Paving the way towards digital materials selection using AI and physical models (DMS-AI)

Important results from the project

The DMS-AI project successfully showed how open databases and combination of AI and physical models can assist in predicting complex material properties. It delivered a literature review, data survey, workshops, and an open-source, user-friendly tool for all partners as example case study. It also provided industry-adapted education and identified physics-informed ML as a promising future approach, laying the foundation for a full-scale project on sustainable, AI-driven material development.

Expected long term effects

The project is expected to accelerate material development and selection by integrating physics informed AI and ML into industrial workflows. Tools like this can connect different part of industrial chain, reduce experimental testing needs, improve efficiency, and support sustainability. By fostering collaboration and data sharing between project partners, it helps build robust databases and digital tools, positioning Sweden as a leader in AI-driven materials innovation.

Approach and implementation

DMS-AI was structured as a feasibility study, literature review, and workshops involving material producers, end-users, institutes, and universities. Two full-day workshops facilitated knowledge exchange and identified industry needs. Literature reviews and data surveys supported the development of a user-friendly feasibility tool focused on corrosion and fatigue as two example cases. Results were shared through report, code, and presentations, laying the groundwork for a full-scale project.