Premium
Operating performance assessment and non‐optimal cause identification for chemical process
Author(s) -
Tao Yang,
Shi Hongbo,
Song Bing,
Tan Shuai
Publication year - 2019
Publication title -
the canadian journal of chemical engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.404
H-Index - 67
eISSN - 1939-019X
pISSN - 0008-4034
DOI - 10.1002/cjce.23401
Subject(s) - principal component analysis , variance (accounting) , identification (biology) , process (computing) , data mining , similarity (geometry) , feature (linguistics) , computer science , plot (graphics) , pattern recognition (psychology) , artificial intelligence , machine learning , mathematics , statistics , linguistics , philosophy , botany , accounting , business , image (mathematics) , biology , operating system
In this paper, a novel operating performance assessment and non‐optimal cause identification approach is proposed for chemical process. Firstly, a modified method that contains both temporal and variance information is developed for grade classification. These obtained grades are divided into three levels and each level would have its specific processing scheme. Secondly, a new weighted principal component analysis (WPCA) is proposed and multiple assessment models are established. WPCA selects the principal components that contain feature information of each grade, rather than based on the variance contribution. Then a similarity index is constructed to assess the performance online. After that, a weighted contribution plot that can highlight the non‐optimal responsible variables is presented. Finally, the efficiency of the proposed method is demonstrated through the Tennessee Eastman process.