Modeling constitutive relationship of BT25 titanium alloy during hot deformation by artificial neural network

In this research, the constitutive relationships of BT25 titanium alloy based on regression and artificial neural network (ANN) methods were established and studied by analyzing the results of hot compression tests. The isothermal compression tests were conducted on a Gleeble 1500 thermo-mechanical simulator in the deformation temperatures ranging from 940 to 1000 °C with an interval of 20 °C and the strain rates of 0.01, 0.1, 1.0, and 10.0 s ^-1 with a height reduction of 60%. The average deformation activation energy of the alloy was derived as 623.26 kJ/mol at strain of 0.7 by using the non-linear regression method and assuming a hyperbolic sine equation between the stress, strain rate, and deformation temperature. On the basis of the experimental data samples, an ANN model was proposed and trained. The hot processing parameters of temperature, strain rate, and strain were used as the input variables and the flow stress as the output variable. The comparison of experimental flow stresses with predicted values by ANN model and calculated value by regression method was carried out. It was found that the predicted results by ANN are in a good agreement with the experimental values, which indicates that the predicted accuracy of the constitutive relationship established by ANN model is higher than that using the multivariable regression method.