development potential and increasing economic benefits over time for FCEV2G applications. Furthermore, İnci et al. [39,40] and Savrun et al. [41] proposed efficient converter interfaces to enable FCEVs to provide ancillary services to the grid. Experimental work by Oldenbroek et al. [42] on a modified Hyundai ix35 FCEV demonstrated that FCEVs could respond significantly faster than traditional fast-reacting gas engines. This rapid response capability makes FCEVs highly promising for delivering frequency control services and stabilizing grid fluctuations caused by renewable energy variability.

References [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41] did not fully address the potential of FCEVs for fast frequency control. While Reference [42] investigated the ramp-up and start-up times of FCEVs and indirectly suggested their suitability as fast frequency control providers, it failed to account for the critical dynamic interactions occurring between FCEVs and the electric grid during power imbalances. Nevertheless, these interactions—between FCEVs and key grid components such as thermal generators, loads, and renewable energy sources—are crucial for evaluating system-wide impacts and ensuring overall grid stability.

Technical insights from battery electric vehicle-to-grid (BEV2G) technology cannot be directly transferred to FCEVs due to their distinct characteristics. Unlike BEVs, FCEVs exhibit slower ramp-up times, larger energy capacities, and the absence of grid-to-vehicle (G2V) capabilities, which prevents them from recharging directly from the grid. These differences necessitate a tailored approach to evaluate FCEVs as fast frequency control providers, distinct from BEVs. This paper addresses this gap by exploring the technical feasibility of FCEVs to supply virtual inertia and primary frequency reserves in low-inertia, non-interconnected islands during periods of high-power imbalances and renewable energy fluctuations. Key considerations include the ramp-up time and reaction delay of FCEVs, with the minimum fleet size required to maintain frequency stability being determined. Additionally, the economic feasibility of deploying FCEVs for fast frequency reserves is analyzed, based on estimated hydrogen consumption, demonstrating the practicality of this innovative solution.

2. Description of the Proposed Concept: Fuel-Cell-to-Grid Fast Frequency Control

Island networks primarily rely on thermal generators such as diesel generators, steam turbines, and gas turbines [43,44,45]. Among these, gas turbines offer the fastest response to rapid changes in power demand. Figure 1a illustrates the electrical system of an isolated island, comprising thermal power plants and wind turbines, both working to meet the island’s energy requirements.

Figure 1b illustrates the internal connection of the FCEV. FCEVs are equipped with essential components like hydrogen storage tanks and fuel cells, which convert stored hydrogen into electricity to power the vehicles’ electric motors. While parked in designated “Parking” areas, these vehicles can be repurposed to support grid operations through a specialized setup. Each FCEV connects to the grid via an external inverter system located at the parking area. This system converts the direct current (DC) generated by the fuel cells into alternating current (AC), compatible with the grid, allowing the vehicles to supply power directly to the network. As shown in Figure 1b, the fuel cell (FC) is linked to the external inverter through a mechanical switch. This switch is controlled by the fuel cell controller, which activates it close to when the FCEV is parked, establishing a connection between the FC and the grid (via the external inverter). The external inverter is directly connected to the grid, enabling the FCEV to provide rapid frequency ancillary services. This support helps thermal generators stabilize grid frequency more effectively. This capability to inject power into the grid from parked vehicles presents an innovative solution for stabilizing grid frequency during sudden power imbalances. The external inverter plays a central role, acting as the system’s “brain”. It continuously monitors real-time grid frequency data and, based on this information, directs the FCEVs’ fuel cells to promptly supply power to aid in frequency stabilization. The proposed control strategy for harnessing the power of FCEVs relies on two components: (a) a derivative term, emulating virtual inertia, and (b) a deviation term, emulating droop control. Both virtual inertia (VI) and droop control are well-established techniques for regulating the output of inverter-based distributed generators (IBDGs), ensuring frequency stability [46].

