Currently there are different methodologies for the estimation of train aerodynamic forces including full-scale measurements, physical modelling techniques and numerical modelling using computational fluid dynamics (CFD) techniques. The aerodynamic assessment of trains was based mainly on physical and CFD modelling and these are the approved methods in the EU standard for train aerodynamic assessment in crosswinds.
From literature it has been found that both physical and CFD modelling are reliable in estimating the side forces in a good accuracy. However, there the two methods provide different values for the lift force and thus for the rolling moment. The aim of this innovative Fellowship was thus to investigate the source of discrepancies between the different methods and in particular to develop an accurate numerical technique based on the steady Reynolds Average Navier Stokes (RANS) capable of accurately predicting the aerodynamic forces. The methodology was based on wind tunnel experiments, moving model testing and different types of steady and unsteady CFD techniques. In addition, the effect of surface roughness on the lift force prediction of a train subjected to cross wind is also investigated. The objectives of this research involved:
• Carrying out wind tunnel tests on an idealised, smoothed roof train model and a rough train model and repeating the tests on a moving model to measure surface pressures and velocity fields at a 30° yaw angle
• To use the results from wind tunnel experiments to develop CFD turbulence models for the flow around roof and the ground for different train models
• To develop CFD models based on different RANS models with and without wall functions
• To investigate the influence of different simulation parameters such as inlet boundary conditions, turbulence modelling, ground movement and discretization schemes on the train surface pressure.
• To analyse the data obtained from the different simulations and physical modelling to investigate the source of discrepancies.
Most trains have irregular surfaces, which can be represented as roughness. It was revealed from the experimental work that added roughness on the roof was able to reduce the minimum surface pressure on the roof and leeward side of the train and thus affecting the aerodynamic lift and side forces. In terms of the numerical research, the choice of turbulence models is a key factor in numerically exploring the flow around trains. Results of the numerical study showed that both, Shear Stress Transport (SST) and Improved Delayed Detached Eddy Simulation (IDDES) turbulence models predict similar trends in the mean flow field around the train with slight differences found in the size of the vortices and the position of separation points. Furthermore, the effects of uniform and non-uniform crosswinds were also explored. It was observed that uniform crosswinds tend to overestimate the pressure coefficients.
