[bmt]

Hi,

I am trying to do a 3D RANS simulation of flow past a sphere at high (up to 1,000,000) Reynolds. I am going to be trying to get data from this for a number of different turbulence models. But let's just stick to k-omega for now. I am also trying to check how well a snappyHexMesh solution will compare to a structured grid solution.

A few questions:

I am having difficulty coming up with a structured 3D sphere by using blockMeshDict -> Does anyone have any idea how to do this? I may post a separate question in the meshing forum.

We are trying to determine the coefficient of drag using a k-omega model (k and omega boundary conditions attached) at Re = 1M and were unsure about what conditions to implement on the sphere to control the surface roughness. Currently we are producing results with drag approximately double the expected for a smooth sphere and suspect that roughness is being introduced that we do not understand and cannot control.

Some people on forums say that nut and nuTilda are used for k-omega and some say they aren't, which is correct and what is an appropriate boundary condition to specify a smooth surface if they are indeed required?

Also, At the moment we are just using the default discretisation schemes for simpleFoam (attached in fvSchemes file).

Thanks for any help

k.txt

omega.txt

fvSchemes.txt