CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Complete beginner with ICEM (

dfitz1000 October 18, 2010 19:11

Complete beginner with ICEM
Hi guys, hopefully somebody can help me here.

I am a complete beginner with ICEM and I'm basically trying to teach myself how to use it which feels like I'm banging my head off a brick wall at this stage.

Ultimately I want to model a 3d wing section in SolidWorks, import it into ICEM and mesh it before bringing it into FLUENT and trying to model the flow over the wing with particular interest in the wing tip vortices generated.

I have some tutorials that I have worked through but have been no help to me as they are just "walkthroughs" and nothing is explained. Does anybody know where I can start to learn from scratch how to mesh 3d parts.

In particular what I cant understand is how to model the flow field, i.e. how do I assign where the air will come in at the wing? Can you just model the wing alone or does it need to be attached to some wall as I have seen done on the wingbody tutorial? Like I said there's lots of tutorials but no basic explanation of the steps involved.

The wing I'd like to model is a NACA-0012 wing with no taper or twist, just a basic straight wing. The tip is flat and the trailing edge is sharp. It's a simulation of a wind tunnel test that I'm going to run.

Sorry for the long post but I really feel quite lost at this stage, hopefully some of you have been in my position at some stage and can give me some help.


PSYMN October 19, 2010 13:08

Sure, you could model it as the wing in the tunnel (simulate the test), where you would have no slip walls, etc. Or you could go beyond the test and put a symmetry boco on the wing root wall and a far field boco on the other walls.

There are several tutorials for the wing in a box. Try to find them in the software under help => tutorials or on the customer sales portal...


dfitz1000 October 29, 2010 13:58

3 Attachment(s)
Hi Simon, thanks for the reply.

I have managed to mesh the wing section and the wind tunnel "box" around it, the pictures are attached.

Bearing in mind that my interest is in the vortices downstream of the wing I think my approach is wrong. Correct me if I'm wrong but should I be meshing the area around the airfoil and inside the wind tunnel "box" instead of the wing surfaces and tunnel walls themselves?

The thing is I can't figure out how to do that at all, does anyone have any ideas?

I know this is basic stuff but no tutorials I've been through deal with this and I'm basically teaching myself.


dfitz1000 October 31, 2010 09:07

So I somehow managed to figure out how to mesh the wind tunnel box using volume mesh and it was actually pretty easy once I figured out how, it's a pity I couldn't find it earlier.

Anyway I've moved onto the analysis in FLUENT and I need help here too! I guess it's best to start a new thread in the FLUENT forum.

PSYMN November 2, 2010 14:32

There are tutorials, sorry you were not able to find them. If you get another chance to look on the customer portal, it may be worth your time.

I think I also posted a link to the entire 11.0 tutorial packet several times on CFD-Online...

The images you showed were very coarse and will not adequately resolve your flow... Hopefully your final mesh was finer. Also, you may want to use a density region to better resolve the flow behind your wing in order to capture the wake, etc.


dfitz1000 January 17, 2011 14:51

1 Attachment(s)
Hi, my final mesh was indeed a bit finer (see the attached picture).

I'm now trying to refine and improve the mesh so that I can get the most accurate results. One thing that is immediately obvious is the mesh on the trailing edge is very poor as you can see from my original pictures. How can I improve this to give a better mesh on the TE which is sharp?

Also, is there anything else I can do to give me the best possible mesh?

Attachment 6085

PSYMN January 17, 2011 17:22

Prism Wake...
3 Attachment(s)
For this sort of configuration, I suggest that you project the trailing edge curve onto the outlet (or at least copy it back a ways), then create a surface between the trailing edge and the new curve. Lets call it "PRISM_WAKE" for fun. Build topology so the new surface gets curves on the symmetry planes...

Then go into PARAMS by Parts and flag "PRISM_WAKE" as an internal wall. Also turn on PRISMS for PRISM_WAKE.

Then Tetra mesh the model.

Set the curve parameters on the curve projected from the trailing edge curve so that it has a larger initial height than you set on the wing (this is to promote a transition in size... Make sure you don't set a prism height on the "PRISM_WAKE" surface its self.

Then generate prism and you should get something like this...

(don't forget to delete the shells in PRISM_WAKE when you are done.

Note... Hexa would still be better. I also posted a combo (Tetra/Prism followed by Tet to Hex conversion).

There are other more detailed conversations about this already on CFD-Online...

All times are GMT -4. The time now is 22:05.