Heatwave Charging: New Strategy Balances Grid and Driver Needs
As scorching summers become the new normal, the dual challenge of soaring electricity demand and the rapid proliferation of electric vehicles (EVs) is pushing urban power grids to their limits. The confluence of air conditioning loads and unpredictable EV charging patterns can create dangerous “peak-on-peak” scenarios, threatening grid stability and leaving drivers searching for an elusive charge. A groundbreaking study from Zhejiang University offers a sophisticated solution, proposing a novel spatial guidance strategy that leverages user psychology and dynamic pricing to turn a potential problem into a powerful grid asset.
The research, published in the prestigious Automation of Electric Power Systems, tackles the core issue: the chaotic and often counterproductive nature of EV charging during extreme heat. Traditional approaches to managing this load have been blunt instruments—either restricting charging times, which frustrates users, or relying on simple time-of-use pricing that fails to account for the complex interplay of human behavior, vehicle energy consumption, and the physical realities of city traffic. The team, led by Han Linyang, a graduate researcher, and Dr. Ye Chengjin, a principal investigator at Zhejiang University, argues that a more nuanced, user-centric model is essential for a resilient and efficient energy future.
Their work begins by acknowledging a fundamental truth often overlooked in technical models: people are not robots. The decision to charge an EV is not made in a vacuum but is deeply intertwined with daily life, travel plans, and personal comfort. The paper introduces a sophisticated framework built on the Theory of Planned Behavior, which posits that a user’s intention to perform a behavior—like charging their car—is shaped by their attitude toward the behavior, their perceived social norms, and their perceived control over the action. In the context of a heatwave, this theory illuminates how external factors dramatically alter user willingness. The discomfort of high temperatures makes non-essential travel less appealing, while the desire for a cool, air-conditioned vehicle makes EV ownership more attractive. This shift in behavior, in turn, changes travel patterns—perhaps leading to earlier commutes to avoid the midday sun—which then dictates when and where a vehicle is available to charge.
To capture this dynamic, the researchers constructed a Markov chain-based travel model. Unlike older methods that rely on rigid “trip chain” assumptions, this model treats a driver’s location at any given moment as a probabilistic state—home, work, or commercial area—with the likelihood of moving to another state depending on the current time and conditions. This elegant approach allows for the simulation of highly diverse and realistic travel trajectories without predefining every possible journey. By analyzing historical travel data from the 2017 National Household Travel Survey (NHTS) alongside meteorological records, the team was able to quantify how user willingness changes across different temperature bands, from comfortable to “scorching hot.” Their analysis revealed distinct patterns: on weekdays, average travel intensity increases as people rely more on their climate-controlled cars for essential commutes, while on holidays, it decreases as people avoid unnecessary outings. This granular understanding of human behavior forms the bedrock of their entire simulation.
The second critical innovation lies in their method for calculating vehicle energy consumption. Conventional models often use a simple “kWh per mile” metric, adjusting it with a correction factor for temperature. This approach, the researchers argue, is fundamentally flawed because it fails to account for the fact that a significant portion of an EV’s energy drain in hot weather comes not from driving, but from running the air conditioning and cooling systems. These systems consume power based on how long they are running, not on the distance traveled. A car idling in traffic on a 40°C day with the AC on full blast will drain its battery far faster than a car covering the same distance on a cool, clear morning.
To address this, the team developed a three-pronged energy calculation model. The first component is the energy used for propulsion, which is influenced by speed, distance, and traffic congestion. The second is the energy for climate control, which is a function of the outdoor temperature and the duration of the trip. The third is the energy for low-voltage accessories like fans and pumps, which also increases in high temperatures. By integrating travel duration and distance as primary inputs, this model provides a far more accurate picture of an EV’s true energy needs, especially during the stressful conditions of a heatwave. Their simulations confirmed that vehicle energy consumption spikes dramatically during hot weather, particularly during peak traffic hours when both congestion and AC usage are high, leading to a significant increase in the frequency and volume of charging required.
