|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Computational Materials Discovery and Optimization
||Learning Grain Boundary Properties from Macroscopic and Microscopic Structural Descriptors
||Ankita Mangal, Ian Chesser, Elizabeth A Holm
|On-Site Speaker (Planned)
The large number of possible grain boundary structures in crystalline materials makes correlation between grain boundary structure and materials properties difficult. Macroscopic crystallographic degrees of freedom of a grain boundary are typically considered separately from microscopic degrees of freedom, which comprise local variations in atomic structure. In this work, we use data mining techniques to construct grain boundary descriptors from both macroscopic crystallographic and local atomic parameters. These descriptors are inputs to a random forest algorithm based machine learning model which predicts grain boundary properties such as energy and mobility. The relative predictive power of different classes of features is compared. This work quantifies the importance of macroscopic versus microscopic parameters for predicting grain boundary energy and mobility.
||Planned: Supplemental Proceedings volume