Physically based constitutive modeling for Ti17 alloy with original basketweave microstructure in β forging: A comparison of three approaches

Based on hot compression tests carried out at 1173–1223 K, 0.001–10 s−1 for Ti17 alloy with original basketweave microstructure in β forging, three physically based constitutive modeling approaches for both peak stress and flow stress considering strain compensation are studied, and the corresponding prediction accuracy and deformation/diffusion mechanisms are analyzed. The established physically based constitutive equation with a fixed creep exponent n = 5 presents poor predictability, which implies that dynamic recovery (DRV) controlled by the glide and climb of dislocations is not the only deformation mechanism in β forging of the alloy. On the other hand, the physically based constitutive equation with exponent n as variable demonstrates good predictability with average absolute relative error (AARE) of 6.18%, correlation coefficient (R) of 0.997 for the peak stress, and AARE of 6.73%, R of 0.995 for the flow stress respectively. This suggests that thermal diffusion mechanism should include lattice diffusion in β forging of the alloy. Furthermore, a revised physically based constitutive model considering the coupling effects of the lattice diffusion and the grain boundary diffusion shows better predictability with AARE of 4.69%, R of 0.994 for the peak stress, and AARE of 4.86%, R of 0.998 for the flow stress respectively. This indicates that thermal diffusion mechanism includes not only lattice diffusion but also grain boundary diffusion in β forging of the alloy. It is believed that the occurrence of DRX in β forging of the alloy is responsible for the obtained exponent n = 2.831, which deviates from the theoretical value of 5. •Physically based constitutive modeling is studied for Ti17 alloy with original basketweave microstructure in β forging.•Three types of physically based constitutive models for both peak stress and strain compensated flow stress are constructed.•The prediction accuracies of the established constitutive equations are compared and analyzed.•The modeling approach considering both deformation and diffusion mechanisms shows the best predictability.