AI for AM – towards more resource-efficient development of additively manufactured light-weight alloys

Research question

Mg is a highly reactive material, but recent advances in AM technology has enabled its production using powder bed fusion with laser beam (PBF-LB), and the PI’s group was one of the first to achieve this on an industrial-scale machine. However, an extensive amount of research is still needed to develop improved microstructures allowing for acceptable corrosion rates. Due to the large number of factors affecting the process, e.g. laser power, beam spot size, scanning rate, hatch distance, powder characteristics, gas flow etc, and the fact that process parameter transfer between machines is very difficult, a more resource-efficient, less material- and energy-consuming way of developing parameters is sought for. The advent of increasingly advanced machine learning models provides a possibility to take leaps forward in not only a faster development of processes, but also to better understand the relationship between process parameters, process mechanisms and the resulting material microstructure.

Sustainability aspects

We address mainly SDG numbers 8: Decent work and economic growth, and 12: Responsible consumption and production. The project will both contribute to both i) goal 8 in terms of resource-efficient use of materials for processes enabling an increased value of (raw) materials; as well as ii) goal 12 through efficient (re)use/recycling of materials for sustainable production/consumption, by a) using less material through the choice of AM as a manufacturing process (once the research phase is complete and parameters are developed); b) enabling a lower material and energy use during process parameter development through progress in machine learning models tailored for the process; c) consisting in a first step towards tools enabling more efficient AM processes also once in production.