Modelling of discontinuous composites using AI

Purpose and goal:

The main goal with this project was to create collaboration and increase knowledge in AI methods within RISE. More specifically to become familiar with the work by Professor Xin Liu and his team; (ii) to apply their innovative approach using AI for modelling discontinuous composites; (iii) establish a long-term collaboration. The objectives have been successfully fulfilled (the collaboration is still ongoing so hopefully it will continue for long time).

Expected results and effects:

The AI developed within this project can learn from the simulations of many composite samples (~20k) that have been subjected to compression along the fibres until final failure. Then, a trained AI was able to predict the behaviour (stress-strain curve) of a new sample. This is the first time that anyone has shown this capability AI in composite materials.

This results have translated into a (i) conference paper was written and will be presented at the ASME IMECE 2022 Conference; (ii) work is ongoing to write a Journal paper; (iii) additional collaboration with GKN Aerospace Sweden AB Approach and implementation:

This type of project was very efficient in term of acquiring new competence in a short time frame. The physical presence in UTA and close discussions with Dr. Xin Liu gave me insights into research that would take years to acquire from the literature. It allows two different researchers to merge their knowledge from different areas and create new knowledge.