Process analysis and abnormal situation detection: from theory to practice

The article discusses the use of latent variable models based on historical data and examines their potential and limitations for improving operations for both batch and continuous processes. The use of these models for multivariate statistical process monitoring, abnormal situation detection, and fault diagnosis is demonstrated. Examples from state-of-the-art major industrial applications currently running online illustrate the tremendous potential of these methods. In this context, an industrial application for abnormal situation detection is defined as "state of the art" if it has been operational several years after it was commissioned, has generated large savings, has been operating safely and/or has improved safety conditions in the plant, and is accepted enthusiastically by the operators. Such an application could be based entirely on known theory, but frequently it includes company proprietary modifications to suit the particular operating characteristics of the process. The article contains an extensive literature review on the subject and practical considerations for the user, as well as warnings about potential pitfalls in areas ranging from data acquisition to modeling to online application.

[1]  S. Wold.  Cross-Validatory Estimation of the Number of Components in Factor and Principal Components Models , 1978 .

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

[3]  Michael J. Piovoso,et al.  Process data chemometrics , 1991 .

[4]  J. Macgregor,et al.  On-Line Latex Particle Size Determination by Dynamic Light Scattering: Design for an Industrial Environment , 1991 .

[5]  J. Macgregor,et al.  MULTIVARIATE IDENTIFICATION: A STUDY OF SEVERAL METHODS , 1992 .

[6]  J. Macgregor,et al.  Composition and Particle Diameter for Styrene/Methyl Methacrylate Copolymer Latex Using UV and NIR Spectroscopy , 1993 .

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

[8]  J.F. MacGregor,et al.  Multi-way PCA applied to an industrial batch process , 1994, Proceedings of 1994 American Control Conference - ACC '94.

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

[10]  Theodora Kourti,et al.  Process analysis, monitoring and diagnosis, using multivariate projection methods , 1995 .

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

[12]  Alison J. Burnham,et al.  Frameworks for latent variable multivariate regression , 1996 .

[13]  Karlene A. Kosanovich,et al.  Improved Process Understanding Using Multiway Principal Component Analysis , 1996 .

[14]  P. A. Taylor,et al.  Missing data methods in PCA and PLS: Score calculations with incomplete observations , 1996 .

[15]  J. Macgregor,et al.  Experiences with industrial applications of projection methods for multivariate statistical process control , 1996 .

[16]  Bhupinder S. Dayal,et al.  Recursive exponentially weighted PLS and its applications to adaptive control and prediction , 1997 .

[17]  Manish H. Bharati,et al.  Multivariate image analysis for real-time process monitoring and control , 1997 .

[18]  S. D. Jong,et al.  Multivariate statistical process control in chromatography , 1997 .

[19]  P. A. Taylor,et al.  Synchronization of batch trajectories using dynamic time warping , 1998 .

[20]  C. E. Schlags,et al.  Multivariate statistical analysis of an emulsion batch process , 1998 .

[21]  J. Macgregor,et al.  Analysis of multiblock and hierarchical PCA and PLS models , 1998 .

[22]  Michael J. Piovoso,et al.  Translating third-order data analysis methods to chemical batch processes , 1999 .

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

[24]  Weihua Li,et al.  Recursive PCA for adaptive process monitoring , 1999 .

[25]  J. Huvenne,et al.  On-Line Monitoring of a Latex Emulsion Polymerization by Fiber-Optic FT-Raman Spectroscopy. Part I: Calibration , 2000 .

[26]  John F. MacGregor,et al.  Product transfer between plants using historical process data , 2000 .

[27]  John F. MacGregor,et al.  Multivariate analysis and optimization of process variable trajectories for batch processes , 2000 .

[28]  Michael S. Dudzic,et al.  On-line applications of multivariate statistics at Dofasco , 2000 .

[29]  Michael J. Piovoso,et al.  On unifying multiblock analysis with application to decentralized process monitoring , 2001 .

[30]  B Lennox,et al.  Process monitoring of an industrial fed-batch fermentation. , 2001, Biotechnology and bioengineering.

[31]  Stephen P. Gurden,et al.  A comparison of multiway regression and scaling methods , 2001 .

[32]  John F. MacGregor,et al.  Texture analysis of images using principal component analysis , 2001, SPIE Optics East.

[33]  Seongkyu Yoon,et al.  Incorporation of External Information into Multivariate PCA/PLS Models , 2001 .

[34]  S. Wold,et al.  The GIFI approach to non‐linear PLS modeling , 2001 .

[35]  M. Dudzic,et al.  Industrial use of multivariate statistical analysis for process monitoring and control , 2002, Proceedings of the 2002 American Control Conference (IEEE Cat. No.CH37301).

[36]  J. Macgregor,et al.  Control of Particle Size Distributions in Emulsion Semibatch Polymerization Using Mid-Course Correction Policies , 2002 .