[S.Rhee]

Hi

I am trying to implement Large Eddy Simulation code for our in-house CFD code for compressible flows.

Most of all, I'm really confused in treating Filtered Navier-Stokes Equations (FANS) whether they are compressible or not since I have only programmed for Inviscid or RANS solver.

Generally, vector for physical variables for FANS is Filtered variables.

Then, can I get solutions from FANS directly using SGS models?

Is Flitering used only for recasting Navier-Stokes equations???

I thought it isn't since various filters are used for LES for incompressible flows

Sorry for my rambled question

Thanks in advance

[FMDenaro]

Originally Posted by S.Rhee

Hi

I am trying to implement Large Eddy Simulation code for our in-house CFD code for compressible flows.

Most of all, I'm really confused in treating Filtered Navier-Stokes Equations (FANS) whether they are compressible or not since I have only programmed for Inviscid or RANS solver.

Generally, vector for physical variables for FANS is Filtered variables.

Then, can I get solutions from FANS directly using SGS models?

Is Flitering used only for recasting Navier-Stokes equations???

I thought it isn't since various filters are used for LES for incompressible flows

Sorry for my rambled question

Thanks in advance

Generally, you can see that a lot of LES codes use the so-called "implicit filter", that is both the computational grid size and discretization of the equations act as filter for the unresolved solution. For example the FVM is used as implicit top-hat filter.
On the other hand, some SGS models such as the dynamic procedure, scale similar, require to code an explicit filter to compute the test-filtered fields.
Since you programmed the code for the inviscid case, I suppose you have a weak formulation, that is a volume integral solution. In such a case you can suppose this is your filter and you have to supply the equations with the SGS model.
Note that often codes for inviscid compressible flows have some flux limiter or similar, there is also a LES community using that as implicit SGS model in the so called ILES.

[S.Rhee]

Originally Posted by FMDenaro

Generally, you can see that a lot of LES codes use the so-called "implicit filter", that is both the computational grid size and discretization of the equations act as filter for the unresolved solution. For example the FVM is used as implicit top-hat filter.
On the other hand, some SGS models such as the dynamic procedure, scale similar, require to code an explicit filter to compute the test-filtered fields.
Since you programmed the code for the inviscid case, I suppose you have a weak formulation, that is a volume integral solution. In such a case you can suppose this is your filter and you have to supply the equations with the SGS model.
Note that often codes for inviscid compressible flows have some flux limiter or similar, there is also a LES community using that as implicit SGS model in the so called ILES.

Thank you for your generous answer. Definitely i have focused on FVM (and using flux limiter as you said ).

Then, FVM is alrealy implicit top-hat filtered equations and if I Solve FVM with SGS models, then LES solution can be obtained???

[FMDenaro]

Originally Posted by S.Rhee

Thank you for your generous answer. Definitely i have focused on FVM (and using flux limiter as you said ).

Then, FVM is alrealy implicit top-hat filtered equations and if I Solve FVM with SGS models, then LES solution can be obtained???

Yes but that means nothing about the real quality of your solution. Adding the SGS to the already existing flux limiter can produce an excess of energy dissipation.
You have to check your solution on well known test-cases.

[S.Rhee]

Originally Posted by FMDenaro

Yes but that means nothing about the real quality of your solution. Adding the SGS to the already existing flux limiter can produce an excess of energy dissipation.
You have to check your solution on well known test-cases.

I sincerely appreciate your answer. now i understand why governing equations are expressed like that way from articles.

so many Thanks ..