About this Abstract |
Meeting |
6th International Congress on 3D Materials Science (3DMS 2022)
|
Symposium
|
6th International Congress on 3D Materials Science (3DMS 2022)
|
Presentation Title |
Machine-learning Model to Identify and Classify Dislocations in Aluminum via 3D Dark Field X-ray Microscopy |
Author(s) |
Pin-Hua Huang |
On-Site Speaker (Planned) |
Pin-Hua Huang |
Abstract Scope |
The mechanical properties of metals strongly depend on their structure dislocation composition at the level of the crystal lattice. Methods like annealing alter mechanical properties through interactions of dislocations to form 3D hierarchical structures. Dark Field X-ray Microscopy (DFXM) is a unique new method that was recently demonstrated as being able to resolve subsurface deformations. 3D DFXM was recently shown by our group to be able to construct 4D scans of space and rocking curves into spatially 3D deformation maps of the crystal, to provide direct maps of dislocations. We can now resolve detailed 3D structures of dislocation networks that span 150x300x300-μm3 volumes with 150-nm spatial resolution in post-annealed single crystalline aluminum. My work applies machine learning methods to identify and sort individual dislocations in these 3D datasets. Our approach demonstrates a progress towards full microstructural characterization over length scales now accessible with the new capabilities afforded by DFXM. |
Proceedings Inclusion? |
Definite: Other |