TL;DR
This paper proposes a comprehensive optimization model for integrated solar-storage-charging stations that moves beyond simple "buy low, sell high" energy arbitrage. By incorporating battery degradation costs, grid capacity charges, and network transmission losses into the objective function, the researchers demonstrate a scheduling method that increases total profitability while doubling battery life and reducing peak grid demand by over 17%.
What Problem Does This Paper Address?
As electric vehicle (EV) adoption surges, integrated stations—combining photovoltaic (PV) generation, battery energy storage systems (BESS), and fast-charging piles—are becoming critical infrastructure. However, managing these stations is a complex balancing act.
Most existing research focuses narrowly on "peak shaving and valley filling" (shifting load to cheaper hours). This often ignores "hidden" costs that can erode a station's long-term viability, such as:
- Battery Degradation: Frequent cycling to chase small price differences can wear out expensive batteries prematurely.
- Capacity Charges: Utilities often charge commercial customers based on their single highest peak of power demand (kW), not just total energy used (kWh).
- Network Losses: Rapid, high-power charging/discharging increases resistive losses in local transmission lines.
The authors aim to create a "multi-benefit" model that optimizes all these factors simultaneously.
Methods
The authors developed a cost-benefit model programmed in MATLAB using the YALMIP toolbox. The core of the paper is an objective function designed to maximize daily net benefit ($I_{total}$):
$$I_{total} = I_{sale} - C_{electric} - C_R - C_s - C_{net}$$
- $I_{sale}$: Revenue from charging services.
- $C_{electric}$: Cost of purchasing energy from the grid based on Time-of-Use (TOU) pricing.
- $C_R$: Capacity Charge, calculated based on the maximum 15-minute average load from the grid.
- $C_s$: Cycle Cost, representing the wear and tear on the battery per charge/discharge cycle.
- $C_{net}$: Network Loss Penalty, which penalizes high-current spikes that cause energy dissipation in transmission lines.
The model is constrained by power balance equations (PV + Grid + Battery = EV Load) and physical limits like the battery's State of Charge (SOC) and converter power ratings.
Datasets
The study uses a simulated case study based on a realistic station configuration:
- Infrastructure: 40 fast-charging piles (35kW each, 1400kW total capacity).
- Storage: 1000kWh battery with a 400kW power limit.
- Input Data: Representative 24-hour curves for PV output and EV charging demand (likely based on typical regional patterns in China, given the author affiliations).
- Pricing: Specific TOU electricity rates and a capacity fee of 48 RMB/kW*month.
Key Insights and Ideas
- The "Anti-Cycling" Effect: When battery degradation costs ($C_s$) are ignored, the system cycles the battery twice a day to catch every minor price dip. Including $C_s$ reduces this to one strategic cycle per day, significantly extending the hardware's lifespan without sacrificing the primary economic goals.
- Smoothing the Spikes: By penalizing network losses ($C_{net}$), the optimizer avoids "slamming" the grid with power during cheap windows. Instead, it smooths the charging curve, which is healthier for the local distribution transformer and reduces energy waste.
- Capacity Fee Mitigation: The model identifies that the highest financial burden often comes from the "capacity charge." By using the battery to "cap" the peak power drawn from the grid (even if the energy itself is cheap at that moment), the station saves more money than it would through simple energy arbitrage.
Key Results
The researchers compared four scenarios, with Scenario 4 being the full proposed model.
- Profitability: Scenario 4 achieved the highest total daily benefit (9047.96 yuan), outperforming simpler models.
- Peak Demand Reduction: Compared to a model that ignores capacity fees (Scenario 2), the proposed method reduced the peak grid load, lowering capacity charges from 907.83 yuan to 742.02 yuan—a 17.08% reduction.
- Operational Efficiency: The daily energy storage loss (cycle cost) was slashed from 1360.42 yuan to 611.38 yuan by optimizing the number of charge/discharge cycles.
- Grid Harmony: The grid power curve in the optimized scenario was significantly "smoother," reducing the network loss cost from 169.7 to 165.8 yuan.
Limitations and Future Work
- Deterministic Assumptions: The model uses fixed PV and load curves. In reality, solar output is intermittent (clouds) and EV arrival is stochastic. Future work could incorporate uncertainty modeling or robust optimization.
- Battery Modeling: The cycle cost is treated as a linear converted unit cost. More sophisticated electrochemical models that account for temperature and depth-of-discharge (DoD) non-linearly could provide even more accurate results.
- Single-Station Focus: The paper looks at one station in isolation. Future research could explore how a network of these stations interacts with the wider distribution grid.
Summary updated