| About this Abstract |
| Meeting |
MS&T22: Materials Science & Technology
|
| Symposium
|
Integration between Modeling and Experiments for Crystalline Metals: From Atomistic to Macroscopic Scales IV
|
| Presentation Title |
Alloying Design and Deep Learning Applications for Concentrated and High-entropy-Driven Ni-based Superalloys |
| Author(s) |
Ridwan Sakidja, Marium Mostafiz Mou |
| On-Site Speaker (Planned) |
Ridwan Sakidja |
| Abstract Scope |
In this work, we have developed an alloying design that is based on the application of Deep Learning to model the mechanical properties of concentrated and/or high-entropy-driven Ni-based Superalloys. In developing the atomistic-based model, we utilized the results from electronic structure calculations generated from ab-initio molecular dynamics including the energy, forces, and virial database, and constructed the interatomic potentials for the multi-component system through Deep Learning algorithms. The compositions used can be strategically sampled and varied to maximize their statistical variations so as to ensure a wide range of compositional/phase transferability of the potentials. The validity of the developed potentials was then tested through molecular dynamics (MD) simulations to model various types of thermomechanical properties that are consistent with the experimental observations. We would like to acknowledge the support from the National Energy Technology Laboratory (NETL) and also National Energy Research Scientific Computing Center (NERSC) with their supercomputer resources. |