Neural Network Regressor for Designing Biomedical Low Elastic Modulus Ti-Zr-Nb-Mo Medium Entropy Alloys

Abstract

The excellent biocompatibility of Ti and Zr alloys makes them the best candidates for orthopedic implantations. The design of high Ti and Zr-containing alloys that show low Young's modulus for implant manufacturing is the objective of this work. Here, a feed-forward-back propagation neural network was used to speed up the design process and optimize alloy composition. The β-typeTi45-Zr39-Nb12-Mo4 alloy is designed and showed promising properties. The alloy showed a low elastic modulus of 78 GPa and a high yield strength of 891 MPa resulting in a high elastic admissible strain that made it suitable for orthopedic applications.