[klilla]

Dear All,

I try to modify the boundary condition which reads a outside (measurement data) timeVaryingMappedFixedValue in order to superimpose perturbations to it. I'd like to use this BC for LES. Up to now I first added random perturbations normalized by the Reynolds stresses. The results were rather discouraging, the velocity field improved, however the pressure field became completely incoherent. Well, this was a well known phenomenon in incompressible solvers by my professor, so he suggested to do the following trick, before going to more sophisticated coherent synthetic turbulent inlet:

1. compute rotPerturbation = rot(perturbation)
2. perturbation = eps*(rotPerturbation), where eps is a scaling parameter to keep the perturbations in the desired range.

What one gains is that in this way the imposed velocity field stays divergence free (which is the property of incompressible flows, coming from the continuity equations), so pressure won’t go crazy.

The pressure not, me yes. So I took the aforementioned BC and renamed, modified. It was not working, fine it never does for me at first. Now there is only 2 lines stayed uncommented:

1. The one where I access the velocity filed: const volVectorField UField = db().lookupObject<volVectorField>("U"); – no problem with this.
2. The one where I compute the vorticity field: const volVectorField vorticity(fvc::curl(UField)); – if this line is there, the pressure field go crazy. Note, I’m not using the vorticity for anything, the rest of the code stays unchanged. I plotted the inlet velocity field, stays constant, I plotted the overall velocity field, nothing remarkable happens after one time step. If I plot the pressure field, it changes a lot.

Can anyone tell me why?

Thanks,
Lilla