[CFDguy99]

Hello folks,

So I am using a compressible flow solver: rhoPimpleFoam and I am a using LES model (Smagorinsky) to measure the turbulent kinetic energy dissipation for the taylor green vortex problem. So initially, I ran my simulation as it is without making any changes to the solver and the peak dissipation that I obtained was at t*=5.5 (for 64^3 cells), t* = 6 (for 96^3 cells) and t* = 7 (for 128^3 cells) which is way away from what's mentioned in literature (t* = 9 for Re > 400, Brachet et al.).

So I made a modification and commented out the following line in the UEqn.H file (which is part of the rhoPimpleFoam solver):

// turbulence->divDevRhoReff(U)

I know that this function is responsible for calculating the shear rate tensor which in turn is used in modelling the Reynolds stresses but by commenting out this particular line, my results improved.

The peak dissipation that I now obtained was at t*=7.5 (for 64^3 cells), t* = 8 (for 96^3 cells) and t* = 8.5 (for 128^3 cells).

So my question is that by commenting out the above function, am I still modelling the subgrid stresses since the LES model has an implicit function for it or am I just ignoring all of the stresses and making my flow inviscid and not modelling the small eddies?

I appreciate any comment and would be grateful if one of you experienced OpenFOAMers could help me out!