| About this Abstract |
| Meeting |
TMS Specialty Congress 2025
|
| Symposium
|
The 7th International Congress on 3D Materials Science (3DMS 2025)
|
| Presentation Title |
Scaling Ductility From Microscale to Bulk by Coupling Crystal Plasticity Simulations With Deep Learning |
| Author(s) |
Laura Z. Vietz, Rebecca G. Divine, Carter K. Cocke, Ashley D. Spear |
| On-Site Speaker (Planned) |
Laura Z. Vietz |
| Abstract Scope |
Characterizing ductile fracture in bulk materials can be challenging and cost-prohibitive, namely for materials used in harsh environments like nuclear reactors. In these cases, experimental characterization is limited to sample sizes below the representative volume element (RVE), called statistical volume elements (SVEs), which exhibit scatter in mechanical properties due to size effects. The RVE is the smallest volume of material above which a property of interest converges to bulk behavior. This research aims to link microstructure-dependent SVE-derived mechanical properties to bulk-scale properties by combining crystal plasticity simulations with deep learning. As a proof-of-concept, a generated 3D dataset trained a convolutional neural network to predict RVE ductility from SVE responses. The dataset is populated with different-sized SVEs and their corresponding stress-strain responses for four materials simulated using a damage-enabled large-strain elasto-viscoplastic FFT model for ductile fracture. This work could enable a cost-effective method for characterizing bulk properties using small-scale test specimens. |
| Proceedings Inclusion? |
Undecided |