EV Aggregators Embrace Smart Bidding to Balance Grid and Profits
As the world accelerates toward electrification, electric vehicles (EVs) are no longer just a means of transportation—they are emerging as dynamic energy assets capable of reshaping the power grid. With over 18 million EVs on China’s roads by late 2023 and growing, the integration of these mobile batteries into the electricity system has become both an opportunity and a challenge. Uncoordinated charging, especially during peak hours, risks overloading the grid, increasing operational costs, and undermining the stability of the power network. However, a new research breakthrough from Wuhan University of Technology offers a promising solution: a multi-agent day-ahead bidding strategy that transforms EV aggregators into intelligent grid partners.
Led by Associate Professor Hou Hui and her team, including postgraduate researcher He Ziyin, the study introduces a sophisticated framework that enables EV aggregators to participate more effectively in the electricity market. Published in the Journal of Global Energy Interconnection, the research presents a novel approach that combines user behavior modeling, demand response potential assessment, and profit-driven bidding strategies to achieve a triple win—grid stability, economic efficiency, and sustainable market development.
The core of the innovation lies in its recognition that not all EV users are the same. Instead of treating the entire fleet as a uniform load, the team developed a clustering methodology based on real-world driving patterns. By analyzing factors such as arrival and departure times, initial state of charge (SOC), and parking duration, they categorized EV users into five distinct clusters. This segmentation allows aggregators to better predict when and how much energy each group can contribute to or draw from the grid.
For instance, one cluster consists of users who park overnight at home, typically arriving in the evening and leaving in the morning. These vehicles have long dwell times, making them ideal candidates for flexible charging and even discharging during peak demand periods. Another cluster includes daytime parkers—such as office workers—who arrive in the morning and leave in the afternoon. Their availability window is shorter, but still valuable for midday load management. A third group comprises commercial fleet operators, whose vehicles may return multiple times a day, offering frequent but shorter charging opportunities. Two additional clusters represent users with irregular schedules and those who charge opportunistically throughout the day.
By modeling these behaviors, the researchers were able to simulate how individual EV owners make charging decisions based on convenience and cost. The model assumes that users sign contracts with aggregators, agreeing to allow controlled charging in exchange for financial incentives or lower electricity rates. This contractual relationship is crucial—it transforms unpredictable consumer behavior into a manageable and predictable resource.
Once the clusters are defined, the next step is assessing their demand response potential. This refers to the amount of energy that can be shifted in time without compromising the user’s mobility needs. The team evaluated this potential by considering two key constraints: the maximum charging power available at each location and the allowable energy window for each vehicle. For example, a car arriving with a 30% SOC and needing to reach 80% before departure has a fixed energy requirement. However, within that range, the aggregator can decide when and how fast to charge, depending on market conditions.
The researchers also incorporated battery degradation and charging efficiency into their model, ensuring that the proposed strategies do not come at the expense of vehicle longevity. By setting upper and lower limits on battery state of charge—typically between 15% and 90%—they prevent deep discharges and overcharging, which can accelerate battery wear. This attention to technical detail enhances the practicality of the approach, making it more likely to be adopted by real-world operators.
With demand response potential quantified, the focus shifts to market participation. The study proposes a day-ahead bidding model in which multiple aggregators compete to offer charging and discharging services to the grid operator. Unlike traditional approaches that prioritize user satisfaction or technical feasibility, this model centers on profitability. The objective is to maximize net profit—the difference between revenue from selling electricity to users and the cost of purchasing it from the wholesale market.
To achieve this, the model accounts for several critical factors. First, it considers the uncertainty of electricity prices. Day-ahead market prices are not fixed; they are determined through an auction process and can fluctuate significantly based on supply and demand. The model uses historical data to simulate various price scenarios, allowing aggregators to hedge against volatility. Second, it incorporates penalties for imbalances. If an aggregator commits to deliver a certain amount of energy but fails to do so—due to unexpected user behavior or forecasting errors—it faces financial penalties. The model includes a tolerance threshold (set at 10% in the study) beyond which penalties apply, encouraging accurate forecasting and conservative bidding.
