Influence of mesh refinement on the accuracy of numerical results for the Next- Generation High-Speed Train Aerodynamics

Abstract

One of the critical issues is the accuracy of numerical simulations using Computational Fluid Dynamics, which is dependent not only on the appropriate flow physics but also on grid generation. In this study, the Richardson extrapolation method was used to investigate the effect of mesh refinement on the accuracy of simulation results for aerodynamic loads and the flow structure around a Next-Generation High Speed train (NG-HST). For the numerical simulation, the commercial Ansys software was used with an SST-based Delayed-Detached Eddy turbulence model. The simulation was carried out with the DLR NG-HST train model. Richardson Extrapolation (RE) and the grid convergence index method were applied using three different mesh resolutions (fine, medium, and coarse). The results obtained from the simulation compared with the previous study. The results of RE showed that the error for drag and moment coefficient was less than 1.5%, while the error for lift was around 7% in terms of predicted value. Furthermore, it was discovered that using a small number of mesh elements improved accuracy over the earlier study. Because it agrees well with previous work, this may conclude that it is possible to predict aerodynamic characteristics on a CFD simulation with a small number of mesh elements.

Analysis of the Aerodynamic Characteristics of Small-Sized Car Vehicles under the Influence of Steady Crosswind

“…Both numerical simulation and mesh are generated by the Commercial Ansys Fluent Software. One of the most critical and important steps in any type of CFD analysis is generating appropriate meshes and achieving grid independence [17], [18]. As a result, a grid independence test was conducted for the current study utilizing a grid convergence study.…”