Advanced simulation and optimization of the entire process route for production of topmost steels

Laboratorij za dinamiko fluidov in termodinamiko

Fakulteta za strojništvo

Univerza v Ljubljani

Aškerčeva 6, SI-1000 Ljubljana

Slovenija

Štore Steel

Železarska cesta 3

3220 Štore, Slovenija

Slovenija

ARRS L2-3173: ADVANCED SIMULATION AND OPTIMIZATION OF THE ENTIRE PROCESS ROUTE FOR THE PRODUCTION OF PREMIUM STEELS.

1.10.2021 – 30.4.2025

FINAL REPORT

B. Šarler, T. Dobravec, J. Dolinar, U. Hanoglu, M. Kovačič, Q. Liu, B. Mavrič, K. Mramor, M. Perpar, R. Vertnik, G. Vuga

March 2026

  Abstract

For competitive performance in demanding international markets, product quality and production efficiency are extremely important, especially in steel production. Consistently achieving these two goals requires understanding and controlling all steps of the manufacturing process, for which numerical models are particularly useful. In the project, we established a prototype simulation environment for the production chain of the leading Slovenian steelmaker Štore Steel. This enabled the use of numerical models along the entire production chain in a way that coherently links the numerical models of individual production steps and allows traceability of product properties.

The project results represent a logical upgrade of our successfully completed applied projects “L2-6775 Simulation of Industrial Solidification Processes under the Influence of Electromagnetic Fields” and “L2-9246 Multiphysics and Multiscale Numerical Modelling for Competitive Continuous Casting,” as well as the related EU framework programme projects. With the help of the acquired knowledge, we equipped the co-financier’s major investments in a new steel plant and rolling mill with advanced numerical models and encouraged the company’s digital transformation toward Industry 4.0. The purpose of the project was the further development of models within the entire process chain for producing premium steels in order to understand, predict, and eliminate production defects such as macrosegregation, inclusions, shape distortions, porosity, hot tearing, reduced surface quality, and cracks.

The first objective was the development of a modular multiscale process model based on the assumption of a travelling slice for the process chain from continuous casting to heat treatment. The slice included thermal, concentration, mechanical, and microstructural models. Information on quality and possible defects was obtained from the combination of these models. The already existing models for continuous casting and hot rolling, built on this principle, were supplemented with modules linking these two processes and extended to heat treatment, thereby establishing a complete process model. In this model, a balance between physical depth and computational time was considered. The model enables the establishment of fundamental relationships among energy consumption, macroscopic fields, microscopic fields, and properties.

The second objective was the development of missing physical modules and the improvement of existing ones without specific constraints regarding computational time. These modules run on high-performance computing platforms. The newly developed models mainly include radiative heat transfer and fracture mechanics, while the improved models include three-dimensional microstructure models. Artificial intelligence was used to adapt simplified models using detailed models and for multi-objective optimization with respect to energy consumption, efficiency, and product quality.

The third objective was the experimental analysis of inclusions based on the water model for continuous casting developed in the previous project, and the experimental analysis of microstructure, product properties, and defects after each process step for validation purposes.

The solution procedures in the simulation system were further developed on the basis of our repeatedly awarded innovative meshless computing technology, continuum mechanics concepts at the macroscopic level, cellular automaton concepts at the mesoscopic level, and phase-field concepts at the microscopic level. The numerical implementation exploited the parallel computing capabilities of modern workstations and supercomputers. The effects of the new knowledge include improved quality, improved process capability, and higher productivity. The results were directly applied in production, published in journals with the highest impact factors in the field, and presented as keynote lectures at major international meetings.