CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ANSYS Meshing] Finned tube Hexa mesh (

aliemadi May 12, 2012 04:20

Finned tube Hexa mesh
2 Attachment(s)

I'm trying to simulate a crossflow heat exchanger. water flow inside pipe and hot gas flow over finned tube. I'm trying to use Hexa mesh to reduce the number of generated mesh and also due to nature of Geom(thin fins and small gaps between them which gas flow through it). at this step just consider a finned tube which is surrounded by cylinder. I tried to use multizone method to mesh Inside water,finned tube and gas between fins(a multibody part) but generated mesh is not compeletly Hexa. how can I mesh such a Geom by Hexa? I'm using Ansys mesh 12

Far May 12, 2012 05:26

I can help you with ICEM CFD for this geometry

aliemadi May 12, 2012 05:41

Hi Far
this Finned tube is just one section of my Geometry. the compelet jeometry consist of burner, U tubes and is very complex.also there are18 finned tube which every tube has 50 fins!I have generated acceptable mesh for other parts and just finned tube is not complete. so I prefer to mesh it in ansys mesh. Is it possible?

Far May 12, 2012 05:44

Ya. It is possible in ICEM CFD, but will take time. Also you can try the ICEM CFD tetra + prism.

But I think for your case AM meshing with multi zone (technology from ICEM-CFD) would be the good choice.

aliemadi May 12, 2012 06:03

I tried to mesh by multizone but I dont know how I have some question:

1- this three body should be in one part or each body as a part?
2- if one part is enough then in choosing method I should select three body as Geometry in scope or I should define three method? as I know the scop of multizone is part not body
3- which surfaces should be selected as soureses. I examined different surfaces but some free blocks created

PSYMN May 14, 2012 19:58

This should be done as one multi-body part. It looks like it is chopped up enough to work. If you can post this "piece" for us to try, we can probably figure it out and tell you how... If we can't I will pass to development.

Best regards,


All times are GMT -4. The time now is 10:38.