This project explores the use of Universal Machine Learning Interatomic Potentials (UMLIPs) in atomistic modeling to develop CALPHAD phase diagrams. The aim is to accurately reproduce known phase diagrams using only calculated phase energies, without relying on experimental thermodynamic or phase data. If successful, this approach could streamline the creation and refinement of phase diagrams, making it particularly valuable for materials development and evaluating new UMLIPs.
The project aims to develop a complete workflow, from UMLIP selection, energy calculations to phase phase diagram generation. By automating this process, the production and validation of new phase diagrams become efficient and scalable.
| 2024–present | MSc in Materials Science, Montanuniversität Leoben, Austria/University of Lorraine, France Tentative thesis title: “Building phase diagrams from atomistic simulations” |
| 09/2020–06/2023 | BSc in Mechanical Engineering at Eindhoven University of Technology, Netherlands Thesis title: “Modelling the Orientation of Fibers during Injection Molding” |