New Method Boosts Accuracy of Grid Capacity for EV Charging
As electric vehicle (EV) adoption accelerates worldwide, power distribution networks face growing pressure to accommodate surging charging demand. Traditional methods for assessing grid capacity often fall short, leading to overly conservative estimates that hinder infrastructure planning and investment. A breakthrough study published in the Journal of Global Energy Interconnection introduces a novel evaluation framework that significantly improves the accuracy of available capacity assessments by accounting for the spatiotemporal correlation between EV charging loads and conventional base loads.
The research, led by Zhao Zijun and a team of engineers from State Grid Hunan Electric Power Co., Ltd., Changsha Power Supply Branch, addresses a critical gap in current grid planning methodologies. Conventional approaches typically rely on historical peak load data without considering the complex interplay between different types of electrical demand across time and space. This oversight can result in underestimating the actual flexibility and resilience of distribution networks, ultimately slowing down the deployment of EV charging infrastructure.
The new model developed by Zhao and colleagues integrates advanced optimization techniques with a refined understanding of load behavior. By recognizing that EV charging patterns are not random but exhibit predictable correlations with household and commercial electricity usage, the researchers have created a more realistic simulation of real-world grid conditions. This approach allows planners to identify opportunities for higher capacity utilization without compromising system reliability.
At the heart of the methodology is a two-stage process. First, the team devised an improved technique for selecting representative “typical days” from historical load data. Unlike conventional methods that treat the entire network as a single entity, this new approach analyzes each node individually, capturing local variations in consumption patterns. The result is a more accurate representation of how loads aggregate across the network throughout the day.
Second, the researchers constructed a robust stochastic bilevel optimization model to evaluate the distribution network’s available capacity—the maximum additional load it can safely support. This model accounts for uncertainties in renewable energy generation, particularly photovoltaic (PV) output, which fluctuates based on weather conditions. By defining a “box-shaped” uncertainty set for PV production and introducing adjustable conservatism parameters, the method ensures that capacity estimates remain reliable even under adverse conditions.
One of the key innovations of the study is its treatment of load correlation. Previous models often assumed independence between EV charging and other forms of electricity demand, leading to worst-case scenario assumptions that unnecessarily limit capacity. In contrast, Zhao’s model quantifies the degree to which EV charging coincides with or offsets base loads, enabling a more nuanced assessment. For instance, if many EV owners charge their vehicles during off-peak hours when household demand is low, the overall stress on the grid is reduced, allowing for greater total capacity.
To validate their approach, the team applied it to a modified IEEE 33-bus test system—a standard benchmark in power systems research. The network was enhanced with realistic configurations, including EV charging stations at nodes 14, 22, and 25, and PV installations at nodes 10, 25, and 26. Historical load data came from the PJM Interconnection, a major U.S. grid operator, while solar irradiance records were sourced from the DKA Solar Centre in Australia. Real-world EV charging data from a station in Hunan Province, China, provided granular insights into actual user behavior, sampled at 15-minute intervals over a full year.
The results demonstrated clear advantages over traditional assessment techniques. When compared to methods that ignore spatiotemporal load correlations, the proposed model yielded higher available capacity values—62.92 MW versus 60.00 MW over a 24-hour period—while maintaining safety margins. More importantly, the new method produced capacity estimates that closely followed the inverse pattern of total load, forming what the authors describe as a “duck curve.” This means that during periods of low demand, such as late at night, the grid can accept more new loads, whereas during peak hours, available capacity naturally decreases. This dynamic responsiveness reflects real operational conditions far better than static, one-size-fits-all estimates.
Further analysis revealed that the improved typical day selection process led to higher data density and smaller radiation radius—two metrics used to evaluate representativeness. In practical terms, this means the selected load profiles are closer to the most frequently observed real-world scenarios, reducing the risk of basing planning decisions on outlier events. The probability density function of distances between original data points and the chosen typical day showed a tighter clustering around zero for the new method, indicating superior fit and reliability.
Another significant finding relates to power losses within the network. The study found that when load correlations are properly accounted for, the resulting capacity estimates lead to lower overall network losses during off-peak hours and higher losses during peak times—mirroring the natural variation in system efficiency. This trade-off between maximizing capacity and minimizing losses is inherent in grid operation, and the new model provides a balanced way to navigate it. By incorporating loss minimization as part of the objective function, the optimization seeks not just raw capacity but also efficient utilization of existing assets.
The implications of this research extend beyond academic interest. For utility companies, regulators, and city planners, having a more accurate tool for assessing grid capacity means faster and more confident decision-making regarding EV infrastructure expansion. It reduces the need for costly upgrades by showing where existing lines and transformers can handle additional load through smarter management. Moreover, it supports the integration of distributed energy resources like rooftop solar, which can offset demand and further enhance local capacity.
