Shear band propagation from a crack tip subjected to shear wave loading

Pressure-shear plate impact experiments have been used to study the response of semi-infinite cracks in 4340 VAR steel subjected to either pure Mode II or pure Mode III loading. The material is heat-treated to achieve a fully-martensitic microstructure, which is confirmed through hardness tests and optical microscope observations of etched samples. Half-through, mid-plane cracks are grown in the target plates by means of fatigue loading. In order to obtain pure shear wave loading, flyer and target plate geometries are designed to ensure that the compressive wave passes through the crack plane before the shear wave arrives. The rear surface motion of the target is monitored by a combined normal displacement interferometer (NDI) and transverse displacement interferometer (TDI). For the case of Mode II loading, an improved TDI using a combined lens-beamsplitter unit is introduced to obtain high quality records in the presence of substantial rotation of the rear surface of the target. Both normal velocity-time profiles and transverse velocity-time profiles are obtained through post-processing from interferometric data. Numerical simulations based on both elastic and elastic viscoplastic response indicate that the deformation is essentially elastic except near the crack tip. Simulated velocity-time profiles for the case of a stationary crack agree quite well with those obtained in the experiments. Numerical simulations based on elastic response are also used to probe the effects of various parameters, e.g. friction coefficients, surface roughness, surface flatness, tilt, and etc., on the velocity-time profiles.; Microscopic examinations of the recovered samples show long shear bands emanating from the tip of the fatigue crack. Numerical simulations using a power law viscoplasticity model are performed on a supercomputer to gain insight into the shear band formation process. Mesh size effects of the FEM calculations are noted and extrapolation is used to obtain temperature distributions for a mesh with zero size. The length of the band and the level of the temperature rise fall far short of what are suggested by microscopical observations. Further simulations following adjustments of power law model parameters to enhance localization fail to produce the shear band features observed in the experiments.; Several closely related investigations are reported in appendices. An analytic approach employing a Green's function is used to estimate a shear band speed and a displacement jump across the band for the case of a mode III crack problem. Further improvement of this approach by including a band of failed material is also considered. Weiner-Hopf methodology is used for the Mode II case to obtain an analytical solution for assessing the accuracy of the FEM simulations. Finally, an attempt to remove the reflected tensile wave by introducing a transparent window is also described.