基于大数据的输电线路故障预警模型设计  

Design of Fault Warning Model of Transmission Line Based on Big Data

作　　者：郑茂然[1] 余江[1] 陈宏山[1] 高宏慧[1] 张静伟[1] 吕梁 刘智勇 ZHENG Maoran YU Jiang CHEN Hongshan GAO Honghui ZHANG Jingwei LU Liang LIU Zhiyong(Power Dispatching Control Center, CSG, Guangzhou 510663, China Dareway Software Co. , Ltd. , Jinan 250100, China)

机构地区：[1]中国南方电网有限责任公司电力调度控制中心,广州510623 [2]山大地纬软件股份有限公司,济南250100

出　　处：《南方电网技术》2017年第4期30-37,共8页Southern Power System Technology

摘　　要：发展输电线路远程故障诊断是建设智能电网的重点内容。针对输电线路的上传数据单一、数量多、利用率低等问题,基于背景平台提供的大数据处理集群技术,以朴素贝叶斯算法结合时间序列相似性故障匹配建立输电线路故障预警模型。模型采用采集到的相关电气量、开关量、事件顺序信息、电网拓扑等数据以及输电线路故障过程中故障录波装置所获取的数据,采用朴素贝叶斯算法挖掘潜在故障发生因素(即故障因子)的发生指数,随后结合时间序列相似性故障匹配对输电线路进行故障预警。案例分析表明,所构建的模型能较好地挖掘出故障因子,预测结果较好。The development of remote fault diagnosis of transmission line is one of the focuses of the construction of smart grid. To solve the problems of homogeneous uploaded data, enormous amount of data, and poor data utilization, naive Bayes algorithm is used together with similarity fault matching by time sequence to establish a transmission line fault warning model, based on the big data pro- cessing cluster technology provided by the background platform. The model uses collected data of electrical and switching information, event sequence information, power network topology, and data acquired by fault recorders during transmission line failure, and applies naive Bayes algorithm to excavate the occurrence factor of potential fault （i. e. , fault factor）, then, by cooperating with similarity fault matching by time sequence, the fault of transmission line can be forecasted. Case analysis shows that the model can excavate fault factors correctly and the forecast result is satisfactory.

关 键 词：输电线路 大数据 朴素贝叶斯算法 时间序列相似性匹配 故障预警模型 

分 类 号：TM76[电气工程—电力系统及自动化]