China’s Multi-Fuel Energy Stations Boost EV Profitability with Dynamic Pricing
A new optimization model is demonstrating how China’s next-generation energy stations—serving electric, hydrogen, and natural gas vehicles—can lift operational profits by over 11 percent using dynamic pricing and intelligent load management. The breakthrough, developed by researchers at NARI Technology and the State Grid Electric Power Research Institute, integrates real-time price signals with flexible energy conversion—electrolytic hydrogen production, power-to-gas, and multi-modal storage—to turn stations into active grid participants rather than passive fueling points. In a simulated system with 2,000 vehicles across three fuel types, dynamic pricing alone improved daily revenue by 5.3 percent; adding coordinated load shifting pushed the gain to 11.3 percent. Crucially, the model maintains driver satisfaction by enforcing energy neutrality—no net loss of charge, hydrogen, or gas per vehicle—while reshaping demand to match renewable generation peaks. This marks a pivotal step toward scalable, grid-responsive refueling infrastructure as China races to deploy over 25 million new-energy vehicles by 2025—and over 160 million by 2035.
The innovation centers on a simple but powerful design shift: treating the vehicle’s filling price not as a fixed tariff, but as a real-time decision variable that co-optimizes with energy procurement, storage dispatch, and load modulation. Unlike prior studies that fixed charging or refueling fees—or treated them as exogenous inputs—this model endogenizes them within a 96-interval daily horizon (15-minute steps), subject only to an average-price anchor that preserves regional price stability. The result is a staircase-like pricing pattern: lower tariffs during midday solar surges, higher rates during evening peaks—mirroring electricity spot market behavior now emerging in Guangdong, Zhejiang, and Sichuan pilot programs. This allows stations to absorb excess wind and solar power when grid prices dip, convert it into hydrogen or synthetic natural gas via electrolysis and methanation, store it, and sell it later at premium hours—while simultaneously nudging drivers to fill when supply is abundant.
Field data from the study show how tightly price and behavior intertwine. When charging fees dropped 18 percent between 10 a.m. and 2 p.m., electric vehicle (EV) load surged by 22 percent—driven not by subsidies, but by transparent, time-differentiated signals. Meanwhile, hydrogen refueling—often criticized for high costs—saw load shift earlier into the day as stations prioritized using surplus solar for electrolysis instead of buying costly grid-powered hydrogen. Natural gas refueling remained more evenly distributed, but its price curve flattened as stations leveraged cheaper off-peak electricity to compress and store gas. The system’s economics hinge on three layered capabilities: flexible conversion, temporal arbitrage via storage, and responsive demand.
First, energy conversion. The station is equipped with an 800 kW wind turbine and 1 MW rooftop photovoltaic array—standard for modern Chinese energy hubs. When renewable output exceeds local demand, excess power feeds two pathways: a 55-percent-efficient electrolyzer producing hydrogen, and a 60-percent-efficient power-to-gas unit generating methane. Hydrogen is prioritized for conversion due to its higher market price (26 RMB/kg vs. 2.6 RMB/m³ for natural gas), yielding greater margin when stored and resold. Only when hydrogen storage nears capacity does surplus power divert to methane synthesis. A compressor—running at 80 percent efficiency with 250:1 pressure ratio—then readies low-pressure gas for vehicle use. This cascading logic ensures every extra kWh of solar or wind displaces the most expensive external purchase.
Second, temporal shifting via storage. Three independent reservoirs—batteries for electricity, high-pressure tanks for hydrogen, and cascaded buffer vessels for natural gas—enable decoupling of production and delivery. Battery cycling is the most active, charging during solar peaks and discharging in early evening to serve EV demand without grid draw. Hydrogen storage sees two major fills: midday (solar-driven electrolysis) and overnight (low-cost grid power). Gas storage, by contrast, operates more steadily, absorbing compressed output from both purchased and synthesized gas, then releasing it during afternoon and evening rushes. Crucially, all storage units enforce cyclical boundary conditions: end-of-day inventory equals start-of-day levels. This prevents “bleed-out” strategies that sacrifice long-term resilience for short-term gains—ensuring sustainability across days, not just within one.
Third, and perhaps most novel, is load management as a revenue tool, not just a grid-service obligation. The model allows up to 20 percent upward or downward adjustment of vehicle demand in any 15-minute window—so an EV could be asked to delay charging by 30 minutes, or a hydrogen truck could accept a slightly slower fill—to align with cheaper supply. Compensation is paid for each adjustment (e.g., 224 RMB/day in one scenario), but the net effect is strongly positive: stations reduce expensive grid imports by 11.5 percent while increasing high-margin self-generated fuel sales. Importantly, total energy delivered per vehicle remains unchanged. No driver receives less juice, hydrogen, or gas—only different timing. This avoids the equity concerns that have plagued some demand-response pilots.
