Artificial intelligence guided steel design towards zero-carbon emission contributing to circular economy

Researchers: Lovisa Åkesson and Wangzhong Mu

Artificial intelligence (AI) is a powerful tool for accelerating materials design and advancing the sustainability of materials development. This project aims to achieve the goal of utilizing fully waste to produce the new steel products without supplying new raw virgin materials. The influence of interstitial elements (e.g. C, N, S, etc.) on microstructure and mechanical property evolution is of vital importance. This is very important to reach sustainability goals. The AI-guided alloy design proposed here focuses on the design of crossover alloys covering a broader range of properties to serve mass markets using full recycled materials. Additionally, the microstructure evolution of these designed steels will be predicted using AI-based methods such as machine learning. The overarching goal of this research is to integrate robust computational tools with experimental characterization to address key challenges in achieving full recyclability of metallic alloys without relying on any virgin materials.

Sustainability aspects

Sustainability is a global priority, and the iron and steel industry alone accounts for around 8% of worldwide CO₂ emissions, highlighting the urgent need for cleaner production routes. Alongside fossil-free ironmaking, increasing the use of recycled scrap offers a powerful pathway toward decarbonization, but requires advanced tools to evaluate and optimize these strategies. This project applies machine learning to design sustainable steels and control their microstructure and properties, contributing to the development of next-generation, eco-friendly materials and manufacturing processes. This project is related to SDGs 7 & 12.

Project team

• Principal applicant: Wangzhong Mu, LTU
• Industrial PI: Joakim Odqvist, Ferritico AB
• Industrial PhD student: Lovisa Åkesson, Ferritico AB/LTU

Project page on the WISE website

• Artificial intelligence guided steel design towards zero-carbon emission contributing to circular economy

Publications related to the project

• Physics-Informed Machine Learning for Steel Development: A Computational Framework and CCT Diagram Modeling
• Integrated machine learning approach for predicting low-temperature embrittlement in duplex stainless steels via ferrite hardening