[Jehan99]

Hi everyone, hope you're doing well.
I initially ran my simulation using a wedge-type mesh, and it worked fine. However, after switching to a full burner mesh, I'm encountering multiple errors, including floating-point exceptions.
I've attached images of my full mesh. What improvements should I make to ensure a high-quality mesh for accurate simulations? Any insights on potential issues that could be causing these errors would be greatly appreciated!
Wedge_mesh.png

CheckMesh_for_wedgeMesh.png

Full_mesh.jpg

CheckMesh_error_for_fullMesh.png

[dlahaye]

Not sure.

Is error caused by merging front (0 degree) and back (360 degrees) patches? Does same error occur revolving over, say, 180 degrees?

Is it more comfortable to move away from blockMesh and use tools such as cfMesh or GMSH, possibly embedded in Blender?

[Jehan99]

Originally Posted by dlahaye

Not sure.

Is error caused by merging front (0 degree) and back (360 degrees) patches? Does same error occur revolving over, say, 180 degrees?

Is it more comfortable to move away from blockMesh and use tools such as cfMesh or GMSH, possibly embedded in Blender?

Thank you very much for your reply. Is any chance for revolve this wedge mesh and make a full cylinder.

[dlahaye]

Again, not sure.

Searching online might

Possibly post number 17 at Generating 3-D axisymmetric mesh using extrudeMesh

Dixit:

If you do a full 360 extrusion, consider using "mergeFaces true;". For a full 360 extrusion with mergeFaces true, as far as I remember, extrudeMesh automatically fixes the zero face areas that are generated if you have cells in the 2D mesh that lie on the axis of rotation. If not, you can use the collapseEdges utility, which will change the cell shape from hex to whatever shape results from collapsing zero-length edges and zero-area faces.

[Jehan99]

Originally Posted by dlahaye

Again, not sure.

Searching online might

Possibly post number 17 at Generating 3-D axisymmetric mesh using extrudeMesh

Dixit:

Thank you very much.
I set the boundary condition to cyclic and run the one-fourth (90 degree) section of a burner. Then it was fine.What is the reason for that?
Screenshot from 2025-02-06 17-46-33.png

[dlahaye]

At less than 360 degrees, there is no need to merge two patches into a single path.

Think of trigonometry where it is customary to identify 360 degree as 0 degrees.

One needs to tell blockMesh that at 360 degrees, two patches collapse into a single path.

Do I make sense?