About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Algorithms Development in Materials Science and Engineering
|
| Presentation Title |
Graph-Theoretical Approach for Defect Detection in Polycrystalline Atomic Specimen |
| Author(s) |
Younggak Shin, Vichhika Moul, Keonwook Kang, Byeongchan Lee |
| On-Site Speaker (Planned) |
Younggak Shin |
| Abstract Scope |
Wigner-Seitz analysis is a widely used method for characterizing point defects in atomistic simulations involving massive defect formation such as collision cascades. However, in polycrystalline systems, structural changes at grain boundaries during defect generation can lead to mismatches between the post-cascade structure and the reference lattice, resulting in significant false positives near grain boundaries. To address this, we propose a defect detection framework that transforms non-crystalline atoms identified via Common Neighbor Analysis into a graph-based representation. This approach enables the separation of evolving grain boundary structures from other defects and facilitates the classification of interstitials and vacancies using pattern recognition techniques. Our method demonstrates accurate identification of grain boundary structures and clustered point defects without relying on a fixed reference lattice which makes it suitable for analyzing complex defect configurations in polycrystalline materials under extreme conditions. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Other, Other, Other |