[segersson]

Hi,
I get a nasty inflow on the outlet boundary when doing LES using buoyantBoussinesqPisoFoam in OF 1.6.x (changed to allow generic turbulence models the same way as pisoFoam). The inflow seems to be caused by a very low pressure in the cells closest to the outlet boundary in the south-east corner of the domain.

outlet_problem_U.jpg outlet_problem_p.jpg

I'm running a test case for a heavy plume from a chimney surrounded by a number of buildings. I have tried three different meshes with different resolution at the outlet. I tried one mesh with very low resolution of perfect hex cells just before the outlet in order to get a diffuse very "nice" outflow.

I have tried different numerical shemes (linear, upwind, limitedLinear). However, the soultion does not blow up. I've kept the courant number below one (approx. 0.5). I am currently trying to use Smagorinsky but have also tried with oneEqEddy.

The bc:s are:
east (outlet) : U inletOutlet, p fixedValue, nuSgs zeroGradient
west(inlet) : U fixedValue, p zeroGradient, nuSgs zeroGradient
north: slip
south:slip
top:slip
ground:wall function
chimney: wall function
building: wall functions

The case is initialized using a converged run with simpleFoam. So far the temperature field is uniform, which should make the solver very similar to pisoFoam.

I have run out of ideas! Has anyone seen a similar behaviour? I have very limited experience of LES, so please tell me if you see something that seems strange in my setup.

Regards
David