Dear all, I had a hard time to model axisymmetric swirling flows. I have a question for people who have done some work in swirling flows, how small the mesh has to be (the manual says the mesh should be very fine)? Is there a criterion for this or just using experiences? Thanks a lot.

yz.

it should be based upon y+ values and use mesh optimazation techniques. Pratap

thanks, would you elaborate a little? or are there some material I can read? yz

if you read the fluent manual it will clear for you. pratap

In strongly swirling flows you typically have a vortex core near the symmetry axis. In this region the velocity gradients are very large, and accordingly a fine grid is required for adequate resolution. I am sad to say, but the y+ concept does not offer much aid in this context as the primary source of turbulence is not wall bounded, but rather in the centre of the flow. Please, beware though, swirling flows are not modelled well with isotropic turbulence models such as k-eps., k-omega etc.