|
[Sponsors] |
February 26, 2021, 08:09 |
Advice on limitations of SHM
|
#1 |
Senior Member
Join Date: Apr 2020
Location: UK
Posts: 745
Rep Power: 14 |
Dear Foamers
I have been battling with SHM to get a high quality, sharp mesh around a REALLY simple geometry - a surface mounted cube. I can generate a snapped mesh, no problem - the cube surfaces in the snapped mesh are all planar and the edges are located where expected. So far so good. Now I try and add the tiniest of layers at the base, and SHM seems to struggle with the edges at the top of the cube - these move, I guess as part of the mesh shrinkage to make room for the layer ... but there's nothing that I can do to encourage the edges to snap back to their proper place. See the attached plots for pictures of this. I know that the distortion is minor, but it bugs me and I want the corners/edges in the right place. So here's my question - am I just inexperienced with SHM (that's true by the way) and have poor settings, or is SHM really this limited in what it can achieve? (I am not dishing SHM ... I just want to know the extent of its capabilities). Re: my settings, I have the following in my surfaceFeaturesDict: Code:
surfaces ("cube.stl"); includedAngle 150; subsetFeatures { nonManifoldEdges no; openEdges yes; } Code:
castellatedMesh true; snap true; addLayers true; geometry { cube { type triSurfaceMesh; file "cube.stl"; } refBox1 { type searchableBox; min (-0.9 -0.9 0); max ( 2.5 0.9 1.5); } refBox2 { type searchableBox; min ( -2 -2 0); max ( 6 2 1.8); } refBox3 { type searchableBox; min ( -3 -3 0); max ( 10 3 2.4); } }; castellatedMeshControls { maxLocalCells 1000000; //default 100000; maxGlobalCells 5000000; //default 2000000; minRefinementCells 10; maxLoadUnbalance 0.10; nCellsBetweenLevels 5; //feature edge refinement - eMesh read from constant/triSurface features ( { file "cube.eMesh"; level 4; } ); //surface refinement refinementSurfaces { cube { level (4 4); patchInfo {type wall;} } } resolveFeatureAngle 30; //volume refinement refinementRegions { refBox1 { mode inside; levels ((3 3)); } refBox2 { mode inside; levels ((2 2)); } refBox3 { mode inside; levels ((1 1)); } } //other mesh parameters locationInMesh (1 0 1); allowFreeStandingZoneFaces true; } snapControls { nSmoothPatch 3; tolerance 2.0; nSolveIter 30; nRelaxIter 5; nFeatureSnapIter 10; implicitFeatureSnap false; explicitFeatureSnap true; multiRegionFeatureSnap false; } addLayersControls { relativeSizes false; expansionRatio 1.2; firstLayerThickness 0.004; //finalLayerThickness 0.136; //thickness 0.25; minThickness 0.001; layers { ground { nSurfaceLayers 1; } } nGrow 0; featureAngle 60; slipFeatureAngle 30; nRelaxIter 3; nSmoothSurfaceNormals 1; nSmoothNormals 3; nSmoothThickness 10; maxFaceThicknessRatio 0.5; maxThicknessToMedialRatio 0.3; minMedianAxisAngle 90; nBufferCellsNoExtrude 0; nLayerIter 50; } meshQualityControls { #include "meshQualityDict" } writeFlags ( scalarLevels layerSets layerFields ); mergeTolerance 1e-6; Any suggestions/comments on how to get SHM to preserve the geometry in the rest of the mesh during layer addition would be gratefully received. Oh - and one more thing - and this is blowing my mind - when I plot the mesh in paraview, load in the cell coordinates (after running writeCellCentres) and pick off the coordinates of the upper edges, paraview reports these as having z<1 (cube top surface is at z=1), despite paraview clearly showing that these vertices are at z>1! What on earth is going on there? |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
[snappyHexMesh] trying to get a good mesh with SHM for about two weeks now | tobiaslindstrom | OpenFOAM Meshing & Mesh Conversion | 1 | November 7, 2016 03:39 |
[snappyHexMesh] SHM not snapping to some surfaces | Swift | OpenFOAM Meshing & Mesh Conversion | 13 | January 4, 2016 02:56 |
[snappyHexMesh] 2D AMI Moving Mesh with sHM; How hard can this be? | ADGlassby | OpenFOAM Meshing & Mesh Conversion | 18 | June 18, 2013 07:07 |
[snappyHexMesh] SHM: feature edge snapping not conforming | thab | OpenFOAM Meshing & Mesh Conversion | 3 | November 26, 2012 15:04 |
[snappyHexMesh] Multi Region Meshing with sHM | marango | OpenFOAM Meshing & Mesh Conversion | 3 | March 27, 2012 01:51 |