Grid independence square cylinder
 

Hi,

I'm trying to simulate the flow past a square cylinder at a Re = 33000 and I'm trying to compare it with experimental results (St, CLrms, CDmean).
I'm doing an unsteady RANS, with Coupled scheme and k-epsilon RNG turbulence modelling.
The problem is that I'm trying to perform a mesh convergence study, but I'm getting poor results: as I refine the mesh, CLrms and CDmean continue to fluctuate around the experimental values, so I don't get a real convergence.

I was thinking: should I refine the time-step as I refine the grid?

Also, I'm a bit confused as how to interpret the Courant No. in FLUENT: I fix the timestep and of course the mesh size is fixed, so what is the meaning of changing the CFL? Does it influence stability and solution convergence?

Thank you so much in advance.

Cfl
 

As the spatial mesh is refined, temporal mesh, i.e., time-step, must be refined as well.

As far as CFL for Coupled method is concerned, that is not Courant number. It is only for stability and convergence of numerical system. You can leave it to default.