Description of FCEV Frequency Control Algorithm

The droop control mechanism operates by dynamically adjusting the output power of the fuel cells in response to changes in the grid’s frequency. When a power imbalance causes the grid frequency to drop, the droop control proportionally increases the power supplied by the FCEVs, helping to stabilize the frequency within an acceptable range. The primary goal is to provide a fast, automated response to balance electricity supply and demand, bridging the gap until secondary controllers take action to maintain frequency stability. The droop control principle is mathematically described in Equation (1) and visually represented in Figure 2.

$$P_{fcev,droop}=\left\{\begin{array}{c}0 if f>49.5 \mathrm{Hz}\\ K_{droop}·\left(49.5-f\right) if 48 \mathrm{Hz}<f<49.5 \mathrm{Hz}\\ P_{max} if f<48 \mathrm{Hz}\end{array}\right.$$

(1)

For minor power imbalances causing small frequency deviations—less than 0.5 Hz—gas turbines are solely responsible for stabilizing the frequency. FCEVs are only activated when the frequency drops below 49.5 Hz, minimizing unnecessary activations of private FCEVs. The FCEVs are configured to deliver their maximum power at 48 Hz [43], which is considered the critical frequency threshold in this study. This threshold is particularly relevant for non-interconnected islands, where frequency fluctuations are more pronounced due to limited system inertia. Once the frequency rises back to 49.5 Hz, driven by the secondary control of thermal units, the FCEVs cease power delivery, indicating the grid’s return to stable operation.

To emulate system inertia, the power output of the fuel cells is determined as a function of the frequency derivative and the synthetic inertia constant $\left(K_{vi}\right)$, as described by the following equation:

$$P_{fcev,vi}=-K_{vi}\times {\displaystyle \frac{\mathrm{d}f}{\mathrm{d}t}}$$

(2)

The total FCEV power is given in (3) as the summation of (1) and (2):

$$P_{fcev}=P_{fcev,droop}+P_{fcev,vi}$$

(3)

The simulation model details in the Laplace domain are illustrated in Figure 3. In this model, the frequency variation is fed back to the gas turbine, which provides primary reserves (3.3 MW/Hz) through droop control. The secondary control of gas turbines is activated 30 s after the disturbance, aiming to restore the island’s frequency to 50 Hz using a PI controller. Load and renewable energy source (RES) variations are represented by $\Delta P_L \mathrm{a}\mathrm{n}\mathrm{d} \Delta P_{wind}$, respectively. The system’s inertia is modeled using the constants $H_{eq}$ and $S_B$. Since RESs are inverter-based, they do not contribute to inertia response. In contrast, FCEVs supply virtual inertia (VI) and primary frequency reserves via droop control, as depicted in the figure. Fuel cells inherently exhibit a startup and ramp-up delay, modeled in the simulation through the terms $e^{-s·T_d}$ and $T_r(s)={\displaystyle \frac{1}{T_r·s+1}}$ for ramp-up behavior [47].

3. Case Study

Dynamic simulations were conducted using MATLAB/Simulink 2020a based on the network configuration depicted in Figure 3. The simulation parameters are detailed in

Table 1. Network and simulation parameters.

Block	Parameter	Value
Gas turbine	$\alpha ,c,b,X,Y,T_{CR},T_F,T_{CD}$	See Figure 4 of [51]
	Technical minimum–maximum	5–20 MW
	Droop control	3.3 MW/Hz
	Secondary control	Activation at 30 sec
Electric Network	Nominal Power	100 MW (island the size of Rhode, Greece)
	$H_{eq}$	1.5 ${\displaystyle \frac{\mathrm{MW}·\mathrm{s}}{\mathrm{MVA}}}$ [7,48,49] (Figure 8 of [49])
	D	0.05 [51]
	$S_B$	40 MVA (2 gas turbines) 60 MVA (3 gas turbines)
FCEV	Start-up delay ($T_d$)	100 ms
	Ramp-up time	2 s [47,50,52,53]
	$P_{max}$	200 kW (per truck) [50,54]
	$K_{droop}$	$200 \mathrm{kW}/1.5 \mathrm{Hz}$
	$K_{vi}$	0.0001
	Efficiency	50% [50,54]
Load	${\Delta P}_L$	0
RESs	${\Delta P}_{wind}$	Scenario 1: ${\displaystyle \frac{{\Delta P}_{wind}}{\Delta t}}={\displaystyle \frac{-27}{2}} (\mathrm{M}\mathrm{W}/\mathrm{s})$ Scenario 2: ${\displaystyle \frac{{\Delta P}_{wind}}{\Delta t}}={\displaystyle \frac{-27}{2}} (\mathrm{M}\mathrm{W}/\mathrm{s})$ Scenario 3: ${\Delta P}_{wind}=-20 \mathrm{M}\mathrm{W}$

