Data-Driven Control-oriented Modelling of Submerged Arc Furnaces

The present thesis focuses on the integration of chemophysical knowledge and data-driven insights to develop control-oriented models for submerged arc furnaces (SAF). The primary objective of this research is to facilitate the metal- lurgical processes involved in the production of ferrosilicon (FeSi). The state-of-art of simulators utilized for submerged arc furnaces is founded upon a static meta-model, that functions as a data-driven substitute for Physic- based Finite Element Methods (FEM) models. The aim of this study is to develop a linear dynamic model utilizing data- driven techniques from the system identification literature, and subsequently try to evaluate its performance in comparison to the meta-model. The current thesis addresses the difficulties associated with this specific task and suggests potential solutions for overcoming them.