EV Drivers Make Smarter Grid Choices with New Decision Model
As electric vehicles (EVs) continue their rapid ascent in global markets, their role is expanding far beyond personal transportation. A new study reveals how EV owners can become active, intelligent participants in power grid operations—specifically in frequency regulation markets—by making decisions that align not only with economic incentives but also with their psychological preferences.
The research, conducted by Wang Liwei, Wang Haotian, and Sun Yingyun from the School of Electrical and Electronic Engineering at North China Electric Power University, introduces a novel day-ahead decision model that leverages prospect theory to better reflect real-world user behavior. Published in Modern Electric Power, the paper presents a framework that helps EV users navigate the complexities of participating in frequency regulation ancillary services, a critical function for maintaining grid stability amid growing renewable energy integration.
With EV adoption accelerating worldwide, the potential for vehicle-to-grid (V2G) technology to support power systems is increasingly recognized. However, actual participation remains limited due to technical complexity, economic uncertainty, and behavioral barriers. Most existing models assume users act as perfectly rational agents, optimizing solely based on financial returns. But in reality, human decision-making is influenced by emotions, expectations, and risk perceptions—factors traditional models often overlook.
This new model addresses that gap by incorporating bounded rationality through the lens of prospect theory, a behavioral economics concept developed by Daniel Kahneman and Amos Tversky. Unlike classical utility theory, which assumes consistent risk preferences, prospect theory acknowledges that people evaluate outcomes relative to a reference point—typically the status quo—and are more sensitive to losses than gains. This “loss aversion” profoundly shapes choices under uncertainty, a condition inherent in electricity markets where prices, demand, and performance outcomes fluctuate unpredictably.
The researchers applied this psychological framework to the specific challenge of day-ahead bidding in frequency regulation markets. In such markets, participants commit capacity and submit price offers one day in advance, knowing that actual compensation depends on real-time performance and market clearing. For EV owners, this means balancing several competing priorities: maximizing income from providing grid services, minimizing battery degradation costs, ensuring sufficient charge for driving needs, and managing the risk of financial loss or service penalties.
Rather than treating these factors as purely technical variables, the model treats the user’s net financial gain and final battery state as psychological outcomes evaluated against a personal reference point. This reference point is derived from a baseline scenario—choosing not to participate in frequency regulation and simply charging at off-peak rates. By comparing potential outcomes to this familiar alternative, the model captures how users perceive gains and losses, even when the absolute financial difference is small.
One of the model’s key innovations is its integration of battery aging into the decision calculus. Frequent charging and discharging, especially at high power levels, accelerates battery wear, which translates into long-term replacement costs. The researchers developed a dynamic battery degradation model that estimates aging based on actual usage patterns during frequency regulation. This cost is then factored into the net benefit calculation, ensuring that short-term earnings are weighed against long-term asset value.
The decision model evaluates three simplified participation modes, making it accessible to non-expert users:
Mode 1: Passive charging only—no grid services provided. This serves as the behavioral reference point.
Mode 2: Bidirectional charging within the charging window—EVs adjust their charging rate up or down in response to grid signals, but do not discharge back to the grid.
Mode 3: Full V2G capability—EVs can reduce charging, increase charging, discharge to the grid, or reduce discharging, offering maximum flexibility and revenue potential.
Each mode offers different trade-offs between revenue potential, battery stress, and complexity. Mode 3, for instance, allows the largest reserve capacity and lowest bid prices, making it more likely to win market contracts. However, it also involves deeper cycling and higher degradation risk. Mode 2 offers moderate benefits with less battery impact, while Mode 1 guarantees no additional wear but forgoes potential income.
The core of the model is the “comprehensive prospect value,” a metric that combines the psychological value of financial outcomes and battery state, weighted by the user’s personal preferences. Users who prioritize immediate earnings will assign higher weight to net revenue, while those concerned about vehicle longevity or daily driving needs will emphasize battery state. The model identifies the participation mode that maximizes this combined prospect value, effectively recommending the choice that best aligns with the user’s individual mindset.
To validate the model, the team conducted simulations using real-world data from the PJM Interconnection frequency regulation market and pricing structures from Guangdong Province, China. They examined five distinct user scenarios, each representing different charging durations and target battery levels upon departure. The results revealed clear patterns in user preferences based on their underlying motivations.
In scenarios where users had lower charging needs and could afford to prioritize income, Mode 3 consistently delivered the highest prospect value. These users were willing to accept greater battery wear for the chance to earn more from grid services. Conversely, when charging time was limited or the desired departure state of charge was high, Mode 1 became optimal—preserving battery health and guaranteeing sufficient charge outweighed potential earnings.
