Virtual Power Plants and EV Fleets Team Up for Smarter Grids
The integration of electric vehicles (EVs) into power systems is no longer just about reducing tailpipe emissions. As the world accelerates toward decarbonization, the role of EVs is rapidly evolving from passive consumers of electricity to active participants in grid management. A groundbreaking new study reveals how virtual power plants (VPPs) can harness the collective power of EV fleets to not only stabilize the grid but also generate substantial economic value. This research, led by Xiao-Zhou Li and colleagues at the Key Laboratory of Power System Operation and Control in Shanxi, offers a comprehensive strategy for optimizing the participation of VPPs in both energy and ancillary service markets, with a particular focus on the synergistic relationship between VPPs and electric vehicle clusters (EVCs).
The concept of a virtual power plant is not new. By aggregating distributed energy resources (DERs)—such as rooftop solar panels, wind turbines, battery storage systems, and flexible loads—VPPs can act as a single, coordinated entity in electricity markets. They can buy and sell power, provide backup services, and help balance supply and demand in real time. However, the increasing penetration of renewable energy sources, which are inherently variable, has made grid stability more challenging than ever. At the same time, the growing number of EVs on the road presents a unique opportunity. These vehicles, when parked, represent a vast, distributed, and largely untapped reservoir of energy storage. The challenge has been how to effectively coordinate this resource in a way that benefits both the grid operator and the vehicle owner.
Li and his team at Taiyuan University of Technology have developed a sophisticated optimization framework that addresses this challenge head-on. Their approach is built on three key pillars: a robust model for assessing the dispatchable capacity of an EV fleet, a game-theoretic mechanism for aligning the interests of the VPP and the EV owners, and a risk-aware bidding strategy that accounts for the uncertainty of wind and solar generation and market prices. The result is a VPP that can not only participate in the main energy market but also compete in the lucrative ancillary services market, particularly frequency regulation.
The first step in this process is understanding what an EV fleet can actually do. Individual EVs have different charging needs, arrival and departure times, and battery capacities. Aggregating thousands of such vehicles into a single, predictable resource is a complex task. The researchers tackled this by developing a “dispatchable domain” assessment model for EVCs. This model defines the boundaries of what is possible for the fleet at any given time, considering both power and energy constraints. It takes into account the maximum charging and discharging power of each vehicle, the state of charge (SOC) when the vehicle arrives, the desired SOC when it leaves, and the time window during which it is connected to the grid. By using a mathematical technique known as the Minkowski sum, the model combines the dispatchable domains of all individual vehicles into a single, unified domain for the entire cluster. This significantly reduces the computational complexity of the problem, making it feasible to manage a large fleet in real time.
One of the most innovative aspects of the study is its use of a Stackelberg game to model the interaction between the VPP and the EVC. In this framework, the VPP acts as the “leader,” setting the price for charging and discharging, while the EVC acts as the “follower,” responding to those prices by optimizing its own charging schedule to minimize costs. This is a departure from traditional models where EV owners are simply given a fixed price or incentive. Instead, it creates a dynamic, two-way market. The VPP must set prices that are attractive enough to encourage EV owners to participate—especially during peak demand periods or when the grid needs additional power—but not so high that they erode the VPP’s own profits. This delicate balance is what the researchers call “interest equilibrium.” It ensures that both parties benefit, creating a win-win scenario that fosters long-term cooperation.
The implications of this are profound. For EV owners, it means they can earn money by allowing their vehicles to be used as a grid resource. For the VPP, it means access to a flexible, responsive, and cost-effective source of power. The study shows that this approach can significantly increase the overall revenue of the VPP compared to scenarios where EVs are not actively managed or where they are managed with fixed pricing. This is particularly important in the context of frequency regulation, a critical ancillary service that helps maintain the grid’s stability by quickly adjusting power output to match changes in demand. Battery energy storage systems (BESS) are ideal for this task due to their rapid response times, but they are expensive. By leveraging the batteries in EVs, VPPs can provide the same service at a lower cost, increasing their competitiveness in the market.
The second major contribution of the research is its focus on risk management. Electricity markets are inherently uncertain. Wind and solar generation can fluctuate dramatically, and market prices can swing wildly based on supply and demand. A VPP that bids too aggressively into the market based on optimistic forecasts can face significant financial losses if those forecasts are wrong. To address this, Li and his team incorporated Conditional Value at Risk (CVaR) into their optimization model. CVaR is a statistical measure that quantifies the expected loss in the worst-case scenarios, beyond a certain confidence level. By minimizing CVaR, the VPP can make more conservative, risk-averse decisions that protect its financial health.
The researchers demonstrated the effectiveness of this approach through a series of simulations based on real-world data from a provincial power market in China. They compared three different scenarios. In the first, the VPP only participates in the energy market, and EVs charge according to a fixed time-of-use tariff. In the second, the VPP still only participates in the energy market, but it uses the Stackelberg game to set dynamic prices for the EVC. In the third, the VPP participates in both the energy market and the frequency regulation market, again using the dynamic pricing model for the EVC.
