Constitutive relationship of TC4 titanium alloy based on back propagating (BP) neural network (NN)

Abstract

Using Gleeble-3800 thermal simulation testing machine, the TC4 titanium alloy was subjected to hot compression experiments under the conditions of deformation temperature of 810 – 950 °C, strain rate of 0.001 - 1s-1. The research shows that the flow stress of TC4 titanium alloy is more sensitive to the deformation temperature and strain rate during thermal deformation, and it increases with the decrease of the deformation temperature and the increase of the strain rate. Based on BP neural network, a constitutive model of TC4 titanium alloy α+β two-phase region is established. The correlation coefficient reaches 0,996, which proves that the model can predict the high temperature flow stress of TC4 titanium alloy.

Keywords

• TC4 titanium alloy
• compression test
• constitutive model
• BP neural network
• stress-strain curve