[doctorWho]

Hi,

I am trying to simulate flows around a circular cylinder at Re=10^6 (refer to Kim and Mohan, PREDICTION OF UNSTEADY LOADING ON A CIRCULAR CYLINDER IN HIGH REYNOLDS NUMBER FLOWS USING LARGE EDDY SIMULATION, OMAE, 2005)
Question is how to estimate characteristic length and time scale of the smallest resolved eddies in order to make sure that time step has turn-over time enough to resolve the smallest resolved eddies.

[FMDenaro]

Originally Posted by doctorWho

Hi,

I am trying to simulate flows around a circular cylinder at Re=10^6 (refer to Kim and Mohan, PREDICTION OF UNSTEADY LOADING ON A CIRCULAR CYLINDER IN HIGH REYNOLDS NUMBER FLOWS USING LARGE EDDY SIMULATION, OMAE, 2005)
Question is how to estimate characteristic length and time scale of the smallest resolved eddies in order to make sure that time step has turn-over time enough to resolve the smallest resolved eddies.

the smallest vortical structure you can solve is dictated by the Nyquist frequency of your computational grid, that is Kc=Pi/h. However, depending on the accuracy order of your discretization, the description of these structures can be affected by smoothing of high resolved frequency. For example, this happend for smooth filtering induced by FV discretization. Spectral methods are conversely suitable to well resolve up to the Nyquist frequency.
The maximum time step is then consequent to stability constraint of your scheme but you can work with smallest time step, until to the Kolmogorov time-scale

[doctorWho]

FMDenaro

Thanks for your reply. I will try with the Kolmogorov time-scale first.