EV Aggregation Strategy Enhances Voltage Stability in Active Distribution Networks
In an era defined by the urgent global push toward decarbonization, the integration of renewable energy sources into power grids has become both a necessity and a challenge. As solar photovoltaics and wind turbines proliferate across urban and rural landscapes, distribution networks—once passive conduits of electricity—are transforming into dynamic, bidirectional systems known as active distribution networks (ADNs). However, this evolution brings with it a host of technical complexities, chief among them voltage instability caused by the intermittent and unpredictable nature of distributed generation.
A groundbreaking study published in Electric Drive offers a compelling solution to this pressing issue by leveraging the untapped potential of electric vehicles (EVs) not just as consumers of electricity, but as active participants in grid stabilization. Authored by Jialin Yu, Jinrui Guo, Xiaodong Tu, Weiliang Liu, and Weidong Zhong, the paper introduces a novel cooperative voltage regulation strategy that integrates EV aggregations, multifunctional photovoltaic inverters, and reactive power compensation devices into a unified control framework. The approach promises to enhance grid resilience, reduce infrastructure costs, and support the safe, efficient operation of modern distribution systems under high renewable penetration.
At the heart of this innovation lies a sophisticated prediction model rooted in the concept of the “travel chain”—a behavioral framework that captures the spatiotemporal patterns of EV usage. Unlike conventional forecasting methods that treat EV charging as a random or isolated event, the travel chain model recognizes that daily vehicle movements form a sequence of interlinked activities: departure from home, commute to work, errands, leisure trips, and eventual return. Each leg of this chain carries probabilistic information about departure times, parking durations, destination types (residential, commercial, recreational), and state of charge (SOC) upon arrival. By modeling these dependencies through Monte Carlo simulations and Markov state transitions, the researchers achieve a significantly higher degree of accuracy in predicting aggregated EV load profiles.
This predictive precision is not merely an academic exercise; it is the foundation upon which effective grid services can be built. For EVs to participate meaningfully in voltage regulation, grid operators must know—hours or even days in advance—how many vehicles will be available at which locations, with what battery capacity, and for how long they can remain connected. The travel chain model delivers exactly that: a high-fidelity forecast of flexible load availability across different functional zones of a city. In comparative tests against traditional methods like the NSGA-II optimization algorithm, the proposed approach demonstrated dramatic improvements in forecasting accuracy, reducing mean absolute percentage error (MAPE) from nearly 25% to just 3.54%. Such reliability transforms EV fleets from uncertain variables into dispatchable grid assets.
With accurate load predictions in hand, the research team designed a hierarchical, sensitivity-based voltage control strategy. Recognizing that low-voltage distribution networks are characterized by high resistance-to-reactance ratios, the authors emphasize that both active and reactive power injections significantly influence node voltages. Their strategy begins by calculating voltage-power sensitivity matrices—quantifying how much a unit change in real or reactive power at one node affects the voltage at another. These sensitivities are then used to prioritize control actions.
The regulation protocol follows a “reactive power first, active power later” sequence, a principle aligned with established power engineering practices that minimize energy losses and preserve battery life. When overvoltage conditions arise—common during midday solar peaks—the system first dispatches reactive power adjustments from available resources: static VAR compensators, smart PV inverters operating in voltage support mode, and, crucially, aggregated EV chargers capable of bidirectional power flow (V2G) or even unidirectional reactive support (as many modern chargers can provide). If voltage deviations persist after reactive measures are exhausted, the system then taps into active power modulation—temporarily reducing EV charging rates or, in advanced scenarios, drawing power from vehicle batteries to absorb excess generation.
This layered approach proved highly effective in simulation scenarios modeled on real-world distribution feeders. In one test case, three adjacent buses simultaneously experienced severe overvoltage—exceeding 15% above nominal levels—due to coincident solar generation and light local load. Conventional reactive compensation devices alone were insufficient to correct the imbalance. However, by engaging EV aggregations connected to those same nodes, the system restored voltages to within the permissible ±5% band in under 0.07 seconds. In another scenario involving a bus without dedicated reactive compensation hardware, the EV aggregation provided the dominant corrective action, demonstrating the strategy’s adaptability to heterogeneous grid configurations.
Critically, the proposed method does not require massive new infrastructure investments. Instead, it maximizes the value of existing assets—solar inverters, EV chargers, and minimal reactive compensation units—through intelligent coordination. This is a significant economic advantage, especially for utilities facing budget constraints while managing rapid renewable integration. Moreover, by prioritizing EV-based regulation over mechanical solutions like on-load tap changers or capacitor banks, the strategy reduces wear and tear on physical equipment and enables faster, more granular responses to voltage fluctuations.
The implications extend beyond technical performance. As EV adoption accelerates worldwide—with global sales surpassing 10 million units in 2022 and projected to reach 40 million annually by 2030—the collective battery capacity parked in garages, office lots, and shopping centers represents a vast, distributed energy reservoir. Harnessing this resource for grid services creates a virtuous cycle: EV owners can earn revenue through participation in ancillary service markets, utilities gain a low-cost voltage regulation tool, and society benefits from a more stable, renewable-friendly grid. The study’s framework provides a practical blueprint for activating this potential.
Importantly, the researchers acknowledge limitations and outline clear paths for future work. The current model assumes widespread availability of smart, controllable EV chargers—a reasonable assumption given ongoing regulatory pushes in the EU, U.S., and China for standardized V1G (unidirectional smart charging) and eventual V2G (bidirectional) capabilities. However, real-world deployment will require robust cybersecurity protocols, standardized communication interfaces (e.g., IEEE 2030.5 or OpenADR), and equitable compensation mechanisms for EV owners. The authors also suggest incorporating additional flexible resources in future iterations, such as distributed battery storage systems and controllable residential loads (e.g., heat pumps, water heaters), to further enhance system flexibility.
From a policy perspective, this research underscores the need for regulatory frameworks that recognize EV aggregators as legitimate grid service providers. Currently, many markets restrict ancillary service participation to large, centralized generators. Updating these rules to accommodate distributed, aggregated resources would unlock immense value and accelerate the energy transition. Pilot programs in California, the UK, and Germany are already exploring such models, and the technical foundation laid by this study could inform their expansion.
The human element remains central. While the algorithms and control strategies are sophisticated, their success hinges on consumer acceptance. Will drivers allow their vehicles to be throttled or discharged during grid events? Evidence from early V2G trials suggests yes—provided users retain control over minimum charge levels, receive fair compensation, and experience no degradation in vehicle performance or battery life. The travel chain model inherently respects user behavior by only scheduling regulation during predictable parking windows, minimizing inconvenience.
In conclusion, this work represents a significant step toward a symbiotic relationship between transportation electrification and power grid modernization. By treating EVs not as passive loads but as intelligent, mobile grid assets, the proposed strategy turns a potential challenge—voltage instability from distributed renewables—into an opportunity for enhanced system efficiency and resilience. As cities worldwide strive to meet net-zero targets, such integrated, data-driven approaches will be essential to building the flexible, responsive, and sustainable energy infrastructure of the future.
Authors: Jialin Yu¹, Jinrui Guo², Xiaodong Tu¹, Weiliang Liu¹, Weidong Zhong¹
Affiliations:
¹ Jiaxing Power Supply Company, State Grid Zhejiang Electric Power Co., Ltd., Jiaxing 314033, China
² College of Automation & College of Artificial Intelligence, Nanjing University of Posts and Telecommunications, Nanjing 210023, China
Published in: Electric Drive, 2024, Vol. 54, No. 9
DOI: 10.19457/j.1001-2095.dqcd24472