Interestingly, an intermediate scenario emerged where Mode 2 was the preferred choice. This suggests a “sweet spot” for users who want to participate in grid services but are risk-averse or have moderate driving requirements. The model’s ability to identify this middle ground demonstrates its sensitivity to nuanced behavioral patterns.
The simulations also showed how user priorities shift the optimal strategy. When net revenue was given higher psychological weight, users gravitated toward Mode 3, even at the cost of faster battery degradation. When battery state was prioritized, Mode 2 or Mode 1 became preferable. This dynamic responsiveness makes the model highly adaptable to diverse user profiles, from cost-conscious commuters to fleet operators managing multiple vehicles.
Perhaps most importantly, the model’s recommendations align closely with observed human behavior. In real-world settings, many EV owners hesitate to engage in V2G programs due to fears of battery damage or inconvenience. The prospect theory-based approach validates these concerns as rational within a psychological framework, rather than dismissing them as irrational barriers to market participation.
By acknowledging and incorporating these behavioral realities, the model offers a more realistic pathway to scaling V2G adoption. Instead of pushing users toward theoretically optimal but psychologically unappealing choices, it guides them toward decisions they feel comfortable with—increasing the likelihood of sustained participation.
The implications extend beyond individual users. Aggregators—companies that pool hundreds or thousands of EVs to bid into wholesale markets—can use this model to design more effective customer engagement strategies. By understanding the psychological drivers behind participation, they can tailor offerings, set realistic expectations, and build trust with users. For example, an aggregator might offer a “conservative” plan focused on battery preservation and moderate earnings, alongside an “aggressive” plan for users seeking maximum revenue, even if it means faster battery wear.
Grid operators also stand to benefit. Wider EV participation in frequency regulation can enhance grid flexibility, reduce reliance on fossil-fuel peaker plants, and lower overall system costs. However, this requires predictable and reliable behavior from distributed resources. Models that account for human psychology can improve forecasting accuracy, helping system operators plan more effectively and integrate EVs as dependable assets.
Policy makers can use insights from this research to design better incentives. Current programs often focus on direct payments or time-of-use pricing, but may fail to address deeper behavioral barriers. Understanding that users evaluate outcomes relative to a reference point suggests that framing matters—presenting V2G income as a “bonus” rather than a risky investment could increase uptake. Similarly, transparent communication about battery degradation costs can help users make informed trade-offs, reducing post-participation regret.
The study also highlights the importance of user education. Many EV owners are unaware of the technical and economic mechanisms behind grid services. Providing clear, personalized feedback—such as projected earnings, battery wear estimates, and comparison to the “do nothing” scenario—can empower users to make better decisions. The model’s structure lends itself to integration into user-facing apps or in-vehicle displays, turning complex market dynamics into intuitive recommendations.
While the model is grounded in Chinese market conditions, its principles are universally applicable. As frequency regulation markets evolve in Europe, North America, and elsewhere, the need for human-centric decision tools will only grow. The increasing complexity of electricity markets—with multiple service types, dynamic pricing, and performance-based compensation—demands approaches that go beyond pure optimization.
The researchers emphasize that their model is not intended to replace technical analysis, but to complement it. Engineers and economists will still need to model physical constraints, market rules, and cost structures. But when it comes to translating those factors into user behavior, psychology must play a central role.
Looking ahead, the team suggests several directions for future work. One is to incorporate real-time learning, allowing the model to adapt to individual user patterns over time. Another is to expand the reference point framework to include social and environmental motivations—such as reducing carbon emissions or supporting renewable energy—which may influence decisions for some users.
Integration with smart charging infrastructure is another promising avenue. As EVs become more connected, the model could be embedded in charging stations or home energy management systems, automatically adjusting bids based on user preferences, current grid conditions, and battery health.
The study also opens questions about equity and access. Will sophisticated decision models primarily benefit tech-savvy, affluent users, leaving others behind? Designing inclusive interfaces and ensuring transparency will be critical to avoiding a “digital divide” in V2G participation.
Despite these challenges, the research represents a significant step forward in bridging the gap between power system engineering and human behavior. By treating users not as passive assets but as active, psychologically complex agents, the model paves the way for a more resilient, flexible, and user-friendly grid.
As the energy transition accelerates, the role of distributed resources like EVs will only grow. Success will depend not just on technology, but on understanding the people behind the devices. This new decision model shows that when we design systems with human nature in mind, everyone—from drivers to grid operators—can benefit.
The findings underscore a fundamental truth: the future of energy is not just about smarter grids, but about smarter choices. And those choices begin with recognizing that people don’t just respond to prices—they respond to perceptions, fears, hopes, and habits. By building models that reflect that reality, we move closer to a truly sustainable and inclusive energy system.
Wang Liwei, Wang Haotian, Sun Yingyun, North China Electric Power University, Modern Electric Power, DOI: 10.19725/j.cnki.1007-2322.2022.0282