[Julian K.]

Hi,

I have a stability problem in my flow simulation. Here's the configuration:

I'd like to simulate the flow air through a convergent-divergent nozzle. The geometry starts at the inlet and stops at the outlet of the nozzle. The flow should be critical in the throat. For this I use a pressure difference of dp=300mbar, whereas at the inlet I apply atmospheric pressure. The the calculation I use the sonicFoam solver with 2nd order, bounded interpolation schemes (limitedLinear with phi=1/0.5 and MUSCL) for the divergence terms.


I set the BCs as follows:
/0/p:

    inlet
    {
        type            totalPressure;
        p0               uniform 1.01325e+05;
        U                U;
        phi              phi;
        rho              none;
        psi              none;
        gamma        1.4; 
        value           uniform 1.01325e+05;
    }

    outlet
    {
        type            fixedValue;
        value           uniform 0.71325e+05;
    }

/0/U:

    inlet
    {
        type            pressureInletVelocity;
        value           uniform (0 0 0);
    }

    outlet
    {
        type            zeroGradient;
    }

    wall
    {
        type            fixedValue;
        value           uniform (0 0 0);
    }

What happens is the following:
The simulation starts fine. The p-, U- and T-fields develop as they should. Shocks occur, which cause a separation of the boundary layer resulting in a creation vortices, which are eventually transported towards the outlet.
However, at some point the solution becomes unstable. At the outlet there is a drastic increase in velocity and pressure (of order > 10^30). Also negative pressure occurs near the outlet. Then the simulation crashes.

I suppose the reason for this effect is the BC at the outlet. What I could do is to fix the pressure, velocity and temperature at the inlet and have 'zeroGradients' at the outlet. However, for my analysis it is necessary to induce the flow via a pressure difference between inlet and outlet.

[Julian K.]

As suggested by prapanj in this thread (http://www.cfd-online.com/Forums/ope...-velocity.html) I used the 'waveTransmissive' BC for the pressure at the outlet. Now, the simulation seem to be stabilized, however it's still running so that, yet, I cannot tell if it works. I hope so .

[girogirozakk]

Hi

I try to solve nozzle-flow and my sample-files are opened on my URL.
http://giropenfoam.web.fc2.com/sample/index_sample.html

This is how you can help?

[blais.bruno]

Other than people trying to steal your topic, how did you results end up?

It seems like you had a bad case of a boundary condition polluting your simulations, which happens a lot for compressible flows.

Originally Posted by Julian K.

As suggested by prapanj in this thread (http://www.cfd-online.com/Forums/ope...-velocity.html) I used the 'waveTransmissive' BC for the pressure at the outlet. Now, the simulation seem to be stabilized, however it's still running so that, yet, I cannot tell if it works. I hope so .