From a policy perspective, the study underscores the importance of data-driven planning in the transition to sustainable transportation. As governments set ambitious targets for EV adoption, ensuring that the supporting electrical infrastructure keeps pace is essential. Overly cautious assessments could delay the rollout of public chargers, discourage private investment, and ultimately slow the shift away from fossil fuels. Conversely, overly optimistic estimates risk overloading circuits and causing service disruptions. The method proposed by Zhao and his team strikes a balance, offering a scientifically grounded approach that enhances both safety and efficiency.
The research also highlights the value of interdisciplinary collaboration. Combining expertise in power systems engineering, statistical modeling, and optimization theory, the team was able to develop a solution that addresses a complex, real-world challenge. Their use of KKT (Karush-Kuhn-Tucker) conditions to transform a complex bilevel optimization problem into a solvable single-layer model exemplifies how theoretical advances can be translated into practical tools for industry.
While the current study focuses on normal operating conditions, the authors acknowledge limitations and suggest directions for future work. Specifically, they note that the model does not yet incorporate N-1 security criteria—the requirement that the grid remain stable even if a single component fails. Extending the framework to include such contingencies would make it even more robust and applicable to real-world reliability standards. Additionally, integrating demand response mechanisms and energy storage systems could further refine capacity estimates by modeling active load management strategies.
Nonetheless, the contribution of this work is substantial. It moves the field beyond simplistic, conservative rules of thumb toward a more sophisticated, dynamic understanding of grid capabilities. By treating EV charging not as an isolated burden but as part of a broader, interconnected load ecosystem, the researchers offer a vision of a more flexible, resilient, and future-ready power system.
For stakeholders across the energy and automotive sectors, the message is clear: accurate capacity assessment is not just a technical detail—it is a foundational element of the clean transportation revolution. With millions of new EVs expected to hit the roads in the coming years, the ability to efficiently utilize existing grid infrastructure will be crucial. Tools like the one developed by Zhao Zijun and his team provide the analytical rigor needed to ensure that the electricity grid evolves in step with the vehicles it powers.
As cities redesign urban landscapes to accommodate electric mobility, planners must rely on models that reflect reality, not worst-case assumptions. This study delivers just that—a method grounded in empirical data, sensitive to temporal and spatial dynamics, and capable of guiding smart, cost-effective investments. Whether siting fast-charging hubs along highways or enabling apartment dwellers to charge overnight, accurate capacity evaluation enables smarter decisions today for a more sustainable tomorrow.
The research also serves as a reminder that technological progress in one domain often depends on advances in another. While much attention focuses on battery chemistry, vehicle design, and charging speed, the unseen backbone of the EV revolution—the power grid—requires equal innovation. Without improvements in how we model, monitor, and manage electricity distribution, the full potential of electric transportation cannot be realized.
In this context, the work of Zhao, Hu, Deng, Li, Lu, Peng, and Yang represents a quiet but vital advancement. It may not capture headlines like a new EV launch, but its impact could be just as profound. By enabling utilities to say “yes” to more connections with confidence, it removes a key barrier to widespread EV adoption. And in doing so, it helps accelerate the transition to a cleaner, quieter, and more sustainable transportation future.
What sets this study apart is not only its technical sophistication but also its practical relevance. The authors are not academic theorists but engineers working within a real utility, confronting the same challenges faced by grid operators around the world. Their solution is therefore not a hypothetical construct but a tool designed for implementation, tested against real data, and validated in a realistic simulation environment.
Moreover, the emphasis on spatiotemporal correlation reflects a deeper understanding of human behavior. EV charging is not purely a function of vehicle ownership; it is shaped by daily routines, work schedules, tariff structures, and social habits. By embedding these realities into the model, the researchers have created a more human-centered approach to grid planning—one that recognizes that electricity demand is not just a technical variable but a reflection of how people live, work, and travel.
Looking ahead, the methodology could be adapted to other types of flexible loads, such as heat pumps, smart appliances, and industrial processes. As the grid becomes increasingly dynamic, the ability to assess capacity in a way that accounts for the timing and location of diverse demand sources will become ever more valuable. The principles established in this paper—correlation-aware modeling, robust optimization, and data-driven scenario selection—could form the foundation of next-generation grid planning tools.
In conclusion, the study published in the Journal of Global Energy Interconnection offers a timely and impactful contribution to the field of power systems engineering. It demonstrates that by rethinking traditional assumptions and leveraging modern computational techniques, it is possible to unlock hidden capacity within existing infrastructure. For the electric vehicle revolution to succeed, such innovations are not just helpful—they are essential.
Zhao Zijun, Hu Xiangwei, Deng Yazhi, Li Yongjian, Lu Xinxing, Peng Qingwen, Yang Xiaodan, State Grid Hunan Electric Power Co., Ltd., Changsha Power Supply Branch, Journal of Global Energy Interconnection, DOI: 10.19705/j.cnki.issn2096-5125.2024.03.005