. The studied network represents a renewable-rich island grid powered by wind turbines and gas turbines, with gas turbines chosen for their rapid response capabilities. Due to technical limitations, only two (2) or three (3) gas turbine units can be connected, each with a technical minimum capacity of 5 MW. This study investigates the role of FCEVs in providing frequency control ancillary services. The model includes variable load and wind inputs, adjusted according to the scenario being analyzed. The equivalent inertia $\left(H_{eq}\right)$ is set to 1.5 ${\displaystyle \frac{\mathrm{MW}·\mathrm{s}}{\mathrm{MVA}}}$, typical for gas turbines [7,48,49] (refer to Figure 8 of [49]), with a system base power $S_B$ of 40 MVA or 60 MVA for two or three connected gas turbines, respectively. Each FCEV, modeled as hydrogen-powered trucks, is assumed capable of delivering up to 200 kW of power [50].

Three scenarios are examined, differentiated by the type of disturbance and the number of connected gas turbines, which affects system inertia. For each scenario, the minimum number of FCEVs required to maintain frequency within acceptable limits is determined:

▪

Scenario 1: The network includes three (3) gas turbines. The disturbance involves a rapid 27 MW drop in wind power over 2 s, simulating the output reduction in a 50-MW wind farm when wind speed decreases from 14 m/s to 10.8 m/s. This wind power variation is modeled in Figure 3 via the input ${\Delta P}_{wind}$.

▪

Scenario 2: The disturbance is identical to Scenario 1. However, only two (2) gas turbines are connected, resulting in reduced system inertia, highlighting the impact of lower inertia on frequency stability.

▪

Scenario 3: The network consists of three (3) gas turbines. The disturbance is a sudden outage of a large 20 MW wind farm, potentially caused by a fault.

For each scenario, the simulation evaluates the ability of FCEVs to stabilize the grid frequency and identifies the minimum number of vehicles required to meet compliance thresholds.

A.

Simulation Results: Scenario 1

In this scenario, a minimum fleet of 70 fuel cell-powered trucks is required to maintain frequency stability. The frequency response immediately following the disturbance (at t = 1 s), both with and without the FCEVs, is shown in Figure 4. The results highlight the critical role of FCEVs, as their absence results in the system frequency quickly dropping below the minimum allowable limit of 48 Hz, risking instability or load shedding. With the contribution of FCEVs, the frequency is stabilized above 48 Hz, without exceeding this critical threshold at all. Subsequently, it is restored to 50 Hz, following the activation of secondary control by thermal units at 30 s.

Figure 5 illustrates the total power output of gas turbines after the disturbance, comparing scenarios with and without FCEVs. In the absence of FCEVs, the primary control of gas turbines increases their output to the maximum capacity of 60 MW to address the power imbalance. When FCEVs are included, their contribution reduces the load on gas turbines, allowing them to operate at lower power levels. After 30 s, the secondary control is activated, restoring the frequency to 50 Hz, at which point the gas turbines resume handling the entire power deficit.

The total power contribution from the FCEVs (70 hydrogen trucks) is shown in Figure 6. The droop control and virtual inertia response of the FCEVs is triggered after a delay of 100 ms, effectively mitigating the system’s power deficit and reducing the rate of change of frequency (ROCOF). The FCEVs provide a steady power output until the secondary control of gas turbines is activated at 30 s. Following this, the FCEVs gradually reduce their power output according to the droop control settings (Equation (1)). Once the frequency returns to 49.5 Hz, the FCEVs stop supplying power, ceasing hydrogen consumption.

Figure 7 displays the hydrogen consumption per FCEV for Scenario 1, which amounts to approximately 0.1 kg of hydrogen per event. Given that such large disturbances (causing frequency drops below 49.5 Hz) are infrequent and may occur only a few times per day (if at all), the hydrogen consumption is negligible, leaving sufficient reserves for the vehicle’s driving needs. In terms of cost, with hydrogen prices ranging between 2 and 10 EUR/kg [45,46,55], the total cost per vehicle per disturbance is estimated to range from EUR 0.2 to 1, making this approach both effective and economical.

B.

