Penetrating elements in extruded mesh
 

Hello, I have a problem with a mesh I created with ICEmCFD5.1. I will briefly describe the mesh: I start off by meshing a plane rectangulare surface with a triangular surface mesh. This 2d-domain is 38x38 Meters big, the element size is 6000mm at each curve. Then, this Mesh is extruded along a vector into the third dimension with 28 layers, each 1.25 mm thick. The resulting 3d-mesh is well-adapted to the surrounding geometry (a cuboid), which means that the extruded surface elements "arrive" exactly at the opposing surface of the cuboid. All resulting shell elements are assigned to the proper geometric families (all shells in different families, all volumic elements in a seperate family).

The problem: Grid check gives "duplicate elements", which appear only at one curve of the original rectangle (the rectangle which was first meshed in 2d) and a few elements (about 10 elements of about 170) of the of the adjactant surface which is produced by the extrusion. Running this mesh in FLUENT causes a divergence because of these elements, who cause abnormal high velocities (10exp7 m/s with an inlet velocity of 40 m/s) and pressures.

By zooming into the region, where the problematic elements are displayed, no actual interference between different surface elements is visible. Plus, all elements have been created by extrusion, so how can they penetrate? What are possible solutions?

Thanks in advance, Michael

Re: Penetrating elements in extruded mesh
 

do you have access to gambit, if you have you can do one simple thing, read the mesh into gambit, if you see some arbit facet delete them and re-export the mesh, this time it shall be clean, (i am assuming that if u can read the mesh into fluent u can also read the same mesh into gambit , by its import).

Re: Penetrating elements in extruded mesh
 

Hi zxaar,

the problem is, that I'm writing a script, which shall create tehe geometry and the mesh. So, unfortunately, its no use to exprot and reimport the mesh with gambit, as it makes the automatisation much more less practicable.

I found by try&error, that some spacings for the extrusion cause the mesh to have these bad elements, while other spacings give good meshes, all on the same geometry and all with the same original 2d-mesh.

however, thanks for the help!