Premium
Development of high predictive soft sensor method and the application to industrial polymer processes
Author(s) -
Kaneko Hiromasa,
Funatsu Kimito
Publication year - 2012
Publication title -
asia‐pacific journal of chemical engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.348
H-Index - 35
eISSN - 1932-2143
pISSN - 1932-2135
DOI - 10.1002/apj.631
Subject(s) - soft sensor , nonlinear system , process (computing) , state variable , computer science , construct (python library) , variable (mathematics) , regression analysis , variables , linear regression , econometrics , mathematics , machine learning , mathematical analysis , physics , quantum mechanics , thermodynamics , programming language , operating system
In industrial plants, soft sensors are widely used to estimate process variables that are difficult to measure online. One of the problems of soft sensors is that their predictive accuracy gradually decreases with changes in the state of the plants. Although regression models are reconstructed with new data to solve this problem, some problems remain in practice. Hence, it is attempted to construct soft sensor models based upon the time difference of an objective variable and that of explanatory variables for reducing the effects of deterioration with age such as the drift and gradual changes in the state of plants. However, the time difference model cannot account for the nonlinearity in process variables. Therefore, to consider the nonlinearity and the effects of changes with age, we have proposed to construct time difference models after modeling nonlinear relationship between and among process variables. Variables obtained by physical models or those calculated by statistical nonlinear regression methods are used to consider the nonlinearity, and then, a time difference model is constructed including these variables. First, we verified the superiority of the proposed methods over traditional ones. Then, we applied these methods to the actual industrial data obtained during an industrial polymer process. The proposed models achieved high predictive accuracy for melt flow rate and density, and the bias of the prediction errors could be eased in both cases by using the proposed methods. We confirmed the usefulness of the proposed methods without reconstruction of soft sensor models. © 2011 Curtin University of Technology and John Wiley & Sons, Ltd.