Parameters of Large Eddy Simulation(LES)

[chenyx]

Dear developers,

Sorry to bother you again.

In code_saturne 7.0.4, I noted that the parameter of LES is different with other sovlers like charLES or some reference[Pope 2000 Turbulent flows] about LES. More specifically, the computation for filtered size is that,

delta = xfil *(xa*volume(iel))**xb,

where xfil defaults to 2.0, xa to 1.0, xb to 1/3, which is different with other references that use xfil to be 1.0

Second, the double filtered size of dynamic model in LES equation is that,

deltaf = xfil2*delta,

where xfil2 = 1.5 ,which is also different with other sovler that use xfil2 to be 2.0

Therefore, I want to figure out why these parameters are set like that which is different with the LES textbook and other sovler.

Best regards,
Yuxuan Chen

[Yvan Fournier]

Hello,

Checking with colleagues, it seems this must also be adjusted based on the numerical implementation/discretization, and I remember assisting an argument between several top-level turbulence specialists regarding the a factor of 2 in the interpretation of the filter width.

There might be more details in S. Benhamadouche's Thesis (https://www.librarysearch.manchester.ac ... c&offset=0),

Best regards,

Yvan