Per: ALISSON PAULO DE OLIVEIRA (NSigma Consulting), antônio augusto gorni (Autônomo), Marcelos arantes rebellato (Eurosport)
Abstract:
This work aims to review the advanced tools, based on Artificial Intelligence resources, used in rolled product development projects and in the optimization of steel processes. These tools have been widely used because they allow to determine, with an excellent degree of precision, the understanding of the various metallurgical phenomena involved in characteristics of interest of steel products as well as in their forecast. With the advent of a series of new technologies, including new architectures of Artificial Neural Networks, Deep Learning, Machine Learning, Genetic Algorithms, among others, these tools have become quite reliable, even allowing to design new products and new processes, yet not known. Numerous recent examples of uses of these tools are displayed as well as the results regarding the use of new technologies. These technologies are already being used on site, for microstructure prediction, for example. In the end, it is demonstrated that Artificial Neural Networks, the basis of many of these technologies, are able to predict with a high degree of precision the historically established metallurgical trends, which proves their effectiveness as tools for studying and understanding the phenomena involved.