LES with multi component systems

[fra-kd]

Why I get LES is not supported for multi-component systems activating a Constant Smagorisky LES model ? Obviously I'm running a case with two species and the relative mechanism but I do not understand why this limitation! I checked both LES.H and LES.cpp scripts together with the relative PelePhysics routines and the properties seem to be correctly computed for the mixture and do not enter in the turbulence model!

[baperry2]

There are some limitations with the LES implementation in PeleC and these warnings are there to ensure that users do not apply LES models to systems where the implementation is incomplete without knowing this. Most notably, the LES implementation includes turbulent fluxes for energy and momentum, but not for species. Additionally, with nonlinearity in thermochemical properties, several additional unclosed terms appear in the LES equations. For example, computing filtered temperature from filtered density, energy, and species is not a closed operation with non-constant specific heats, and no additional treatment is implemented for these additional unclosed terms in PeleC. For say, a mixture of N2 and O2 with no major temperature variation, that approximation is fine. For reacting flows with large temperature variation, it is perhaps not. You can see more details in the notes at the top of LES.cpp.

If you are aware of these limitations and do not believe they are important in your simulations, there should be no issues with applying the LES models in PeleC. In that case, the LES terms should work fine as implemented and you can ignore the warnings or simply comment them out in your code.

[fra-kd]

Thanks so much for your answer Bruce, really appreciated! I'm studying a mixture of H2 and N2 so I think my only issue would be the inclusion of the turbulent fluxes in the species equations... From the online documentation I saw that the diffusive terms are included in the specie equation so, basically, I "just" need to implement the turbulent stresses as source terms (as done for momentum and energy)? Am I right?

[baperry2]

Yeah if you want turbulent fluxes for species, you'd have to modify the functions encapsulated by getLESTerm to also include species fluxes where momentum and energy fluxes are computed. If you do this, please consider submitting a pull request as this would be handy capability for other users as well.

[fra-kd]

Sure ! I will do it ! but let me ask you another question… within the variable flx can I define the flux term for specie equations by just changing the second parameter ? Like flx(iv, UFS + n) ? Thanks again

[baperry2]

Yes, that should be correct.

[fra-kd]

Great! Thanks again!