[Steefan]

Hi guys!

I am a new OpenFOAM user (using v2012 version) and I'm working on the flow around a moving train. I am working on a simplified case with overPimpleDyMFoam (overset mesh), to check that the case is set in the correct way. The mesh is made by merging (mergeMeshes) a background mesh with an overset mesh; the screenshot is attached. I tried to run the simulation in serial and it seemed to work, but when I decompose the domain and try to run in parallel the simulation starts but remains stuck at the first iteration, precisely when the pressure equations are solved (no error appears, the processors work but the simulation does not procede).


Parallel case log:
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create mesh for time = 0.7

Selecting dynamicFvMesh dynamicOversetFvMesh
Selecting motion solver: multiSolidBodyMotionSolver
Applying solid body motion to entire mesh
Selecting solid-body motion function linearMotion
Applying solid body motion linearMotion to 15411 points of cellZone oversetZone

PIMPLE: no residual control data found. Calculations will employ 10 corrector loops

Reading field p

Reading field U

Reading/calculating face flux field phi

Creating cellMask field to block out hole cells

--> FOAM Warning :
From bool Foam::oversetPolyPatch::master() const
in file oversetPolyPatch/oversetPolyPatch.C at line 151
The master overset patch is not the first patch. Generally the first patch should be an overset patch to guarantee consistent operation.
Creating interpolatedCells field

Selecting incompressible transport model Newtonian
Selecting turbulence model type RAS
Selecting RAS turbulence model kEpsilon
RAS
{
RASModel kEpsilon;
turbulence on;
printCoeffs on;
Cmu 0.09;
C1 1.44;
C2 1.92;
C3 0;
sigmak 1;
sigmaEps 1.3;
}

Reading/calculating face velocity Uf

No MRF models present

No finite volume options present
Courant Number mean: 0.000929265723051 max: 0.0706134337253

Starting time loop

Courant Number mean: 0.000929265723051 max: 0.0706134337253
deltaT = 0.000123915737299
Time = 0.700124

inverseDistance : detected 2 mesh regions
zone:0 nCells:24000 voxels:(22 22 22) bb:(42.9999919961 4.99999199609 3.74999199609) (48.0000080039 10.0000080039 7.50000800391)
zone:1 nCells:11666 voxels:(22 22 22) bb:(22.5009964727 -3.00018015675 -4.50032015675) (39.5087767862 3.00042015675 4.50108015675)
Overset analysis : nCells : 35666
calculated : 33048
interpolated : 2310 (interpolated from local:0 mixed local/remote:0 remote:2310)
hole : 308

DICPCG: Solving for pcorr, Initial residual = 1, Final residual = 0.0143876348735, No Iterations 60
PIMPLE: iteration 1
smoothSolver: Solving for Ux, Initial residual = 0.00212050916343, Final residual = 9.48591688485e-06, No Iterations 4
smoothSolver: Solving for Uy, Initial residual = 0.00110282091759, Final residual = 5.20908926019e-06, No Iterations 4
smoothSolver: Solving for Uz, Initial residual = 0.00716174027334, Final residual = 3.28175448585e-05, No Iterations 4





I also attached the Serial case log and the system folder to let you have more information.

I think the problem is in the parallel run, so I tried to change decomposition method (from Hierarchical to Scotch) without results; since the simulation seems to stop at pressure calculation, I also tried to change the pressure solver (and modify tolerances) in fvSolution from PBiCGStab to PCG or GAMG, but even in this case the parallel simulation still gets stuck at the same point.

If someone have experienced this problem or know how to help, please let me know!



Thanks a lot,


Stefano


MeshScreenshot.jpg

Attachments.zip

[Steefan]

Hi again,



I found the reason of the problem with the parallel run: in fvSolutions, PIMPLE section, I had the "massFluxInterpolation" activated and I discovered that this setting caused the parallel simulation to be stuck, while the serial case was running with no problems.

I am not still sure why, but deactivating this massFluxInterpolation the problem seems to be solved.


Hope this may help someone.


Stefano

[Himanshu_B]

Thank You! It Worked.


Himanshu