Evaluation of RANS, SAS and IDDES models for the simulation of the flow around a high-speed train subjected to crosswind

This paper compares the performance of EARSM, SAS and IDDES turbulence models for the aerodynamic problem of a high-speed train subjected to crosswind. These approaches, as well as the effect of the curvature correction for conventional RANS models (standard k - epsilon and k - omega SST), are analyzed in terms of aerodynamic force coefficients and pressure distribution at the train surface, time-averaged and instantaneous flow structures, vortices location and viscosity ratio, and computational cost. The turbulence models are also evaluated against wind tunnel experimental data and published numerical results. It has been shown that SAS, which is a turbulence model that is formally a BANS model, can allow the generation of LES content, and so it is found to be a good alternative for LES or DES in this aerodynamic problem. Besides, when a significant save in computational cost is required, EARSM has shown to give reliable results. Consequently, this turbulence model results into a good choice for expensive aerodynamic optimization problems as it balances computational cost and prediction accuracy.