Electric Vehicle Charging Dataset

This electric vehicle (EV) charging dataset includes the EV connection time, charging duration, energy consumption, and day number each corresponding respectively to connectionTime_decimal, chargingDuration, kWhDelivered, dayIndicator columns in the dataset for an EV charging parking lot. The data is generated using conditional tabular generative adversarial networks (CTGAN) and kernel density estimation (KDE) from the Caltech dataset to maintain a realistic load profile, accurately model EV owner behaviors, and preserve the relationship between the columns of the dataset. This dataset includes EV data for 29,600 days while the original Caltech dataset includes data for only 185 days. This data can be useful in training machine learning algorithms, specifically, reinforcement learning algorithms. The connection time range is 0-24. The unit for charging duration is hour and energy consumption is in kWh.

本电动汽车（EV）充电数据集涵盖电动汽车连接时间、充电时长、能耗及对应日期编号四类信息，分别对应该电动汽车充电停车场数据集中的connectionTime_decimal、chargingDuration、kWhDelivered、dayIndicator字段。本数据集基于加州理工学院（Caltech）数据集，通过条件表格生成对抗网络（conditional tabular generative adversarial networks，CTGAN）与核密度估计（kernel density estimation，KDE）生成，旨在保留真实的负荷曲线、精准建模电动汽车车主的充电行为，并维持数据集各字段间的关联关系。本数据集包含29600天的电动汽车充电数据，而原始加州理工学院数据集仅涵盖185天的数据。该数据集可用于训练机器学习算法，尤其适用于强化学习算法的训练。连接时间的取值范围为0至24，充电时长的单位为小时，能耗单位为千瓦时（kWh）。

搜集汇总

背景与挑战

背景概述

该数据集是一个电动汽车充电数据集，通过生成对抗网络和核密度估计从Caltech数据集扩展生成，模拟真实充电行为，包含连接时间、充电时长和能耗等关键列。数据规模扩展到29,600天，适用于机器学习特别是强化学习算法的训练，具有较高的实用性和研究价值。