On the use of non-cylindrical specimens in a split-Hopkinson pressure bar

Despite the practice of using cylindrical specimens in a split-Hopkinson pressure bar (Kolsky bar) experiment, the use of non-cylindrical prismatic specimens is not uncommon. This is convenient when testing extra-soft materials like brain tissues, muscles, or samples that are brittle and cannot be machined to an exact cylindrical shape (like bone-samples). The use of a non-cylindrical sample with a flat surface also renders the specimen amenable to a two-dimensional image correlation algorithm. This research aims to show the feasibility of using non-cylindrical specimens in a Kolsky bar. For this, experiments were conducted with a model material for different model cross-sections at a nearly constant strain rate in the split-Hopkinson pressure bar. The findings suggest the use of a suitable characteristic cross-section dimension of the specimen to determine the critical slenderness ratio while selecting a non-cylindrical prismatic specimen. It has been shown that if the specimen design is governed by the suggested slenderness criterion, then there is no effect of specimen length or cross-sectional shape on the stress-strain curve of the material. Through the use of a computational code, the research also shows the effect of non-uniform axial stress distribution along the cross-section of the specimen, resulting due to specimen geometry. On quantification of the stress non-uniformity along the cross-section of the specimen, the findings indicate that the magnitude of the non-uniformity is both small and temporary.