Electric Vehicles’ Charging Behavior Reshaping Industrial Park Energy Systems
As electric vehicles (EVs) become increasingly common on city streets and highways, their impact extends beyond transportation. A new study published in Electric Power Engineering Technology explores how the charging and discharging behavior of EVs in industrial parks can be optimized to improve energy efficiency, reduce costs, and enhance grid stability. The research, led by Feng Yemu from the School of Electric Power Engineering at Nanjing Institute of Technology, introduces a novel two-layer optimization scheduling model that takes into account drivers’ willingness to charge and discharge their vehicles.
The proliferation of electric vehicles is one of the most significant transformations in the automotive industry in recent decades. According to data cited in the study, China had 13.1 million EVs on its roads by the end of 2022, supported by 5.21 million charging infrastructure units. Projections suggest that global EV numbers could reach 220 million by 2030. While this shift toward electrification is crucial for reducing carbon emissions and combating climate change, it also presents new challenges for power systems, particularly in concentrated areas like industrial parks where hundreds or thousands of employees may plug in their vehicles during the workday.
Uncontrolled charging of EVs can lead to significant spikes in electricity demand, exacerbating peak load issues and increasing strain on the grid. This phenomenon, known as “peak load amplification,” can result in higher energy costs, reduced system reliability, and the need for costly infrastructure upgrades. To address these challenges, researchers have been exploring ways to manage EV charging through smart scheduling, demand response programs, and vehicle-to-grid (V2G) technologies that allow EVs to not only draw power from the grid but also feed it back during periods of high demand.
The study by Feng Yemu and colleagues offers a sophisticated approach to this problem by developing a comprehensive model that considers multiple factors influencing driver behavior. At the heart of their methodology is the concept of “charging and discharging willingness,” which recognizes that EV owners make decisions based on more than just the immediate need to power their vehicles. Factors such as electricity pricing, battery state of charge (SOC), potential battery degradation from frequent cycling, and financial incentives for participating in V2G programs all play a role in determining whether a driver will charge, discharge, or do nothing at any given time.
One of the key innovations in this research is the use of a dynamic real-time pricing mechanism that goes beyond traditional time-of-use tariffs. While conventional peak and off-peak pricing schemes divide the day into broad periods with fixed rates, the dynamic pricing model proposed in this study adjusts electricity prices more frequently and responsively based on actual grid conditions. Specifically, the price for the next time period is determined by the ratio of current load to the daily average load, creating a feedback loop that encourages load shifting. When demand is high relative to the average, prices increase, discouraging charging and encouraging discharging. Conversely, when demand is low, prices decrease, incentivizing charging activity. This approach aims to smooth out load curves and prevent the formation of new charging peaks that could undermine the benefits of time-based pricing.
However, the researchers recognize that price signals alone are insufficient to achieve optimal outcomes. Battery degradation is a major concern for EV owners, as each charge-discharge cycle contributes to wear and tear on the battery pack, potentially reducing its lifespan and resale value. To address this, the model incorporates a battery loss compensation cost that quantifies the financial impact of additional cycling due to V2G participation. This cost is factored into the overall calculation of charging and discharging willingness, ensuring that drivers are not penalized for contributing to grid stability. Additionally, the model includes an extra participation incentive, set at 30% of the electricity price, to further encourage engagement in V2G programs.
The integration of these various factors into a single framework allows for a more realistic simulation of driver behavior. For example, an EV owner arriving at work with a battery SOC of 40% might be willing to charge if the price is low and no immediate trip is planned, but unwilling to discharge even at a high price if the battery level is already below a comfortable threshold. Similarly, an owner with a nearly full battery might be more open to discharging during peak hours, especially if the financial incentive is attractive and the compensation for battery wear is adequate.
To operationalize this complex set of interactions, the researchers employ a two-layer optimization structure. The outer layer represents the perspective of the park integrated energy system (PIES), which seeks to minimize its total operating cost. This includes expenses related to purchasing electricity, natural gas, and thermal energy from external sources, as well as the costs of operating and maintaining on-site generation and storage equipment. The inner layer represents the individual EV user, whose objective is to minimize their personal charging costs. These two objectives are not always aligned, as actions that benefit the system (such as discharging during peak hours) may come at a cost to the individual (in terms of battery wear or inconvenience).
