Data Evaluation and Modeling of Billet Characteristics in the Steel Industry

This study presents a methodology to analyze and model the continuous mill process of billets in steel plants. A reliable billet tracking system able to track each workpiece from the furnace entrance to the exit area of the rolling mill stands is designed based on information derived from rolling mill signals processed through advanced data science algorithms. From the stored information, two issues are solved: the data of the thermal sensors (of a thermographic camera or pyrometers) and the current linked to the absorption of the rolling mill stands; and a mathematical modelisation of the temperature of the billets along its way through the rolling mill is elaborated. Based on this data analysis, we concluded that some hardware changes were needed: the IR camera was moved to eliminate the interference of the scaly formation of the sensor with the data collected. The modelization step served as a foundation for future applications of control and/or diagnosis that will make use of a temperature decay model.