| About this Abstract |
| Meeting |
2023 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advances in Multi-Principal Element Alloys II
|
| Presentation Title |
J-7: A Neural Network Model for High Entropy Alloy Design |
| Author(s) |
Jaemin Wang, Hyeonseok Kwon, Hyoung Seop Kim, Byeong-Joo Lee |
| On-Site Speaker (Planned) |
Jaemin Wang |
| Abstract Scope |
We suggest a novel and robust neural network model to relieve the burden of searching vast compositional space of high entropy alloys (HEAs) and confirm the superiority of the model by comparing its performance with several other models. The reasons for the high accuracy of the model are discussed around thermodynamic descriptors calculated by thermodynamic modeling and the effect of the convolution neural network in the model. Furthermore, we inverse-predicted using the model to design HEAs with good mechanical properties and conducted experimental verification of the designed HEAs to prove the validity of the model and alloy design method. The strengthening mechanism of the designed HEAs is further discussed by analyzing microstructure and calculating the lattice distortion effect. We calculated the lattice distortion effect by performing molecular dynamics simulation using a 2NN MEAM interatomic potential. This study demonstrates the reliability of the alloy design approach with the machine learning model. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, High-Entropy Alloys, Computational Materials Science & Engineering |