Another innovative aspect of the model is its time resolution. While most studies use hourly intervals, this research divides each hour into four 15-minute segments, resulting in 96 time slots per day. This finer granularity allows for more precise scheduling and better alignment with real-time grid operations. It also enables aggregators to exploit short-term price differences—buying low during off-peak hours and selling high during peak periods—a practice known as peak shaving and valley filling.
The optimization problem is solved using Gurobi, a powerful commercial solver capable of handling complex mixed-integer linear programming tasks. By integrating this tool with MATLAB and YALMIP, the researchers created a robust computational framework that can process large datasets and generate optimal bidding strategies in a reasonable time frame.
To validate their approach, the team conducted a case study using real-world data from a district in Wuhan, Hubei Province. The simulation involved 3,000 EVs and 5,000 residential consumers, with two types of aggregators: one specializing in slow charging (6.6 kW) and the other in fast charging (22 kW). The results were compelling. Under the proposed bidding strategy, the overall load profile showed a significant reduction in peak demand and an increase in off-peak consumption. This flattening of the load curve translates directly into reduced stress on the grid, lower transmission losses, and deferred infrastructure investments.
Moreover, the economic analysis revealed substantial profits for both aggregators. The slow-charging aggregator (EVA1) achieved a net profit of over 11,600 yuan per day, while the fast-charging aggregator (EVA2) earned nearly 7,350 yuan. These figures demonstrate that demand response is not just a technical or environmental benefit—it is also a viable business model.
Interestingly, the study found that slow chargers played a more dominant role in load shifting, primarily because their users tended to park for longer durations and were more responsive to price signals. Fast chargers, while essential for convenience, were used more during peak hours, limiting their flexibility. As a result, the slow-charging aggregator submitted lower bids during peak times and higher bids during off-peak hours, effectively incentivizing users to charge when it was most beneficial for the grid.
This price-based incentive mechanism is a key innovation. Rather than relying solely on direct control or rigid schedules, the model uses market signals to influence user behavior. When electricity prices are low, users are encouraged to charge; when prices are high, they are discouraged. This creates a self-regulating system that aligns individual interests with collective goals.
The implications of this research extend beyond China. As countries around the world push for deeper decarbonization, the role of EVs in energy systems will only grow. In Europe, for example, the European Union’s “Fit for 55” package includes provisions for vehicle-to-grid (V2G) integration. In the United States, the Biden administration has set a goal of 50% EV sales by 2030, accompanied by major investments in charging infrastructure. In both regions, aggregators will play a critical role in managing the influx of EVs.
However, several challenges remain. One is market access. In many jurisdictions, EV aggregators are still not recognized as formal market participants, limiting their ability to bid in wholesale markets or provide ancillary services. Regulatory frameworks need to evolve to accommodate these new actors. Another challenge is user trust. For the model to work, users must be willing to cede some control over their charging process. Transparent contracts, fair compensation, and reliable performance are essential to building that trust.
Technology readiness is also a factor. While the study assumes a high level of connectivity and control, not all EVs and charging stations are equipped with the necessary communication and control systems. Standards such as ISO 15118 and OpenADR are helping to bridge this gap, but widespread deployment will take time.
Despite these hurdles, the trajectory is clear: EVs are becoming active participants in the energy ecosystem. The work by Hou Hui, He Ziyin, and their colleagues provides a blueprint for how this transition can be managed in a way that benefits everyone—utilities, aggregators, consumers, and the environment.
The study also opens the door to future research. The authors suggest exploring real-time bidding strategies, multi-market participation (such as combining energy and reserve markets), and more nuanced models of user preferences. They also highlight the importance of platform-based aggregation, where multiple aggregators compete on a shared digital marketplace, fostering innovation and efficiency.
In conclusion, the integration of EVs into the power grid is no longer a question of if, but how. This research demonstrates that with the right strategies, EV aggregators can become powerful tools for grid optimization, turning a potential liability into a valuable asset. As the energy transition accelerates, innovations like this will be essential to building a smarter, cleaner, and more resilient power system.
Hou Hui, He Ziyin et al., Journal of Global Energy Interconnection, DOI: 10.19705/j.cnki.issn2096-5125.2024.02.011