Simulation Results: Scenario 2

In this scenario, a minimum fleet of 99 fuel cell-powered trucks is necessary to maintain frequency stability. Figure 8 illustrates the frequency response with and without the FCEVs. Without FCEVs, the frequency collapses entirely, resulting in a system blackout because the two gas turbines, with their limited reserves of 20 MW, cannot compensate for the significant 27 MW reduction in wind power. Consequently, as shown in Figure 9, the gas turbines reach their maximum output but fail to prevent the frequency drop. On the other hand, as shown in Figure 10, this shortfall is effectively mitigated by the FCEVs, which provide 16 MW of primary frequency reserves, preventing frequency collapse. However, in this scenario, FCEVs continue supplying power even after secondary control because the gas turbines reach their maximum capacity (40 MW) at around 45 s (Figure 9) and cannot restore the frequency to 50 Hz. The resulting secondary reserve deficit of 7 MW is compensated by the FCEVs (Figure 10), which increases hydrogen consumption, as shown in Figure 11. If left unaddressed, this condition could deplete the vehicles’ hydrogen reservoirs entirely. To prevent complete depletion, additional secondary reserves should be deployed by the system. This scenario highlights the need for FCEVs to be mainly used for virtual inertia and primary reserves, and not for secondary reserves, to minimize hydrogen consumption.

C.

Simulation Results: Scenario 3

In this scenario, a larger fleet of 122 fuel cell-powered trucks is required to maintain frequency stability due to the instantaneous outage of a generator, which results in a higher rate of change of frequency (ROCOF). Unlike the previous scenarios where wind power reductions occurred gradually over 2 s, this disturbance is immediate. The frequency response, with and without FCEVs, is shown in Figure 12. Without the support of FCEVs, the system frequency falls below 48 Hz, necessitating load shedding. In contrast, the integration of FCEVs effectively prevents unacceptable frequency drops. Following the activation of the secondary control of gas turbines at t = 30 s, the frequency is restored to 50 Hz. Figure 13 illustrates the total power output of gas turbines after the disturbance. With the participation of FCEVs, the primary control of the gas turbines requires less power output due to the assistance from the FCEVs. Once the secondary control is activated at 30 s, the gas turbines take over the entire power deficit. The total power contribution from the FCEVs is shown in Figure 14. Their droop control and virtual inertia mechanisms are activated with a delay of 100 ms, supplying virtual inertia and droop-controlled power to reduce the ROCOF. The FCEVs maintain a constant power output until the secondary control of the gas turbines takes over at t = 30 s. After this point, the FCEV power output is gradually reduced according to the droop control settings. Once the frequency stabilizes at 49.5 Hz, the FCEVs cease power delivery, as described by Equation (1). Finally, Figure 15 presents the hydrogen consumption per FCEV for Scenario 3, amounting to approximately 0.05 kg of hydrogen per vehicle, per disturbance. With hydrogen prices ranging from 2 to 10 EUR/kg, this translates to a cost of EUR 0.1 to 0.5 per disturbance per vehicle, demonstrating the economic feasibility of the FCEV2G approach.

4. Discussion

In the examined island, it is estimated that an average of 100 FCEVs are required to stabilize grid frequency. The proposed V2G concept necessitates parking FC trucks in designated areas. It is important to emphasize that this number refers to the cumulative parking areas across the entire island, not a single location. To assess the technical feasibility of concentrating this number of parked trucks on the entire island, a general ratio based on typical truck-to-population distributions was used. For developed regions, there are approximately 4–8 trucks per 1000 people, not including the number of other heavy vehicles such as buses. The examined island, with a base power of 100 MW, corresponds to an island the size of Rhodes, Greece—a developed, touristic island with a population of 125,000 people. Based on these assumptions, the estimated number of trucks on the island ranges from 500 to 1000. Considering that trucks spend a significant portion of time parked (for loading, unloading, driver rest, etc.), it is concluded that achieving the required number of 100 parked FC trucks, while challenging, is a realistic goal for the future.

A preliminary techno-economic evaluation of the proposed concept is presented below. Based on Figure 7, Figure 11 and Figure 15, the hydrogen consumption per disturbance per vehicle averages 0.1 kg ${\mathrm{H}}_2$. In assuming approximately that three disturbances per day trigger the FCEVs (i.e., the frequency drops below 49.5 Hz, as defined in Equation (1)), each FCEV consumes about 0.3 kg ${\mathrm{H}}_2$ per day. While hydrogen costs are currently high, projections indicate that prices will drop below 5 EUR/kg in the near future [55,56,57,58]. Consequently, the daily cost of providing frequency ancillary services using FCEVs is estimated at EUR 1.5 per vehicle. In the case study, an average of 100 FCEVs are required to stabilize frequency, resulting in a total daily cost of approximately EUR 150 or an annual cost of EUR 54,750.

