State Key Laboratory of Traction Power, Southwest Jiaotong University, Chengdu610031, China
School of Engineering, University of Birmingham, Birmingham B15 2TT, U.K.
College of Mechanical and Electrical Engineering, Qingdao University, Qingdao266071, China
Currently, there are different computational fluid dynamic (CFD) techniques used to obtain the flow around trains. One of these techniques is the Reynolds-averaged Navier-Stokes (RANS), which is commonly and widely used by industry to obtain the mean flow field around trains in different operating conditions. In order to assess the performance of RANS turbulence modelling for train aerodynamics, five different common RANS modelling have been used in this paper to obtain the flow and the surface pressure around a simplified train model subjected to crosswind; the standard k- model, the realisable k- model, the Re-Normalisation Group (RNG) k- model, the standard k- model and Shear Stress Transport (SST) k- model. The train model was stationary and subjected to crosswind with a 90o yaw angle. The effects of mesh size and spatial discretization scheme on the aerodynamic characteristics of the train were also investigated. The results obtained from the different RANS models were compared to those from published experimental data. In general, all the RANS models provided the pressure distribution trend. However, all k- models overestimate the surface pressure on the train body except the bottom face. The standard k- model underestimates the surface pressure on the train body except the streamwise face. It was shown that the simulation using SST k- model together with a second order discretization scheme provides the closest results to the experimental surface pressure. It could be concluded from the present study that the SST k-model with a second order discretization scheme and y+ around 1.0 is the most appropriate RANS model for simulating the flow around trains subjected to crosswinds.