Modelling of spectroscopic batch process data using grey models to incorporate external information

In both analytical and process chemistry, one common aim is to build models describing measured data. In cases where additional information about the chemical system is available, this can be incorporated into the model with the aim of improving model fit and interpretability. A model which consists of a ‘hard’ or ‘white’ part describing known sources of variation and a ‘soft’ or' black' part describing unknown sources of variation is called a ‘grey’ model. In this paper the use of a grey model is demonstrated using data from a first‐order chemical batch reaction monitored by UV‐vis spectroscopy. The resultant three‐way data matrix is modelled using a Tucker3 structure, and external information about the spectroscopically active compounds is incorporated in the form of constraints on the model parameters with additional restrictions on the Tucker3 core matrix. The grey model is then used to analyse new batches. Different approaches to building grey models are described and some of their properties discussed. Copyright © 2000 John Wiley & Sons, Ltd.

[1]  H. Law.  Research methods for multimode data analysis , 1984 .

[2]  N. Sidiropoulos,et al.  Least squares algorithms under unimodality and non‐negativity constraints , 1998 .

[3]  B. Kowalski,et al.  Partial least-squares regression: a tutorial , 1986 .

[4]  J-R Lalanne,et al.  UV-visible Absorption Spectroscopy , 1999 .

[5]  J. Macgregor,et al.  Monitoring batch processes using multiway principal component analysis , 1994 .

[6]  A. Smilde,et al.  Multivariate statistical process control of batch processes based on three-way models , 2000 .

[7]  A. Smilde,et al.  Estimating rate constants and pure UV‐vis spectra of a two‐step reaction using trilinear models , 1999 .

[8]  R. Bro.  PARAFAC. Tutorial and applications , 1997 .

[9]  K. S. Banerjee.  Generalized Inverse of Matrices and Its Applications , 1973 .

[10]  Theodora Kourti,et al.  Comparing alternative approaches for multivariate statistical analysis of batch process data , 1999 .

[11]  F. Mosteller,et al.  Understanding robust and exploratory data analysis , 1985 .

[12]  Y. Takane,et al.  Principal component analysis with external information on both subjects and variables , 1991 .

[13]  L. Tucker,et al.  Some mathematical notes on three-mode factor analysis , 1966, Psychometrika.

[14]  Forrest W. Young,et al.  Additive structure in qualitative data: An alternating least squares method with optimal scaling features , 1976 .

[15]  Paul Geladi,et al.  Principal Component Analysis , 1987, Comprehensive Chemometrics.

[16]  John F. MacGregor,et al.  Multivariate SPC charts for monitoring batch processes , 1995 .

[17]  Henk B. Verbruggen,et al.  Semi-mechanistic modeling of chemical processes with neural networks , 1998 .

[18]  Barry M. Wise,et al.  The process chemometrics approach to process monitoring and fault detection , 1995 .

[19]  G. Box.  Some Theorems on Quadratic Forms Applied in the Study of Analysis of Variance Problems, I. Effect of Inequality of Variance in the One-Way Classification , 1954 .

[20]  A. Höskuldsson.  PLS regression methods , 1988 .

[21]  J. Kruskal,et al.  Candelinc: A general approach to multidimensional analysis of many-way arrays with linear constraints on parameters , 1980 .

[22]  A. J. Morris,et al.  The introduction of process chemometrics into an industrial pilot plant laboratory , 1998 .

[23]  Theodora Kourti,et al.  Analysis, monitoring and fault diagnosis of batch processes using multiblock and multiway PLS , 1995 .

[24]  J. Leeuw,et al.  Principal component analysis of three-mode data by means of alternating least squares algorithms , 1980 .

[25]  R. P. McDonald,et al.  A simple comprehensive model for the analysis of covariance structures: Some remarks on applications , 1980 .

[26]  Age K. Smilde,et al.  Constrained three‐mode factor analysis as a tool for parameter estimation with second‐order instrumental data , 1998 .

[27]  J. E. Jackson,et al.  Control Procedures for Residuals Associated With Principal Component Analysis , 1979 .

[28]  J. Kruskal,et al.  How 3-MFA data can cause degenerate parafac solutions, among other relationships , 1989 .

[29]  John F. MacGregor,et al.  Process monitoring and diagnosis by multiblock PLS methods , 1994 .

[30]  W. Rayens,et al.  Two-factor degeneracies and a stabilization of PARAFAC , 1997 .

[31]  Sue Leurgans,et al.  [27] Component resolution using multilinear models , 1995 .

[32]  George E. P. Box,et al.  Empirical Model‐Building and Response Surfaces , 1988 .

[33]  J. Chang,et al.  Analysis of individual differences in multidimensional scaling via an n-way generalization of “Eckart-Young” decomposition , 1970 .

[34]  Age K. Smilde,et al.  Monitoring and diagnosing batch processes with multiway covariates regression models , 1999 .

[35]  B. Kowalski,et al.  Theory of medium‐rank second‐order calibration with restricted‐Tucker models , 1994 .

[36]  Yoshio Takane,et al.  The analysis of multitrait-multimethod matrices via constrained components analysis , 1996 .

[37]  Yu-Long Xie,et al.  Mixed multiway analysis of airborne particle composition data , 1999, Optics East.

[38]  Jan de Leeuw,et al.  Component analysis with different sets of constraints on different dimensions , 1995 .

[39]  Thomas E. Marlin,et al.  Multivariate statistical monitoring of process operating performance , 1991 .

[40]  Rimvydas Simutis,et al.  How to increase the performance of models for process optimization and control , 1997 .