Heavy-Duty EVs with V2G Capability

• Heavy-duty EVs with V2G capability are electric vehicles equipped with bidirectional chargers that can absorb power and inject energy to support grid frequency stability.
• Simulation studies indicate that bidirectional V2G control reduces settling time by up to 40% and improves frequency nadir by 20–30% compared to unidirectional control.
• Charging strategies such as constant-minimum-power optimize grid support by maintaining consistent reserve capacity while incurring minimal battery degradation.

Heavy-duty electric vehicles (EVs) with vehicle-to-grid (V2G) capability are increasingly recognized as a significant source of fast, distributed flexibility for power system frequency control, particularly during periods of low inertia and high renewable penetration. V2G leverages bidirectional chargers and control schemes that allow EVs not only to absorb power (charging) but also to inject stored energy back into the grid in response to frequency events. Recent studies using detailed simulations for the California power grid, characterized by substantial renewable integration, have systematically quantified the capacity, response speed, and deployment strategies of aggregated heavy-duty EV fleets under multiple operational modes and charging strategies, highlighting their technical viability and system-level impact (Tao et al., 14 Dec 2025, Tao et al., 16 Dec 2025).

1. Heavy-Duty EV Aggregation and System Model

Each heavy-duty EV is modeled with a representative battery capacity $C_\mathrm{bat}$ in the range of $200\;\text{kWh}$ to $600\;\text{kWh}$, with a nominal value of $E_i=300\;\text{kWh}$ used in fleet simulations. The maximum inverter-limited charge and discharge power $P_\mathrm{max}$ is $100\;\text{kW}$ for immediate/delayed charging or $50\;\text{kW}$ for constant-minimum-power charging. Charger and battery roundtrip efficiencies are set at $$\eta_\mathrm{ch} \approx \eta_\mathrm{dis} \approx 0.95$$. For each vehicle $i$, the state-of-charge (SoC) evolves as:

$$\frac{d\,\mathrm{SoC}_i}{dt} = \frac{\eta_\mathrm{ch} P_{\mathrm{ch},i}(t) - \frac{1}{\eta_\mathrm{dis}} P_{\mathrm{dis},i}(t)}{C_\mathrm{bat}}$$

At the fleet level, $N_\mathrm{HDV}$ vehicles may be available according to their operational duty cycle, and a fraction $\alpha \in \{0.2, 0.4, 0.6, 0.8, 1.0\}$ participates in frequency response. The aggregate EV power delivered or absorbed by the grid is $P_\mathrm{EV}(t) = \sum_i P_{\mathrm{EV},i}(t)$. SoC bounds guarantee sufficient energy for scheduled mobility.

2. Frequency Response Control Schemes

EVs respond to grid frequency deviations via droop-based controllers analogous to those used for synchronous generators. For the system-level deviation $\Delta f(t) = f(t) - f_0$ with $f_0 = 60\,\text{Hz}$, the following control laws are implemented:

• V1G (unidirectional control): Charging is paused when $f(t) < f_\mathrm{th}=59.7\,\text{Hz}$. EVs cease charging but do not inject power.
• V2G (bidirectional control): Proportional droop control is deployed,

$$P_{\text{ctrl}}(t) = K_{\mathrm{EV}}\cdot (f_0 - f(t))$$

subject to $\pm P_\mathrm{max}$ and SoC constraints. $K_\mathrm{EV}$ is chosen such that the fleet delivers full bidirectional rating at $\Delta f = -0.3$ to $-0.5\,\text{Hz}$. For a notional depot of $1000$ vehicles, this corresponds to up to $20\,\text{MW}$.

The aggregate frequency dynamics are simulated using the swing equation with effective inertia $H_\mathrm{eff}=6.4\,\text{s}$ for a $19.83\,\text{GW}$ California grid snapshot, including generator droop (with $R_\mathrm{gen}=0.05\,\mathrm{pu/Hz}$), turbine and governor dynamics, and load damping.

3. Charging Strategies and Dynamic Reserve Availability

Three practically motivated charging strategies are analyzed:

Strategy	Peak Power	Charging Window	Duration (h)	V2G Reserve (Up/Down)
Immediate	100 kW	16:00–23:00	7	±100 kW $\times \alpha N_\mathrm{HDV}$
Delayed	100 kW	23:00–06:00	7	±100 kW $\times \alpha N_\mathrm{HDV}$
Constant-minimum	50 kW	16:00–06:00	14	±50 kW $\times \alpha N_\mathrm{HDV}$

• Immediate charging: Vehicles commence charging at maximum rate upon return, yielding a sharp evening peak, and the available upward reserve (V1G) is limited due to full power consumption.
• Delayed charging: Charging starts later (overnight), maximizing upward V1G reserve during 16:00–23:00; downward V2G is maintained during charging intervals.
• Constant-minimum-power: Vehicles charge at a uniform low rate, optimizing grid support by flattening depot demand and enabling consistent V2G reserve capacity throughout the off-shift period.

