Premium
Ethylene compressor monitoring using model‐based PCA
Author(s) -
Rotem Y.,
Wachs A.,
Lewin D. R.
Publication year - 2000
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.690460911
Subject(s) - principal component analysis , fault detection and isolation , multivariate statistics , variance (accounting) , parametric statistics , process (computing) , statistics , forcing (mathematics) , gas compressor , resolution (logic) , computer science , mathematics , data mining , engineering , artificial intelligence , mechanical engineering , business , accounting , actuator , operating system , mathematical analysis
Principal component analysis (PCA), which is widely used in process monitoring, performs best when the system variables are linearly correlated. In practice, however the variables are often nonlinearly related and may be subject to periodic forcing, both of which compromise the performance of conventional PCA. In model‐based PCA (MBPCA), multivariate statistics are used to analyze the portion of the observed variance that cannot be predicted using a model of the process and thus significantly enhances the attainable diagnostic resolution. Here, MBPCA is used for fault‐detection monitoring of an ethylene compressor, which operates under a significant periodic disturbance caused by the ambient temperature. An analytical expression is derived to predict the limits of identifiable faults given bounds on the parametric model uncertainty.