About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Algorithms Development in Materials Science and Engineering
|
| Presentation Title |
Data-Driven Analysis of Atomistic Simulations in a Symmetry-Adapted Basis |
| Author(s) |
Nithin Mathew, Avanish Mishra, Edward Kober |
| On-Site Speaker (Planned) |
Nithin Mathew |
| Abstract Scope |
Atomistic simulations have played a central role in investigation of defect properties and unit mechanisms pertaining to physical processes in materials. We will present a method to characterize atomistic simulations using a complete and symmetry-adapted set of atomic environment descriptors, namely the Strain functional descriptors (SFDs), in conjunction with machine learning methods. The SFDs are derived from invariants of nth order central moments of local number density, based on a Gaussian kernel. We will demonstrate that SFDs, which characterize shape, size, and orientation of local atomic environments, can be mapped to second and higher order deformation metrics as defined in continuum mechanics and represent a complete basis for describing deformation. In conjunction with data-driven analysis methods, we will demonstrate the power of SFDs to characterize grain boundaries, identify defects in molecular dynamics (MD) simulations, and describe high strain-rate deformation behavior of high entropy alloys from MD. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Machine Learning, Mechanical Properties |