Bi-Directional EV Charging

• Bi-Directional EV Charging is a technology enabling electric vehicles to both draw power for charging and discharge stored energy to the grid, serving as distributed energy resources.
• Converter technologies like Dual Active Bridge facilitate efficient power transfer through precise phase-shift modulation, ensuring safe and reliable mode transitions.
• Integrated economic models and optimization frameworks focus on balancing grid services, dynamic pricing, and battery degradation costs to enhance overall system profitability.

Bi-directional electric vehicle (EV) charging—commonly termed Vehicle-to-Grid (V2G)—is the process whereby EVs are equipped not only to draw power for battery recharging but also to dispatch stored energy back to the grid or to other loads. This dynamic two-way interaction positions EVs as distributed energy resources capable of providing grid services, optimizing energy costs, and participating actively in power system operation. The advent of V2G extends both economic and operational optimization challenges across actors—EV owners, aggregators, charging station operators, and system operators—necessitating rigorous models for system coordination, market integration, and control architecture. The deployment of bi-directional EV charging is accelerated by advances in power electronics, communication protocols, and the evolution of smart grid paradigms.

1. Converter Technologies and Charging Configurations

Bidirectional charging systems are fundamentally distinct from unidirectional systems: they support both grid-to-vehicle (G2V) and vehicle-to-grid (V2G) modes by permitting energy to flow in either direction. This technical capability is enabled by specialized converter topologies, most notably the Dual Active Bridge (DAB) and various resonant DC–DC/AC–DC converter designs. The DAB configuration features two full-bridge converters, with controlled phase shift, yielding a power transfer function

$$P_{\text{avg}} \approx \frac{V_1 V_2}{\omega L} \varphi,$$

where $V_1$, $V_2$ are the DC voltages on each side, $\varphi$ is the bridge phase shift, and $\omega L$ is determined by transformer leakage inductance and switching frequency. Bidirectional operation is achieved by adjusting $\varphi$: positive for charging, negative for discharging. Additional implementation alternatives include modified buck/boost, Vienna, and Swiss-converter topologies.

Effective bi-directional systems require robust control strategies, such as phase-shift modulation or closed-loop phase-current tracking, to maintain output quality and safety across mode transitions. Converter architecture trade-offs reflect priorities for isolation, power density, cost, and system reliability (Acharige et al., 2022).

Interface standards—CHAdeMO (supporting V2G operation), CCS, SAE J1772, IEC 61851—and communication protocols (ISO 15118) underpin interoperability and safety. Integrated onboard/offboard frameworks and grid interconnection via IEEE 1547 and UL1741 standards are mandatory for secure V2G deployment.

2. Economic and Incentive Mechanism Design

The economic operation of bi-directional charging incorporates both static and dynamic market models. Static allocation mechanisms distribute energy based on utility maximization or cost minimization using classical or quadratic utility functions, such as

$$\theta(x) = \begin{cases} -ax^2 + bx, & x \leq b/(2a) \ b^2/(4a), & x > b/(2a) \end{cases}$$

where $x$ is the allocated energy and $a,b$ are agent-specific coefficients (Shuai et al., 2016).

Dynamic models, necessary for real-world scheduling under uncertainty, leverage repeated auctions, Stackelberg games (aggregator-leader, EV-follower), and iterative tâtonnement or revelation schemes. In V2G contexts, game-theoretic models determine energy allocation and revenue sharing, with aggregators often employing mechanisms such as the Shapley value for profit distribution across fleets delivering ancillary services (regulation, reserve).

Auction-based systems, notably those built upon Vickrey–Clarke–Groves (VCG) principles, support efficient, incentive-compatible allocation by soliciting (unit price, quantity) bids from participating EVs, subject to constraints such as transformer and line limits.

Major economic findings indicate that V2G participation for ancillary services—frequency regulation, for example—is frequently more attractive than pure energy arbitrage, though the profitability critically depends on electricity price spreads, revenue sharing schemes, and battery degradation costs (Shuai et al., 2016, Schwenk et al., 2020).

3. Optimization, Control, and Scheduling Frameworks

Optimal scheduling under bi-directional operation is formulated at different system layers:

• Decentralized real-time optimization: Charging stations receive baseload and user behavior predictions from a control center, apply a real-time control signal $c^i(t) = \lambda [B(t) + \sum_n p_n^i(t)]$, and locally solve for $p_n(t)$ to minimize composite objectives—load flatness, cost, and smoothness (Xiong et al., 2017). Constraints include power bounds, plug-in intervals, and energy satisfaction.
• Multi-objective optimization (MOO): Coordination incorporates objectives for EVCS profit, grid loss minimization, and BEV owner cost—including battery wear and carbon credits—balanced via evolutionary algorithms (e.g., NSGA-II) and dynamically adjusted pricing for service provision (Mehrnia et al., 22 Sep 2025).
• Integration with grid operation: For national-scale fleets, large MIQP or QP programs coordinate charging/discharging under flexibility constraints, employing ADMM for decomposition and relaxation of bilinear (non-simultaneity) constraints (Nespoli et al., 2022).
• Peer-to-peer and collective schemes: Beyond traditional aggregator-centered frameworks, peer-to-peer (P2P) protocols enable direct EV-to-EV energy transfers coordinated via cloud systems, with mobile charging stations (MoCS) providing network-wide balancing (Chakraborty et al., 2020).

