Three different mathematical models to predict the hot deformation behavior of TA32 titanium alloy

To characterize the TA32 titanium alloy (Ti–5.5Al–3.5Sn–3.0Zr–0.7Mo–0.3Si–0.4Nb–0.4Ta, wt%) hot deformation behavior, the isothermal constant strain rate tensile tests in range of temperature 600–720 ℃ with strain rate 0.0001–0.1 s −1 were performed. Based on strain compensation of Arrhenius constitutive, physically based constitutive and BP neural network constitutive model were used to regress the high-temperature flow stress and the prediction accuracy were evaluated by correlation coefficient and average relative error. A cross-validation approach was used to evaluate the predictability of the three constitutive models. The result show that the tensile strength is more sensitive to strain rate than deformation temperature, according to the fitting results and sixteen cross-validation tests indicate that the BP model can provide the most accurate prediction for TA32 hot flow behavior. As revealed by micrographs, both high temperature and low strain rate are benefited for dynamic recrystallization and spheroidization with β phase. Graphical abstract Cross-validation of three models.