[AvaShahrokhi]

Hi every body

I am doing wind resource assessment in a urban area around a couple of buildings. I will be using LES for turbulence simulation and upstream wind data from meteorological measurements including wind profile and length scale.

The dimensions of the computational domain that I am simulating is 212 mx163m x 89m (in streamwise, spanwise and normal directions). This is while the length scale profile that I have been given changes from around 2 (close to the ground) up to 90 m.

Do you think that such a big length scale can cause numerical instabilities (and probably divergence) in my solution)?

my other question is that I am trying to generate structured mesh using ansys/icem around the buildings as I guess it is more robust and can provide more accurate results for LES simulations but it is really time consuming. Will I pay a big penalty if I switch to unstructured mesh? would unstructured mesh influence the convergence of the solution?

thank you very