timestep for time dependent RANS
 

Hi,

I wonder if there is a size limit of the smallest time step for unsteady RANS models. I read that the averaging time T should be T>> Tf , where the turbulent fluctuations have the time scale Tf.

Where is the problem setting T=Tf?

Maybe, someone has an explanation.

Regards! fab

Re: timestep for time dependent RANS
 

The problem with T=O(Tf) is that in performing the Reynolds averaging you have "averaged out" these time scales. Furthermore in oder to justify retaining the d/dt term in the averaged equations there must be scale separation (i.e. T >> Tf ). In unsteady RANS the unsteadiness should be viewed as a slow modulation of the more rapid turbulent fluctuations.

Re: timestep for time dependent RANS
 

Hi,

thanks! I found for Tf: Tf= max (k/epsilon, alpha sqrt(nu/epsilon)) with alpha=6.

So should the limit for T be: T > 10*Tf ?

Regards! fab

Re: timestep for time dependent RANS
 

Read the following paper, G. Iaccarino,A.Ooi, P. A. Durbin, M. Behnia, "Reynolds averaged simulation of unsteady separated flow", International Journal of Heat and Fluid Flow,Volume 24, 2003, pp.147-156. Then you can get answer for your question.

Re: timestep for time dependent RANS
 

Hi,

thanks! I will try to get a copy.

Greetings! fab