Systematic estimation of state noise statistics for extended Kalman filters

The successful application of model-based control depends on the information about the states of the dynamic system. State-estimation methods, like extended Kalman filters (EKF), are useful for obtaining reliable estimates of the states from a limited number of measurements. They also can handle the model uncertainties and the effect of unmeasured disturbances. The main issue in applying EKF remains that one needs to specify the confidence in the model in terms of process noise covariance matrix. The information about the model uncertainties can effectively and systematically calculate the process noise covariance matrix for an EKF. Two systematic approaches are used for this calculation. The first is based on a Taylor series expansion of the nonlinear equations around the nominal parameter values, while the second accounts for the nonlinear dependence of the system on the fitted parameters by Monte Carlo simulations that can easily be performed on-line. The value of the process noise covariance matrix obtained is not limited to a diagonal form and depends on the current state of the dynamic system. Thus the a-priori information regarding the uncertainty in the model is utilized and the need for extensive tuning of the EKF is eliminated. The application of these techniques to example processes is also discussed (Transesterification process; Methyl methacrylate polymerization; and Semibatch system using tendency models). The accuracy of this methodology is compared very favorably with the traditional methods of trial-and-error tuning of EKF.

[1]  F. Lewis.  Optimal Estimation: With an Introduction to Stochastic Control Theory , 1986 .

[2]  Jake Fotopoulos,et al.  Use of tendency models and their uncertainty in the design of state estimators for batch reactors 1 This contribution is dedicated to the remembrance of Professor Jacques Villermaux. 1 , 1998 .

[3]  A. Jazwinski.  Stochastic Processes and Filtering Theory , 1970 .

[4]  M. Agarwal,et al.  A discretized nonlinear state estimator for batch processes , 1995 .

[5]  B. Tapley,et al.  Adaptive sequential estimation with unknown noise statistics , 1976 .

[6]  Masoud Soroush,et al.  Nonlinear State Estimation in a Polymerization Reactor , 1997 .

[7]  Christos Georgakis,et al.  Uses of state estimation for statistical process control , 1994 .

[8]  S. Bankoff,et al.  Digital monitorina and estimation of polymerization reactors , 1976 .

[9]  John F. MacGregor,et al.  State estimation for semi-batch polymerization reactors , 1992 .

[10]  Sirish L. Shah,et al.  Adaptive multirate state and parameter estimation strategies with application to a bioreactor , 1995 .

[11]  W. Harmon Ray,et al.  The dynamic behavior of continuous polymerization reactors—I: Isothermal solution polymerization in a CSTR , 1981 .

[12]  R. Mehra.  Approaches to adaptive filtering , 1972 .

[13]  K. Ravindranath,et al.  Modeling of poly(ethylene terephthalate) reactors. I. A semibatch ester interchange reactor , 1981 .

[14]  Marco S. Caceci,et al.  Estimating error limits in parametric curve fitting , 1989 .

[15]  Dominique Bonvin,et al.  Application of Estimation Techniques to Batch Reactors , 1991 .

[16]  R. H. Luecke,et al.  Estimation of the co variances of the process noise and measurement noise for a linear discrete dynamic system , 1995 .

[17]  D. Wilson,et al.  Experiences implementing the extended Kalman filter on an industrial batch reactor , 1998 .

[18]  Jake Fotopoulos,et al.  The identification of kinetic expressions and the evolutionary optimization of specialty chemical batch reactors using tendency models , 1992 .

[19]  Kyu Yong Choi,et al.  Optimal State Estimation in the Transesterification Stage of a Continuous Polyethylene Terephthalate Condensation Polymerization Process. , 1988 .

[20]  Derinola K. Adebekun,et al.  Continuous solution polymerization reactor control. 2. Estimation and nonlinear reference control during methyl methacrylate polymerization , 1989 .

[21]  Andrew Klein,et al.  Dynamic modeling and state estimation for an emulsion copolymerization reactor , 1989 .

[22]  Mark A. Kramer,et al.  The simultaneous solution and sensitivity analysis of systems described by ordinary differential equations , 1988, TOMS.

[23]  Mukul Agarwal,et al.  Batch unit optimization with imperfect modelling: a survey , 1994 .

[24]  Jay H. Lee,et al.  Extended Kalman Filter Based Nonlinear Model Predictive Control , 1993, 1993 American Control Conference.

[25]  M. Soroush.  Nonlinear state-observer design with application to reactors , 1997 .

[26]  Claudio Scali,et al.  Control of the quality of polymer products in continuous reactors: comparison of performance of state estimators with and without updating of parameters , 1997 .

[27]  T. Crowley,et al.  Experimental studies on optimal molecular weight distribution control in a batch-free radical polymerization process , 1998 .

[28]  Shi-Shang Jang,et al.  Comparison of two approaches to on-line parameter and state estimation of nonlinear systems , 1986 .