摘要
随着我国配电网信息数据的不断增加,传统方法已经难以及时发现配电网运维中出现的问题。为了深入挖掘配电网运维相关数据,找出配电网运维过程中的隐含规律,进而提升我国配电网的运维水平,提出将大数据挖掘技术、人工智能与雷达图相结合,进行配电网运维的共性挖掘及定量评价。分析了K-means聚类分析法与雷达图,并创新性地引人数据密度思想,进行K-means聚类分析法的改进;并将其与雷达图结合,实现对配电网运维的分析。研究结果表明,改进后的综合评价法能够对配电网运维中的问题进行检测,还能够定量地对配电网运维情况进行可视化评价。该研究能为配电网运维相关风险管理的研究提供参考。
With the increasing information data of distribution network in China,it is difficult to detect the problems in the operation and maintenance of distribution network in time by traditional methods.In order to deeply mine the data related to distribution network operation and maintenance,find out the hidden rules in the process of distribution network operation and maintenance,and then improve the operation and maintenance level of distribution network in China,the combination of big data mining technology,artificial intelligence and radar chart is proposed to mine the commonness of distribution network operation and maintenance,and quantitatively evaluate distribution network operation and maintenance.K-means clustering analysis method and radar chart are studied and analyzed.The idea of data density is innovatively introduced to improve the K-means clustering analysis method,and it is combined with radar chart to realize the analysis of distribution network operation and maintenance.The results show that the improved comprehensive evaluation method can detect the problems in the operation and maintenance of distribution network,and can also quantitatively evaluate the operation and maintenance of distribution network visually.This study can provide some reference for the research of risk management related to distribution network operation and maintenance.
作者
冯建伟
王联智
谢敏
吴海杰
孟超
FENG Jianwei;WANG Lianzhi;XIE Min;WU Haijie;MENG Chao(China Southern Power Grid Hainan Digital Power Grid Research Institute Co.,Ltd.,Haikou 570100,China)
出处
《自动化仪表》
CAS
2021年第7期42-46,共5页
Process Automation Instrumentation
关键词
雷达图
K-means聚类分析法
配电网
运维
数据密度
综合评价
大数据
可视化
轮廓系数
Radar chart
K-means clustering analysis
Distribution network
Operation and maintenance
Data density
Comprehensive evaluation
Big data
Visualization
Contour coefficient