|About this Abstract
||2019 TMS Annual Meeting & Exhibition
||Computational Approaches for Big Data, Artificial Intelligence and Uncertainty Quantification in Computational Materials Science
||Reduced Order Crystal Plasticity Modelling for ICME Using a Machine Learning Approach
||Mengfei Yuan, Sean Paradiso, Bryce Meredig, Stephen R. Niezgoda
|On-Site Speaker (Planned)
The application of complex materials property and processing models within an Integrated Computational Material Engineering (ICME) framework is often too computationally expensive. What is needed is a set of reduced order, fast acting models which can be rapidly exercised and inverted for use within an optimization framework. This work applies a machine learning approach to establish reduced order crystal plasticity modelling for both cubic and noncubic polycrystalline materials. The development and calibration of the model is discussed, and the machine learning model is demonstrated to correctly predict the plastic stress-strain behaviors and the texture evolution under a range of deformation conditions and strain rates. The proposed model can also be inverted to predict optimal crystal plasticity single crystal hardening parameters given the macroscale stress-strain response. Application of the framework for ICME, design, and construction of materials databases will be discussed.
||Planned: Supplemental Proceedings volume