CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Programming & Development (http://www.cfd-online.com/Forums/openfoam-programming-development/)

 babakflame February 24, 2013 10:31

LES with Simple Algorithm for steady situations

Dear Foamers
Hi

I have a question about LES. Can we have a LES solution with Simple Algorithm rather than Pimple. I mean if we use a steady algorithm like Simple for a steady problem like a free jet, Is it possible or even right? :rolleyes::confused::rolleyes:

Regards
Bobi

 atmcfd March 18, 2013 15:12

No. LES is for solving turbulence problems, which are inherently Unsteady. Therefore you should not use a steady solver for these cases. You have to run the unsteady case for a certain time period, and then average the instantaneous solutions to obtain the turbulent flow statistics.

 babakflame March 19, 2013 02:22

Dear ATM
Hi
with your response I conclude that LES is not applicable with Simple algorithm and always I should employ the Pimple one. Nevertheless, RAS models can use Simple algorithm although they characterize turbulent flows. They average the solutions during marching to find the favre solution. My intension was that can we use the same strategy for LES? I mean instead of having real time marching in Pimple algorithms, employing unreal marching in Simple solvers and then averaging the instantaneous solutions to find the ultimate solution. As we can do in RAS model.:rolleyes:

Regards
Bobi

 atmcfd March 19, 2013 02:35

Bobi,

You are right about RAS., and LES runs on a modified SIMPLE method. However, in LES you are resolving much smaller length scales than you do in RAS. Hence time stepping is very important as it should be very small. Since you resolve very small scales , their associated time scales are very small too. So, to obtain good statistics, you have to have a sufficiently small timestep. A courant no. of <0.5 is usually recommended.

Therefore, even pimpleFoam would not be a good choice since its a "Large time-step transient solver for incompressible flow using the PIMPLE (merged PISO-SIMPLE) algorithm" (OF website)

My best bet would be in dedicated transient solvers like pisoFoam, channelFoam etc. for LES.A Steady State solver is unphysical for LES.

 babakflame March 19, 2013 03:26

Dear ATM
Hi
Thnks for ur suggestions buddy.:)

Regards
Bobi

 JimKnopf March 21, 2013 16:46

This is not a question of piso, simple or pimple, which names the way the pressure coupling is done, but more a question of spartial and time resolution.

Afaik you can use SIMPLE als pressure coupling for solving Equations with a time derivative.

But if you are using LES - which means Large Eddy Simulation - you have to bear in mind that you have to resolve these large eddies. Afaik they are commonly the eddy with a length scale above the isotropic ones (but I'm not sure if this is a correct formulation). Also you have to bear in mind that a lenght scale is coupled with a time scale via the velocity. The CFL-criterium gives the time step size if you know length and velocity scales, is has to be 1 for the smalest eddies.

In RANS or RAS you think of turbulence as of a viscosity, whicht provides you with the energie dissipation you get in LES via the breakup of eddies. If you use a to coarse mesh or a to large time step size, it does not mean you get divergence or no result. It does only mean you get a not correct result, what ever correct is. LES is not a easy to use tool.

Greets
Jim

 babakflame March 22, 2013 03:11

Dear jim
Hi

Thnks for your suggestions. My question is not about time steps or grid sizes, cause as u know there are some limitations for LES. As Pope has mentioned in his book, For instance LES should resolve about 80% of the turbulent kinetic energy. I was wondering that if we had a steady turbulence condition, can we use SIMPLE algorithm for coupling between pressure and velocity in conjunction with LES turbulence model.
Regards
Bobi

 JimKnopf March 22, 2013 12:58

Hi Bobi,

but steady or unsteady is not a question of pressure coupling, it's a question of the Equations you're solving. What I mean is that you can't infer from SIMPLE that it's steady, without time resolution. You are right that most (all?) of the esi/OF delivered solvers with SIMPLE algorithm are steady solvers but there are enough unsteady solvers with SIMPLE pressure coupling.

And definetly you are right with saying that LES and steady does not work or say does not bring meaningful results.

Regards,
Jim

 All times are GMT -4. The time now is 19:22.