I tested it with a sphere which has a drag coefficient of around 0. 47. In my simulations I got one of 0.36. I already have a large external flow volume. Are there any settings I can tweak to have a more accurate result?
Hi @Moehrenbaum , drag (particularly for external aerodynamics and FVM) is sensitive to mesh and turbulence models, particularly mesh type and layer inflation.
Here are a few things to use as a checklist:
- Turbulence model: K-omega SST
- Mesh type: Hex-dominant (Tet is known to struggle for drag calcs)
- Mesh layer inflation: to achieve Y+~=1
- Mesh independency
- Numerics: Divergence scheme for velocity, 2nd order i.e. gause linear upwind v nabla U. This should be the default, but it should be checked as viscous effects are extremely sensitive to this.
Professional users are encouraged to try the LBM solver for low Reynolds number aerodynamic problems (turbulent but not in the order of millions).
I hope this helps,
Darren
In addition I now tried the same simulation with a cube. The result was about 0.1 off again but this time it was higher than it should be. So I guess its not a problem of the mesh but just inconsitancy comming with computer simulations?
But thank you very much for your tips. I cant find where all of them are but I really appreciate the help.
Thank you very much! I set the mesh type to hex-dominant but thats the only thing I can find. Are the others under simulations controls and it is just me beeing blind or do I have to search somewhere else?