Size effect on contact behavior in DEM simple shear tests

The 2–2.5 times the simulated sand diameter is widely accepted in giving reasonable DEM simulation results for geotechnical testing. However, it neglects the effect of a specimen height to maximum particle diameter ratio in a specific laboratory test, which may lead to a strong stress concentration and flawed simulations. This study compared laboratory simple shear tests with corresponding DEM simulations with different particle sizes. The DEM model used clump rings to simulate physical rings in the test, and decreased the additional stress applied by the widely used wall-type rings. Results showed that (1) DEM models with tested particle size and twofold sand particle size (1D and 2D tests) can better capture the tested stress–strain behavior, volumetric changes, and noncoaxiality, the 4D model has an asymmetrical distribution of contact force and contact number, indicating the specimen is inhomogeneous and has a strong stress concentration. (2) a specimen height to maximum particle diameter ratio smaller than 10 (it is greater than 10 in the ASTM D6528) could provide reasonable macro-meso mechanical behaviors. Similar studies should be carried out after trial tests on determining a reasonable specimen height to maximum particle diameter ratio under the guidance of ASTM D6528.