COMPUTATIONAL MODELLING OF CROSS-WIND STABILITY OF HIGH- SPEED TRAINS

This paper presents first results of a joint experimental and computational investigation into the prediction of the cross-wind stability of high-speed trains (HST). Attention is confined to the aerodynamic loads on the first car of a generic HST when exposed to a range of yaw angles (0°<β<60°) at Re-numbers of Re = 0.6~1.4�10 6 . The focal points of the study are CFD investigations referring to open wind-tunnel experiments performed for a 1:10 simplified ICE2 steering car model. Accordingly, results are reported for measured forces. Comparisons are made with regards to different levels of geometric complexity, addressing the issue of bogies, spoilers and moving ground. Computational results are obtained from the commercial pressure-based finite-volume code STAR-CD, utilising a co- located hexahedral grid arrangement with up to 6.5 million cells. Various computational modelling approaches scrutinising the impact of turbulence modelling, numerical approximations and meshing strategies are investigated. The paper aims to assess the predictive prospects of CFD - which presently is still considered to be immature for the prediction of cross-wind stability - and tries to give best-practice recommendations for the investigated type of application. Supplementary the related consequences on the mechanical behaviour, i.e. the stability of the vehicle, are briefly addressed by means of a quasi-static mechanical analysis. Results of the present study indicate that - in terms of the maximum allowable cross-wind speed - the predictive accuracy returned from a careful model setup is about 0.5m/s, which corresponds roughly to 2% of the actual value.