Effects of varying twist and twist rate sensitivities on the interpretation of torsion testing data

The Nadai and Fields and Backofen expressions are widely used to convert the torque/twist data obtained during torsion testing into shear stress/shear strain curves as well as von Mises equivalent stress/equivalent strain curves. However, when employed in the conventional manner using average values of the twist hardening exponent N, they overestimate the critical strains for the initiation of twinning, dynamic transformation and dynamic recrystallization by comparison with the values determined using compression testing. By contrast, when the local or instantaneous values of the exponent are employed, the torsion and compression results are in good agreement. Another feature of the corrected curves is that they indicate that considerably more dynamic softening takes place during the high temperature deformation of austenite than suggested by the average N flow curves. It is shown that, despite the lack of work conjugacy between the torque-twist and stress–strain curves, the above expressions always lead to the correct constitutive behavior at the external radius.