[openThanos]

Good morning,

Im new to the forum, thank you for having me. I anticipate that this is not a previously covered topic because Ive done my due diligence searching high and low. I apologise in advance if Im repeating.

I'm new to openFoam; at the stage where I've challenged myself to model my own geometry from scratch (on the framework of the motorbike tutorial). Using the motorbike tutorial as a framework and following Calum Douglas' tutorial (https://www.youtube.com/watch?v=ObsFQUiVi1U) for guidance with my own geometry (grabcad sourced exhaust manfd.)

I am able to 'mesh' and start the run making it a few iterations before diverging. I suspect this is due to my mesh. More specifically, when looking at the mesh in paraView, the inlet and outlet boundaries are patch and discontinuous. That is it appears that not every node on the boundary is associated with the inlet region. Please see attached photo. I suspect this needs to be fixed - and have been working at it for a few days now. Ive been trying a few different parameters in my shmDict, surfFeatExtractDict, surfaceFeaturesDict, blockmeshDict. They do influence the outcome, but not drastically.
Can anyone help me figure out why this is happening and help me fix it?


Notes on the geometry:
The geometry was imported to solidworks and exported as three asci STL files; one solid domain encapsulating all geometry as a solid [volume.stl], and two surface STLs for the inlet [inlet.stl] and outlet [outlet.stl] . I created a fourth merged geometry by concatenating these STLs [merge.stl] and am using that for my 'surfaceFeaturesDict' but not my 'surfaceFeatureExtractDict'. I should not that using volume.stl instead does not change the outcome.

Having exported these stl files independently, I suspected that variations in the exported geometry may be causing this. I used blender to create new inlet and outlet meshes as carbon copies of the volume mesh in the respective regions. This did not change the outcome. The regions are still patchy.


Do you folks have any insight or recommendations? Happy to elaborate or provide the case files.

Many thanks!
Thanos

Please see attached for shmDict, surfFeatExtractDict, surfaceFeaturesDict, and blockmeshDict

[MuriloHenriqueMoreira]

Exactly the same problem here!
I'm so close to get what I want hahaha.

[simrego]

Hi!


Make sure that the "volume" stl doesn't contain the inlet/outlet patches. If there are 2 stls (boundary and inlet), snappy won't know what patch should be these faces, since they are the same.
Based on my experiences also try to avoid a perfectly fitted background mesh because in this case you can have the patch on some faces from the background mesh.


Also check your featureEdges. You can set "writeObj yes;" and you can open the obj files into paraview. It seems like snappy didn't snapped to the features (edges).

[openThanos]

simrego,

You nailed it. Volume stl included the inlet and outlet. I split the surfaces out and used them as my inlet and outlet and got everything working.

In your foam experience, is it important to snappy to have the STL vertices intersect between inlet and volume bodies? I.E. is it best practice to export the volume stl from cad with the Inlet and Outlet boundaries, and subsequently break them out using a program like blender? Or just export them individually?

Thank you all,
I hope this thread helps someone in the future.

[simrego]

It is not necessary to have coincident vertices. At least based on my experiences. I usually just export the face groups from salome with a simple export as stl for each patch, so the neighbouring stls usually have different vertices/edges. And snappy works perfectly. Sometimes (maybe had to do once or twice) I have to increase tolerance when I export the stl (to capture the surface more precisely if it there are a really difficult surfaces near each other in different patches), but most of the time it is working perfectly without bleeding.