Neural network-based intrinsic structure relationship of TC20 titanium alloy for medical applications

Abstract

Isothermal constant strain rate compression experiments were carried out on TC20 titanium alloy using a Gleeble- 1500 thermal simulation tester to investigate its high temperature flow behaviour at deformation temperatures of 750 - 900 °C and strain rates of 0,001 - 1 s-1. The results show that the flow stress basically decreases with increasing deformation temperature and increases with increasing strain rate. The correlation coefficients and mean relative errors were 0,998 and 5,06 % respectively, proving that the BP neural network-based intrinsic structure model is effective in predicting the flow stress of the alloy.

Keywords

• TC20 titanium alloy
• hot compression
• stress-strain curve
• constitutive model
• BP neural networks