[mnabi]

Hello all,

I am solving incompressible flow by LES. The model is based on Runge-Kutta-3 for the velocities. I am able to give my time step relatively large, which is almost 10 times larger than the kolmogorovtime scale for this calculation.

However, in LES, as a filter is used, the time scale is larger than DNS (kolmogorov). How can we give the time step in such a way it does not exceed the physical time scale in LES. Is there any way to estimate it?

Thanks

[FMDenaro]

Originally Posted by mnabi

Hello all,

I am solving incompressible flow by LES. The model is based on Runge-Kutta-3 for the velocities. I am able to give my time step relatively large, which is almost 10 times larger than the kolmogorovtime scale for this calculation.

However, in LES, as a filter is used, the time scale is larger than DNS (kolmogorov). How can we give the time step in such a way it does not exceed the physical time scale in LES. Is there any way to estimate it?

Thanks

In principle, the numerical stability constraint is already an useful indication for using the correct time-step. Furthermore, if LES filter you want to use is only spatial, it is suggested to use a dt that is comparable to the Kolmogorov time-scale in such a way you perform a DNS-like in time. I think that if you are no more than 10 times the Kolmogorov time-scale your dt is correct