Sensitivity analysis reveals where returns plateau. Expanding the allowable price swing from ±5 percent to ±25 percent boosts revenue steadily—but gains taper sharply beyond 25 percent, suggesting regulatory caps around this range may be economically rational. Similarly, increasing load flexibility from 10 percent to 30 percent of instantaneous demand yields accelerating returns; beyond 30 percent, marginal benefit declines, hinting at behavioral or technical ceilings on how much drivers or vehicle systems will tolerate rescheduling. These thresholds offer practical guidance for policymakers: mandate moderate price variation and modest delay tolerances, and let market forces do the rest.
Strategically, the model aligns with Beijing’s dual mandates: accelerate clean transport and integrate renewables. By 2025, China expects over 25 million new-energy vehicles on roads—up from 16.2 million in mid-2023. Yet charging infrastructure growth has lagged, especially for hydrogen and LNG trucks. One bottleneck is economics: standalone hydrogen stations often lose money due to low utilization and high capex. This integrated approach changes the equation. A single site serving EVs, fuel-cell buses, and LNG logistics trucks can cross-subsidize: high-margin EV fast-charging funds hydrogen electrolysis; excess solar offsets compressor electricity; flexible pricing smooths cash flow. It turns infrastructure from a cost center into a profit node in the energy transition.
Implementation is already underway. In Jiangsu Province—home to the study’s authors—NARI Group has piloted similar multi-energy stations in Nanjing and Suzhou, linking them to provincial spot market trials. Early results show 8–12 percent higher asset utilization versus conventional mono-fuel sites. Nationally, the National Development and Reform Commission (NDRC) has signaled openness to dynamic pricing in its 2024 guidelines for “intelligent energy infrastructure,” provided average tariffs stay within government-guided bands. That’s precisely the constraint built into this model: it optimizes within stability guardrails, satisfying both market efficiency and social fairness.
For international observers, the implications extend beyond China. Europe’s Alternative Fuels Infrastructure Regulation (AFIR) and the U.S. NEVI program both emphasize interoperability and grid integration—but lack pricing mechanisms to incentivize it. This Chinese model offers a template: embed price responsiveness at the station level, use storage and conversion as buffers, and let drivers “vote with their wallets” for cleaner, cheaper energy timing. It doesn’t require subsidies—just smarter tariff architecture. As automakers like Stellantis and Hyundai roll out hydrogen trucks, and Tesla and BYD expand fast-charging networks, the question isn’t whether to build multi-fuel hubs—but how to operate them profitably. China’s answer: price dynamically, store flexibly, and shift demand intelligently.
Critically, the approach avoids politicized terminology. There’s no mention of “common prosperity” or “dual circulation.” Instead, it’s framed in universal engineering and economics terms: revenue maximization, constraint satisfaction, marginal cost alignment. This enhances global applicability. A station in Hamburg or Houston could adopt the same logic—swap yuan for euros, RMB/kg for USD/kg—and see comparable gains, assuming similar hardware and market structures. That translatability is key for investor confidence: this isn’t a China-specific policy play, but a replicable operational upgrade.
Looking ahead, the team acknowledges two frontiers. First, uncertainty: real-world EV arrival times, hydrogen truck schedules, and renewable forecasts are stochastic—not deterministic as in the base case. Future work will integrate probabilistic programming or robust optimization to hedge against variability without sacrificing too much upside. Second, spatial coordination: optimizing one station is valuable, but optimizing a network—allowing load to shift between nearby sites via app-based incentives—could unlock system-wide efficiencies. Early simulations suggest 15–18 percent gains at the cluster level, especially in dense urban corridors like the Yangtze River Delta.
For now, the core message is clear: dynamic pricing isn’t just for electricity markets. When applied to vehicle energy retail—with careful design to protect user welfare—it becomes a powerful lever for station profitability, renewable integration, and supply chain resilience. As China’s EV fleet balloons toward 100 million units this decade, getting the economics of refueling right may prove as crucial as battery chemistry or motor efficiency. This research shows how.
Author: Wang Jun, Wang Xin, Zhu Jinda, Du Wei
Affiliation: NARI Technology Co., Ltd.; NARI Group Corporation (State Grid Electric Power Research Institute); National Key Laboratory of Power Grid Operation Risk Prevention Technology and Equipment, Nanjing, China
Journal: Automation of Electric Power Systems, Vol. 48, No. 22, November 25, 2024
DOI: 10.7500/AEPS20230808007