Statistical performance monitoring of dynamic multivariate processes using state space modelling

Abstract There is increasing interest in extending the concept of multivariate statistical process control to incorporate system dynamics into the process performance monitoring representation. A methodology has previously been proposed where the system dynamics and the correlation structure of the process data are captured within a state space representation. The system states and the state space model parameters are identified using the multivariate statistical projection techniques of canonical variate analysis (CVA) and partial least squares (PLS). A number of metrics based on Hotelling's T 2 statistic are proposed for the monitoring of the state of the system. Control limits for these metrics are calculated using the empirical reference distribution and assuming the metrics follow the known, theoretically derived, probability distributions. Two model forms are proposed. In the first the number of inputs and outputs in the model are constant for all variables, whilst the second approach assumes that number of inputs and outputs can vary. The modelling and process performance monitoring ability of the CVA and PLS state space representations, and the sensitivity of the various metrics in identifying faults, is investigated using a comprehensive simulation of a continuous stirred tank co-polymerisation reactor.

[1]  H. Akaike.  Markovian Representation of Stochastic Processes by Canonical Variables , 1975 .

[2]  P. Bühlmann.  The blockwise bootstrap for general empirical processes of stationary sequences , 1995 .

[3]  W. Larimore.  System Identification, Reduced-Order Filtering and Modeling via Canonical Variate Analysis , 1983, 1983 American Control Conference.

[4]  Layth C. Alwan,et al.  Time-Series Modeling for Statistical Process Control , 1988 .

[5]  A. Vasilopoulos,et al.  Modification of Control Chart Limits in the Presence of Data Correlation , 1978 .

[6]  Dale E. Seborg,et al.  Reduced-Rank ARX and Subspace System Identification for Process Control , 1998 .

[7]  Tohru Katayama,et al.  Statistical Methods in Control and Signal Processing , 1997 .

[8]  Theodora Kourti,et al.  Statistical Process Control of Multivariate Processes , 1994 .

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

[10]  B. Efron,et al.  Bootstrap confidence intervals , 1996 .

[11]  Wenceslao González-Manteiga,et al.  The bootstrap -A review , 1992 .

[12]  A. Negiz,et al.  Statistical monitoring of multivariable dynamic processes with state-space models , 1997 .

[13]  A. Çinar,et al.  PLS, balanced, and canonical variate realization techniques for identifying VARMA models in state space , 1997 .

[14]  A. Simoglou,et al.  Dynamic multivariate statistical process control using partial least squares and canonical variate analysis , 1999 .

[15]  T. W. Anderson.  An Introduction to Multivariate Statistical Analysis , 1959 .

[16]  H. Akaike,et al.  Information Theory and an Extension of the Maximum Likelihood Principle , 1973 .

[17]  Costas Kiparissides,et al.  On the validity of the steady-state approximations in high conversion diffusion-controlled free-radical copolymerization reactions , 1994 .

[18]  Jen Tang,et al.  Control Charts for Dependent and Independent Measurements Based on Bootstrap Methods , 1996 .

[19]  Heinz Unbehauen,et al.  Structure identification of nonlinear dynamic systems - A survey on input/output approaches , 1990, Autom..

[20]  A. J. Morris,et al.  Non-parametric confidence bounds for process performance monitoring charts☆ , 1996 .

[21]  Christos Georgakis,et al.  Disturbance detection and isolation by dynamic principal component analysis , 1995 .

[22]  George C. Runger,et al.  Designing control charts using an empirical reference distribution , 1996 .

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