A comparison study at the flow stress prediction of Ti-5Al-5Mo-5V-3Cr-1Zr alloy based on BP-ANN and Arrhenius model

A series of hot compression tests were conducted on a Gleeble-3500 isothermal simulator to obtain the hot flow curves of Ti-5Al-5Mo-5V-3Cr- lZr alloy and the specimens were compressed with the height reductions of 60% under the deformation temperatures of 973, 1023, 1073, 1123 K and the strain rates of 0.001, 0.01, 0.1, 1 s(-1) . The corresponding back-propagation artificial neural network (BP-ANN) model and the Arrhenius model for this alloy were constructed on the basis of the obtained flow curves for flow stress prediction. Subsequently, the constructed BP-ANN model was proved to be better by comparing the prediction accuracy with the developed Arrhenius model according to statistic calculations. The relative error and the standard deviation for BP-ANN model were calculated to be 1.4714% and 2.2271%, while for Arrhenius model, the corresponding values were 1.2213% and 5.3641%, respectively. Besides, the correlation coefficient of BP-ANN model is 0.9949 and it is 0.9761 for Arrhenius model. The average absolute relative error for BP-ANN model is 2.2836% and it is 23.4527% for Arrhenius model. Finally, the flow curves were extended on the basis of the BP-ANN model, which is believed to be helpful to achieve high accuracy in finite element simulation.