Hot deformation behavior and artificial neural network modeling of β-γ TiAl alloy containing high content of Nb

Abstract

In this work, the hot deformation behavior of high Nb-containing TiAl alloy with β + γ phases was investigated. Elongation, twisting and bending of the microstructure are the primary deformation phenomena at low temperatures and high strain rates, whereas high temperatures and low strain rates are beneficial for dynamic recrystallization, and the grains are equiaxed. Based on the experimental flow stress data, the Arrhenius constitutive modeling was established. In addition, an artificial neural network was innovatively utilized to propose a high temperature constitutive modeling. The results denote that the accuracy of artificial neural network is higher than the Arrhenius equation. The proposed artificial neural network modeling provides high accuracy, and the average absolute relative error and correlation coefficient are 1.28 % and 0.99988, respectively. Thus, this is a new approach to establish a constitutive modeling for TiAl alloys.