[Roie_E]

Hello all,

I use a buoyantBoussinesqPimpleFoam solver to model transient turbulent flow in a 2-D square salty lake which is open to the atmosphere at the top. I try to only look at the water phase without simulating the air above. Left, right and bottom boundaries are walls.

I would like to pose a fixedFluxPressure for p_rgh at the walls (left right and bottom) and a zeroGradient boundary condition for p_rgh at the top but this field starts to diverge from the bottom left corner and convect the whole regime.

It looks like this after the first time step (0.005s):


I could gladly add the source code I just don't know how to do it.


Other boundary conditions used:

left, right and bottom walls:
epsilon --> epsilonWallFunction
k --> kqRWallFunction
alphat --> alphaJayatillekeWallFunction
nut --> nutkWallFunction
p --> calculated
p_rgh --> fixedFluxPressure of rho=rhok and value of uniform 0
T --> zeroGradient
U --> noSlip

top patch:
epsilon --> slip
k --> slip
alphat --> slip
nut --> slip
p --> calculated
p_rgh --> zeroGradient
T --> fixedValue of 300
U --> slip

Once I pose a totalPressure=0 boundary condition for p_rgh at the top, the bottom left corner stops acting weird so I think this might be a better boundary condition - is it actually?

One last thing - I would like to mention that added some extra scalar fields such as salinity to the solver with the same equations as T but with different diffusion coefficient. I did not mention them here but they will be available in the source code if necessarry.

Thanks a lot,
Roie