Relation between ASTM E606 specimen geometry and misalignment in strain-controlled fatigue testing

Abstract

In strain-controlled fatigue testing, a conventional extensometer is usually used to measure the average strain between its two blades. However, the actual strain distribution in the gauge section cannot be revealed by the extensometer. To measure this distribution during strain-controlled fatigue testing of 304 stainless steel, stereo Digital Image Correlation (stereo-DIC) was used in this study. Three different geometries with two of them satisfying ASTM standard E606/E606M-12 were tested in strain-controlled fatigue testing with two representative strain amplitudes of ±1.0% and ±1.5% on a properly aligned testing machine. At both strain amplitudes, the two geometries in line with ASTM standard E606/E606M-12 exhibit evident nonuniform strain distribution under compression already from the very first cycle while the geometry proposed by the authors (not according to the ASTM standard E606/E606M-12) has better uniformity in axial strain distribution.Next, a novel fixture is designed to monitor the misalignment of the specimens in the fatigue testing. Benefiting from the novel fixture and the stereo-DIC technique, not only the deformation in the gauge section of the specimen can be captured but also the misalignment between both ends of the specimen can be extracted by post-processing. The measured misalignment has been further used as boundary conditions in simulating the mechanical response of those three geometries under cyclic loading. Numerical simulation has successfully reproduced the nonuniform strain distribution under compression, which validates the stereo-DIC measurements. It suggests that while even following the ASTM standard E606/E606M-12, uniformity in strain distribution under compression cannot be ensured and thus might render the constitutive laws calibrated from strain-controlled fatigue testing invalid, since uniform stress and strain across the gauge section of the specimens were falsely assumed.