[WhiteW]

Dear all,
I'm experiencing some issues in the convergence of a HVAC analysis.
The domain is quite simple: a room with an inlet (a massflow is entering) and an outlet (external pressure condition), as in the attached picture.
Furthermore, since I have a certain difference of temperature, I need to use rhoSimpleFOAM. Turbulence model kOmegaSST.

The problem is that there are always instabilities that originates from the outlet. More specifically, in a single cell on the border between the outlet and a wall the velocity start to increase (from a certain iteration, 300 in this case) and then this unphysical velocity start to spread from that zone (at iteration 2500 the problematic cells are 10k). In the pictures I'm showing the threshold of velocity higher then 40m/s.

Sometimes I can fix this behaviour moving the outlet surface far form the room (for example using the spheric outlet in the picture) but in this case this is not working.
The mesh created with snappy has not particular problems, the boundary layer has not been completely generated in that zone, but I don’t think it is needed.

I don’t ‘understandwhat is wrong… I would like to use the outlet as in the reality, placing it near the wall of the room but in this case the instabilities always occur.
Here are my boundary conditions, are they correct?
Thanks!
WalterlW


U:

   inlet
    {
        type                flowRateInletVelocity;
        volumetricFlowRate  5.833;
        value               uniform (0 0 0);
        rho                 rho;
        rhoInlet            1.042;
    }

    outlet
    {
        type        pressureInletOutletVelocity;
        value       $internalField;
        inletValue  uniform (0 0 0);
    }

    walls
    {
        type    noSlip;
    }

p

   inlet
    {
        type    fixedFluxPressure;
        value   $internalField;
    }

    outlet
    {
        type    fixedValue;
        value   $internalField;
    }
    walls
    {
        type    zeroGradient;
    }

T

    inlet
    {
        type    fixedValue;
        value   uniform 351.25;
    }

    outlet
    {
        type    zeroGradient;
    }

    walls
    {
        type    zeroGradient; 
    }

At the moment I'm also using a velocity damping constraint for velocity at 40m/s, it helps a bit but the instabilities remain.

[wstapel]

I have a suspicion what is wrong.

Can you create two plots for the inlet?
1. average pressure over iterations
2. average density over iterations

[WhiteW]

Hi, thanks for the reply.
Attached are the plots of area average of velocity, density and pressure on both inlet and outlet.
I also plotted the min and maximum values.
WalterW

[wstapel]

Hi,


sorry for the late reply. Actually the plots do not show what I expected. I thought the problem was a constant pressure-density osciallation at the inlet which then led to instability.



Can you show your fvSchemes file? I suppose you use a limited scheme for div(phi, U) ?