Premium
Dissimilarity analysis based batch process monitoring using moving windows
Author(s) -
Zhao Chunhui,
Wang Fuli,
Jia Mingxing
Publication year - 2007
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.11164
Subject(s) - batch processing , process (computing) , computer science , series (stratigraphy) , fault detection and isolation , index of dissimilarity , variable (mathematics) , data mining , process engineering , engineering , mathematics , artificial intelligence , statistics , paleontology , mathematical analysis , actuator , biology , programming language , operating system
In recent years, a novel MSPC method known as DISSIM, which is based on the dissimilarity of process time series data, has been developed focusing on continuous and stable processes. However, its further application is hindered because of its unsuitability to batch processes which lie at the heart of many industries. In the present work, combined with variable moving windows, an important extension of the novel DISSIM method, termed EDISSIM, is made for its practical application to batch processes monitoring as well as the theoretical analysis basis for the determination of control limit. Moreover, contribution plots of dissimilarity index are used to identify the variables that contribute significantly to the out‐of‐control state. The applications of the proposed EDISSIM method are illustrated with respect to simulated data collected from both a simple 2 × 2 numerical process and a fed‐batch penicillin fermentation industrial batch process. The results clearly show that it functions very well to successfully detect and diagnose fault in batch processes, which implies the significant investigation potential of the proposed method. © 2007 American Institute of Chemical Engineers AIChE J, 2007