RESUMEN
In this paper, a unified level 2 Advanced Process Control system for steel billets reheating furnaces is proposed. The system is capable of managing all process conditions that can occur in different types of furnaces, e.g., walking beam and pusher type. A multi-mode Model Predictive Control approach is proposed together with a virtual sensor and a control mode selector. The virtual sensor provides billet tracking, together with updated process and billet information; the control mode selector module defines online the best control mode to be applied. The control mode selector uses a tailored activation matrix and, in each control mode, a different subset of controlled variables and specifications are considered. All furnace conditions (production, planned/unplanned shutdowns/downtimes, and restarts) are managed and optimized. The reliability of the proposed approach is proven by the different installations in various European steel industries. Significant energy efficiency and process control results were obtained after the commissioning of the designed system on the real plants, replacing operators' manual conduction and/or previous level 2 systems control.
RESUMEN
This study proposes a data analysis and modelization method for the rolling mill process of billets in steel plants. By exploiting rolling mill signals and advanced data processing algorithms, a reliable billet tracking system is designed, which tracks each workpiece from the furnace entrance to the rolling mill stands' exit area. Based on the stored information, two problems are addressed: the data analysis of the temperature sensors (a thermal imaging camera and pyrometers) and the current that is related to the rolling mill stands' absorption, and subsequently, a mathematical modelization of the billets' temperature along their path in the rolling mill is produced. The data analysis suggested that we should perform hardware modifications: the thermal imaging camera was repositioned to avoid the effect of scale formation on the temperature measurements. The modelization phase provided the basis for future control and/or diagnosis applications that will exploit a temperature decay model.