View Issue Details

IDProjectCategoryView StatusLast Update
0001976OpenFOAMBugpublic2017-06-30 10:08
Reporterroland Assigned Tochris  
PrioritynormalSeverityminorReproducibilityhave not tried
Status closedResolutionfixed 
PlatformGNU/LinuxOSOpenSuSEOS Version13.1
Summary0001976: Reconstruction of a mesh with sliding cyclicAMI patches breaks the mesh
DescriptionReconstruction of a mesh with sliding cyclicAMI patches breaks the mesh, see attached figure.

The solver that was used is pimpleDyMFoam with a solidBodyMotionFvMesh dynamicFvMesh using rotatingMotion. The simulation ran without any problems.

The relevent section from the boundary file:
    AMI_uitlaat
    {
        type cyclicAMI;
        inGroups 1(cyclicAMI);
        nFaces 63084;
        startFace 13287681;
        matchTolerance 0.0001;
        transform noOrdering;
        neighbourPatch AMI_uitlaat_shadow;
    }
    AMI_uitlaat_shadow
    {
        type cyclicAMI;
        inGroups 1(cyclicAMI);
        nFaces 63084;
        startFace 13350765;
        matchTolerance 0.0001;
        transform noOrdering;
        neighbourPatch AMI_uitlaat;
    }
Steps To ReproduceReconstruct parallel mesh with sliding cyclicAMI patches. View in ParaView.
Additional InformationFor the simulation OpenFOAM version 3.0.x was used (build 3.0.x-f5fbd395e495).
For the reconstruction different versions were tried (2.4.x, 3.0.x and dev) but they all created the same mesh. V2.4.x did return a lot of these warnings that were not present in the output of the other two versions:
--> FOAM Warning :
    From function void Foam::faceAreaWeightAMI<SourcePatch, TargetPatch>::interArea(const label, const label) const
    in file lnInclude/faceAreaWeightAMI.C at line 359
    Invalid normal for source face 5903 points 3((0.0248276 -0.16843 0.437) (0.132566 -0.121466 0.437) (0.014873 -0.170696 0.437)) target face 5823 points 3((0.0113854 -0.180326 0.437) (0.136461 -0.131661 0.437) (0.132566 -0.121466 0.437

ANSA was used to create the mesh
TagsNo tags attached.

Activities

roland

2016-01-20 09:24

reporter  

reconstructError.png (104,633 bytes)   
reconstructError.png (104,633 bytes)   

MattijsJ

2016-06-10 16:58

reporter   ~0006426

Have you run checkMesh -allTopology -allGeometry in parallel? Also postprocess the individual domains (eg 'paraFoam -case processor0')

chris

2017-06-30 10:08

manager   ~0008284

A workaround has been provided. A more elegant solution could appear in a future development.

Issue History

Date Modified Username Field Change
2016-01-20 09:24 roland New Issue
2016-01-20 09:24 roland File Added: reconstructError.png
2016-06-10 16:58 MattijsJ Note Added: 0006426
2017-06-30 10:08 chris Assigned To => chris
2017-06-30 10:08 chris Status new => closed
2017-06-30 10:08 chris Resolution open => fixed
2017-06-30 10:08 chris Note Added: 0008284