When compared to a 10 MW grid-scale lithium-ion battery system designed to deliver frequency response services, the battery’s capital expenditure (CAPEX) is estimated at around EUR 1,000,000, assuming a levelized CAPEX of 100 EUR/kW. To provide the same level of frequency response as the FCEV fleet, the battery would require approximately 500 kWh of electricity daily, incurring an energy cost of 75 EUR/day (assuming an electricity price of 0.15 EUR/kWh). This results in an annual energy cost of EUR 27,375. When considering a battery lifetime of 10 years and an inflation rate of 5%, the annualized CAPEX for the battery system—calculated using the capital recovery factor—is approximately EUR 130,000. Thus, the total annual cost of the battery system amounts to EUR 157,375, which is nearly three times the annual cost of the proposed FCEV-based approach.

Table 2. Economic comparison with a grid-scale battery system.

FCEVs	Consumption per disturbance per vehicle	0.1 kg H2
	Number of disturbances per day	3
	Cost of hydrogen	5 EUR/kg
	Number of vehicles	100
	Total annual cost	150 EUR/day or 54,750 EUR/year
Grid-scale battery	Levelized CAPEX	100,000 EUR/MW
	Capacity	10 MW
	Capital recovery factor	0.13
	Annualized CAPEX	130,000 EUR/year
	Electricity consumption	500 kWh/day
	Levelized electricity cost	0.15 EUR/kWh
	Annual electricity cost	75 EUR/day or 27,375 EUR/year
	Total annual cost	157,375 EUR/year

summarizes this comparison, emphasizing the economic viability of utilizing FCEVs for fast frequency response services in non-interconnected, low-inertia island grids.

The adoption of FCEVs to provide fast-frequency response to the grid necessitates the standardization of the service, considering both grid requirements and FCEV characteristics. In this paper, we proposed clear guidelines for this standardization, as outlined in Figure 2 and the accompanying table, which are summarized below:

✓

Frequency Threshold: The FCEV response is triggered when the frequency drops below 49.5 Hz.

✓

Droop Slope: ${\displaystyle \frac{dp}{df}}={\displaystyle \frac{P_{max}}{1.5}} (\mathrm{kW}/\mathrm{Hz})$, where $P_{max}$ represents the maximum power provided by the FCEV.

✓

Maximum Reaction Delay: The response delay should be less than 100 ms.

✓

Ramp-Up Time: The ramp-up time from 0 to $P_{max}$ is recommended to be less than 2 s.

These recommendations aim to ensure the effective integration of FCEVs into grid-support operations considering both the network’s requirements and the FCEV response capability.

5. Conclusions

This paper explores an innovative approach to fast frequency control in electric grids utilizing parked FCEVs, specifically heavy vehicles like trucks. These vehicles, equipped with hydrogen storage tanks and fuel cells, can be repurposed as dynamic grid-support assets while parked in designated areas. Through an external cable and inverter system, FCEVs inject power directly into the grid by converting DC from fuel cells into AC comp4atible with grid requirements. This capability addresses sudden power imbalances, providing a rapid and efficient solution for frequency stabilization. The system’s external inverter acts as a central control hub, continuously monitoring real-time grid frequency and directing FCEVs to supply virtual inertia and primary (droop-controlled) power reserves when needed. By harnessing these capabilities, the proposed approach supports thermal generators in stabilizing grid frequency more effectively, enhancing the resilience of low-inertia grids. Specifically, three scenarios were analyzed, each involving different power imbalances and system inertia levels. The results reveal that, without the support of FCEVs, two out of the three scenarios lead to frequency drops approaching 46 Hz—well below the acceptable threshold of 48 Hz. Furthermore, one scenario results in a complete system collapse. In contrast, when FCEVs are utilized as dynamic grid-support assets, the frequency remains above 48 Hz in all scenarios, demonstrating the effectiveness of the proposed concept.

Regarding the economic feasibility of the proposed concept, the daily hydrogen consumption per vehicle is estimated at approximately 0.3 kg, corresponding to a cost of around EUR 1.5 per vehicle. For the examined island with a nominal power of 100 MW (an island the size of Rhode, Greece), the case study indicates that maintaining frequency stability requires a fleet of 100 FCEVs, resulting in a total daily cost of EUR 150. Comparatively, the proposed FCEV-based solution offers a cost advantage, with expenses up to three times lower than those of a grid-scale battery system designed to deliver equivalent frequency response services. Namely, the annual cost of the proposed concept amounts to EUR 54,750, compared to the annualized cost of a grid-scale battery, which was calculated at EUR 157,375.