Boundary effects in DEM simulations of ballast under cyclic simple shear

Abstract

Ballast is a key component of railway infrastructure, and its mechanical response can be assessed through laboratory testing to support construction and maintenance strategies. Due to the large particle size and the dependence on boundary measurements to characterise the mechanical behaviour of the ballast, the influence of boundary conditions on test results is critical. This study investigates the influence of the number of particles in contact with the top loading cap on the response of ballast subjected to cyclic simple shear using DEM. The initial contact condition is systematically varied to assess its effect on macroscopic behaviour and micromechanical evolution of the sample. Results show that the number of top boundary contacts strongly affects the macroscopic response during the first loading cycles. However, these differences progressively diminish with cycling, and samples with more than 50% of the maximum achievable contacts exhibit comparable macroscopic behaviour. At the micromechanical level, the coordination number in the central and bottom regions is largely unaffected by the top boundary condition. Furthermore, the top boundary effect primarily influences the strong normal force network, while the weak network remains largely isotropic. Finally, bulk force anisotropy is higher and less reversible than bulk fabric anisotropy. These findings highlight the importance of carefully performing the sample preparation, as it strongly influences the boundary conditions, therefore, the interpretation of laboratory measurements.