The brilliance of the two-layer approach lies in its ability to reconcile these competing interests through a mathematical technique known as the Karush-Kuhn-Tucker (KKT) conditions. By transforming the inner optimization problem (the user’s cost minimization) into a set of constraints for the outer problem (the system’s cost minimization), the researchers can solve the entire system as a single, unified model. This not only simplifies the computational process but also ensures that the solution is stable and converges to an optimal point where both system and user objectives are satisfied as much as possible.
To validate their model, the team conducted a series of simulations comparing three different scenarios. The first scenario represents a baseline case of uncoordinated charging, where EVs begin charging immediately upon arrival at the park and continue until fully charged, regardless of price or system conditions. The second scenario uses a traditional charging and discharging willingness model that considers only price and SOC, without accounting for battery loss compensation or additional incentives. The third and final scenario implements the full proposed model with all its components.
The results of these simulations are striking. In the uncoordinated charging scenario, EV load peaks at 12.9 MW between 7:00 and 11:00, coinciding with the morning arrival of employees and significantly increasing the total electrical load on the system. This creates a substantial burden on the grid and forces the PIES to rely more heavily on expensive peak-period electricity purchases or on-site generation.
In contrast, the traditional willingness model reduces the morning charging peak by 73%, shifting much of the load to off-peak hours when electricity is cheaper and more abundant. Furthermore, it enables some EVs to discharge during peak periods, providing up to 2.7 MW of power back to the grid and helping to alleviate supply pressure. However, the most impressive results come from the full proposed model. By incorporating battery loss compensation and participation incentives, the model achieves even greater load flexibility while also reducing costs for both the system and individual users.
According to the study’s findings, the total operating cost of the PIES decreases by 11.48% when moving from uncoordinated charging to the traditional willingness model, and by an additional 4.03% when adopting the full proposed model. Similarly, the average charging cost for EV users drops by 19.16% in the traditional model and by a further 15.02% in the enhanced model. These savings are achieved through a combination of reduced energy purchases during peak periods, increased utilization of renewable energy sources like wind and solar, and more efficient operation of on-site generation and storage assets.
The implications of this research extend far beyond the confines of a single industrial park. As EV adoption continues to grow, the principles and methodologies developed in this study could be applied to a wide range of settings, including commercial complexes, residential communities, and public charging networks. The success of such applications will depend on several factors, including the availability of smart charging infrastructure, the willingness of utilities and grid operators to implement dynamic pricing schemes, and the development of regulatory frameworks that support V2G and other demand response initiatives.
From a technological standpoint, the widespread deployment of these optimization strategies will require robust communication networks, advanced metering infrastructure, and sophisticated software platforms capable of processing vast amounts of data in real time. Vehicle manufacturers, charging equipment providers, and energy service companies will need to collaborate closely to ensure interoperability and security. Moreover, consumer education and engagement will be critical, as the effectiveness of any demand response program ultimately depends on the participation and cooperation of individual EV owners.
The study also highlights the importance of considering the human element in energy system design. Traditional engineering approaches often focus on technical efficiency and cost minimization, sometimes at the expense of user experience and satisfaction. By explicitly modeling charging and discharging willingness, this research acknowledges that people are not passive recipients of energy services but active participants in the energy ecosystem. Their decisions, shaped by economic, social, and psychological factors, have a profound impact on system performance.
This human-centered perspective aligns with broader trends in sustainable development and smart city planning, which emphasize the need for inclusive, equitable, and user-friendly solutions. It also underscores the value of interdisciplinary research, combining insights from electrical engineering, economics, behavioral science, and computer science to tackle complex real-world problems.
Looking ahead, the researchers suggest several avenues for future work. These include expanding the model to account for uncertainties in renewable energy generation and load forecasting, incorporating more detailed representations of battery aging and health, and exploring the potential for peer-to-peer energy trading among EV owners and other prosumers. Additionally, there is a need for empirical validation of the model through field trials and pilot projects that can provide real-world data on driver behavior and system performance.
In conclusion, the research by Feng Yemu and his team represents a significant step forward in the integration of electric vehicles into modern energy systems. By developing a comprehensive model that balances the interests of both system operators and individual users, they have demonstrated a practical pathway toward more efficient, resilient, and sustainable energy management. As the world continues its transition to a low-carbon future, studies like this will play a vital role in ensuring that the benefits of electrification are realized not just in terms of environmental impact, but also in terms of economic value and social well-being.
Feng Yemu, Lyu Ganyun, Shi Mingming, Zhu Zhiying, Wang Haoyu, Chen Guangyu, School of Electric Power Engineering, Nanjing Institute of Technology, Electric Power Engineering Technology, DOI: 10.12158/j.2096-3203.2024.02.015