Investigating side-wind stability of high speed trains using high resolution large eddy simulations and hybrid models

Abstract

Crosswind flow over high speed trains can pose serious safety concerns for rail transport. Methodologies for evaluating the aerodynamic forces exerted on the train include full-scale measurements, physical modeling using wind-tunnel experiments and numerical modeling using computational fluid mechanics (CFD). Although CFD presents the most cost-effective approach, it faces severe uncertainties in the predicted forces, most of which are related to the turbulence modeling technique employed. In here we investigate the influence of various turbulence modeling approaches on crosswind flow simulations and calculated force coefficients. In particular, we perform URANS, LES and DDES simulations utilizing the DLR Next Generation Train 2 model geometry. Particular emphasis is laid on simulating a wind angle of 30 degrees and Reynolds number of 225,000 for which validation data is provided by wind tunnel measurements. We confirm that a major vortex system on the leeward side of the train develops, which mainly drives the overturning force and moment of the train. The lift force is determined mainly by the underbody flow, which is characterized by unsteady vortex shedding. Due to its dual ability to properly model the roof boundary layer on the one hand and to resolve small-scale turbulent eddies in the underfloor region on the other, the DDES approach is found to give the most accurate force predictions. LES overpredicts the overturning force and moment, while URANS overpredicts the lift force.