文章摘要
吴润泽,陈文伟,唐良瑞,等.基于高风险模式树挖掘方法的电力系统风险设备集分析[J].电力系统自动化,2017,41(18):137-145. DOI: 10.7500/AEPS20161228011.
WU Runze,CHEN Wenwei,TANG Liangrui, et al.High Risk Tree Mining Method for Analysis of Power System Risk Device Set[J].Automation of Electric Power Systems,2017,41(18):137-145. DOI: 10.7500/AEPS20161228011.
基于高风险模式树挖掘方法的电力系统风险设备集分析
High Risk Tree Mining Method for Analysis of Power System Risk Device Set
DOI:10.7500/AEPS20161228011
关键词: 大数据  数据挖掘  风险影响度  高风险模式树(HRT)
KeyWords: big data  data mining  risk influence degree  high risk tree
上网日期:2017-06-27
基金项目:国家自然科学基金资助项目(51507063);国家电网公司科技项目(B34681150152)
作者单位E-mail
吴润泽 华北电力大学电气与电子工程学院, 北京市 102206  
陈文伟 北京国电通网络技术有限公司, 北京市 100070 572451718@qq.com 
唐良瑞 华北电力大学电气与电子工程学院, 北京市 102206  
范军丽 北京国电通网络技术有限公司, 北京市 100070  
摘要:
      迅速积累的调度控制系统大数据为电网设备风险分析提供了充足的条件,在分析调度控制系统大数据特性的基础上,给出了具有普遍适用性的调度控制系统大数据分析的总体架构,并针对在电网风险管控中的应用,提出一种基于高风险模式树(HRT)的高风险设备集挖掘方法。通过分析电力系统中设备的风险诱发因素,定义了设备风险影响度,用于量化设备发生告警或故障后对电网运行的影响程度,并提出设备风险影响度计算指标体系,通过融合设备故障发生频次,计算设备风险值。以设备风险值为目标进行高风险设备集挖掘,通过构建HRT,保留原始事务数据库中各设备风险值及设备风险先验知识信息,根据HRT的路径信息输出满足一定风险阈值的高风险设备集。以调度控制系统的海量历史告警数据为基础进行了仿真,结果表明,HRT可以在告警数据中迅速挖掘出满足条件的高风险设备集,并且能够反映出高风险设备组合之间存在的潜在关联性。
Abstract:
      The rapid accumulation of big data in the power grid dispatching and control system provides a sufficient condition for the risk analysis of power grid equipment. Based on the analysis of characteristics of big data in the dispatching and control system, a universal analysis framework of big data in power dispatching and control system is built, and on basis of the application of risk management and control in power grid, a data mining method for high risk equipment based on high risk tree(HRT)is proposed. From the perspective of multiple factor analysis of equipment risk, considering impact of equipment importance and equipment hidden danger on equipment risk, equipment risk influence degree is defined, equipment importance indicators and equipment hidden danger indicators are proposed. The equipment risk value as the target of mining high risk equipment, the construction of HRT retaining the original database of equipment risk value and equipment risk prior knowledge information is for mining high risk equipment set, specifically the high risk equipment risk set meeting the threshold will be obtained according to the HRT path information. Finally, the proposed method is simulated based on massive historical alarm data in dispatching and control system. The simulation results show that HRT can quickly deal with the alarm data to get high risk equipment set meeting the conditions, and it can reflect the potential association between high risk equipment, so it will provide reference for the follow-up of power grid risk management and control.
查看全文(Free!)   查看附录   查看/发表评论  下载PDF阅读器