With a realistic model of where people go and how much energy they use, the final piece of the puzzle is understanding how they choose to charge. The researchers moved beyond the simplistic rule of “charge when battery is low.” Instead, they propose a “rational choice model” where drivers are seen as economic agents making complex decisions to minimize their total cost of ownership. This cost is not just the price of electricity; it also includes the cost of battery degradation from deep discharges or overcharging, and what they term “trouble cost”—a quantifiable measure of the inconvenience of charging at a particular location or time. A driver might prefer to pay a slightly higher rate to charge at a convenient, secure spot near their home rather than a cheaper but remote or less safe public station.
This holistic view of user decision-making allows the model to predict not just if a car will charge, but when and where. It explains why, in their simulations, charging loads shift from the evening hours in residential areas to the daytime in workplaces and commercial districts when the model accounts for user preferences for convenience and cost. This level of detail is crucial for grid operators, who need to know not just the total load, but its precise location on the network.
Armed with this highly accurate simulation tool, the researchers then turned to the core of their work: the development of a “spatial guidance strategy.” Their solution is an elegant, market-based mechanism they call “low-valley charging discounts.” The concept is simple in principle but sophisticated in execution. Each day, the distribution network operator would use their simulation model to predict the next day’s load profile. By analyzing the node marginal prices—the cost of supplying an additional unit of power at each point in the grid—they can identify specific times and locations where there is an abundance of electricity, often during the afternoon when solar generation is high but overall demand is lower.
The operator then offers a targeted discount on the charging price at these “low-valley” nodes and times. This information is communicated to drivers in advance, perhaps through a dedicated app or their vehicle’s navigation system. Faced with a financial incentive, drivers are encouraged to adjust their charging plans. A driver who would normally charge at home in the evening might instead choose to charge at their workplace during the afternoon, or seek out a discounted public charger on their way home. This is not a command; it is a nudge, leveraging the driver’s own economic rationality to achieve a system-wide benefit.
The brilliance of this strategy is that it solves multiple problems simultaneously. It reduces the peak load on the grid by shifting demand away from critical evening hours, thereby lowering the risk of blackouts and the need for expensive and polluting “peaker” power plants. It increases the utilization of renewable energy by creating demand during periods of high solar output. And crucially, it does so without inconveniencing users; in fact, it saves them money. The study’s simulation results are compelling. When applied to a model city during a “scorching hot” scenario, the implementation of low-valley charging discounts reduced the overall load peak-to-valley difference by a remarkable 38.1% and cut the need for emergency load shedding by 75.7%. The grid’s operating costs decreased, leading to a net increase in system revenue. For the drivers, the total cost of vehicle ownership—the sum of charging fees, battery wear, and inconvenience—dropped by 5.57%, a significant saving for a large fleet of vehicles.
This research represents a significant leap forward in the field of vehicle-to-grid integration. It moves beyond the technical silos of power engineering and transportation modeling to create a truly integrated, human-centered approach. By recognizing that drivers are rational actors with complex needs and preferences, the strategy transforms EVs from a passive, unpredictable load into an active, responsive, and valuable resource for grid stability. It offers a blueprint for a future where the energy system is not just smart, but also empathetic, working in harmony with the people it serves.
The implications of this work extend far beyond the immediate challenge of summer heatwaves. The same principles could be applied to manage charging during other extreme weather events, to integrate higher levels of wind and solar power, or to optimize the use of local energy storage. It provides a powerful tool for city planners and utility companies as they navigate the complex transition to a sustainable, electrified transportation system. As the number of EVs on the road continues to grow exponentially, strategies like this one will be essential for ensuring that the promise of clean transportation does not come at the cost of a fragile and unreliable power grid. The work of Han Linyang, Ye Chengjin, Zhu Chao, Gao Qiang, and Yu Haiyue demonstrates that with the right combination of advanced modeling and behavioral insight, we can build an energy system that is not only robust and efficient but also fair and user-friendly.
Han Linyang, Ye Chengjin, Zhu Chao, Gao Qiang, Yu Haiyue, Zhejiang University. Automation of Electric Power Systems. DOI: 10.7500/AEPS20230731005