This work presents aerodynamic results of crosswind stability obtained numerically and experimentally for the leading control unit (class 808) of Deutsche Bahn AG's high-speed train Inter-CityExpress 2. The train model is on top of a 6 m high embankment in accordance with the proposed European code for interoperable trains, the so-called technical specifications for interoperability. The purpose of the study is to convey the predictive accuracy that typical steady-state computational fluid dynamics-Reynolds average Navier-Stokes methods (industry standard) return and to contribute to the understanding of the aerodynamics for the current application.Attention is drawn to the aerodynamics around the train and embankment when subjected to a steady block profile crosswind of 30 • yaw angle on the basis of the onset velocity far upstream the embankment. The Re (Reynolds number) of the embankment cases is 4.6 × 10 6 . Calculated results are obtained with the commercial code STAR-CD, with exclusively hexahedral meshes with a total cell count of 13.5 × 10 6 . Results are obtained when the train stands on the windward and leeward tracks on top of the embankment. These results are first compared with a flat ground case from a previous study.Then experimental data are obtained in a high-pressure wind tunnel with a model scale of 1:100. Re effects are compensated by raising the ambient pressure by a factor of 60, which increases the air density and thus the Re by a similar factor. Calculated results are in fair agreement with the experiments, where both the calculations and the experiments predict the leeward case to be the more critical one.In addition, the related consequences on the mechanical behaviour, i.e. the stability of the car, are briefly addressed by means of a quasi-static mechanical analysis. The results of the present study indicate that the 6 m high embankment concerning the current train reduces the permissible crosswind speed with approximately 20 per cent.

This work addresses crosswind stability exemplified for the German Railway Deutsche Bahn AG high-speed train ICE 2. The scope of the work is to describe the flow by means of computational fluid dynamics past the leading two cars of the train for yaw angles in the range 12.2–40.0°. Three track formations are utilized. The basic results are the set of independent aerodynamic coefficients for the lead and subsequent cars. The results are to some extent compared with experimental data for ICE 2 and also with data obtained for the Swedish high-speed train X2000. A numerical sensitivity study is undertaken to quantify differences in the above results dependent on the grid density and quality, turbulence model, numerical scheme, location of inlet and outlet boundaries, turbulence intensity and flow simulation software.

This work describes a quasi-static tool developed to assess the performance swiftly of crosswind stability for three types of rolling stock with conventional, semi-trailer and Jacobs bogie running gear configurations. The prediction accuracy of the results returned by the tool for the quasistatic assumption is fair in comparison with results of more advanced multibody simulation software that is commercially available. The codes, which are based on steady equilibrium equations for the wheels and axles, bogie frames and vehicle body/bodies, handle arbitrarily canted embankments and circular curves. To a large extent the accuracy hinges on the bodies' lateral displacements relative to the contact points between the wheels and rails; therefore proper modelling of the suspension systems and bump stops are found to be important. Examples are given of the limitations associated with the quasi-static approach, studying the following: (a) the combined wind and track scenario in Deutsche Bahn AG's guideline, (b) the effects of typical track irregularities for high-speed transportation as a function of train speed and (c) the effects of oscillating crosswind. It has also been found relevant to demonstrate some of the large differences regarding provisions regulating crosswind safety. To this extent the present results are compared with those derived with the British Group Standard and also with results presented in the guideline of Deutsche Bahn AG. In addition, examples are given of the differences found of the permissible crosswind speed using calculated (with CFD-RANS) and experimentally obtained aerodynamic loads.

Rail vehicles in everyday operation experience large lateral influences from curves and track imperfections, yielding large suspension deflections and displacements of the carbody relative to the track. Aerodynamic loads caused by crosswind may deteriorate the conditions that can result in vehicle overturning. This study investigates the influence of crosswind on a high-speed rail vehicle negotiating a curve. A multi-body simulation model of a high-speed rail vehicle is subjected to unsteady aerodynamic loads. The vehicle response is studied for different gusts, and variations of some vehicle parameters are performed.

This work compares aerodynamic results of crosswind stability obtained experimentally and numerically for a typical regional train model. The experimental work focuses attention on relatively high yaw angles of 30°, 41.5°, 50°, and 60°, whereas the numerical results are confined to 30° and 41.5° that are typically critical for regional trains. In particular, two vehicle models (VM1 and VM2) with different roof configurations are exploited numerically. VM1 has a flat roof with two boxy roof equipments. VM2 has a flat roof where the roof equipments are streamlined. The numerical work addresses the ability to resolve the flow field around the vehicles subjected to relatively high yaw angles with Reynolds average navier-stokes (RANS) and Delayed-Detached Eddy Simulation (D-DES) by employing the commercial code STAR-CD in conjunction with automatically generated polyhedral meshes of CCM+. High- and low-Reynolds meshes are exploited for this purpose. The massively separated flow, for the higher yaw angles, on the leeward side of the vehicles justifies the use of D-DES, where the results show good agreement with the experi-mental work. The results of RANS are fair, and the ability to predict the challenging lift force is again found to be problematic.