Predicting Electrical Energy Consumption in Electric Arc Furnace Steelmaking: A Data-Driven Approach

Optimizing energy consumption in industrial processes, such as steel production, is crucial to reducing operational expenses and achieving sustainability goals. This thesis focuses on predicting the electrical energy consumption in the Electric Arc Furnace (EAF), a key component of modern steelmaking. It aims to provide a framework for analyzing raw material input, process parameters, and operational conditions, as well as their influence on energy demand. Data preprocessing techniques, including outlier detection, feature selection and principal component analysis (PCA), are implemented alongside statistical models to enhance model reliability and interpretability. Throughout this thesis, the role of steel scrap in energy consumption variability is explored, along with an evaluation of different predictive modeling approaches, such as Multiple Linear Regression (MLR), Partial Least Squares (PLS) regression, to determine their effectiveness in an industrial setting. The findings provide a foundation for enhancing predictive accuracy, enabling more informed operational decisions in EAF steelmaking, and promoting energy-efficient manufacturing practices.