[MrHyperNova]

Hi, im very new to openfoam and am struggling to get my head round some problems.

for context, im working on a masters Diss covering altitude compensating nozzles and i need to be able to run simulations of a bell nozzle and aerospike at different altitudes/atmo pressure, including vacuum opperation.

currently i have played with sonicfoam, rhopimplefoam, and rhocenteral foam but each has given different errors.

so a couple of questions i have are:
-for my case, what would be the best solver?
-and how should i go about setting this up, i.e. boundary conditions etc.

the first nozzle is based off a chamber pressure of 9720000Pa at a temp of 3600k and designed for 3km so roughly 70000Pa atmo, at 270k.

see attached the mesh and domain that i created in fluent and converted over.

[Tobermory]

I have run axisymmetric sonic jet simulations with rhoSimpleFoam, but this took quite a lot of effort to make it work. rhoPimpleFoam would probably have been more stable, but I didn't want to run a transient solver. In the end, I found the following boundary set-up worked for me, along with upwind schemes to start and some relaxation:

pressure:

    upStream {
        type            uniformTotalPressure;
        p0              table (
                            (       0   101325)
                            (  1000   150000)
                            (  2000   500000)
                            (  8000   $pInlet)
                            (999999   $pInlet)
                        );
    }

    downStream {
        type            totalPressure;
        p0              101325;
        value           uniform 101325;
    }

pressure:

    upStream
    {
        type            zeroGradient;
    }
    
    downStream {
        type            pressureInletOutletVelocity;
        value           uniform (0 0 0);
    }

I also had to avoid the realizableKE model for the initial start-up, since that model is less stable than kEpsilon. Hope it helps - good luck.


PS the key was to ramp up the upstream pressure rather than just applying the full pressure at the start - that gave the pressure solver an easier time. Let us know how you get on, Ewan.