电工技术学报  2017, Vol. 32 Issue (9): 199-210    DOI:
电力系统及其自动化 |
ODPS平台下的电力设备监测大数据存储与并行处理方法
朱永利, 李莉, 宋亚奇, 王刘旺
华北电力大学控制与计算机工程学院 保定 071003
Storage and Parallel Processing of Big Data of Power Equipment Condition Monitoring on ODPS Platform
Zhu Yongli, Li Li, Song Yaqi, Wang Liuwang
School of Control and Computer Engineering North China Electric Power University Baoding 071003 China
全文: PDF (3837 KB)  
输出: BibTeX | EndNote (RIS)      
摘要 计算性能是制约电力大数据应用(基于大数据的故障诊断、预测等)的关键问题。利用分布式存储、并行计算加速此类数据密集型应用是目前较有效的手段。尝试利用阿里云开放数据处理服务(ODPS)存储并加速电力设备监测大数据分析过程。以变压器局部放电(PD)数据相位图谱分析(PRPD)为例,提出了适合高采样率、时序性强的局部放电信号数据存储方法。采用ODPS扩展MapReduce模型(MR2)设计了“Map-Reduce-Reduce”方式的PD信号宏观特征提取方法,提出了并行化PRPD分析算法(ODPS-PRPD),实现了大量PD信号的并行基本参数提取、统计特征计算与放电类型识别。在实验室中构造了4种放电模型并采集了大量PD信号,分别在ODPS平台上和实验室自建的Hadoop平台上进行了性能评估和成本分析。实验分析和结果表明,ODPS-PRPD将大量的中间过程数据(PD谱图数据等)一直保存在内存中,相比自建Hadoop MapReduce平台性能明显提升,并在数据可靠性、服务可用性以及成本方面具有明显优势。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
朱永利
李莉
宋亚奇
王刘旺
关键词 电力大数据公有云开放数据处理服务扩展MapReduce模型局部放电局部放电相位图谱分析    
Abstract:Computing performance is one of the key issues existing in the applications of big power data,such as fault diagnosis and prediction.Distributed storage and parallel computing are currently as the effective measures to accelerate the data-intensive applications.This paper describes an open distributed processing service(ODPS)from Ali Cloud,is used to store and accelerate the analytic process of monitoring big data about electrical equipment.Taking the phase resolved partial discharge(PRPD)processing of a partial discharge(PD)signal as example,a method for storing the signal with high sampling rate and time series data,and extracting the feature of the signal through the extended MapReduce model(MR2)of ODPS is proposed in this paper.The paralleled PRPD procedure(ODPS-PRPD)implements amounts of PD signals parallel basic parameters calculation and discharge type recognition,statistics features.To verify the effectiveness of the proposed method,a large number of partial discharge signals of four types from laboratory tests are respectively analyzed on ODPS and Hadoop.Because ODPS-PRPD stores the large amounts of middle data in the primary memory,its computing procedure is much faster.The results show that ODPS-PRPD has obviously better performance in data reliabltity,service anailabilty and cost than that of Hadoop.
Key wordsBig power data    public cloud    open distributed processing service(ODPS)    extended MapReduce model(MR2)    partial discharge    phase resolved partial discharge   
收稿日期: 2016-04-18      出版日期: 2017-05-12
PACS: TM764  
基金资助:国家自然科学基金项目(51677072)、河北省自然科学基金项目(F2014502069)和中央高校基本科研业务费专项资金(2016MS116,2016MS117)资助
通讯作者: 朱永利 男,1963年生,教授,博士生导师,研究方向为网络化监控与智能信息处理。E-mail:yonglipw@163.com   
作者简介: 李 莉 女,1980年生,博士研究生,研究方向为现代信号处理方法在电力系统故障诊断等方面的应用。E-mail:haolily12@163.com
引用本文:   
朱永利, 李莉, 宋亚奇, 王刘旺. ODPS平台下的电力设备监测大数据存储与并行处理方法[J]. 电工技术学报, 2017, 32(9): 199-210. Zhu Yongli, Li Li, Song Yaqi, Wang Liuwang. Storage and Parallel Processing of Big Data of Power Equipment Condition Monitoring on ODPS Platform. Transactions of China Electrotechnical Society, 2017, 32(9): 199-210.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/Y2017/V32/I9/199