Remote Energy Self-Sufficiency Strategy Boosts EV Charging on Highways
As electric vehicles (EVs) continue to gain traction across China, the infrastructure supporting them—especially in remote regions—faces mounting pressure. One of the most pressing challenges lies in ensuring reliable, sustainable, and cost-effective energy supply for EV charging stations located far from urban centers and main power grids. In response to this growing need, a groundbreaking study from Sichuan University introduces a novel energy self-sufficiency framework tailored specifically for remote transportation corridors, with a primary focus on highway-based EV charging systems.
Published in the journal Power System Technology, the research led by Junhang Wu, Jichun Liu, and Yiyang Wu presents a coordinated source-storage capacity configuration method designed to overcome two critical barriers in remote energy planning: the lack of historical meteorological data and the frequent disconnection from the main grid. These issues have long hindered the deployment of resilient renewable energy systems in rural and isolated areas, where traditional microgrid models—developed for urban environments—fall short.
The team’s approach diverges from conventional methods by integrating natural resource assessment with operational resilience, creating a dual-layer optimization model that balances economic efficiency and energy reliability. At its core, the strategy enables remote energy systems to function independently while minimizing dependence on external power sources, a crucial advancement for regions where grid connectivity is weak or unreliable.
One of the most innovative aspects of the study is its solution to the absence of long-term weather records in remote zones. Instead of relying on historical solar irradiance or wind speed data—which are often unavailable—the researchers developed a computational model based on geographical and astronomical parameters such as latitude, solar position, and seasonal variations. This allows planners to simulate full-year solar intensity profiles for any given location without needing physical measurement stations. For wind resources, the team applied a Weibull distribution model calibrated using nearby operational wind farms, enabling accurate wind power output estimation even in data-scarce environments.
These simulated renewable profiles are then used to generate a set of typical time-series scenarios that reflect real-world variability, including adverse conditions like low sunlight and weak winds. By focusing on worst-case and high-variability scenarios, the model ensures that the resulting energy system remains robust under challenging weather patterns—a vital consideration for maintaining continuous EV charging services.
The proposed system architecture is built around a hybrid AC/DC microgrid, which enhances compatibility with diverse energy sources and loads. Photovoltaic panels are directly connected to the DC bus to reduce conversion losses, while wind turbines feed into the AC side through inverters. Energy storage systems play a central role in stabilizing supply, absorbing excess generation during peak production and discharging during periods of low renewable output or high demand.
A key load in this context is fast-charging EV infrastructure, which exhibits distinct behavioral patterns compared to urban charging stations. On highways, EV charging demand tends to cluster around midday and evening hours, coinciding with meal breaks and rest stops. Unlike city drivers who may charge overnight, highway travelers require rapid charging, placing sudden and significant power demands on the local grid. To capture these dynamics, the researchers employed a Monte Carlo simulation approach based on vehicle arrival statistics and battery characteristics, generating a realistic daily load profile that reflects actual highway usage.
This load model was validated using data from a representative highway segment, revealing two pronounced peaks—one around noon and another in the early evening—aligned with typical travel patterns. Additionally, the system must support other critical infrastructure, including tunnel lighting, surveillance systems, and service area operations, all of which contribute to a relatively high baseline energy demand, particularly at night when solar generation is absent.
To address these complexities, the research team developed a bi-level optimization framework that integrates both planning and operational decisions. The upper layer determines the optimal capacities of wind, solar, and energy storage components by minimizing the total annualized system cost, which includes investment, maintenance, and replacement expenses over the equipment’s lifespan. The lower layer simulates system operation across multiple weather scenarios, optimizing dispatch strategies to minimize operating costs while enhancing self-sufficiency.
What sets this model apart is its ability to simultaneously account for both grid-connected and islanded (off-grid) modes of operation. In many remote areas, the connection to the main power grid is fragile, with frequent outages due to extreme weather, aging infrastructure, or limited transmission capacity. Rather than treating these states as separate cases, the model incorporates an “annual failure off-grid coefficient” that quantifies the likelihood and duration of grid disconnection. This parameter allows the optimization process to weigh the trade-offs between relying on external power and maintaining internal resilience.
In grid-connected mode, the system can import electricity during shortages and export surplus generation, taking advantage of time-of-use pricing to reduce costs. However, in islanded mode, all energy must be locally sourced, making storage and generation adequacy paramount. The operational layer includes multiple objectives: minimizing annual operating costs, reducing grid dependency (measured by power exchange volume), and lowering load shedding and curtailment rates during off-grid events.
