Investigation of Failure Criterion in Dynamic Torsion Tests with Solid Cylindrical Specimens

When investigating the limiting states of materials under dynamic loading conditions, it’s important to specify the dependency of plastic failure strain on the stress state. Usually, such dependence is build upon the experimental data obtained from dynamic tests in tension and compression of solid cylindrical specimens with different working part geometry, followed by a monotonic extrapolation. In the recent studies [1] the existence of complex, non-monotonic dependence of failure strain on the stress state parameters is shown for a number of materials. In these cases, a mentioned set of tests is not enough to construct a reliable criterion relations. In statics, one of the most informative experiments for the failure criterion construction is a torsion test on solid or thick-walled cylindrical specimens. Although a nonuniform stress state arises in the sample in this case, effective methods of its interpretation are developed [2,3]. The theory of this experiment conformably to the dynamic processes at large plastic strains has not yet been developed. Using the LS-DYNA implemented virtual test bench, the experimental setup for the solid cylinder torsion test with high strain rates and methods of its stress state identification are 2 4discussed. It is shown that for the strain rate range of 10 -10 1/s the kinematic hypotheses that are taken in the quasi-static torsion are valid, that allows the effective use of known methods of the sample’s stress state decoding.