流程工业仪表故障快速检测方法  

Rapid Sensor Fault Detection in Process Industry

作　　者：白小琴[1] 戴连奎[1] 

机构地区：[1]浙江大学工业控制技术国家重点实验室,杭州310027

出　　处：《化工自动化及仪表》2008年第1期36-39,共4页Control and Instruments in Chemical Industry

基　　金：国家高技术研究发展计划项目(2006AA04Z169)

摘　　要：通过对大量真实工业数据的统计分析,发现仪表测量信号变化量的方差能很好地表征仪表性能的稳定状态,由此提出了一种基于信号方差分析的快速仪表故障检测算法,并在此基础上开发了一套在线仪表故障检测系统。该系统以诊断数据库为核心,包括实时数据采集、仪表故障在线诊断、故障数据的保存与查询、诊断参数自学习以及远程访问功能。工业应用结果表明,这种故障检测方法能够对仪表早期的故障作出正确的判断,且检测迅速,物理意义明确,适用于大规模的流程工业过程。According to statistic analysis of large real industry data, the variance of measured data can be a good characterization of the sensor＇ s steady-state performance. So a rapid sensor fault detection algorithm was presented, and on the basis of the algorithm developed an online sensor fault detection system. Detection database was the core of the system,which included real-time data acquisition, online fault detection, preservation and query of fault data, self-study of detection parameters, remote access and so on. The application results in actual industrial process show that this fault detection method can rapidly detect early fault with clear physical meaning. It applies to large-scale industrial processes.

关 键 词：流程工业 仪表故障检测 方差分析 

分 类 号：TH86[机械工程—仪器科学与技术]