Nonlinear model predictive control of the strip temperature in an annealing furnace. (December 2016)

• Record Type: Journal Article
• Authors: Niederer, M.; Strommer, S.; Steinboeck, A.; Kugi, A.
• Is Part Of: Journal of process control. Volume 48(2016:Dec.)
• Page Start: 1
• Page End: 13
• Publication Date: 2016-12
• Subjects: Steel industry -- Annealing furnace -- Direct- and indirect-fired sections -- Nonlinear model predictive control -- Receding horizon control -- Vector-valued Lagrangian -- Adjoint-based method -- Gauss–Newton method; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2016.09.012
• Languages: English
• ISSNs: 0959-1524

Abstract

Abstract : Highlights: A model predictive temperature controller is designed for a strip annealing furnace. The controller is based on a nonlinear first-principles model the furnace. A dynamic optimization problem with equality constraints is solved using the Gauss–Newton method. The required gradient and the approximated Hessian matrix are analytically determined using an adjoint based method. The material throughput is systematically optimized. Abstract: A nonlinear model predictive controller is designed for the strip temperature in a combined direct- and indirect-fired strip annealing furnace. Based on a tailored first-principles dynamical model and the estimated current system state, the receding horizon controller selects optimal trajectories for both the fuel supply and the strip velocity so that the strip temperature is controlled to its desired target temperature. The controller additionally maximizes the throughput and minimizes the energy consumption. In the control algorithm, the dynamic optimization problem with equality constraints is numerically solved by using the Gauss–Newton method. The gradient and the approximated Hessian matrix of the objective function are analytically computed using an adjoint-based method. The capabilities of the proposed controller are demonstrated for a validated high-fidelity simulation model of an industrial annealing furnace.