In the context of global energy transition and carbon neutrality goals, the rapid electrification of the transportation sector, primarily driven by battery electric vehicles (EVs), presents both a significant challenge and a pivotal opportunity for power systems. The proliferation of battery EV car necessitates the development of smart charging infrastructure. Photovoltaic-Storage-Charging Stations (PSCS), which integrate local renewable generation (PV), energy storage systems (ESS), and charging facilities, represent a promising solution. However, the carbon emissions within a PSCS originate from multiple sources, including electricity purchased from the grid, discharging of the internal ESS, and the discharging of battery EV cars themselves. A lack of coordinated scheduling strategies that differentiate between clean and carbon-intensive energy flows often leads to the underutilization of the station’s inherent low-carbon resources. This paper addresses this gap by proposing a novel two-stage joint low-carbon scheduling strategy between PSCSs and battery EV car clusters, explicitly considering a dynamic electricity price mechanism.
The core of our strategy lies in a fine-grained carbon accounting model and a corresponding price signal designed to guide the charging and discharging behavior of battery EV car clusters. We first classify the electrical energy within the PSCS into two virtual streams: “Clean Electricity” (primarily from PV) and “Carbon Electricity” (from grid purchases and ESS discharge). A dynamic pricing scheme is then formulated based on this “Clean-Carbon Electricity Coordination”. The price for consuming Clean Electricity $$c_{t}^{ch,PV+}$$ is set equal to the feed-in tariff, incentivizing battery EV car to absorb local PV surplus. The price for consuming Carbon Electricity $$c_{t}^{CO_2+}$$ is constructed by adding a carbon surcharge to the standard Time-of-Use (TOU) price:
$$c_{t}^{CO_2+} = c^{TOU} + \Delta c_{t}^{CO_2+}$$
$$\Delta c_{t}^{CO_2+} = E_{N,i,t}^* \cdot \Delta c_{max}^{CO_2+}$$
$$E_{N,i,t}^* = \frac{E_{N,i,t} – \min(E_{N,i,t})}{\max(E_{N,i,t}) – \min(E_{N,i,t})}$$
where $$\Delta c_{t}^{CO_2+}$$ is the carbon surcharge, $$E_{N,i,t}$$ is the carbon intensity (carbon emission per kWh) at the grid connection node of the PSCS, and $$E_{N,i,t}^*$$ is its normalized value. This price effectively internalizes the carbon cost, guiding the battery EV car cluster to charge during periods of lower grid carbon intensity. The discharging price for the battery EV car is set as a multiple (α > 1) of $$c_{t}^{CO_2+}$$ to ensure economic viability for vehicle owners.
Secondly, we establish a meticulous carbon emission model for the PSCS. Beyond the straightforward carbon emissions from grid purchases ($$E_{C,i,t}^{buy} = E_{N,i,t} \cdot P_{i,t}^{buy} \cdot \Delta t$$), we model the dynamic carbon intensity of both the stationary ESS and the mobile battery EV car cluster. When a battery EV car charges, it stores not only electricity but also the associated carbon emissions of the electricity source. This stored carbon is released when the battery EV car discharges. The carbon intensity of the EV cluster ($$E_{I,t}^{ev}$$) is dynamically updated based on its state of charge (SOC) and the carbon intensity of the electricity it absorbs:
During charging: $$E_{I,t}^{ev} = \frac{E_{I,t-1}^{ev} S_{t-1}^{ev} – E_{C,t}^{ev-} + E_{C,t}^{ev+} + E_{C,t}^{ch,ev}}{S_{t}^{ev}}$$
During discharging: $$E_{I,t}^{ev} = \frac{E_{I,t-1}^{ev} S_{t-1}^{ev} – E_{C,t}^{ev-} + E_{C,t}^{ev+} – E_{I,t-1}^{ev*} (P_{t}^{dc,ev} \Delta t / \eta^{ev})}{S_{t}^{ev}}$$
Here, $$S_{t}^{ev}$$ is the aggregate energy of the battery EV car cluster, $$E_{C,t}^{ev-}$$ and $$E_{C,t}^{ev+}$$ represent carbon carried by departing and arriving battery EV car, and $$E_{C,t}^{ch,ev}$$ is the carbon stored during charging, calculated based on the proportion of energy drawn from PV, grid, and ESS discharge. A similar model is applied to the stationary ESS. This approach allows for accurate tracking of the carbon footprint associated with the mobile energy storage provided by battery EV car.
