|About this Abstract
||2022 TMS Annual Meeting & Exhibition
||Environmental Degradation of Multiple Principal Component Materials
||Machine Learning Potential for High Entropy Alloys
||Qiang Zhu, Yanxon Howard, Pedro Santos, Xiaoxiang Yu, Yunjiang Wang
|On-Site Speaker (Planned)
In the past, the simulation of high entropy alloys (HEA) was largely based on the classical interatomic potentials. In this talk, we present PyXtal_FF—a package based on Python programming language—for developing machine learning potentials (MLPs) for more accurate modeling that is close to the level quantum mechanic simulation. The aim of PyXtal_FF is to promote the application of atomistic simulations through providing several choices of atom-centered descriptors and machine learning regressions in one platform. In particular, PyXtal_FF can train MLPs with neural network models, by simultaneously minimizing the errors of energy/forces/stress tensors in comparison with the data from ab-initio simulations. The trained MLP model from PyXtal_FF can be interfaced with the ASE and LAMMPS packages for large scale atomistic simulations. We will illustrate the performance of PyXtal_FF by applying it to investigate the BCC NbMoTaW, as well as several other prototypical binary systems.
||High-Entropy Alloys, Computational Materials Science & Engineering, Machine Learning