To achieve a balanced solution, the researchers employed an improved multi-objective evolutionary algorithm called AGE-MOEA-II, which uses geometric estimation techniques to enhance convergence and diversity in the solution space. This algorithm identifies a Pareto frontier of non-dominated solutions, from which a compromise solution is selected using entropy-based weighting and achievement scalarizing functions. This ensures that the final configuration is not biased toward any single objective and reflects a holistic balance between cost, reliability, and sustainability.
The effectiveness of the proposed method was demonstrated through a case study based on a real-world highway corridor in a remote region of China. Using local geographic data and estimated traffic patterns, the team simulated six representative weather scenarios, ranging from light rain and low wind to overcast skies with moderate breezes. Each scenario was assigned a probability based on regional climate statistics, allowing the model to prioritize robustness without over-investing in underutilized capacity.
The results showed that the integrated bi-level approach significantly outperformed conventional single-scenario or single-mode designs. When compared to configurations optimized for only the best or average weather conditions, the proposed method reduced annual operational costs by up to 15% while improving energy self-sufficiency by over 20%. More importantly, it drastically lowered the risk of load shedding during extended grid outages, a critical factor for maintaining public confidence in EV travel.
Among the various design alternatives tested, a configuration featuring 540 kW of wind capacity, 360 kW of photovoltaic capacity, and 720 kWh of battery storage emerged as the most balanced solution. While wind capacity exceeded solar in this setup, the choice was driven by the need for nighttime generation and smoother output profiles, especially during winter months when daylight hours are short. The larger-than-expected battery size reflects the importance of bridging multi-hour gaps between renewable availability and peak demand.
Further analysis revealed that ignoring off-grid risks leads to underinvestment in storage and backup generation, resulting in frequent blackouts during grid failures. Conversely, designing solely for islanded operation leads to oversized and costly installations that rarely operate at full capacity. The proposed model strikes a pragmatic balance, increasing upfront investment slightly but delivering superior long-term performance and reliability.
An important finding from the sensitivity analysis was the role of the annual failure off-grid coefficient in shaping the optimal design. As this value increases—indicating more frequent or longer grid outages—the model naturally shifts toward greater self-reliance, increasing storage capacity and reducing reliance on external power. This flexibility makes the framework adaptable to different regions, from mountainous highways with frequent landslides to desert routes prone to sandstorms.
From a policy perspective, the study highlights the importance of location-specific planning in national efforts to expand clean transportation infrastructure. Blanket deployment strategies that assume uniform grid access and weather patterns are ill-suited for China’s vast and diverse geography. Instead, planners should adopt adaptive frameworks that consider local environmental, technical, and economic conditions.
The implications extend beyond highways. The same methodology could be applied to remote villages, border outposts, or industrial sites where reliable power is essential but grid access is limited. As China pushes forward with its dual carbon goals—peaking carbon emissions before 2030 and achieving carbon neutrality by 2060—decentralized, renewable-powered microgrids will play an increasingly important role in decarbonizing hard-to-reach sectors.
Moreover, the integration of EV charging into these systems creates synergies between transportation and energy transitions. By aligning vehicle charging patterns with renewable generation, planners can maximize the utilization of clean energy and minimize the need for fossil-fueled peaking plants. Smart charging strategies, such as delaying non-urgent charging until solar output peaks, can further enhance system efficiency.
Looking ahead, the research team acknowledges that their current model focuses primarily on source and storage coordination. Future work will incorporate network topology optimization and carbon emission constraints, paving the way for comprehensive source-grid-storage planning. As distributed energy resources become more prevalent, the design of physical infrastructure—such as cable routing, substation placement, and protection schemes—will also influence overall system performance.
Nonetheless, the current study represents a significant leap forward in remote energy system design. It provides a practical, data-driven methodology for building resilient, cost-effective, and environmentally friendly power systems in some of the country’s most challenging environments. For highway operators and energy planners alike, it offers a roadmap for ensuring that the electric mobility revolution does not stall at the city limits.
The success of EV adoption in China hinges not just on vehicle sales but on the strength and intelligence of the supporting infrastructure. In remote areas, where every kilowatt-hour counts, the ability to generate, store, and manage energy locally is no longer a luxury—it is a necessity. This research from Sichuan University delivers a powerful tool for meeting that challenge, combining advanced modeling techniques with real-world applicability to chart a sustainable path forward.
As the nation continues to expand its expressway network and electrify its transport sector, solutions like this will be essential for bridging the energy gap between urban centers and the vast expanses beyond. With intelligent planning and innovative engineering, even the most isolated stretches of road can become nodes in a cleaner, more resilient national energy ecosystem.
Junhang Wu, Jichun Liu, Yiyang Wu, College of Electrical Engineering, Sichuan University. Published in Power System Technology. DOI: 10.13335/j.1000-3673.pst.2023.2008