[Chaoticzz]

Hi all,

Just a quick question. I'm studying an expansion flow for a circular axisymmetric duct. When setting up a grid on the boundary, should I setup the optimum node distribution (i.e. towards the wall and expansion region) before doing a grid independence test (coarse, medium, fine) or vice versa?

Thanks

[lcarasik]

Originally Posted by Chaoticzz

Hi all,

Just a quick question. I'm studying an expansion flow for a circular axisymmetric duct. When setting up a grid on the boundary, should I setup the optimum node distribution (i.e. towards the wall and expansion region) before doing a grid independence test (coarse, medium, fine) or vice versa?

Thanks

Hey James, just for clarification, did you mean have your near wall mesh set up to be valid and then do the mesh convergence study?

If so, yes I recommend an approach that ensures your near wall damping functions (near wall treatment/wall functions/etc) is properly resolved (this being used loosely). Then, leaving your near wall meshing alone, refine your bulk region of flow.

[Chaoticzz]

Originally Posted by lcarasik

Hey James, just for clarification, did you mean have your near wall mesh set up to be valid and then do the mesh convergence study?

If so, yes I recommend an approach that ensures your near wall damping functions (near wall treatment/wall functions/etc) is properly resolved (this being used loosely). Then, leaving your near wall meshing alone, refine your bulk region of flow.

Thanks for the reply Lane!

I'm restricted the the near wall treatment & wall functions I can use, say standard wall treatment for k-epsilon. I was talking more about the distribution of nodes. For example, shifting the nodes on the outlet pipe towards the expansion zone and also shifting them outwards towards the wall to produce a finer mesh to cover the vortices occuring on expansion.

Following this, then increase the No. of cells to determine independence.

Thanks