DEM analysis of railway ballast behavior under shaking table tests

Railway ballast, as a typical unbound granular material, forms the core of the ballast bed and plays a pivotal role in ballasted track systems. Under dynamic excitations, the unbound granular ballast is prone to particle rotation, sliding, and even flow, potentially compromising the stability and serviceability of the track as an engineering structure. To investigate these issues, Zacher and Baeßler (2008) conducted shaking table experiments to investigate the vertical and lateral behavior of ballast aggregates; however, their numerical simulations were limited to the use of the commercial software PFC2D (Particle Flow Code 2D). As a result, the inherently three-dimensional nature of ballast particles, including their complex shapes and angularity, was not considered, and the dynamic behavior of ballast assemblies remains insufficiently understood. To overcome these limitations, this study employs advanced methods, combining laser 3D scanning and image processing with PFC3D to accurately capture the complex geometries of ballast particles. In accordance with ballast particle grading specifications, a specified number of particles is prepared and compacted to form ballast bed with unbound granular structure. Small-scale sleeper is then positioned at designated locations within the box. A 3D discrete element model of the sleeper–ballast bed is subsequently established and rigorously validated against empirical data from prior shaking table experiments. The validated model is employed to systematically investigate the dynamic response of the ballast–sleeper system under controlled vertical and lateral excitation, allowing systematic variation of excitation levels to identify the mechanisms and thresholds governing ballast assembly instability.