Emerging models also co-optimize charging incentives and grid stability metrics, for example, by minimizing the $\mathcal{H}_2$-norm of grid transfer functions subject to user incentivization constraints, using state-feedback laws with LQR synthesis (Podder et al., 2 May 2024).

4. Grid Impact and Ancillary Service Integration

Bi-directional EV charging offers both opportunities and grid integration challenges:

• Peak shaving and valley filling: Optimally scheduled V2G discharging can reduce peak load by up to 30%, lower voltage drops by 2%, and decrease transmission line currents by 40%, as demonstrated in substantiated power flow studies (Xiong et al., 2017).
• Congestion and voltage control: Incorporation of voltage and congestion constraints in low-voltage (LV) distribution network optimization models shows that dynamic tariffs, tailored to transformer loading, combined with V2G charging, reduce the frequency of line congestion and decrease operational costs for all stakeholders (Verbist et al., 2023).
• Ancillary services and flexibility: Fleet-scale V2G scheduling determines the upward/downward flexibility envelope, enabling informed bidding in ancillary service markets, with rapid ADMM-based solution schemes scalable to thousands of vehicles (Nespoli et al., 2022).

A plausible implication is that, as fleet mobilization grows, co-optimization and real-time control architectures are critical for quantifying and deploying EV flexibility at the grid edge.

5. Battery Aging, Degradation, and Profitability

Battery aging, both cyclic and calendar, is a dominant factor in the economic viability of V2G:

• Degradation models incorporated into charging optimization use empirical relationships—e.g., $$\Delta H_{\text{cyc},n} = \beta_A|\Delta E_n|^{\beta_B}$$, with cost scaled by battery value at end-of-life.
• Failure to include aging costs underestimates lifetime EV operating costs by ∼30%; for V2G profitability, the required ratio of sell/buy electricity price ($\gamma^*$) must be sufficiently high to offset both energy losses and accelerated capacity fade (theoretical threshold $\gamma^*$ frequently exceeds values typical of retail tariff spreads) (Schwenk et al., 2020).
• Thermal models, especially ANN-based, become necessary as charging rates exceed 7 kW; precise battery temperature tracking is essential to limit unwarranted degradation and optimize dynamic charging/discharging profiles.

6. Communication, Cybersecurity, and Privacy Requirements

Efficient and secure bi-directional operation is impossible without robust two-way communication:

• Communication architectures: Layered topologies support information flow (preferences, pricing, regulation signals), with message types and frequencies tailored to use case (iterative pricing, real-time regulation, day-ahead scheduling) (Shuai et al., 2016).
• Cybersecurity: V2G presents expanded attack surfaces. Security measures leverage well-established standards (AES, TLS), PKI for authentication, and are challenged by scalability and legacy hardware (Metere et al., 2021). Risk assessments use quantifications such as $R = P \times I$ (probability × impact). Mitigation strategies include secure channels, intrusion detection, and flexible cryptographic governance.
• Privacy: V2G operation entails collection of user-sensitive data (travel, willingness-to-pay), requiring encryption, anonymization, and adherence to regulatory frameworks (e.g., GDPR).

7. Systemic Challenges, Future Directions, and Open Research Problems

Despite progress, several research gaps persist:

• Dynamic mechanism design: Truthful, incentive-compatible mechanisms under uncertainty and strategic user adaptation remain unresolved, especially in online/dynamic auction settings (Shuai et al., 2016).
• Scalability and computation: Incorporation of spatial–temporal power constraints and dynamic driving patterns necessitates advanced MILP/QP decomposition and distributed optimization (Lee et al., 17 Jul 2025, Nespoli et al., 2022).
• Distributed generation and reactive support: Joint planning models now address both active and reactive bi-directional power flows, using sequential MILP/MISOCP decomposition for grid planning and EV control; V2G charging stations offering reactive power support can reduce voltage fluctuations by ∼17.6% and lower voltage variance by 28.6% (Wang et al., 2023).
• Data-driven scheduling and learning: Incorporation of demand response via ML, reinforcement learning, and clustering for modeling EV user behavior is a topic of current exploration (Tang et al., 2022).
• Battery degradation cost offsetting: Economic frameworks that include carbon revenue (via emissions programs) are now deployed to compensate for battery degradation in intensive V2G regimes, leveraging dynamic economic dispatch with aggregated objectives (Mehrnia et al., 22 Sep 2025).
• Peer-to-peer and mobility integration: Scaling P2P, spatial arbitrage, and mobile grid services (e.g., with delivery fleets) demands further paper of spatial price forecasting, infrastructure constraints, and full integration with delivery/routing operations (Mohammadian et al., 2023, Du et al., 25 Jun 2025).

These directions highlight the multi-disciplinary, intersectional future for bi-directional EV charging: control, economics, communication, cybersecurity, and grid planning co-evolve to anchor EVs as critical, self-optimizing energy resources in modern power systems.