Novel domestic building energy consumption dataset: 1D timeseries and 2D Gramian Angular Fields representation

This dataset, collected in 2022 in a domestic household in the UK, provides appliance-level power consumption data and ambient environmental conditions as a timeseries and as a collection of 2D images created using Gramian Angular Fields (GAF). The importance of the dataset lies in (a) providing the research community with a dataset that combines appliance-level data coupled with important contextual information for the surrounding environment; (b) presents energy data summaries as 2D images to help obtain novel insights from the data using data visualization and Machine Learning (ML). The methodology involves installing smart plugs to a number of domestic appliances, environmental and occupancy sensors, and connecting the plus and the sensors to a High-Performance Edge Computing (HPEC) system to privately store, pre-process, and post-process data. The heterogenous data include several parameters, including power consumption (W), voltage (V), current (A), ambient indoor temperature (°C), relative indoor humidity (RH%), and occupancy (binary). The dataset also includes outdoor weather conditions based on data from MET Norway including temperature (°C), outdoor humidity (RH%), barometric pressure (hPA), wind bearing (deg), and windspeed (m/s). This dataset is valuable for energy efficiency researchers, electrical engineers, and computer scientists to develop, validate, and deploy and computer vision and data-driven energy efficiency systems.

本数据集于2022年在英国一户家庭中采集，提供了设备级功耗数据与周边环境条件的时序数据，以及通过格拉姆角场（Gramian Angular Fields, GAF）生成的二维图像集。该数据集的价值在于：其一，为研究社区提供了兼具设备级能耗数据与周边环境关键上下文信息的数据集；其二，将能耗数据以二维图像形式进行摘要呈现，可借助数据可视化与机器学习（Machine Learning, ML）技术从数据中挖掘全新研究洞察。本数据集的采集流程为：为多款家用设备、环境传感器与人员占用传感器加装智能插头，并将插头与传感器连接至高性能边缘计算（High-Performance Edge Computing, HPEC）系统，以实现数据的私有存储、预处理与后处理。该异构数据集包含多项参数，涵盖功耗（W）、电压（V）、电流（A）、室内环境温度（°C）、室内相对湿度（RH%）、人员占用状态（二进制）。此外，数据集还包含基于挪威气象研究所（MET Norway）公开数据得到的室外气象信息，具体包括室外温度（°C）、室外湿度（RH%）、大气压强（hPA）、风向角度（deg）与风速（m/s）。本数据集可助力能效研究人员、电气工程师与计算机科学家开发、验证并部署计算机视觉与数据驱动的能效系统。