摘要
主元分析 (PCA)是一种有效的多元统计过程监测方法。PCA监测方法不依赖于过程的精确数学模型 ,这使得其难以对故障的可检测性问题进行系统的研究。基于故障子空间的描述方式 ,本文在主元子空间和残差子空间中分别讨论了 PCA故障可检测性的充分和必要条件 ,并提出了临界故障幅值的概念。通过对双效蒸发过程的仿真故障检测 ,表明所获得的结果能较好地刻画 PCA的故障检测行为。
Principal component analysis (PCA) is an effective multivariate statistical process monitoring approach and substantial industrial applications have been reported in recent years. The most significant advantage of PCA is that no precise process model is needed. Nevertheless, PCA based process monitoring methods show difficulties in systematically analyzing the issue of fault detectability. Based on the fault description method of fault subspace, the sufficient and necessary conditions of fault detectability in the principal component (PC) space and residual space are presented respectively. A conception of critical fault magnitude is introduced and used to analyze the detected behavior of faults. The acquired results are then illustrated and verified by fault detection examples of a simulated double effective evaporator.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2002年第3期232-235,240,共5页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金 (2 0 0 0 760 4 0 )
青岛市工业信息化技术重点实验室资助项目
