CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   Problem with geometry (http://www.cfd-online.com/Forums/ansys-meshing/91296-problem-geometry.html)

sheth August 5, 2011 23:50

Problem with geometry
 
Hi all,

I am doing analysis of truck body. Here is the pic of truck that I am making in ICEM. I have geometry constructed in CATIA. Then I import in ICEM.

The problem is to reduce analysis time I want to create symmetry plane. SO I want to cut the body in half. Is there any way to do such operation in ICEM?

and I want to create meshing near the surface very small and the region far away coarse so How do I do such meshing?

Thanks in advance. :)

Mohankumarg12 August 8, 2011 02:39

Create a surface bigger than your geometry at the center and use the option "surface surface interaction" in "create/Modify curve". It will create a curve at the intersection then using that curve split the surface into two using the command " Segment/Trim surface" in "Create/Modify surface. Remove the unwanted surface and close the open surfaces using the option " Simple surface" option in "Create/Modify surface" to make it water-tight.

For getting a finer mesh near the geometry and coarser mesh in the far field
Specify a minimum size which is required to capture your geometry and give a maximum size that can be possible to put it in your geometry from the "Mesh"------"Surface mesh setup".

Regards,
Mohan.G

sheth August 8, 2011 03:00

Mohan, thanks for reply.

That cleared my doubt.

I wanted to know if I just setup min and max parameters in surface mesh as you said, won't it create a 2d mesh? I need a 3d mesh.

Mohankumarg12 August 8, 2011 03:26

The size that you are mentioning is for the surface mesh(2D) and the 3D mesh will depend upon your surface mesh.

I think I am right but I need clarification from an expert.

Mohankumarg12 August 10, 2011 03:26

Simon, Can I have your comment on this thread?

sheth August 15, 2011 12:50

One more thing. Will cutting geometry in half affect results?

stuart23 August 16, 2011 06:23

In order to get the correct mesh refinement near the surface your most interested in, you will need to assign different surface mesh sizings to the surfaces. The easiest way of doing this is to create different families for your inlet, outlet, symmetry, ground, far bounds and truck surfaces. Once these have been established you can set the required mesh sizing using Part Mesh Setup. If you wish to further refine specific surfaces, you can use Surface Mesh Setup to assign mesh sizings to specific surfaces.

You can then use mesh densities to refine surface and volume meshes in specific areas of interest without having to subdivide your domain and mesh volumes individually. For a complicated model such as yours, you may need to use any number of the above methods to create a mesh that will give you usable results.

For a steady state model, you could split the geometry along the symmetry plane. However I know people currently studying truck wakes, and I can assure you that the flow is very transient dominated. In order to accurately model the flow behavior, you should do transient simulations using both halves of the model.

Good Luck

sheth August 17, 2011 09:30

Quote:

Originally Posted by stuart23 (Post 320245)
In order to get the correct mesh refinement near the surface your most interested in, you will need to assign different surface mesh sizings to the surfaces. The easiest way of doing this is to create different families for your inlet, outlet, symmetry, ground, far bounds and truck surfaces. Once these have been established you can set the required mesh sizing using Part Mesh Setup. If you wish to further refine specific surfaces, you can use Surface Mesh Setup to assign mesh sizings to specific surfaces.

You can then use mesh densities to refine surface and volume meshes in specific areas of interest without having to subdivide your domain and mesh volumes individually. For a complicated model such as yours, you may need to use any number of the above methods to create a mesh that will give you usable results.

For a steady state model, you could split the geometry along the symmetry plane. However I know people currently studying truck wakes, and I can assure you that the flow is very transient dominated. In order to accurately model the flow behavior, you should do transient simulations using both halves of the model.

Good Luck

Thanks a lot for giving proper guidance.

I would like to ask, if I am finding drag co-efficient , should I do steady state analysis or should I go for transient analysis.


All times are GMT -4. The time now is 21:07.