Thirdly, we formulate a two-stage optimization framework. The day-ahead stage involves a bi-level model. The upper level is the Distribution System Operator (DSO) which minimizes its operation cost (including gas turbine cost and power purchase cost from the main grid) while ensuring power flow and security constraints, and calculates the nodal carbon intensity $$E_{N,i,t}$$. The lower level is a joint scheduling model between multiple PSCSs and the commuting battery EV car cluster. The PSCS operator’s objective is to minimize total cost, including electricity purchase cost, carbon emission cost, PV O&M cost, minus revenue from selling electricity:
$$\min C^{CS} = \sum_{i} \left( \sum_t c^{TOU}P_{i,t}^{buy} + c^{CO_2}E_{C,i,t}^{buy} + c^{PV}P_{i,t}^{PV} – \sum_t c_{t}^{sell}P_{i,t}^{sell} \right)$$
The battery EV car cluster’s objective is to minimize its net energy cost:
$$\min C^{ev} = \sum_t \left( c_{t}^{ch,PV+} P_{t}^{ch,PV+} + c_{t}^{CO_2+} P_{t}^{ch,CO_2+} \right) – \sum_t \left( \alpha \cdot c_{t}^{CO_2+} \cdot P_{t}^{dc,ev} \right)$$
The constraints include EV cluster charging/discharging limits (e.g., $$0 \le P_{t}^{ch,PV+} \le (1-\delta^{ev}) \min(P_{t,left}^{PV}, P_{t,max}^{ev})$$), energy balance ($$P_{i,t}^{PV} + P_{i,t}^{dc,es} + P_{t}^{dc,ev} + P_{i,t}^{buy} = P_{t}^{FC} + P_{i,t}^{ch,es} + P_{t}^{ch,ev} + P_{i,t}^{sell}$$), and ESS constraints. The intra-day stage performs a rolling-horizon adjustment to correct deviations from the day-ahead schedule caused by forecast errors in fast-charging load and PV generation, minimizing adjustment costs.
We conduct simulations based on a modified IEEE 33-bus distribution system with PSCSs located in residential and commercial areas. The performance of the proposed dynamic pricing strategy (Scenario 3) is compared against a traditional TOU price (Scenario 1) and a partial strategy with only clean-energy pricing (Scenario 2). The key results are summarized below.
| Scenario | Description | PSCS Operator Cost ($) | Battery EV Car Cluster Cost ($) | Total Cost ($) | Grid Purchase Carbon (kgCO₂) | EV Discharge Carbon (kgCO₂) | Total Carbon (kgCO₂) |
|---|---|---|---|---|---|---|---|
| 1 | TOU Price Only | 6,064 | 5 | 6,069 | 4,172.3 | 541.7 | 4,865.0 |
| 2 | Clean-Elec. Price Only | 6,335 | -632 | 5,706 | 3,913.6 | 438.3 | 4,520.8 |
| 3 | Proposed Dynamic Price | 5,988 | -482 | 5,506 | 3,768.6 | 391.1 | 4,305.8 |
The results demonstrate the effectiveness of our strategy. Scenario 2, with clean-energy pricing, successfully guides the battery EV car cluster to prioritize charging during PV surplus periods, increasing local PV consumption and reducing the cluster’s discharge carbon intensity compared to Scenario 1. Scenario 3, with the full dynamic price, further reduces total carbon emissions by 11.5% compared to Scenario 1. This is achieved because the carbon surcharge $$c_{t}^{CO_2+}$$ suppresses charging during periods with high grid carbon intensity (e.g., hour 16), which is not achieved in Scenario 2. Economically, the total cost in Scenario 3 is 9.3% lower than in Scenario 1, benefiting both the station operator (reduced purchase and carbon costs) and the battery EV car owners (increased discharging revenue).
We further analyze the impact of the mobility of the battery EV car. Scenarios where the battery EV car cluster is only scheduled at the residential PSCS (Scenario 4) or only at the commercial PSCS (Scenario 5) are tested.
| Scenario | Scheduling Mode | Total Cost ($) | Total Carbon (kgCO₂) |
|---|---|---|---|
| 3 | Two-Station Commuting | 5,506 | 4,305.8 |
| 4 | Residential Station Only | 6,140 | 4,020.9 |
| 5 | Commercial Station Only | 6,271 | 3,910.8 |
While single-station scheduling (Scenarios 4 & 5) can achieve slightly lower carbon emissions due to reduced overall charging activity, it leads to a significant increase (11.5%-13.9%) in total cost. This is because it fails to leverage the spatial flexibility of the battery EV car, increasing the purchase pressure on a single station and reducing the economic benefit for vehicle owners from “charging low and discharging high” across different price zones. This highlights the importance of coordinating scheduling across multiple PSCSs to balance economic and low-carbon objectives effectively.
The intra-day rolling optimization effectively manages forecast uncertainties. The real-time adjustment keeps the total carbon emission very close to the day-ahead plan, with a deviation of only 1.5%, and the real-time carbon-electricity price closely tracks the day-ahead values.
Sensitivity analyses are conducted regarding PV penetration and the number of battery EV car. The proposed strategy consistently outperforms the TOU基准across all tested conditions. For instance, with varying PV output levels, the reductions in total cost and total carbon of Scenario 3 over Scenario 1 range from 3.2%-14.6% and 4.0%-17.4%, respectively. Similarly, with different numbers of battery EV car, the proposed strategy maintains its advantages.
In conclusion, we propose a comprehensive two-stage low-carbon scheduling strategy for integrated PSCSs and battery EV car clusters. The core innovations include: 1) A clean-carbon electricity coordinated dynamic pricing mechanism that provides precise economic signals for low-carbon charging and discharging of battery EV car; 2) A fine-grained carbon emission model that accurately accounts for the dynamic carbon intensity of both stationary storage and mobile battery EV car energy storage; 3) A two-station commuting scheduling framework that harnesses the spatiotemporal flexibility of battery EV car to enhance both economic and environmental performance. Simulation results verify that the strategy significantly improves local PV consumption, reduces the overall carbon footprint of the PSCS operation, and achieves cost savings for all participants, demonstrating a viable pathway for the synergistic development of renewable energy and electrified transportation.