[Gmsh] 100 % under-determined cells with gmsh

Hillie · August 31, 2015, 20:55

Dear Foamers

I am quite new to the world of openfoam, and opensource mesh generation, so bear with me if my problem is trivial. I have been playing around with gmsh to replace icem which i am currently using.

I am currently looking at square tutorial which is on the openfoam wiki (https://openfoamwiki.net/index.php/2...ial_using_GMSH). If I mesh the .geo file of that tutoial, convert that mesh to openfoam and then run checkMesh -allGeometry i get the following problem:

Cell determinant (wellposedness) : minimum: 0 average: 4.84831e-35
***Cells with small determinant (< 0.001) found, number of cells: 940

The entire mesh consists of 940 cells, so all of them are underdetermined. I have checked the mesh with the solvers that I want to use, and they do complain about this, so I do need a solution.

Had anybody seen this before/ has an idea of how to solve it?

Hillie · September 3, 2015, 21:48

I have made a bit of progress on this. If after the meshing step i run a refine step (gmsh mymesh.msh -refine) the tutorial mesh seems oke.

If i do the same for the geometry that i am interested in I am still left with about 10 bad cells. If I run the refine step again it will create the double amount of bad cells.

Does anybody have a suggestion on how to remove the bad cells without increasing the cell count by a great deal?
(I set the cells size really small I can reduce the number of bad cells by a couple but this leads to huge meshes which aren't necessary for what i want to do with it )

August 31, 2015, 20:55	100 % under-determined cells with gmsh	#1
Hillie Member Hilbert Join Date: Aug 2015 Location: Australia Posts: 50 Rep Power: 10	Dear Foamers I am quite new to the world of openfoam, and opensource mesh generation, so bear with me if my problem is trivial. I have been playing around with gmsh to replace icem which i am currently using. I am currently looking at square tutorial which is on the openfoam wiki (https://openfoamwiki.net/index.php/2...ial_using_GMSH). If I mesh the .geo file of that tutoial, convert that mesh to openfoam and then run checkMesh -allGeometry i get the following problem: Cell determinant (wellposedness) : minimum: 0 average: 4.84831e-35 ***Cells with small determinant (< 0.001) found, number of cells: 940 The entire mesh consists of 940 cells, so all of them are underdetermined. I have checked the mesh with the solvers that I want to use, and they do complain about this, so I do need a solution. Had anybody seen this before/ has an idea of how to solve it?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] SnappyHexMesh running killed!	Mark JIN	OpenFOAM Meshing & Mesh Conversion	7	June 14, 2022 01:37
cellZone not taking all the cells inside	rahulksoni	OpenFOAM	0	January 16, 2019 01:16
Cht tutorial in 15	braennstroem	OpenFOAM Running, Solving & CFD	197	June 10, 2015 03:02
formatted point data- icem	HSK	CFX	12	August 11, 2011 21:25
[Gmsh] Import problem	ARC	OpenFOAM Meshing & Mesh Conversion	0	February 27, 2010 10:56

September 3, 2015, 21:48		#2
Hillie Member Hilbert Join Date: Aug 2015 Location: Australia Posts: 50 Rep Power: 10	I have made a bit of progress on this. If after the meshing step i run a refine step (gmsh mymesh.msh -refine) the tutorial mesh seems oke. If i do the same for the geometry that i am interested in I am still left with about 10 bad cells. If I run the refine step again it will create the double amount of bad cells. Does anybody have a suggestion on how to remove the bad cells without increasing the cell count by a great deal? (I set the cells size really small I can reduce the number of bad cells by a couple but this leads to huge meshes which aren't necessary for what i want to do with it )