[SadBoySquad]

Greetings to all FOAM'ers here!

I am still trying to find a proper setup for a natural convection + buoyancy case with multiple free openings that the flow can enter and exit based on buoyancy-induced currents inside the domain.

I think I've read each and every one of the relevant threads I found here regarding these types of cases, the most relevant being:

• Common setup for p_rgh in complete open domains
• buoyantSimpleFoam and watertank
• Correct boundary conditions for p_rgh (special for vertical patches)

(I seem to remember a couple more threads that had relevant info and that I may have gathered some additional tips, although I cannot seem to find them right now)

In the following link you can get my full setup directory for a simple buoyant box:

• a low opening on the left and a high opening on the right
• bottom wall takes a fixed temperature or a heat flux
• flow should enter from the left opening (low) and exit through the right opening (high)

https://www.dropbox.com/scl/fi/e9x1y...prw7k6vmo&dl=0


Short BC info:

U:

• walls noSlip
• openings pressureInletOutletVelocity

p_rgh:

• walls fixedFluxPressure
• openings prghTotalHydrostaticPressure

ph_rgh:

• fixedFluxPressure everywhere
• function that initializes ph_rgh in controlDict

epsilon|omega|k:

• wall functions on walls
• inletOutlet on openings

T:

• same as above, with only the bottom wall having a fixedValue or an externalWallHeatFluxTemperature BC

Overall things that I've learned trying to run this case:

• You have to initialize ph_rgh accordingly, because otherwise the fluid will produce flow due its own gravity force
• Even with 0 temperature difference (which should be a domain with 0 velocity), the solver produces some definitely non-zero velocities (they are in the order of magnitude 1e-02)
• The case needs to be run with buoyantPimpleFoam because buoyantSimpleFoam produces high artificial velocities (even without temperature differences)

The thing that bugs me is that the flow takes literally forever to converge into a steadyState-like state. Moreover, I am not quite sure that my setup is correct and I cannot move to more complex cases (which is my final goal).



Is this BC setup the best thing we can do in openFOAM? Should I try something different?