Gunnar Mathiason
Associate Professor of Informatics
EKOPRO is to reduce scrap, energy consumption and carbon dioxide emissions in Swedish steel and metal production through AI-driven real-time analysis.
"We want to use AI to detect problems earlier in the process, so that we can reduce emissions, save energy and make better use of resources. For one of our partners, a reduction of just one per cent in scrap would mean annual savings of over SEK 30 million, while at the same time making the supply of materials more sustainable."
Gunnar Mathiason, Associate Professor of Informatics
Steel production is a highly energy-intensive process. When material is scrapped, this not only results in a direct loss of material but also means that large amounts of energy have already been consumed unnecessarily.
This leads to increased emissions, higher production costs, and reduced resource efficiency throughout the value chain.
Reduced scrap volumes for enhanced competitiveness
Reducing the amount of scrap is therefore a key factor both for the industry’s competitiveness and for achieving climate and sustainability targets.
The project’s approach shifts quality control from reactive end-of-line inspection to a proactive system that identifies and addresses problems early in the process. This enables significant reductions in scrap, energy use, and CO₂ emissions, while maintaining the high precision and quality standards that characterize Swedish metals production.
AI in steel and metal production
The EKOPRO project develops and implements AI-driven solutions for the early detection of quality deviations in steel and metals production. Building on Sweden’s strong digital infrastructure for process monitoring, the project takes a significant step forward by integrating real-time analysis with AI across the entire production chain.
This goes beyond previous initiatives in Sweden and internationally, which have typically focused on retrospective data analysis or optimization of individual process steps.
Project management
- Peter Lundin, Swerim
- Gunnar Mathiason, University of Skövde