[mabinty]

dear all!

I m about to simulate a round turbulent buoyant plume with an axis-symmetric set up. My problem is the inlet-velocity BC: the value I prescribe in the 0/U dictionary does not correspond with the results of "probeLocations". the probed inlet velocity gives values about 10% lower than the prescribed one. Does anybody have experience with this kind of problem??

I greatly appreciate your comments!

best regards,
aram

[mabinty]

Hi!

I investigated a bit about the problem I described in the last post. I used a simple pipe flow with the solver of interest (i.e. buoyantPisoFoam) and tried different boundary conditions (fixedValue, flowRateInletVelocity) as well as different grid sizes but the problem remains. Interestingly things look better when the inlet velocity is increased by a factor of 10 (see attached graphs). Furthermore the problem disapears for a 2D set-up (empty front and back). Woud be interesting to know how inlet velocities are prescribed when using a axis-symmetric set up. I ll keep on digging!

Appreciate your comments!! Thx in advance,
Aram

[mabinty]

Dear all!!

Finally I could solve my problem described above: changing the pressure BC at inlet from "zeroGradient" to "buoyantPressure" did it. When looking into the code of the "buoyantPressure" BCs, it says:

gradient() = rho*(g.value() & patch().nf());

Hence the value for the pressure gradient at inlet in direction of the face normal (here equal to gravity) is set to a non- zero value, compared to zero for the "zeroGradient" BC.

Cheers,
Aram