Machine Learning-Based Models Applied for Improving Electric Steel Works' Sustainability

Steelmaking by electric steelworks is expected to increase due to their pivotal role in the steel sector transition towards C-lean processes. Therefore, several efforts are made to increase their sustainability. The European project ALCHIMIA provides an exemplar investigation aimed at decreasing energy and resource consumption in electric steelworks. The application of a decision support system is envisaged based on federated learning and advanced optimization algorithms to support operators in the selection of the best input material mix. The contribution focuses on the application of machine learning-based models computing the characteristics of scrap stock and estimating steel composition and temperature after secondary metallurgy.