The results were striking. The second scenario, with dynamic pricing, increased the VPP’s total revenue by 5.2% compared to the first, even though its revenue from the energy market was slightly lower. This was because the lower charging prices incentivized more EV owners to participate, leading to better load-shaping and higher real-time market gains. The third scenario, which included frequency regulation, delivered the highest overall benefit, increasing total revenue by 7.4% compared to the second scenario. This demonstrates that the additional revenue from providing ancillary services far outweighs the marginal increase in operational costs. The study also found that the VPP’s ability to track its day-ahead bidding plan in real time was excellent, with minimal deviations, thanks to the accurate forecasting of the EVC’s dispatchable domain.
The success of this strategy hinges on several key factors. First, it requires a high degree of coordination and communication between the VPP and the EV owners. This necessitates robust IT infrastructure and secure data-sharing protocols. Second, it relies on accurate forecasting of both the availability of the EV fleet and the future state of the power system. The researchers used advanced machine learning techniques, including Long Short-Term Memory (LSTM) networks, to predict the EVC’s dispatchable domain based on historical data. Third, it depends on a supportive regulatory environment. The study is based on the rules of a specific provincial market in China that allows VPPs to participate in both energy and ancillary service markets. Not all markets have such progressive regulations, and policymakers will need to adapt to unlock the full potential of VPPs and EVCs.
The findings of this research have significant implications for the future of the power grid. As the share of renewable energy continues to grow, the need for flexible, responsive resources will become even more acute. EVs, with their massive collective battery capacity, are uniquely positioned to fill this gap. However, realizing this potential requires more than just technology; it requires smart market design and innovative business models. The work of Li and his team provides a blueprint for how this can be done. By treating EVs not as a problem to be managed, but as an asset to be leveraged, VPPs can play a central role in building a more resilient, efficient, and sustainable energy system.
Moreover, this approach has the potential to accelerate the adoption of EVs. If vehicle owners can earn a reliable income from their cars simply by plugging them in, the total cost of ownership will decrease, making EVs more attractive to a wider range of consumers. This could create a virtuous cycle: more EVs on the road lead to more grid flexibility, which in turn makes the grid more reliable and able to accommodate even more renewable energy. The study also highlights the importance of battery storage. While the focus is on using EV batteries, the principles apply equally to stationary battery systems. The integration of BESS into the VPP’s operations is crucial for providing frequency regulation services, and the research shows that the revenue from these services can justify the investment in storage technology.
Another important takeaway is the emphasis on risk management. The energy transition is not just a technological challenge; it is a financial one. Utilities, VPP operators, and investors need tools to navigate the uncertainties of a rapidly changing market. The use of CVaR in this model provides a rigorous, quantitative framework for making risk-informed decisions. It allows VPP operators to tailor their bidding strategies to their own risk tolerance, whether they are aggressive, seeking maximum returns, or conservative, prioritizing stability. This level of sophistication is essential for building investor confidence and attracting the capital needed to fund the grid of the future.
The study also underscores the importance of multi-agent collaboration. The VPP in this model doesn’t just manage EVs; it also coordinates with gas turbines, controllable loads, and photovoltaic systems. This holistic approach, which the researchers describe as a “source-grid-load-storage integration,” is key to maximizing the overall system efficiency. For example, the model shows how the VPP can shift controllable loads away from peak periods, reducing the need for expensive and polluting peaker plants. It can also use its gas turbines to provide a stable baseload of power, while relying on BESS and EVCs for rapid, short-term adjustments. This kind of intelligent coordination is what separates a true VPP from a simple aggregation of resources.
In conclusion, the research by Xiao-Zhou Li, Wen-Ping Qin, Xiang Jing, Zhi-Long Zhu, Rui-Peng Lu, and Xiao-Qing Han from the Key Laboratory of Power System Operation and Control at Taiyuan University of Technology represents a significant step forward in the field of smart grid management. Their comprehensive optimization strategy for VPPs, which integrates EV fleets, battery storage, and risk management, offers a practical and profitable pathway for integrating distributed energy resources into the power system. By demonstrating the economic and operational benefits of a coordinated, multi-market approach, this work provides valuable insights for grid operators, policymakers, and technology developers around the world. As the energy landscape continues to evolve, the lessons learned from this study will be essential for building a cleaner, more flexible, and more resilient grid.
Xiao-Zhou Li, Wen-Ping Qin, Xiang Jing, Zhi-Long Zhu, Rui-Peng Lu, Xiao-Qing Han, Key Laboratory of Power System Operation and Control, Taiyuan University of Technology, Power System Technology, DOI:10.13335/j.1000-3673.pst.2023.1309