Predictive control algorithms with guaranteed stability and asymptotic tracking

A model of the continuous steel casting process is developed and validated using open loop identification. Modifications are made to the identification algorithm to enhance accuracy of the estimated model. The GPC (generalized predictive control) algorithm is then implemented for mold level regulation of the casting process. Compensation of measurement noise by the standard GPC algorithm results in excessive controller activity. A modification to the GPC cost function is suggested to account for measurement disturbances by dynamically filtering the predicted free response of the process model before the total future response is computed and weighted in the GPC cost function. Experimental results indicate that regulation performance of the modified GPC in the presence of load disturbances is much better than conventional PI control. A lower bound on the costing horizon that results in closed loop stability under GPC is not known a priori. Sufficient conditions are presented for stability of closed loop systems that result from implementing solutions of the finite horizon LQ (linear quadratic) problem for arbitrary fixed costing horizons. On this basis, a class of predictive control laws referred to as SPC (stabilizing predictive control) that ensures stability of the closed loop system is proposed. For tracking reference signals with step changes, a controller structure is derived that achieves asymptotic tracking while preserving stability of the closed loop system. Simulation results that illustrate the stabilizing and asymptotic tracking properties of SPC for both minimum phase and nonminimum phase unstable plants are presented.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Mechanical Science and Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2011

Author dc:creator

• Manayathara, Thomas Jolly

• Engineering, Electronics and Electrical
• Engineering, Mechanical

Statement dc:rights

• Copyright 1994 Manayathara, Thomas Jolly

Language dc:language

eng

Identifier

AAI9416401
(UMI)AAI9416401

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/19873