Linear Fundamental Control Problem

Prett, David M.; García, Camino López

doi:10.1016/b978-0-409-90082-8.50014-x

Cited by 5 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The upper end of the complexity spectrum includes both linear and nonlinear model predictive control (MPC). MPC algorithms include a process model and use optimization routines to generate control moves that minimize a selected performance objective (see Prett and Garcıa 14 for an overview of MPC algorithms). The specific linear MPC algorithm used was standard dynamic matrix control (DMC).…”

Section: Control Algorithmsmentioning

confidence: 99%

“…ISE results are then collected for a variety of other disturbances of different magnitudes and sign under the same tunings. Following common practice 14 to avoid excessive numbers of computations, the MPC algorithms use prediction horizons of 30 coefficients and control move horizons of 10. Note that, by fixing the prediction horizons and by assuming that the length of the prediction horizon should be equal to the time to approximately 99% of steady state (based on the linearized model), the controller step size is set.…”

Section: Control Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Correlation of Process Nonlinearity with Closed-Loop Disturbance Rejection

Hernjak

Doyle

2003

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Section: Control Algorithmsmentioning

confidence: 99%

Section: Control Algorithmsmentioning

confidence: 99%

Correlation of Process Nonlinearity with Closed-Loop Disturbance Rejection

Hernjak

Doyle

2003

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

“…21 In this simulation, the prediction horizon, P, is chosen to be 10 and the control horizon, M, to be 4. Equal weight among the inputs (weight ) 1) is used, implying no excessive penalties on the inputs.…”

Section: Multiple Modelsmentioning

confidence: 99%

“…The selection of MPC parameters need not be optimal, but they should follow the general MPC design rules. 21 In this simulation, the prediction horizon, P, is chosen to be 10 and the control horizon, M, to be 4. Equal weight among the inputs (weight ) 1) is used, implying no excessive penalties on the inputs.…”

Section: Multiple Modelsmentioning

confidence: 99%

Multiple Model-Based Control of the Tennessee−Eastman Process

Tian

Hoo

2005

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

This work addresses the model-based control of the Tennessee-Eastman (TE) challenge problem using a state-shared model. A detailed mechanistic nonlinear model is developed and validated with data taken from the original Downs and Vogel work [Downs, J. J.; Vogel, E. F. Comput. Chem. Eng. 1993, 17 (3), 245-255] and their accompanying Fortran programs. Two plantwide model predictive control (MPC) strategies that differ in the types of models used are applied to address regulation and transition control. The first employs multiple fixed-parameter models, identified at known grades, and the second uses an adaptive state-shared model Hoo, K. A. Comput. Chem. Eng. 2003, 27 (11), 1641-1656] constructed from three fixed-parameter models and one specialized adaptive model. The first MPC strategy demonstrates closed-loop regulation and grade-transition performance especially when the worst disturbance, the loss of the A feed, is present. The second MPC strategy provides a better closed-loop performance for grade transitions when compared with the first MPC strategy.

show abstract

“…As far as the identification problem is concerned, the scope of analysis as well as applications seem to be quite restricted. As pointed out in [4], identification remains to be the most time consuming step and requires expertise from the user in the implementation of DMC.…”

Section: Introductionmentioning

confidence: 99%

On parameter estimation of the DMC models

Yin

Asbjørnsen

1991

Stochastic Analysis and Applications

View full text Add to dashboard Cite

Keywords. dynamic matrix control, consistency, asynlptotic normality, confidence regions.In recent years, much effort has been devoted to the study of the Dynamic Matrix Control (DMC) model. Such a model is essentially a predictive controller that computes moves on manipulated variables to create changes in the output. In a wide range of applications, the coefficients (control transfer coefficients) of the input-output model are generally unknown. Thus, in order to design the desired predictive controller, the first important task is to identify these parameters. In this work, a recursive algorithm for the aforementioned task is developed. Some asymptotic properties of such an algorithm is obtained. It is shown that the algorithm is strongly consistent and a suitably scaled error sequence satisfies a functional invariance principle. The asymptotic normality is used to build up interval (confidence region) estimates. Moreover, a useful and easily implementable stopping rule is also developed.

show abstract

Linear Fundamental Control Problem

Cited by 5 publications

References 0 publications

Correlation of Process Nonlinearity with Closed-Loop Disturbance Rejection

Correlation of Process Nonlinearity with Closed-Loop Disturbance Rejection

Multiple Model-Based Control of the Tennessee−Eastman Process

On parameter estimation of the DMC models

Contact Info

Product

Resources

About