CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Large Eddy Simulation - LES (https://www.cfd-online.com/Forums/main/115834-large-eddy-simulation-les.html)

 stix April 8, 2013 05:18

Large Eddy Simulation - LES

Hello all,

I want to used a LES to simulate the transport of smoke in an air flow. I know that LES is always unsteady. I have read in "Turbulence Modeling for CFD" by David C. Wilcox and could not find out why LES only can be used unsteady. Could anyone of you please tell me why it's not possible to use LES steady?

Thanks

 RodriguezFatz April 8, 2013 07:10

The LES turbulence model is derived by modeling the small scales of the fluid. Formally, it looks like the RANS model, but it has nothing to do with that. Since no time integration / averaging is involved in LES, it just does not behave like that and thus results in transient solutions.

 FMDenaro April 8, 2013 07:28

Quote:
 Originally Posted by RodriguezFatz (Post 419057) The LES turbulence model is derived by modeling the small scales of the fluid. Formally, it looks like the RANS model, but it has nothing to do with that. Since no time integration / averaging is involved in LES, it just does not behave like that and thus results in transient solutions.
I agree, it is simple to see the LES equation when you start from the NS equation and apply to each term a local filter. As a result, the time derivative of each conserved variable, e.g., mass, momentum, etc., does not vanish.

 stix April 8, 2013 08:05

So because filtering is not equal to time averaging, the equation is time dependent and i have to regard the time derivative?

 RodriguezFatz April 8, 2013 08:19

Yes. Better think of LES as a slight relaxation of computational efford (compared to DNS), that is bought by a loss of precision. But it's not too different from DNS.

 stix April 8, 2013 08:23

Ok, thank you very much. I think i got it.

 All times are GMT -4. The time now is 20:02.