A computational fluid dynamics study of the influence of sleeper shape and ballast depth on ballast flight during passage of a simplified train

The paper assesses the effect on the air flow regime underneath a simplified high-speed train of changing the ballast depth and the sleeper shape, with regard to its potential for causing ballast flight or pickup. The study was carried out numerically using the commercial Computational Fluid Dynamics (CFD) software AnSys Fluent. The flow profile beneath the underbody of the train was generated by means of a moving wall above the track. The Delayed Detached Eddy Simulation (DDES) with the SST [Formula: see text] turbulence model was used to simulate turbulent flow, and the ballast bed roughness was applied parametrically using the wall roughness feature when resolving the boundary layer. CFD simulations were validated for flow over a cube, showing good agreement with experimental results. Up to three different depths to the ballast surface and three different sleeper profiles were investigated. Velocity profiles and aerodynamic forces on cubes placed between or on top of the sleeper blocks were used to assess the propensity of individual ballast grains for movement. For a standard G44 sleeper, increasing the ballast depth and/or the ballast bed roughness was found to reduce aerodynamic loads on an individual ballast grain. A ballast grain on top of the sleeper is more prone to uplift than a grain on the surface of the ballast bed in the crib. A curved upper surface to the sleeper is beneficial in that it prevents ballast from settling on top, the most vulnerable position. However, the reduced flow separation associated with the curved top may increase the likelihood of ballast pickup from the crib. Hence new sleeper shapes intended to reduce the potential for ballast flight should not only prevent ballast from settling on top, but also increase flow separation through the provision of a sharp surface. A prismatic sleeper shape that achieves both is suggested.