CFD Online Discussion Forums - [ICEM] tubes with different diameter

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)

- ANSYS Meshing & Geometry (https://www.cfd-online.com/Forums/ansys-meshing/)

- - [ICEM] tubes with different diameter - Ogrid questions (https://www.cfd-online.com/Forums/ansys-meshing/159181-tubes-different-diameter-ogrid-questions.html)

tubes with different diameter - Ogrid questions

Hi guys,

I'm teaching myself some ICEM meshing and after doing most of the tutorials, I'm getting the idea of how it works. For a bigger project, I would like to mesh two tubes/cynlinders with a chamfer between them. The bigger one will later be an inlet/outlet.

Picture 0 shows the blocks before the vertex association:
http://s8.postimg.org/b4f8zedht/ICEM...fore_Ogrid.jpg

Picture 1 shows my blocking and vertex/curve association:
http://s24.postimg.org/5z6jwxntt/ICEM_Help_3.jpg

Picture 2 shows my blocking after applying Ogrid:
http://s2.postimg.org/6fgumecp1/ICEM_Help_2_blocks.jpg

Picture 3 shows my completed mesh
http://s27.postimg.org/yjx4he0dr/ICEM_Help_1.jpg

The problem is obviously that either Ogrid cannot split my blocking of both tubes at once, or I cannot find the right options to set it up correctly.

1) Do I have to apply the Ogrid split for each of the tubes individually?
2) Which faces do I have to select for the Ogrid? Right now, I am selecting all flat surfaces (top and bottom, not the cylinder surfaces).
3) Do I have to set the initial associations or blocking differently?
4) Regarding the first picture I posted: Is there a way how to merge the front faces of the bigger tube so that Ogrid doesn't split the bigger tube in a weird way?

Of course I can provide more screenshots or information if needed.

Thanks for your advices.

Best regards,
Maxim
http://postimg.org/image/yjx4he0dr/

Quote:

Originally Posted by -Maxim- (Post 563330)

Hi guys,

I'm teaching myself some ICEM meshing and after doing most of the tutorials, I'm getting the idea of how it works. For a bigger project, I would like to mesh two tubes with a cone which will later be an inlet/outlet.

Picture 0 shows the blocks before the vertex association:

Picture 1 shows my blocking and vertex/curve association:

Picture 2 shows my blocking after applying Ogrid:

Picture 3 shows my completed mesh

The problem is obviously that either Ogrid cannot split my blocking of both tubes at once, or I cannot find the right options to set it up correctly.

1) Do I have to apply the Ogrid split for each of the tubes individually?
2) Which faces do I have to select for the Ogrid? Right now, I am selecting all flat surfaces (top and bottom, not the cylinder surfaces).
3) Do I have to set the initial associations or blocking differently?
4) Regarding the first picture I posted: Is there a way how to merge the front faces of the bigger tube so that Ogrid doesn't split the bigger tube in a weird way?

Of course I can provide more screenshots or information if needed.

Thanks for your advices.

Best regards,
Maxim
http://postimg.org/image/yjx4he0dr/

One of possible ways:
1. Make block for big cylinder.
2. Make o-grid passing through circular faces.
3. Extrude central face of o-grid to make block for small cylinder.
4. Make o-grid passing through small and big circular faces.

Quote:

Originally Posted by -Maxim- (Post 563330)

Hi guys,

I'm teaching myself some ICEM meshing and after doing most of the tutorials, I'm getting the idea of how it works. For a bigger project, I would like to mesh two tubes with a cone which will later be an inlet/outlet.

Picture 0 shows the blocks before the vertex association:
http://s8.postimg.org/b4f8zedht/ICEM...fore_Ogrid.jpg

Picture 1 shows my blocking and vertex/curve association:
http://s24.postimg.org/5z6jwxntt/ICEM_Help_3.jpg

Picture 2 shows my blocking after applying Ogrid:
http://s2.postimg.org/6fgumecp1/ICEM_Help_2_blocks.jpg

Picture 3 shows my completed mesh
http://s27.postimg.org/yjx4he0dr/ICEM_Help_1.jpg

The problem is obviously that either Ogrid cannot split my blocking of both tubes at once, or I cannot find the right options to set it up correctly.

1) Do I have to apply the Ogrid split for each of the tubes individually?
2) Which faces do I have to select for the Ogrid? Right now, I am selecting all flat surfaces (top and bottom, not the cylinder surfaces).
3) Do I have to set the initial associations or blocking differently?
4) Regarding the first picture I posted: Is there a way how to merge the front faces of the bigger tube so that Ogrid doesn't split the bigger tube in a weird way?

Of course I can provide more screenshots or information if needed.

Thanks for your advices.

Best regards,
Maxim
http://postimg.org/image/yjx4he0dr/

Hi!
are you looking for it?
http://i.imgur.com/MVdAo6l.jpg

associated files regarding that pic are:
https://drive.google.com/file/d/0B4Y...ew?usp=sharing

Thank you two for your answers.
I was trying to make to top-down method going to work but extruding a block from an Ogrid face is really helpful in this case!

The difficult part though is the chamfer between the two cylinders. I just took it as a third cylinder and extruded the face again. So far so good. Now I'm trying to fix those cells around the chamfer with very little angles.
Is there an easy way to model chamfers in ICEM?

Quote:

Originally Posted by -Maxim- (Post 563437)

Thank you two for your answers. I will try to make two Ogrids - one for each cylinder.
The difficult part though will be the chamfer between the two cylinders. I guess I will have to model that like another cylinder? So three Ogrids?

2 o-grids, for ex.:

thank you again Antanas, that worked great so far.
My chamfer looks like that at the moment:
http://s22.postimg.org/d1dt96a7x/ICEM_chamfer.jpg

which gives me cells with bad angles:
http://s12.postimg.org/67tgawvmx/ICE..._bad_cells.jpg

Can I correct that with cell size settings etc or do I have to set up my Ogrid differently in the first place?

I merged the blocks of the chamfer like this:
http://s14.postimg.org/y13r3ln9p/ice...cks_merged.jpg
and now the bad angle cells are gone! Woot woot.

Now onto mesh refinement :)
Thanks again