| About this Abstract |
| Meeting |
2020 TMS Annual Meeting & Exhibition
|
| Symposium
|
Low-cost Titanium: 'Affordable Ti'
|
| Presentation Title |
Machine Learning Assisted Discovery of Affordable Biomedical Ti Alloy |
| Author(s) |
Chunte Wu, Hsiao-Tzu Chang, Shi-Wei Chen, Sih-Ying Huang, Yeong-Tsuen Pan, Joshua Chou, Hung-Wei Yen |
| On-Site Speaker (Planned) |
Chunte Wu |
| Abstract Scope |
In this work, machine learning has led to a revolutionary discovery of biomedical Ti alloy with bone-like modulus and affordable price. Young’s modulus is a critical property for alloy design in orthopedic implant field. However, the present low modulus Ti alloy compositions were constrained in certain intervals with high beta stabilizer content. The constraint prevented discovered alloys from having affordable price. To address this problem, a machine learning based alloy discovery system, named βLow, was developed. It consists of two artificial neural networks for phase stability and Young’s modulus predictions of beta Ti alloys. Following the predictions of βLow, new alloys were produced for model validation and alloy exploration. Surprisingly, βLow not only performed high consistency to experimental results, but also guided the discovery of new alloy with low material price. In conclusion, βLow has opened up a new path for affordable biomedical Ti alloy design. |
| Proceedings Inclusion? |
Planned: Supplemental Proceedings volume |