Smart Meter Data-Driven Evaluation of Operational Demand Response Potential of Residential Air Conditioning Loads

The ground truth data used in this paper are obtained from three different areas to verify the effectiveness and robustness of the proposed methods. The detailed data information is described as follows: 1) The ground truth data from the Pecan Street dataset are collected from real households in the Muller project in Austin, TX, USA. Muller project funded by the U.S. Department of Energy and the U.S. National Science Foundation are located on the site of the Austin’s former municipal airport, close to central Austin.The selected homes in the project received monitoring equipment that captures electricity use on less than or equal to 1 min intervals for the whole home and 6 to 22 major appliances. Data over one year from August 2015 to July 2016 are analyzed, which contain the application-level and the whole-house energy consumption data. The corresponding 1-hour level temperature data are collected from the nearest Mueller weather station. We down-sample the energy consumption data to the 1-hour level to maintain consistency with the resolution of the temperature data. After data cleaning, customers without air conditioners or with missing readings are omitted, and the data of 119 residential customers are selected for accuracy analysis. 2) The ground truth data from smart home dataset are collected from real households in the Smart Home project in the Western Massachusetts, USA. The goal of this project is to optimize home energy consumption. The project involves several different types of dataset, including apartment dataset of 114 single-family, home dataset of 7 household and solar panel dataset, etc. However, the apartment dataset only contains the aggregated electrical data which can not be used to verify the accuracy of the load decomposition. Therefore, data over one year from January 2016 to December 2016 of home B and home G with individual ACLs monitor are selected for robustness analysis. 3) The ground truth data from low voltage distribution area are collected from low voltage distribution boxes in a developed city, Jiangsu province, China. Power and corresponding temperature data over one year from 2017 to 2018 are used for local DR programs. The dataset involves different distribution areas (i.e., different aggregated DR customers), including garment factory, hotel, rural neighborhoods, etc. However, the sub-meter data of all the ACLs are unavailable, thus it will only be used for aggregated DR potential analysis.

本文所用的真实基准数据源自三个不同领域，用于验证所提方法的有效性与鲁棒性。详细数据信息说明如下：
1) 佩坎街数据集（Pecan Street dataset）的真实基准数据，采集自美国德克萨斯州奥斯汀市穆勒项目（Muller project）中的真实住户。该项目由美国能源部与美国国家科学基金会资助，选址于奥斯汀前市政机场旧址，紧邻奥斯汀市中心。项目中选定的住宅搭载了监测设备，可采集全屋用电数据及6至22台主要家电的用电数据，采样间隔不超过1分钟。本次分析使用2015年8月至2016年7月共计一年的数据集，涵盖应用级与全屋能耗数据。对应的1小时分辨率气温数据采集自最近的Mueller气象站。我们将能耗数据下采样至1小时分辨率，以匹配气温数据的时间粒度。经数据清洗后，剔除了未安装空调或存在读数缺失的住户，最终选取119户居民的数据集用于精度分析。
2) 智能家居数据集（Smart Home dataset）的真实基准数据，采集自美国马萨诸塞州西部智能家居项目中的真实住户。该项目旨在优化家庭能耗管理，涵盖多类不同数据集，包括114套独栋住宅公寓数据集、7户家庭数据集以及太阳能板数据集等。但其中的公寓数据集仅包含聚合用电数据，无法用于验证负荷分解的精度。因此本次鲁棒性分析选取了搭载独立ACL监测设备的B户与G户2016年1月至2016年12月共计一年的数据集。
3) 低压配电区域数据集的真实基准数据，采集自中国江苏省某发达城市的低压配电箱。本次使用2017年至2018年共计一年的用电与对应气温数据，用于本地需求响应（Demand Response, DR）项目研究。该数据集涵盖不同配电区域（即不同聚合型需求响应用户），包括服装厂、酒店、乡村居民区等。但由于所有ACL的分户计量数据不可用，因此该数据集仅可用于聚合型需求响应潜力分析。

搜集汇总

背景与挑战

背景概述

该数据集包含来自美国和中国三个不同地区的住宅空调负荷智能电表数据，用于评估操作需求响应潜力。数据涵盖家庭和主要电器的用电量及温度信息，时间跨度从2015年至2018年，旨在验证方法的有效性和鲁棒性。