摘要
针对电力多元数据集群中由于访问集中等原因引起集群性能下降的热点问题,提出基于智能感知的电力多元数据集群热点处理方法。在电力系统中安装智能感知设备,并利用该设备采集基础电力数据。以采集并完成预处理的电力数据为基础,利用数据集成整合的方式,得到电力多元数据集群,并在其中挖掘出热点信息数据。最终通过聚类、迁移等方式,实现对电力多元数据集群中热点数据的处理。通过对比实验得出结论:设计的热点处理方法可以有效降低数据集群中的热点负载,从而提高电力数据管理的吞吐量,与传统处理方法相比,应用设计的处理方法吞吐量提高了58.33%。
In view of the hot issues of the performance degradation of power multi data cluster caused by the centralized access and other reasons,this paper proposes a hot spot processing method of power multi data cluster based on intelligent perception.The intelligent sensing device is installed in the power system and used to collect basic power data.Based on the power data collected and preprocessed,using the way of data integration and integration,the power multi data cluster is obtained,and the hot information data is mined out.Finally,through clustering,migration and other ways to achieve the hot data processing in power multi data cluster.Through the comparative experiment,the conclusion is drawn that the designed hot spot processing method can effectively reduce the hot spot load in the data cluster,so as to improve the throughput of power data management.Compared with the traditional processing method,the throughput of the designed processing method is increased by 58.33%.
作者
姜一
刘菁
姚嵘
JIANG Yi;LIU Jing;YAO Rong(State Grid Shanghai Municipal Electric Power Company,Shanghai 200120,China)
出处
《电子设计工程》
2020年第21期20-23,28,共5页
Electronic Design Engineering
基金
上海市科技计划项目(201804010204)。
关键词
智能感知
电力数据
多元数据集群
热点处理
intelligent perception
power data
multivariate data cluster
hot spot processing
