[slaners]

Hello ,

I am trying to simulate a channel with pins inside. I defined the pins with searchable cylinders. SHM creates the castellated mesh. However, when it comes to snapping, SHM does not snap surfaces of the pins.
I attached a simple version of the geometry (using one pin for the moment to solve this issue).

Thank you in advance for your help !

[simrego]

Hi!

This 10.0 tolerance is pretty huge, you can leave it on the default 2.0 or somewhere between 1.2 and 2.0.
Do you resolve the surfaces sufficiently? Can you attach an image?

A guess: For me the implicitSurfaceSnap usually results badly snapped edges, and even if it can snap, the quality of my mesh is poor. Try to use explicitFeatureSnap.

[slaners]

Thanks for the answer simrego,
I changed the tolerance but it did not have any influence on the final mesh. I added a snapshot of a z-normal slice in the mesh. As you can see, the mesh is castellated and there was no snapping on the pin. Any thoughts ?

[simrego]

Wow. That's the snapped one? Nice.

You ran with snappyHexMesh -overwrite? If not, check the other time directories.
Have you got any warnings in the log?

[slaners]

Yeah it is. I did use -overwrite.
I have one warning :

"FOAM Warning : Displacement (-1.620465057e-08 0 0) at mesh point 1503539 coord (-0.004444983795 -0.003175 1.257804343e-05) points through the surrounding patch faces"

But the most concerning one is :

" faces with face pyramid volume < 1e-13 : 1542251 "

Even with more cells, I still get this castellated kind of mesh and there seem to be no snapping on the cylinder surface.

[simrego]

Hi!


I tried your dicts and works perfectly for me. But you should increase the number of the elements in the x direction (i tried with 120), and then you should have something like the attached image.
I tried it with OpenFOAM 4.0.

[slaners]

Hi simrego !

Thanks for the answer. You're right, the correct mesh is in another time folder !
Do you also have the "faces with face pyramid volume < 1e-13" issue ?

[simrego]

Run snappyHexMesh as:
snappyHexMesh -overwrite
and you will have the final mesh in the constant time (0).

Yes i have. Honestly i don't know the real reason for that, but with implicitFeatureSnap I usually have many of these bad faces. Maybe you can still use this mesh. But with explicitFeatureSnap I usually get rid of these errors.
I hope some professional SHM user could explain this error.