These strategies influence the distribution and timing of upward (load-shedding) and downward (power-injection) reserves, affecting the primary frequency response.

4. Simulation Scenarios and Performance Metrics

System-level simulations assess EV fleet response to a loss-of-generation event ($\Delta P_L = 1800\,\text{MW}$) during a low-inertia, high-renewable hour (renewable share $11.8\%$; $H_\mathrm{eff}=6.4\,\text{s}$). Key frequency-security metrics include:

• Frequency nadir ($f_\text{nadir}$): Minimum frequency following the disturbance.
• Rate-of-change-of-frequency (RoCoF): Maximum $|\frac{d\Delta f}{dt}|$ to $t_\text{nadir}$.
• Overshoot: Maximum $f(t) - f_0$ post-disturbance.
• Settling time ($t_\text{settle}$): Time to return within $|f - f_0| \leq 0.02\,\text{Hz}$.

For $100\%$ fleet participation under immediate charging, results are summarized:

Mode	$f_\text{nadir}$ (Hz)	RoCoF (Hz/s)	Overshoot (Hz)	$t_\text{settle}$ (s)
None	59.20	–0.16	0	32
V1G	59.55	–0.14	0	28
V2G	59.75	–0.12	+0.08	20

Constant-minimum-power charging achieves the highest nadir (59.80 Hz) and fastest settling (18 s) in V2G mode. Improvements scale quasi-linearly with participation: each additional 10% fleet increment yields ≈0.05 Hz improvement in $f_\text{nadir}$ in V2G (Tao et al., 14 Dec 2025, Tao et al., 16 Dec 2025).

5. Comparative Assessment of Control Modes and Strategies

• V2G consistently outperforms V1G, providing both faster and deeper stabilization following large disturbances. V2G reduces settling times by up to 40% and raises nadir by 20–30% compared to V1G.
• Constant-minimum-power charging best balances ancillary-service readiness and grid stress mitigation, improving the frequency nadir by ≈0.03 Hz over immediate charging and reducing feeder congestion.
• Delayed charging supports off-peak electricity pricing (potentially reducing depot bills by ~20%) but can reduce readiness if a disturbance occurs prior to charging initiation—making V2G indispensable in such windows.
• Control gain $K_\mathrm{EV}$ should be set to deliver full rated power for moderate frequency excursions ($\Delta f = -0.3\,\text{Hz}$), ensuring prompt and proportional response.

V2G cycling per frequency response event is modest ($\sim$2 kWh/vehicle, $\Delta$SoC $\sim$0.7%), corresponding to less than 1% of a full cycle, with degradation costs below $0.20/vehicle/event$, which can be offset by frequency control market revenues ($1$–$3/vehicle/event$) (Tao et al., 16 Dec 2025).

6. System Integration and Practical Implications

Heavy-duty EV fleets with V2G capability can supply primary response capacity on the order of gigawatts when fully coordinated at scale, significantly reducing reliance on traditional spinning reserves during renewable-dominated, low-inertia periods. This operational flexibility is robust to fleet participation rates, with meaningful contributions (0.2 Hz nadir improvement) even at moderate levels (40% participation).

Integrated management platforms are essential for SoC buffering, fleet scheduling, and reliable communications across depots. Regulatory structures must adapt to recognize and fairly compensate distributed V2G ancillary services, reflecting both their system stability value and battery degradation costs.

7. Limitations and Forward Directions

Principal technical limitations include constraints on vehicle availability (due to duty cycles), charger hardware (bidirectional equipment required), and SoC management for mobility assurance. System-level participation is ultimately limited by the size of the off-shift fleet, network hosting capacities, and communication latency. The implementation of real-time aggregation and economic incentive structures, in conjunction with dynamic scheduling that aligns grid need with depot mobility requirements, remains an active area of development.

Future research directions recommended in the referenced studies include demonstration of real-time aggregator controls, assessment of network impacts under high spatial EV clustering, co-optimization with other distributed resources, and refinement of market products to suitably reward dynamically dispatched distributed frequency support (Tao et al., 14 Dec 2025, Tao et al., 16 Dec 2025).