摘要
针对复杂过程数据的非线性、动态性和中心漂移等特征,提出了基于时空近邻标准化和局部离群因子的故障检测方法(TSNS–LOF).首先使用训练样本在时空两个方向上的近邻集来标准化训练样本;然后在标准样本集上计算样本的局部离群因子,并确定其上分位点作为检测控制限,进行在线故障检测.时空近邻标准化解决了复杂过程数据的非线性、动态性和中心漂移的问题;局部离群因子通过度量样本的相似度实现了故障样本和正常样本的分离.将TSNS–LOF应用于田纳西–伊斯曼过程(TE)过程进行故障检测实验,结果表明相对于主元分析、动态主元分析、k近邻、局部离群因子等方法, TSNS–LOF对故障预警更加及时且具有更高的故障检测率.理论分析和仿真实验说明TSNS–LOF方法适用于具有动态性或多模态特性或两者兼具的过程故障检测,能够更好地保障生产过程的安全性和产品的高质量.
A fault detection method based on time-space nearest neighborhood standardization and local outlier factor(TSNS–LOF) was proposed to deal with the problem of nonlinear, dynamic and mean drift of complex process data. Firstly,training samples are normalized by using the time-space nearest neighborhood set;then the local outlier factor of samples are calculated on standard sample set, and the upper quantile is determined as the detection control limit, and the online fault detection is performed. The time-space nearest neighborhood standardization overcomes the difficulties of the nonlinearity,dynamics and mean drift. The local outlier factor measures the similarity of samples, to separate the fault samples and the normal samples. The fault detection experiment of Tennessee Eastman Process was carried out. The results showed that TSNS–LOF is timelier for the early fault warning, and has higher detection rate than principal component analysis,dynamic principal component analysis, k nearest neighbor rule and local outlier factor methods. The theoretical analysis and simulation experiments showed that the TSNS–LOF method is suitable for fault detection of dynamics or multiple or both operating faults and ensures the safety of the production process and high quality of products.
作者
冯立伟
李元
张成
谢彦红
FENG Li-wei;LI Yuan;ZHANG Cheng;XIE Yan-hong(College of Science,Shenyang University of Chemical Technology,Shenyang Liaoning 110142,China;Research Center for Technical Process Fault Diagnosis and Safety,Shenyang University of Chemical Technology,Shenyang Liaoning 110142,China)
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2020年第3期651-657,共7页
Control Theory & Applications
基金
国家自然科学基金重点项目(61490701)
国家自然科学基金项目(61673279)资助.
关键词
时空近邻标准化
局部离群因子
模型
主元分析
过程控制
time-space neighborhood standardization
local outlier factor
model
principal component analysis
process control
