TY - GEN

T1 - Optimization based tuning approach for offset free MPC

A1 - Olesen,Daniel Haugård

A1 - Huusom,Jakob Kjøbsted

A1 - Jørgensen,John Bagterp

AU - Olesen,Daniel Haugård

AU - Huusom,Jakob Kjøbsted

AU - Jørgensen,John Bagterp

PB - Technical University of Denmark

PY - 2012

Y1 - 2012

N2 - We present an optimization based tuning procedure with certain robustness properties for an offset free Model Predictive Controller (MPC). The MPC is designed for multivariate processes that can be represented by an ARX model. The advantage of ARX model representations is that standard system identifiation techniques using convex optimization can be used for identification of such models from input-output data. The stochastic model of the ARX model identified from input-output data is modified with an ARMA model designed as part of the MPC-design procedure to ensure offset-free control. The ARMAX model description resulting from the extension can be realized as a state space model in innovation form. The MPC is designed and implemented based on this state space model in innovation form. Expressions for the closed-loop dynamics of the unconstrained system is used to derive the sensitivity function of this system. The closed-loop expressions are also used to numerically evaluate absolute integral performance measures. Due to the closed-loop expressions these evaluations can be done relative quickly. Consequently, the tuning may be performed by numerical minimization of the integrated absolute error subject to a constraint on the maximum of the sensitivity function. The latter constraint provides a robustness measure that is essential for the procedure. The method is demonstrated on two simulated examples: A Wood-Berry distillation column example and a cement mill example.

AB - We present an optimization based tuning procedure with certain robustness properties for an offset free Model Predictive Controller (MPC). The MPC is designed for multivariate processes that can be represented by an ARX model. The advantage of ARX model representations is that standard system identifiation techniques using convex optimization can be used for identification of such models from input-output data. The stochastic model of the ARX model identified from input-output data is modified with an ARMA model designed as part of the MPC-design procedure to ensure offset-free control. The ARMAX model description resulting from the extension can be realized as a state space model in innovation form. The MPC is designed and implemented based on this state space model in innovation form. Expressions for the closed-loop dynamics of the unconstrained system is used to derive the sensitivity function of this system. The closed-loop expressions are also used to numerically evaluate absolute integral performance measures. Due to the closed-loop expressions these evaluations can be done relative quickly. Consequently, the tuning may be performed by numerical minimization of the integrated absolute error subject to a constraint on the maximum of the sensitivity function. The latter constraint provides a robustness measure that is essential for the procedure. The method is demonstrated on two simulated examples: A Wood-Berry distillation column example and a cement mill example.

KW - Model Predictive Control

KW - Controller tuning

KW - Multivariate processes

KW - Autoregressive models

KW - Optimization

BT - The 10th European Workshop on Advanced Control and Diagnosis

T2 - The 10th European Workshop on Advanced Control and Diagnosis

ER -