EVs Cut Rail Energy Costs by 20% in Smart Grid Integration
A groundbreaking study reveals that integrating electric vehicles (EVs) into urban rail traction power systems can significantly reduce electricity costs while enhancing energy efficiency and sustainability. Conducted by Tang Zhaoxiang and Xu Wantao from CRRC Qingdao Sifang Co., alongside Deng Hao and Lu Wenjie from the School of Electrical Engineering at Southwest Jiaotong University, the research demonstrates how coordinated energy management between EVs, supercapacitors, and railway infrastructure can transform parking lots near rail stations into dynamic energy hubs.
As cities expand and urban sprawl intensifies, regional rail networks have become essential arteries connecting city centers with suburbs and satellite towns. These systems offer fast, high-capacity transit solutions that ease congestion and reduce reliance on private automobiles. However, the growing demand for sustainable transportation has placed increasing pressure on rail operators to minimize energy consumption and carbon emissions. At the same time, the rapid rise of electric mobility has created new opportunities for synergy between transportation modes.
One such opportunity lies in the widespread adoption of park-and-ride (P&R) schemes, where commuters drive their EVs to suburban rail stations, park, and continue their journey via train. While this model promotes multimodal commuting, it also presents a unique challenge: how to manage the charging demands of hundreds or even thousands of parked EVs without overloading the local power grid. The research team saw not a challenge, but a hidden asset—each parked EV represents a mobile battery that could be tapped to support the very rail system it connects to.
The study focuses on co-phase AC traction power supply systems, which are increasingly being adopted for modern regional railways due to their superior power delivery, reduced electrical complexity, and elimination of phase separation issues common in traditional AC systems. Unlike older DC-based systems, which suffer from stray current corrosion and limited regenerative braking utilization, 25 kV AC systems offer greater efficiency and scalability. By integrating power flow controllers (PFCs), these systems can manage reactive power, suppress harmonics, and facilitate bidirectional energy transfer—making them ideal platforms for incorporating energy storage.
What sets this research apart is its innovative use of chance-constrained programming to address the inherent uncertainty in EV behavior. Traditional optimization models assume predictable arrival and departure times, fixed initial battery states, and consistent charging patterns. In reality, EV owners arrive at different times, leave unpredictably, and begin charging with varying levels of battery depletion. Attempting to optimize energy dispatch under such uncertainty using deterministic methods often leads to overly conservative or infeasible solutions.
To overcome this, the researchers employed a stochastic modeling approach based on probabilistic constraints. Instead of requiring that all EVs meet their charging needs with 100% certainty—an impractical goal—they allowed for a small, predefined risk of constraint violation, say 5%, while ensuring that the overall system operates reliably within acceptable confidence levels. This method, known as chance-constrained programming, enables more flexible and realistic scheduling of EV charging and discharging activities.
The model was formulated as a mixed-integer linear program (MILP), a type of optimization framework well-suited for large-scale decision-making problems involving discrete choices (e.g., whether an EV charges or discharges) and continuous variables (e.g., power levels). Using the CPLEX solver, a powerful commercial optimization engine, the team simulated a full day of operations for a representative urban rail line equipped with both EV parking facilities and onboard supercapacitors.
Supercapacitors play a crucial role in the proposed architecture. While EV batteries offer high energy capacity, they are not designed for rapid, frequent charge-discharge cycles, which can accelerate degradation. Supercapacitors, on the other hand, excel at handling short bursts of high power, making them ideal for smoothing out transient loads caused by train acceleration and braking. By combining both technologies, the system achieves a balance between energy capacity (from EVs) and power responsiveness (from supercapacitors), maximizing overall performance while preserving battery health.
The simulation results were striking. Compared to a baseline scenario without any energy storage integration, the optimized system reduced the daily electricity cost for the traction substation by 20.37%. This translates into substantial savings for rail operators, especially when scaled across entire networks. The cost reduction stems from three main factors: lower energy consumption during peak tariff periods, reduced demand charges, and revenue generated from selling excess energy back to the grid.
Demand charges—fees based on the highest average power drawn over a 15-minute interval during the billing month—are a significant component of utility bills for industrial and transportation customers. In the unoptimized case, the peak demand reached 9.3 MW, triggering high charges under China’s two-part tariff structure. With the integrated EV and supercapacitor system actively managing load profiles, the peak was shaved down to 5.4 MW—a reduction of 41.9%. This dramatic flattening of the load curve not only cuts costs but also enhances grid stability and reduces stress on transformers and switchgear.
Moreover, the system achieved a regenerative braking energy utilization rate of 96.68%, far exceeding typical values of 30–50% seen in conventional systems. When trains brake, they generate electricity that can be fed back into the overhead lines. Without adequate storage or nearby trains to absorb this energy, it is often dissipated as heat through resistors or lost due to voltage rise. In this model, however, the recovered energy is either stored in supercapacitors for immediate reuse or directed to charge parked EVs, effectively turning waste into value.
Another key finding was the impact of EV penetration levels on cost savings. As the number of connected EVs increased, total electricity expenses declined steadily, approaching a lower bound of approximately 71,000 yuan per day. This saturation effect occurs because, beyond a certain point, the additional charging demand from more EVs begins to offset the benefits of load leveling and energy resale. Nevertheless, the study confirms that even moderate EV adoption—around 100 to 200 vehicles per station—can yield significant economic returns.
