Assessing the effect of friction on compression split Hopkinson pressure bar tests

An enhanced understanding of material behaviour during rapid loading allows designers to improve the safety and efficiency of industrial processes and commercial products. These improvements are due, in part, to the use of increasingly powerful and sophisticated numerical simulation codes, which in tum, depend on accurate high strain rate material data. A Split Hopkinson Pressure Bar (SHPB) facility had previously been developed at the University of Cape Town to allow high strain rate testing of materials in compression. However, uncertainties regarding aspects of the test method that can affect the accuracy of the results, such as interface friction and specimen inertia, still required further clarification. This report details a thorough experimental and numerical investigation into the effect of friction on SHPB test data. The objectives are to assess the magnitude of the effect of friction under various conditions and to suggest strategies for reducing friction error in SHPB tests to an acceptable level. The ring compression test was used to obtain experimental friction factors. The effect of surface finish, lubricant, and strain rate on the friction experienced by mild steel, copper and aluminium samples was investigated. Numerical simulation was used to assess an energy-based analytical solution by Avitzur [1], and in particular to establish the effect of neglecting barrelling. Avitzur's analytical solution [l] Was then used to interpret the experimental results. The tested specimen microstructure was examined and used to estimate the stress distribution in the specimen during deformation. Uneven deformation and fold-over diminished at higher strain rates. Optimal surface finish and lubricant conditions were found for which experimentally measured coulomb friction coefficients lay between J.l = 0.04 and 0.08, with copper samples exhibiting marginally higher friction. By Avitzur's analytical solution [1] the error in SHPB tests under these recommended conditions was estimated to lie between 1% and 2%. The results show that roughened compression and specimen surfaces, lubricated with a suitable grease containing molybdenum disulphide, are useful in effectively reducing the error in SHPB tests due to friction effects