Article: LMI approach for robust predictive control using orthonormal basis functions Journal: International Journal of Modelling, Identification and Control (IJMIC) 2015 Vol.24 No.3 pp.244 - 256 Abstract: This paper is focused on the problem of process control by using robust model predictive control (RMPC) and orthonormal basis functions (OBF). A relevant class of RMPC algorithms is the one characterised by the use of the LMI framework. Most of them assume a polytopic representation of the process uncertainties and require full-state feedback. On the other hand, several works describing the theory and applicability of OBF in identification and control fields can be found in the literature. The present paper proposes the use of OBF for modelling time-varying processes and for synthesis of RMPC algorithm. Using this proposed approach, the process model can be written in terms of new states, which are known or measurable. This represents an advantage in real applications where frequently the actual states are not known. Simulation results, one of them based on a debutaniser column model, illustrate the proposed methodology. The examples highlight the applicability of the method in the sense that a state feedback RMPC strategy is applied in a context where the actual states are unknown. Inderscience Publishers - linking academia, business and industry through research

Title: LMI approach for robust predictive control using orthonormal basis functions

Authors: Humberto X. Araújo; Rafael R. De Araújo; Gustavo H.C. Oliveira

Addresses: Electrical Engineering Graduate Program (PPGEE), Federal University of Bahia (UFBA), Rua Aristides Novis, 02, Zip Code: 40210-630 – Salvador, BA, Brazil ' Braskem S.A., Rua Eteno, 1561, Basic Petrochemicals Unit 1, Zip code: 42810-000, Camaçari, BA, Brazil ' Department of Electrical Engineering, Federal University of Paraná (UFPR), Zip Code: 80215-901, Curitiba, PR, Brazil

Abstract: This paper is focused on the problem of process control by using robust model predictive control (RMPC) and orthonormal basis functions (OBF). A relevant class of RMPC algorithms is the one characterised by the use of the LMI framework. Most of them assume a polytopic representation of the process uncertainties and require full-state feedback. On the other hand, several works describing the theory and applicability of OBF in identification and control fields can be found in the literature. The present paper proposes the use of OBF for modelling time-varying processes and for synthesis of RMPC algorithm. Using this proposed approach, the process model can be written in terms of new states, which are known or measurable. This represents an advantage in real applications where frequently the actual states are not known. Simulation results, one of them based on a debutaniser column model, illustrate the proposed methodology. The examples highlight the applicability of the method in the sense that a state feedback RMPC strategy is applied in a context where the actual states are unknown.

Keywords: model predictive control; MPC; robust control; orthonormal basis functions; OBF; closed-loop stability; input constraints; linear matrix inequalities; LMIs; process control; modelling; feedback control; simulation; debutaniser column model.

DOI: 10.1504/IJMIC.2015.072613

International Journal of Modelling, Identification and Control, 2015 Vol.24 No.3, pp.244 - 256

Received: 09 Dec 2014
Accepted: 11 Mar 2015
Published online: 22 Oct 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: LMI approach for robust predictive control using orthonormal basis functions

Keep up-to-date