About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Modeling, AI Applications, and Method Development in Nuclear Materials
|
| Presentation Title |
Data-driven representations of dislocation networks at the atomic and mesoscale |
| Author(s) |
Thomas D. Swinburne, Sylvain Queyreau |
| On-Site Speaker (Planned) |
Thomas D. Swinburne |
| Abstract Scope |
Under sustained irradiation dense dislocation networks form in most metallic alloys. Understanding how these networks respond to stress is a major modelling challenge. Central to all AI/ML driven modelling efforts is representation of dislocation network structures in a form suitable for abstract analysis. While local order parameters typically used to build models of energies and forces work well at the atomic scale, a true mesoscale-to-continuum representation must respect known compatibility laws, e.g. incompressibility of the Kroner-Nye tensor. I will discuss recent work blending these two approaches for a unified description. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Machine Learning, Modeling and Simulation |