<b>MP-EVData: An AI-Augmented Dataset of Multi-Prototype Electric Vehicle Charging </b><b>Load Profile</b><b>s in China</b>

Large-scale electric vehicle integration poses significant challenges to power grid operation, demanding high-fidelity and diversified datasets for in-depth research. To address this need, we introduce MP-EVData, a comprehensive dataset <b>of </b>station-level charging load profiles from a major Chinese metropolis, covering the full 2024 year. The core value of MP-EVData <b>is </b>providing charging load data for 10 stations in the same geographical location during the same time period, representing five distinct prototypes: taxi demonstration stations, bus depots, residential charging stations, battery swapping stations and heavy-duty truck stations. This unique structure eliminates the disturbances of external variables such as geographical, climatic, and policy, enabling controlled comparative analysis of their load characteristics. Furthermore, the dataset is augmented with a parallel, high-fidelity synthetic dataset generated using advanced generative AI models to support data-intensive research. Technical validation reveals highly distinct daily, weekly, and annual temporal patterns across prototypes and demonstrates clear price-responsive charging behavior under time-of-use pricing. MP-EVData provides a crucial benchmark for advancing researches in load forecasting, smart charging algorithms and urban infrastructure planning.大规模电动汽车并网对电网运行带来了显著挑战，亟需高保真、多样化的数据集以支撑深入研究。为应对这一需求，我们推出MP-EVData数据集：该数据集源自中国某特大城市的场站级充电负荷曲线，覆盖完整的2024年度。MP-EVData的核心价值在于，其提供了同一地理位置、同一时间段内10个场站的充电负荷数据，涵盖五类不同的场站原型：出租车示范充电站、公交场站、居民充电站、换电站以及重型卡车充电站。这一独特的数据集结构消除了地理、气候、政策等外部变量的干扰，可实现对各类场站负荷特性的可控对比分析。此外，本数据集还辅以基于先进生成式AI（Generative AI）模型生成的并行高保真合成数据集，以支撑数据密集型研究。技术验证结果表明，各类场站原型的日、周、年时间维度负荷模式存在显著差异，且在峰谷分时电价机制下呈现出明确的电价响应充电行为。MP-EVData可为负荷预测、智能充电算法以及城市基础设施规划领域的前沿研究提供关键基准数据集。