In complex industrial environments such as steel production, many key processes take place deep inside large-scale systems, far from direct measurement. Understanding and optimizing these systems is crucial for operational stability, energy efficiency and competitiveness. This was the starting point for the research project PreMoBAF – Predicting and Monitoring the Behavior of Blast Furnace and Electric Arc Furnace, now successfully completed by Pro²Future and its academic partners.
Challenge
Blast furnaces (BF) and electric arc furnaces (EAF) are essential to modern steel production, yet their inner dynamics are governed by complex physical and chemical reactions under extreme conditions. These processes are barely accessible to direct observation, and traditional models struggle to represent their behavior accurately or transparently. Especially in industrial settings, explainability and trust in AI-based solutions are crucial to enable adoption.
Solution
PreMoBAF addressed this challenge by developing a hybrid approach that integrates deep learning, explainable artificial intelligence (xAI) and causal discovery techniques. Domain knowledge from steelmaking was incorporated into the models to improve robustness and interpretability. This framework enables advanced monitoring and prediction of furnace behavior, even in the absence of direct sensor data.
The project was carried out in Area 3 – Cognitive Decision Making at Pro²Future and conducted in collaboration with the Institute for Application-Oriented Knowledge Processing (FAW), JKU Linz, and the Business Analytics and Data Science Center (BANDAS), University of Graz.
Impact
PreMoBAF successfully delivered data-driven, yet interpretable AI models for real-world industrial application. These models support informed decision-making, transparent analysis, and process optimization in the steel industry. By combining machine learning with expert-driven understanding, the project contributes to building trustworthy and actionable AI — a key enabler of the digital transformation in high-temperature production environments.
For more details download the Success Story or contact Matej Vukovic directly!