Scientific machine learning for advancing the understanding of the 4D microstructural evolution in metals (SciML4D)

This project aims to develop Scientifically Informed Generative Artificial Intelligence (SIG-AI) methods to analyze and predict the four-dimensional (x, y, z, time) microstructural evolution in industrially relevant aluminum alloys. By integrating advanced synchrotron X-ray nanotomography data with physics-based modeling. To overcome these issues, we will develop Scientifically Informed Generative Artificial Intelligence (SIG-AI) methods, inspired from the emerging research fields of Scientific Machine Learning (SciML), Generative AI (GenAI), and new powerful data driven methods, such as Conditional Flow Matching (CFM). We will develop and demonstrate the SIG-AI capabilities for 4D (x,y,z,time) studies of the microstructural evolution in industrially relevant aluminium alloys. If successful, the concept may easily be expanded to other metallic systems and processes, and thus bring the field to a whole new advanced level.