家庭电器设备电力负载感知数据集

目前存在的两种类型的数据集可以分为两种类型：实验室测量数据集和来自真实世界环境的数据集。一般来说，现有的实验室测量数据提供了单个设备的数据；这些方法对总体基准测试的用途有限。多个设备同时活动的测量，只存在于真实场景数据集中。
目前已有的真实场景数据集数据量大，但实际发生的事件(设备改变状态)较少。Home Equipment Laboratory Dataset（HELD1家庭电器设备电力负载感知数据集有多个设备单独和/或同时运行的多个开启或关闭事件。由于单个设备可以单独控制，因此参考数据的质量非常高。因此，可以在很短的测量时间内产生大量的事件。此外，该数据集包含不同数量的设备的不同复杂场景。该数据集的目的是提供一个比提高个别NILM方法之间的可比性更好的基础。

Two primary categories of datasets are currently available: laboratory measurement datasets and real-world datasets collected from actual environments. Generally, existing laboratory measurement datasets only provide data from single devices, which have limited utility for overall benchmarking. Measurements involving multiple active devices simultaneously are exclusive to real-world scene datasets.

Existing real-world scene datasets have large data volumes but relatively few actual events (i.e., device state changes). The Home Equipment Laboratory Dataset (HELD1, a power load monitoring dataset for household electrical appliances) contains multiple turn-on and turn-off events where individual devices operate either alone or concurrently. Since individual devices can be controlled independently, the quality of the reference data is exceptionally high. As a result, a large number of events can be generated within a very short measurement period. Additionally, this dataset includes diverse complex scenarios with varying numbers of devices. The purpose of this dataset is to provide a superior foundation for enhancing the comparability between individual NILM methods.

搜集汇总

背景与挑战

背景概述

家庭电器设备电力负载感知数据集是一个包含多个设备单独或同时运行的电力负载数据的高质量数据集，旨在提高非侵入性负荷监测方法的可比性。数据集包含42.8GB的数据，61个文件，适用于电力感知、电器识别等计算机应用领域。