|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Advanced Characterization Techniques for Quantifying and Modeling Deformation
||Understanding Effect of Texture and Topology on Stress Hotspots Using Machine Learning
||Ankita Mangal, Elizabeth A. Holm
|On-Site Speaker (Planned)
An applied stress is distributed inhomogeneously in a material, and the regions of high stress in a microstructure are called stress hotspots. Such regions are related to the nucleation of voids in ductile fracture. The effect of grain neighborhoods and orientation on stress hotspot formation is studied by keeping one effect constant while varying the other during crystal plasticity simulations. Microstructural fingerprints are then extracted to develop features which are used by a machine learning algorithm to predict stress hotspot formation as well as, delineate the local microstructure descriptors causing it, thus paving the way for materials design. The results demonstrate the power of combining traditional simulations with data driven methods to analyze structure-property relationships.
||Planned: Supplemental Proceedings volume