An AI-Augmented Dataset of Multi-Prototype Electric Vehicle Charging Load Profiles in China

Area of Science:

• Electrical Engineering
• Data Science
• Urban Planning

Background:

• Large-scale electric vehicle (EV) integration presents significant challenges to power grid stability and operation.
• High-fidelity and diverse datasets are essential for effective research into EV charging impacts.
• Existing datasets often lack the controlled comparative analysis capabilities needed for nuanced research.

Purpose of the Study:

• To introduce MP-EVData, a novel dataset of station-level EV charging load profiles.
• To provide a controlled environment for comparative analysis of different EV charging station prototypes.
• To support data-intensive research in areas like load forecasting and smart charging algorithms.

Main Methods:

• Collected station-level charging load data from 10 distinct EV charging stations in a major Chinese metropolis in 2024.
• Categorized stations into five prototypes: taxi, bus, residential, battery swapping, and heavy-duty truck.
• Generated a parallel, high-fidelity synthetic dataset using advanced generative AI models.

Main Results:

• MP-EVData exhibits distinct daily, weekly, and annual temporal patterns across different charging station prototypes.
• Demonstrated clear price-responsive charging behavior under time-of-use electricity pricing.
• Validated the dataset's utility for comparative analysis by isolating geographical, climatic, and policy variables.

Conclusions:

• MP-EVData serves as a crucial benchmark for advancing research in EV grid integration.
• The dataset facilitates improved load forecasting, smart charging algorithm development, and urban infrastructure planning.
• The inclusion of synthetic data enhances its applicability for data-intensive research methodologies.