[max.pow]

Hello foamers,

for my thesis I'm currently working on a simulation of the NASA CRM model (https://commonresearchmodel.larc.nasa.gov/).
I'm trying to create the mesh using snappyHexMesh. For reaching my yPlus values, I have to add Layers, but there seems to be some problems with snappyHexMesh:
The layers collapse before they reach the tip of the leading edge. I've tried to vary the addLayersControls in my snappyHexMeshDict, with no better results than in the attached screenshot.





Here you can see my snappyHexMeshDict:

FoamFile
{
version 2.0;
format ascii;
class dictionary;
location system;
object snappyHexMeshDict;
}

castellatedMesh true;
snap true;
addLayers true;
geometry
{
plane.stl
{
type triSurfaceMesh;
name plane;

}


refinementFar
{
type searchableCylinder;
point1 (-100 0 5 );
point2 (200 0 5);
radius 100;
}

refinementFuselage
{
type searchableCylinder;
point1 (-1 0 5 );
point2 (68 0 5);
radius 10;
}

refinementWing
{
type searchableCylinder;
point1 (30 0 5 );
point2 (48 31 7);
radius 7;
}

refinementHTP
{
type searchableCylinder;
point1 (60 0 6 );
point2 (65 15 8);
radius 4;
}

refinementLE1
{
type searchableCylinder;
point1 (25.22 3.05 4.47 );
point2 (31.1135 10.871 4.599);
radius 0.3;
}

refinementLE2
{
type searchableCylinder;
point1 (31.1135 10.871 4.599);
point2 (38.019 19.9 5.279);
radius 0.3;
}


refinementLE3
{
type searchableCylinder;
point1 (38.019 19.9 5.279);
point2 (42.06751 25.227 6.011);
radius 0.3;
}


refinementLE4
{
type searchableCylinder;
point1 (42.06751 25.227 6.011 );
point2 (45.27 29.4 6.7);
radius 0.3;
}


}

castellatedMeshControls
{
features
(

{
file "plane.eMesh";
level 5;

}


);
refinementSurfaces
{
plane
{
level ( 9 9 );
patchInfo
{
type wall;
inGroups (planeGroup);
}
}



}

refinementRegions
{
plane
{
mode distance;
levels (( 0.002 9 ) ( 0.05 8 ) ( 0.1 7 ));
}

refinementFar
{
mode inside;
levels ((1E15 2));
}

refinementFuselage
{
mode inside;
levels ((1E15 6));
}
refinementWing
{
mode inside;
levels ((1E15 6));
}
refinementHTP
{
mode inside;
levels ((1E15 6));
}
refinementLE1
{
mode inside;
levels ((1E15 10));
}
refinementLE2
{
mode inside;
levels ((1E15 10));
}
refinementLE3
{
mode inside;
levels ((1E15 10));
}
refinementLE4
{
mode inside;
levels ((1E15 10));
}
}

locationInMesh (-400.0 50.0 -400.0 );
maxLocalCells 10000000;
maxGlobalCells 200000000;
minRefinementCells 2;
nCellsBetweenLevels 3;
resolveFeatureAngle 30.0;
allowFreeStandingZoneFaces true;
planarAngle 30.0;
maxLoadUnbalance 0.01; //was 0.1
}

snapControls
{
nSolveIter 100;
nSmoothPatch 5;
tolerance 2.0;
nRelaxIter 5;
nFeatureSnapIter 10;
implicitFeatureSnap false;
explicitFeatureSnap true;
multiRegionFeatureSnap false;
}

addLayersControls
{
layers
{
plane
{
nSurfaceLayers 15;//was 15
expansionRatio 1.1; //was 1.25
finalLayerThickness 0.01;
minThickness 1.0E-6;
}
}
meshShrinker displacementMotionSolver;
solver displacementLaplacian;
displacementLaplacianCoeffs
{
diffusivity quadratic inverseDistance 1(plane);
}

relativeSizes false;
expansionRatio 1.25;
finalLayerThickness 0.01;
minThickness 1e-06;
nGrow 0; //was 0
featureAngle 60.0; //was 120
slipFeatureAngle 30.0;
nRelaxIter 5;
nSmoothSurfaceNormals 1;
nSmoothNormals 5;
nSmoothThickness 10;
maxFaceThicknessRatio 0.5; // was 0.5;
maxThicknessToMedialRatio 0.3; // was 0.3
minMedialAxisAngle 90;
nBufferCellsNoExtrude 0;
nLayerIter 50;
nRelaxedIter 20;
nMedialAxisIter 10;
nSmoothDisplacement 90;
detectExtrusionIsland true;
mergePatchFacesAngle 45;
}

meshQualityControls
{
maxNonOrtho 65.0;
maxBoundarySkewness 20;
maxInternalSkewness 4;
maxConcave 70.0;
minFlatness 0.5;
minVol 1.0E-13;
minTetQuality 1.0E-15; //was 1.0E-15
minArea -1.0;
minTwist 0.02;
minDeterminant 0.001; // was 0.001
minFaceWeight 0.05;
minVolRatio 0.01;
minTriangleTwist -1.0;
nSmoothScale 4;
errorReduction 0.65;
relaxed
{
maxNonOrtho 75;
}

}

debug 0;
mergeTolerance 1.0E-6;
autoBlockMesh false;

Is there any possiblity to avoid collapsing layers?

Thanks for your help,

Max

[sati]

Hi,
Snappy doesn't work well when the size difference between the last cell of the boundary layer and the first cell of the rest of your mesh is too high.

Try to refine a bit more the cells close to the geometry, and it should be ok.

[hokhay]

Try to set featureAngle in addLayersControl to 180

Sent from my LG-H818 using CFD Online Forum mobile app

[bennn]

Try cfmesh. It'll work, guaranteed

[me3840]

Originally Posted by bennn

Try cfmesh. It'll work, guaranteed

I would not use CfMesh for this, it doesn't really make good prism layers. For a very boundary layer dominated problem like this it's better to go with the better prism generator (albeit more frustrating..)

[bennn]

My experience is exactly the opposite of what you're saying, but I would be very curious to see examples... thanks !

[cosmictrex]

Hi,


Have you been able to successfully generate a good CRM mesh with either snappy or cfMesh? I'm starting to be curious if this is possible at all. One report from the AIAA Drag Prediction Workshop even mentioned that "snappyHexMesh is too poor" for even the coarsest mesh requirement.