基于大数据的变压器油色谱分析可视化诊断平台设计  

Design of Visual Diagnosis Platform for Transformer OilChromatography Analysis Based on Big Data

作　　者：郑博文 王秋雯 牛斌 问尤龙 ZHENG Bo-wen;WANG Qiu-wen;NIU Bin;WEN You-long(Yuxi Power Supply Bureau of Yunnan Power Grid Co.,Ltd.Yuxi Yunnan 653100,China;Kunming Nengxun Technology Co.,Ltd.,Kunming Yunnan 650217,Ghina)

机构地区：[1]云南电网有限责任公司玉溪供电局,云南玉溪653100 [2]昆明能讯科技有限责任公司,云南昆明650217

出　　处：《粘接》2020年第5期99-102,共4页Adhesion

摘　　要：针对当前电力服务的智能化需要,结合当前大数据技术和可视化技术,构建变压器油色谱分析的大数据诊断平台。为实现该平台,首先采用Sqoop和HDFS完成数据的采集与分布式存储;然后通过Storm+CUDA计算架构完成对数据的并行化计算,并引入BP、智能算法;采用Three.js完成数据的三维化展示。结果表明,通过该方案,可实时展示各个区域变压器的油色谱信息,并具有较高的计算效率,为实现玉溪供电局大数据与可视化的深入融合提供了借鉴。According to the intelligence needs of the current power service,combined with the current big data technology and visualization technology,a big data diagnosis platform for transformer oil chromatography analy⁃sis is built.To realize the platform,Sqoop and HDFS are used to collect and store data.Then,the parallel computa⁃tion of data is completed through the Storm+CUDA computing architecture,and introduce BP and intelligent al⁃gorithms.Finally,the regional power supply reliability calculation index and intelligent display screen are set.Fi⁃nally,Three.Js is used to complete the three-dimensional display of the data.The results show that,through this scheme,the oil chromatographic information of transformers in various regions can be displayed in real time,and it has high calculation efficiency,which provides a reference for the in-depth integration of big data and visual⁃ization of Yuxi Power Supply Bureau.

关 键 词：Three.js Storm+CUDA架构 HDFS分布式文件 

分 类 号：O657[理学—分析化学]