Speaker
Description
In a hot-dip galvanizing line for the production of zinc-coated steel strips, the heat treatment of the strip is an important process step. In order to achieve the desired material and surface properties, the strip has to be heated according to a predefined temperature trajectory. From the viewpoint of control, this is a challenging task, in particular in transient furnace operation, e.g. when a welded joint traverses the furnace or when the strip velocity changes. Because the product diversity, the demands on the product quality, and the desire to minimize the energy consumption are steadily increasing, there is a need for an advanced process control concept that accounts for all these challenges. In this paper, a nonlinear model predictive controller for the strip temperature in a combined direct- and indirect-fired annealing furnace is described. The controller was implemented at a hot-dip galvanizing line of voestalpine Stahl GmbH, Linz, Austria and measurement results from three years operational experience are presented.
The basis for the model predictive controller is a first-principles dynamical model of the furnace, which is characterized by moderate complexity and which captures the essential dynamical behavior of the real furnace. The model incorporates sub-models describing the flue gas, the wall, the radiant tubes, the rolls, the strip, and the relevant heat transfer mechanisms. Using this furnace model and the estimated current system state, the model predictive controller selects optimal trajectories for the fuel supply so that the strip temperature reaches its desired target temperature. In the control algorithm, a tailored constrained nonlinear optimization problem is numerically solved by the Levenberg-Marquardt method. The gradient and the approximated Hessian matrix of the objective function are analytically computed using an adjoint-based approach. This computationally highly efficient algorithm ensures that the controller can be executed in real time.
Measurements from the implementation at the industrial plant of voestalpine demonstrate the excellent performance of the developed control concept. It ensures compliance with all temperature constraints and achieves accurate strip temperature control in both steady-state and non-steady-state furnace operation. A long-term analysis shows the significant improvement of control performance in terms of accuracy and homogeneity of the strip temperature compared to the previous implemented PI temperature controllers. Encouraged by excellent feedback from three years operational experience, the developed control concept is currently transferred to the other hot-dip galvanizing lines of voestalpine.