• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2014, Vol. 36 ›› Issue (07): 1417-1422.

• 论文 • 上一篇    

基于数据的流程工业关联报警的识别

张益农1,周进1,杨帆2,何自凭2,吴奕男2   

  1. (1. 北京联合大学自动化学院,北京 100101;2. 清华大学自动化系,北京 100084)
  • 收稿日期:2013-08-01 修回日期:2013-09-21 出版日期:2014-07-25 发布日期:2014-07-25
  • 基金资助:

    国家863计划资助项目(2013AA040702);北京市教委科技计划面上项目(KM201111417011);轻工过程先进控制教育部重点实验室开放课题;北京市属高等学校人才强教计划项目(PHR201008348);北京联合大学人才强校项目(BPHR2012C06)

Data-based identification of the
related alarms in the process industries     

ZHANG Yinong1,ZHOU Jin1,YANG Fan2,HE Ziping2,WU Yinan2   

  1. (1.School of Automation,Beijing Union University,Beijing 100101;
    2.Department of Automation,Tsinghua University,Beijing 100084,China)
  • Received:2013-08-01 Revised:2013-09-21 Online:2014-07-25 Published:2014-07-25

摘要:

报警在保证流程工业安全运行方面起到了重要的提示作用,但大量的报警信号也给操作员带来了困扰,使其无法抓住核心信息并作出判断,以便采取适当的行动。多个报警信号之间并非独立,而是存在有关联关系,提出了一种从历史报警数据中识别这些关系的方法。根据两个报警变量的报警发生时间,确定报警变量之间是否存在关联关系,并确定时间上的顺序和因果强度。基于每对报警变量之间的关联关系,构建多个报警之间的关联拓扑图。该方法可用于智能报警管理,为提高报警信息的指导价值提供技术支持。

关键词: 报警, 智能报警管理, 关联性, 拓扑

Abstract:

Alarming plays an important role of reminding to ensure a safety operation in the process industries. However, overwhelming alarm signals bring confusion to operators and prevent them from obtaining key information and making decisions to take proper actions. Multiple alarm signals are not independent, there exists relationship among them. A method of identifying this relationship from historical alarm data is proposed. According to the occurrence time of two alarms, the relationship between them can be determined as well as the sequence and causal intensity. Based on the relationship between each pair of alarm variables, a relationship topology can be built to describe the relationship among multiple alarms. This method can be used for smart alarm management and provide technical support to improve the guiding value of alarm signals.

Key words: alarm;smart alarm management;relationship;topology