Numerical study of load transfer in conventional and X-shaped sleeper-ballast systems under impact loading
Abstract
To investigate the dynamic response and load transfer in ballasted tracks under impact loading, this study developed full-scale three-dimensional discrete element-multi-body dynamics (DEM-MBD) coupled models for I-shaped and X-shaped sleeper systems. Macroscopic responses, including wheel-rail impact forces, sleeper and subgrade surface accelerations, and subgrade stress distribution, were evaluated, while mesoscopic behavior focused on sleeper-ballast contact distribution, ballast force chains, and particle motion. Results indicate that increasing the wheelset drop height from 10 mm to 30 mm nonlinearly amplified key responses: peak sleeper loads increased from 110.6 kN to 188.8 kN, mean ballast contact forces rose from 38.37 N to 72.16 N, and subgrade surface accelerations escalated from 33.4 m/s² to 78.0 m/s². Compared with conventional I-shaped sleepers, the X-shaped design reduced peak sleeper acceleration by 23% and average ballast contact forces by 40%. Mesoscopic analysis revealed that impact loads were transmitted through concentrated force chains beneath the sleepers, forming a trapezoidal stress diffusion zone. The X-shaped sleeper fundamentally optimizes load transfer by transforming sparse, high-intensity force chains (I-shaped) into a dense network of finer force chains. This enhances particle interlocking, promotes multi-directional stress diffusion, and reduces localized stress concentrations in both the ballast and subgrade, leading to superior impact energy dissipation and overall bearing stability. These findings demonstrate that the X-shaped sleeper improves load transfer efficiency, impact resistance, and vibration mitigation in ballasted track systems.
Fields of Study & Subjects
Dynamics and Control of Mechanical Systems
Fluid Dynamics Simulations and Interactions
Cellular and Composite Structures
Publication details
- Journal / source
- Journal of Railway Science and Technology
- Publisher
- Elsevier BV
- Volume
- 2
- Issue
- 1
- Pages
- 27–40
- Publication date
- 28 February 2026
- Language
- EN