摘要
重点研究了智能电网异常分析过程中的大数据分析功能的分层架构。通过对当前电网大数据的数据结构进行分析,同时分析了当前智能电网大数据系统的惯用分析方法,最终形成了4个智能电网大数据分析过程的分析分层架构。第一层直接层可以通过对单节点数据的分析找到单一设备的直接故障,第二层关联层可以通过对多个节点的关联性特征进行时域比较分析找到多个设备的兼容性故障,第三层特诊层可以得到智能电网中的微故障并给状态检修提出数据警告,第四层挖掘层可以通过机器学习过程得到更深入或者更微小的智能电网系统故障信息。每一层的数据分析过程都需要前一层数据分析结果作为数据支持。不同层次下可以对不同层次的故障异常进行针对性的分析处理,可以较大程度减少相关大数据系统的开发量,并实现不同系统之间的数据相互支持和数据分析结果复用。
This paper focuses on the hierarchical structure of big data analysis function in the process of smart grid anomaly analysis.By analyzing the data structure of big data in the current power grid,and the conventional analysis methods of big data system in the current smart grid,four hierarchical analysis architectures of big data analysis process in the smart grid are finally formed.Among them:the first layer can find the direct fault of a single device by analyzing the data of a single node;the second layer can find the compatibility fault of multiple devices by comparing the correlation characteristics of multiple nodes in the time domain;the third layer can get the micro fault in the smart grid and give the data alarm to the state maintenance;and the fourth layer can find the data alarm in the mining layer.Through the machine learning process,we can get more in-depth or smaller fault information of smart grid system.The data analysis process of each layer needs the data analysis results of the previous layer as data support.This paper considers that different levels of fault exceptions can be analyzed and processed in different levels,which can greatly reduce the amount of development of related big data systems,and realize data mutual support and data analysis results reuse between different systems.
作者
王振达
范晟
吴福疆
方志丹
陈爽
WANG Zhenda;FAN Sheng;WU Fujiang;FANG Zhidan;CHEN Shuang(Shantou Power Supply Bureau,Guangdong Power Grid Co.,Ltd.,Shantou 515000,China)
出处
《微型电脑应用》
2021年第2期157-160,共4页
Microcomputer Applications
关键词
大数据
智能电网
异常检测
分层架构
big data
smart grid
anomaly detection
hierarchical architecture
