| 基于异常点检测和改进K-means算法的台区用户相别辨识方法 |
| |
| 引用本文： | 张 然,孙晓璐,何仲潇,薛莉思,陈维民,徐严军,连利波.基于异常点检测和改进K-means算法的台区用户相别辨识方法[J].陕西电力,2020,0(1):91-96. |
| |
| 作者姓名： | 张 然 孙晓璐 何仲潇 薛莉思 陈维民 徐严军 连利波 |
| |
| 作者单位： | （1. 国网四川省电力公司计量中心，四川 成都 610000；2. 清华四川能源互联网研究院，四川 成都 610000） |
| |
| 摘 要： | 解决配电台区用户线变不匹配问题是推进配电网智能化管理的关键一步。大数据技术的快速普及为实现低成本、高效率的台区用户相别辨识提供了可能。提出了基于异常点检测和改进K-means算法的台区用户相别辨识方法。首先通过局部因子算法对聚类分析数据进行预处理，剔除不属于待分析台区的用户数据。然后，根据实际应用场景特点对K-means算法进行改进，包括确定聚类个数、初始质心，并选用相关系数作为评估样本相似度的指标。最后利用改进的K-means算法对预处理后的数据进行聚类分析，实现低压台区用户相别的精准辨识。算例分析表明，所提方法能够有效提升用户辨识准确率，且在不同的数据环境中可保持较高的稳定性。
|
| 关 键 词： | 配电网 台区 相别辨识 局部异常因子算法 改进K-means |
Phase Identification Method for Distribution Area Users Based on Outlier Detection and Improved K-means Algorithm |
| |
| ZHANG Ran,SUN Xiaolu,HE Zhongxiao,XUE Lisi,CHEN Weimin,XU Yanjun,LIAN Libo.Phase Identification Method for Distribution Area Users Based on Outlier Detection and Improved K-means Algorithm[J].Shanxi Electric Power,2020,0(1):91-96. |
| |
| Authors: | ZHANG Ran SUN Xiaolu HE Zhongxiao XUE Lisi CHEN Weimin XU Yanjun LIAN Libo |
| |
| Affiliation: | （1. State Grid Sichuan Electric Power Corporation Metering Center，Chengdu 610000，China；2. Tsinghua Sichuan Energy Internet Research Institute，Chengdu 610000，China） |
| |
| Abstract: | Solving the problem of the mismatch of subscriber line in distribution area is a key step to promote the intelligent management of distribution network. The rapid popularization of big data technology makes it possible to realize distribution area user identification with low-cost and high-efficiency. The paper proposes a phase identification method for the users based on outlier detection and improved K-means algorithm. Firstly， the data of clustering analysis is preprocessed by local factor algorithm， and the user data that aren’t part of the area data to be analyzed is deleted. Then， K-means algorithm is improved according to the characteristics of practical application scenarios， determining the number of clustering， initial centroid and selecting correlation coefficient as the index of evaluating sample similarity. Finally， the improved K-means algorithm is used to cluster the pre-processed data to realize the accurate identification of the users in low voltage station area. Examples show that the proposed method can ef-fectively improve the accuracy of the user identification and maintain high stability in different data environments. |
| |
| Keywords: | distribution network station area phase identification local outlier factor algorithm improved K-means |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《陕西电力》浏览原始摘要信息 |
|
点击此处可从《陕西电力》下载全文 |
