CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Mesh Generation & Pre-Processing (http://www.cfd-online.com/Forums/mesh-generation/)
-   -   Internally finned pipe meshing software (http://www.cfd-online.com/Forums/mesh-generation/119473-internally-finned-pipe-meshing-software.html)

dvcauwe June 18, 2013 05:37

Internally finned pipe meshing software
 
Hello all,

Our group is currently investigating the effect of internal fins on the performance of tubular chemical reactors.

Up to this point, the fins were continuous and hence everything could simply be extruded (with or without twist) from a 2D cross-section using Gambit/Ansys Mesher. As we are not entirely satisfied with the grids this produced, and we're looking to move on to intermittent fins such as in the picture below, I would like to know of any meshing programs you might recommend from personal experience. Main requirement is the very fine meshing of a surface boundary layer.

http://i41.tinypic.com/35b80w3.png

Thank you very much in advance for any tips you can offer!

Best regards,

David

-mAx- June 18, 2013 08:29

I don't know what you did with Gambit/Ansys Meshing, but maybe you could get more than what you achieved.
In other words, if you are open to reconsider your grid with Gambit/Ansys Meshing, I could move your thread to the right subforum, and you can get some feedbacks.
I am curious what you did with Gambit, and what doesn't get your satisfaction.
Please attach pictures.

dvcauwe June 19, 2013 11:43

Dear Maxime,

You are right that perhaps I should reconsider my method of working. In the past we've been working on continuous fins with much smaller helix angles and there we were able to simply extrude from a 2D cross-section. As soon as we started using angles of greater than 40 however, we had problems with skewed cells and negative volumes and so on.

For this case we have non-continuous fins (and angles of over 65) so extrusion is not possible anymore afaik. I have attached a more detailed picture of the geometry, although the full case would be about 500 times repeats of what is shown here.

http://i42.tinypic.com/2reh5xc.png

At the moment I'm thinking the best way is probably to create an inner cylinder in which I can use a simple hex mesh and then use tets on the complex outer ring... But doing it this way is still giving me a lot of skewed and inverted elements. Things can probably also be made much easier by simply meshing a single fin and then copying this whole zone a few hundred times...

Seeing your profile picture I am certain you have handled much more complex geometries so I would be very grateful to hear how you would handle this problem!

Best regards,

David

-mAx- June 20, 2013 02:08

Do you have 3d pictures of what you tried to mesh (I woul like to see what you want exactly)?
Do you want full hexa mesh, or are you open to get hybrid mesh?
I would first build your model, and then start splitting and meshing.
Do you have the possibility to import your geometry from a cad package?

sameben July 25, 2013 09:10

1 Attachment(s)
Hi David,
Are you are looking for something like the one I have attached?

Attachment 23788

Far July 27, 2013 16:08

Gridpro is good package. But for difficult models I think best method is to use hexa + tetra mesh.


All times are GMT -4. The time now is 07:38.