摘要
随着我国电网智能化水平提高以及智能量测终端设备的普及,为电力企业带来了海量的用户侧用电数据。为增强电力企业对用户的了解,基于数据挖掘技术提出一种考虑负荷季节特性的电力用户用电行为画像的方法。首先,对原始负荷数据进行数据清洗和预处理,并利用方差过滤和特征过滤进行特征筛选;然后,根据季节性基础负荷相互独立的特点,将处理后的负荷数据分解成季节性基础负荷和受其他因素影响的敏感负荷;其次,分别对基础负荷和敏感负荷的相关性系数依次聚类分析,得到双重聚类标签结果,最后得到两类标签形成的用户用电行为画像。在算例分析部分利用30个电力用户负荷数据验证了所提出用户画像方案的可行性。
With the continuous improvement of the intelligent level of my country′s power grid,the popularization of intelligent measurement terminal equipment has brought massive amounts of user-side electricity consumption data to electric power companies.In order to enhance the power companies′understanding of customers,based on data mining technology,a method of power users′behavioral portraits considering the characteristics of load seasons is proposed.Firstly,the original load data is cleaned and preprocessed,and the variance filter and feature filter are used for feature screening.Then,according to the characteristic of independent seasonal typical load,the processed load data are decomposed into seasonal basic load and sensitive load affected by other factors.Secondly,the correlation coefficients of the load characteristics of the base load and the sensitive load are clustered in turn to obtain the results of double clustering labels,and the customer′s electricity behavior portraits formed by the two types of labels are established.Finally,load data of 30 customers is utilized to verify the feasibility of the proposed customer profile method in the example analysis part.
作者
万伟
刘红旗
杜单单
郭航源
甄颖
李英超
孙伟卿
WAN Wei;LIU Hongqi;DU Dandan;GUO Hangyuan;ZHEN Ying;LI Yingchao;SUN Weiqing(Heze Power Supply Company of State Grid Shandong Electric Power Company,Heze 274000,China;Department of Electrical Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《哈尔滨理工大学学报》
CAS
北大核心
2023年第3期45-55,共11页
Journal of Harbin University of Science and Technology
基金
国家自然科学基金(5177726)
国网山东省电力公司科技项目(2020A-061)。
关键词
数据挖掘
用户画像
数据分析
负荷特性
季节性基础负荷
敏感负荷
聚类分析
data mining
customer profile
data analysis
load characteristics
seasonal typical load
sensitive load
clustering analysis
