Premium
Improved PCA methods for process disturbance and failure identification
Author(s) -
Wachs Amir,
Lewin Daniel R.
Publication year - 1999
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.690450808
Subject(s) - disturbance (geology) , principal component analysis , identification (biology) , process (computing) , computer science , control theory (sociology) , mathematics , data mining , artificial intelligence , control (management) , paleontology , botany , biology , operating system
Principal component analysis (PCA) is a powerful technique for constructing reduced‐order models based on process measurements, obtained by the rotation of the measurement space. These models can be subsequently utilized for chemical‐process monitoring, particularly for disturbance and failure diagnosis. Since the standard PCA procedure does not account for the time‐dependent relationships among the process variables, this leads to poorer disturbance isolation capability in dynamic applications. A simple idea, in which the last s PCA scores are recursively summed and used to construct descriptive statistics for process monitoring, is presented. Analytically, it is shown that the disturbance resolution afforded is enhanced as a result. Resolution is improved further through the use of an algorithm that enhances the correlations between the input and output variables through optimal time shifting. An overall strategy for on‐line monitoring developed includes disturbance identification through mapping. The approach is demonstrated by two industrially relevant case studies.