ABSORBED ENERGY MODELING OF HYPOEUTECTOID STEELS USING LINEAR REGRESSION AND GENETIC PROGRAMMING TAKING INTO ACCOUNT CONTINUOUS CASTING PARAMETERS

Štore Steel Ltd. is one of the major flat spring steel producers in Europe, producing several hundred steel grades with different chemical compositions. (180 × 180) mm billets are cast using a continuous casting machine with a 9-m radius. The macrostructure of the billets influence the mechanical properties of rolled material. For impact testing, standard test pieces are obtained from the rolled material at a defined location and orientation. Accordingly, the reduction ratio is also essential. Based on 2243 V-notch test pieces obtained from rolled materials produced in several batches between January 2019 and May 2024, models for predicting the absorbed energy of the rolled material at temperatures ranging from –60 °C to 20 °C were developed using linear regression and genetic programming. The models incorporated parameters such as chemical composition (contents of C, Si, Mn, S, Cr, Mo, Ni, Al, and V), casting conditions (average casting temperature, mold water flow, average difference between input and output mold cooling water temperature, average cooling water flow and pressure in three zones of secondary cooling), as well as the width and thickness of the rolled bars and test temperature. The genetic programming model outperformed the linear regression model.