The researchers categorized EV users into two groups based on commuting behavior: Type A, consisting of urban workers who park at suburban stations during the day before returning home in the evening; and Type B, local residents who work in the suburbs and charge their vehicles overnight at P&R lots. Each group exhibits distinct temporal and behavioral patterns, which were modeled using normal distributions for arrival/departure times and initial state of charge (SOC), and uniform distributions for desired SOC upon departure.
Monte Carlo sampling was used to generate 500 stochastic scenarios reflecting real-world variability. These scenarios were then incorporated into the optimization model through sample average approximation (SAA), a technique that converts probabilistic constraints into deterministic equivalents by averaging over multiple realizations. This allowed the chance-constrained model to be solved efficiently while preserving statistical rigor.
From a policy perspective, the findings support the development of integrated transport and energy planning frameworks. Municipalities investing in regional rail should consider co-locating EV charging infrastructure not just as a convenience for passengers, but as a strategic energy asset. Utilities, too, stand to benefit from reduced peak loads and improved grid resilience. Furthermore, rail operators could explore new business models, such as offering discounted parking or priority boarding in exchange for participation in vehicle-to-grid (V2G) programs.
The technical implementation relies on bidirectional DC/DC converters linking EV batteries and supercapacitors to the DC bus of the PFC. This eliminates the need for separate rectification stages and allows seamless power exchange between the grid, trains, and storage units. The control system continuously monitors grid prices, train schedules, and EV availability to determine optimal charging/discharging strategies on a minute-by-minute basis.
Cybersecurity and user privacy are critical considerations in any V2G deployment. The model assumes secure communication protocols and anonymized data handling to protect user information. Participation in energy dispatch would likely be opt-in, with users setting preferences for minimum battery levels and departure times. Incentives such as reduced charging fees or loyalty points could encourage broader adoption.
From an environmental standpoint, the increased use of regenerative braking energy reduces the need for fossil-fuel-generated electricity, lowering greenhouse gas emissions. Even when the grid mix includes coal or natural gas, displacing marginal generation during peak hours contributes to cleaner air and reduced pollution. Over time, as grids become greener, the climate benefits will only grow.
The study also highlights the importance of coordinated control algorithms. Simply allowing EVs to charge whenever they arrive—so-called “dumb charging”—can exacerbate peak loads and increase costs. Smart charging, guided by real-time pricing signals and system constraints, ensures that energy is drawn when it is cheapest and most abundant, and injected back when it is most valuable.
While the current model focuses on day-ahead optimization, future work could extend to real-time adjustments using model predictive control (MPC) or reinforcement learning techniques. Integrating renewable sources such as solar canopies over parking lots would further enhance sustainability, creating fully self-sufficient transit hubs.
One potential limitation is battery degradation from frequent cycling. Although the model avoids deep discharges and limits power rates, long-term impacts on EV battery lifespan require empirical validation. However, given that most EV owners do not fully deplete their batteries daily, and many drive less than 50 km per day, there is ample spare capacity for grid services without compromising mobility needs.
In practice, pilot projects could begin with small fleets of company-owned EVs or municipal vehicles, providing a controlled environment for testing and refinement. As confidence grows, public participation could be gradually expanded.
The implications extend beyond rail systems. Similar principles could apply to bus depots, airport parking facilities, or university campuses with electrified fleets. Any location where EVs park for extended periods alongside significant energy loads represents a potential site for integrated energy management.
Ultimately, this research exemplifies the convergence of transportation and energy systems in the smart city era. It shows that EVs are not merely replacements for internal combustion engine vehicles but active participants in a larger, more intelligent energy ecosystem. By leveraging the collective battery capacity of parked cars, cities can build more resilient, efficient, and sustainable infrastructure.
The integration of EVs into rail power systems is not just a technical innovation—it is a paradigm shift in how we think about mobility and energy. Rather than viewing parked vehicles as idle assets, they become mobile storage units contributing to grid stability and cost savings. This dual function enhances the economic viability of both electric transportation and public transit, creating a virtuous cycle of sustainability.
As governments worldwide push for decarbonization and electrification, studies like this provide actionable pathways forward. They demonstrate that with the right modeling tools, control strategies, and institutional coordination, the transition to clean energy can be both technically feasible and economically advantageous.
The success of this approach depends on collaboration across sectors—rail operators, utilities, automakers, and policymakers must align incentives and standards. Interoperability, data sharing, and regulatory clarity will be essential to scaling these solutions beyond isolated demonstrations.
Nevertheless, the evidence is clear: parked EVs, when intelligently managed, can play a pivotal role in reducing the energy footprint of urban transportation. This study offers a compelling blueprint for turning parking lots into power plants and commuters into contributors to a smarter, greener grid.
Tang Zhaoxiang, Xu Wantao, Deng Hao, Lu Wenjie. Optimal operation of urban railway traction power supply system with electric vehicles based on chance-constrained programming. Energy Storage Science and Technology, 2024. DOI: 10.19799/j.cnki.2095-4239.2023.0487