[NoahA]

Hello everybody,

i am currently trying run an transient RANS simulation of an oscillating flow in a simple environment with a structured grid. Unfortunately i am confronted with problems while running the simulation.

The Problem:
The simulation is crashing at around 280-290 seconds. (flow reversing starts at 219.3; oscillating sin period is 438.6). The results until this point are fine, usable and uncompromised.

The Mesh:
The Mesh is a small cylindrical inlet tube into a cylindrical open room with an o-grid setup. The surface of the room, where the inlet tube enters, is treated as a closed wall. All other room surfaces are open. (see attachments)

The Simulation:
I am running an incompressible RAS Simulation with pimpleFoam in PISO mode with 2 correctors, no underrelaxation and the standard kepsilon turbulence model. The maximum Re.-Num. is approx. ~ 2000.

My BC
patch outlet = outlet form the inlet tube into the cylindrical room
patch outletRoom = open surfaces of the room

k

internalField   uniform 0.015;

boundaryField
{
    inlet
    {
       type 		turbulentIntensityKineticEnergyInlet;
       intensity 	0.1;
       value 		$internalField;
    }
    inletWall
    {
        type 		kqRWallFunction;
        value           $internalField;
    }
    outlet
    {
        type            zeroGradient;
    }
    roomOutlet
    {
        type            zeroGradient;
    }
    roomWall
    {
        type 		kqRWallFunction;
        value           $internalField;
    }
}

epsilon

internalField   uniform 4.312e-3;

boundaryField
{
    inlet
    {
        type		turbulentMixingLengthDissipationRateInlet;
        mixingLength	0.07;
        value 		$internalField;
    }
    inletWall
    {
        type            epsilonWallFunction;
        value           $internalField;
    }
    outlet
    {
        type            zeroGradient;
    }
    roomOutlet
    {
        type            zeroGradient;
    }
    roomWall
    {
        type            epsilonWallFunction;
        value           $internalField;
    }
}

nut

internalField   uniform 0;

boundaryField
{
    inlet
    {
        type            calculated;
        value           uniform 0;
    }
    inletWall
    {
    	type            nutkWallFunction;
    	value           $internalField;
    }
    outlet
    {
        type            zeroGradient;
    }
    roomOutlet
    {
        type            zeroGradient;    
    }
    roomWall
    {
        type            nutkWallFunction;
        value           $internalField;
    }

U

• Inlet with FlowRateInletVelocity with a Sine profile
• noSlip on Walls
• zeroGrad at outlets

p

• totPressure = 0 at outletRoom
• zeroGrad. everywhere else

-------
Add. Inf.
I've run some laminar sim. with this setup without any Problems. I have as well tried k- and nutLowRe, which did not show any significant differences. I would still be guessing that this is a BC Problem?

I am hoping anybody cloud help me with some advise. Please let me know if add. Information is needed. Thanks in advance.
Noah