Modeling of sand particle crushing in split Hopkinson pressure bar tests using the discrete element method

Abstract

The discrete element method (DEM) has been extensively used to study the micro-mechanical behavior of sand subjected to quasi-static strain rates. In most of these studies, the sand specimens were prepared with an upscaled particle size distribution (PSD) to reduce the computational cost. However, the effectiveness of this upscaling approach in replicating the dynamic sand response is not well understood, especially at high stresses where significant particle breakage occurs. In this paper, several split Hopkinson pressure bar (SHPB) tests reported in literature are modeled in DEM using two methodologies: 1) applying the reported strain rates directly to the DEM specimen without accounting for wave propagation in the incident and transmission bars; and 2) modeling the complete SHPB test setup including the incident and transmission bars. The results show that with well-calibrated parameters for contact behavior and particle crushing, specimens with upscaled PSD provide similar dynamic stress-strain response and PSD evolution as those reported in literature. This study shows the importance of particle breakage in the stress-strain response under high stresses as the specimens without particle breakage provide a much stiffer response compared to the specimens with particle breakage. The developed DEM model may be a useful tool to model the complete SHPB test setup as the incident, reflected and transmitted stress waves can be accurately replicated.