Machine learning design of printable high-temperature strength Aluminum alloys

The properties and processability of metals govern the design and performance of structural components in our cars, aircraft, and buildings. The advent of additive manufacturing (AM) with new processing conditions and a potential ability to tailor the alloy composition and microstructure at the voxel size resolution has opened novel routes for alloy design to achieve unprecedented properties. However, exploiting all these benefits requires transformation of design concepts and development of new numerical tools tailored to AM. In this talk, I will present how, by benefiting from rapid solidification and local melting in AM, and combining ICME techniques and machine learning tools, we designed a record high-strength, high-temperature creep-resistant printable Al alloy that outperforms conventionally processed alternatives. I will show how the design will be changed for high-demand industries when costs and sustainability metrics are important factors. Furthermore, I will show how incorporation of advanced machine learning algorithms such as LLMs will make the design more efficiently. The proposed hybrid frameworks offer new perspectives for the discovery of next-generation structural metallic materials to significantly transform industrial applications ranging from aerospace, construction, infrastructure, automotive, and energy sectors to microelectronic devices and biomedical implants.