基于MACD指标与聚类算法的电力用户用电行为分析  

Research on Electricity Consumption Behavior of Electric Power Users Based on MACD Index and Clustering Algorithm

作　　者：武灵耀 郭贺宏[2] 赵庆生[1] 梁定康 王旭平[1] 程昱舒[2] WU Lingyao;GUO Hehong;ZHAO Qingsheng;LIANG Dingkang;WANG Xuping;CHENG Yushu(Taiyuan University of Technology,Taiyuan,Shanxi 030024,China;State Grid Shanxi Power Electric Company,Taiyuan,Shanxi 030021,China)

机构地区：[1]太原理工大学,山西太原030024 [2]国网山西省电力公司,山西太原030021

出　　处：《上海电力大学学报》2023年第2期105-111,共7页Journal of Shanghai University of Electric Power

基　　金：国家自然科学基金青年项目(51907138)。

摘　　要：深入挖掘用户用电行为是电力大数据背景下电力市场精细化发展的迫切需求。为满足该需求,提出了一种基于平滑异同移动平均线(MACD)指标提取特征的聚类分析方法。该方法首先计算用户用电量的MACD指标;然后以MACD指标为特征,采用K-means聚类算法对用户进行分类;最后利用分析股票的思想分析每一类用户的用电行为。对美国某一地区的实测居民用电量数据进行了算例分析,结果表明所提方法与传统方法相比具有更好的聚类效果,并且拓展了用户用电行为分析方式。Digging deeper into the user's electricity consumption behavior is a urgent need for the refinement of the power market in the context of big data.To meet this need,a clustering analysis method based on moving average convergence and divergence(MACD)index extraction feature is proposed.Firstly,the MACD index of user power consumption is calculated.Then,taking MACD index as the feature,K-means clustering algorithm is used to classify users.Finally,the idea of analyzing stock is used to analyze the power consumption behavior of each type of users.Using the measured residential power consumption data in a certain area of the United States as an example analysis,the result shows that the proposed method has better clustering effect than the traditional method,and expands the analysis method of user power consumption behavior.

关 键 词：用电行为 平滑异同移动平均线 基于特征聚类 K-MEANS聚类算法 

分 类 号：TM73[电气工程—电力系统及自动化]