摘要
在泛在电力物联网的建设中,电力企业针对客户的用电行为进行分析是必不可少的。在以往的研究中,k均值聚类算法是常用的客户用电行为分析方法之一,然而由于初始质心采用随机选择的方式,使得其容易陷入局部最优且难以收敛到全局最小值。针对该问题,提出了基于改进的动态粒子群算法优化的K-means算法(DPSO-Kmeans),并将其用于客户用电行为的分析中。在实验中,通过对312个家庭用户的用电消费行为记录进行聚类分析,结果证明DPSO-Kmeans相对于传统的K-means算法具有更好的聚类效果,能够提取更为典型的客户用电行为模式。
In the construction of the ubiquitous power Internet of Things,it is indispensable to analyze customers′electricity consumption behavior for power companies.In previous studies,the K-means clustering algorithm is one of the commonly used methods for analyzing customer electricity consumption behavior.However,because the initial centroid is randomly selected,it is easy to fall into a local optimum and difficult to converge to a global minimum.To this problem,an improved K-means algorithm(DPSO-Kmeans)based on an improved dynamic particle swarm optimization algorithm is proposed and used in the analysis of customers′electricity consumption behavior.In the experiment,the electricity consumption behavior records of 312 household users were used for cluster analysis.The results prove that DPSO-Kmeans has a better clustering effect than the traditional K-means algorithm,and can extract more typical customers′electrical behavior pattern.
作者
王莹
项雯
张群
高秀云
WANG Ying;XIANG Wen;ZHANG Qun;GAO Xiuyun(College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin 150001, China;College of Electrical and Information, Northeast Agriculture University, Harbin, 150038, China;Economic and Technological Research Institute of State Grid, Heilongjiang Electric Power Co., Ltd., Harbin 150036,China)
出处
《哈尔滨理工大学学报》
CAS
北大核心
2022年第2期106-113,共8页
Journal of Harbin University of Science and Technology
基金
国家自然科学基金(61633008,51609046)
黑龙江省科学基金(F2015035).
关键词
用电行为分析
K-MEANS聚类算法
初始质心
动态粒子群算法
用电行为模式
Analysis of electricity consumption
K-means clustering algorithm
initial centroid
dynamic particle swarm algorithm
electricity usage behavior model
