[mbahaa]

Hello,

I have previously asked about corner inflation in structured multi-body part in 2D, and I found very useful answers here, however, I am wondering how to handle the inflation at corners for structured 3D meshes.

I am trying to inflate the corner between "Wall #1" and "Wall #2" in the attached screenshot of the geometry. the mesh is structured so all bodies should be connected.

Notice that the o-grid structure at inlet may seems unconventional but this is because I had to give inlet face its own mesh since the inflow is not spanning the whole radius at inlet, and also because the boundary at outlet has different topology so the o-grid structure had to account for that, but I think that's irrelevant to the question anyways.

Thanks in advance

[Far]

the fan on the axis should be dividied as you have done. faces above this face should be circular

[mbahaa]

Hello @Far, thanks for your response, I managed to block the mesh according to what you are suggesting, but designmodeler is giving me a hard time when I try to perform "Boolean > Unite" because although the two bodies are actually united in a single body, faces are still being splitted, which doesn't help the mesher acquire the structured hexahedral mesh.

I manually merged these faces using facial merge under "Tools", the model looks fine on Designmodeler, but it's broken when I open it for meshing, and the mesher complains about some missing faces:

Mass properties (volume, surface area, centroid, moments of inertia) may be inaccurate due to missing face facets: Part

[Far]

I would suggest ICEM CFD. Much simpler approach. However learning curve may be steep

[Far]

This may help you

https://www.youtube.com/watch?v=InyeCmEuUVM

[mbahaa]

@Far

Thanks a lot for your helpful input, I was trying to block it on ICEM using the same 2D approach discussed before in this thread, and now I believe that ICEM manual blocking approach is actually the best method for meshing this geometry as you suggested, I have created this sample geometry for demonstration of the technique, however, I am unable to slice/merge the blocks on ICEM consistently to represent this geometry, I always end up with connectivity inconsistent with the geometry and/or degenerated blocks, I would appreciate your help regarding this.

I went through the following:
1. slice the initial block into 9 blocks
2. delete block #7
3. merge block #5 and #8
4. merge blocks #6 and #9
5. associated edges and vertices accordingly

Thanks in advance