Modeling the constitutive relationship of Ti-22Al-25Nb alloy using artificial neural network

Constitutive equation reflected the highly nonlinear relationship of flow stress as function of strain, strain rate and temperature is a necessary mathematical model used to describe basic information of materials deformation and finite simulation. In this paper, hot compression experiments on Ti-22Al-25Nb alloy were conducted with Gleeble 1500 thermal simulator at different temperatures and strain rats. High temperature constitutive predicting model is developed using BP neural network method based on the data obtained from experiments (deformation temperature 940°C~1030°C and strain rate 0.001s^-1~1s^-1), and compared with the traditional regression method. It is found that the neural network model provides a better representation of the test data than the commonly used traditional mathematics model. Moreover, in that the complicated nonlinear relationship of thermodynamical parameters can well be described by the network model when the alloy is deformed at high temperature, it is not only a convenient but also effective way to establish the model of constitutive equations for alloys.