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/)
-   -   [ICEM] ICEM Tetra-Hexa (http://www.cfd-online.com/Forums/ansys-meshing/119968-icem-tetra-hexa.html)

pps June 27, 2013 11:56

ICEM Tetra-Hexa
 
1 Attachment(s)
Hey!

Because of complexity of the geometry I decided to mesh the difficult part of my spiral with tetra.
To merge it with the hexa-mesh, I load and merge it with "Edit Mesh-->Merge Nodes-->Merge Meshes".

How you can see in the file below, there are some mistakes. Some nodes don't merge.

Any ideas?
Thx
Peter

diamondx June 27, 2013 12:06

make sure the size are similar. Project sharing via dropbox is welcomed

FJSJ June 27, 2013 12:09

Ratio size must be 1.3

Try again changing hexa and tetra size to get something similar to this relation.

pps June 28, 2013 06:53

Thx for the replies.
I'll try!

pps July 1, 2013 05:46

Changing the Height ratio to 1.3 finally worked.
Thx!!

ankur_kr July 24, 2013 22:11

Too many cells in mesh
 
Hi Ali,

I needed a few suggestions regarding meshing in ICEMcfd. Currently I use unstructured tetrahedral mesh with following options, All Tri Shell Meshing, Tetra/Mixed Volume mesh type and Quick Delaunay Volume mesh method. I wanted to know the following

1) My geometry has a rectangular cuboid of dimension 12m X 3m X 3m with a couple of un-meshed cylindrical tubes (12m & 0.15m dia) through it. I entered the maximum size of mesh cell to be 0.3m. I end up getting about 3,000,000 cells and similar no. of faces!! which really slows down my simulations. [ I do have some inlet/outlet surfaces of dimensions 15 cm over which I applied 2 prism layers of size 3 cm ]. Is there a way I can reduce the no. of cells considerably ?

2) Currently I create surface mesh first and then volume and prism together. Is this the correct sequence or directly computing volume mesh without first creating surface mesh is better ?

3) Does structured mesh gives considerable advantage over un-structured mesh ?

Thanks,
Ankur

Quote:

Originally Posted by diamondx (Post 436383)
make sure the size are similar. Project sharing via dropbox is welcomed


diamondx July 25, 2013 10:58

Quote:

Is there a way I can reduce the no. of cells considerably ?
When generating a delaunay, check the hexa-core, see if it reduces the size a little bit...
Quote:

Is this the correct sequence or directly computing volume mesh without first creating surface mesh is better
it is a good sequence, you can also try an octree mesh then converting it to delaunay, see what kind of mesh you get there.
Quote:

Does structured mesh gives considerable advantage over un-structured mesh ?
Structured mesh does really help in reducing the size, it also gives a better quality if blocking strategy is good. share some picture, we can tell you then if it worth it to block it or not

ankur_kr July 25, 2013 16:50

Geometry and Mesh files
 
2 Attachment(s)
Hi Ali,

Thanks for the reply. Please find attached the pictures of my geometry.
Also, I have shared my mesh files on google-drive if you would like to see them [here the tubes are also meshed].

https://drive.google.com/folderview?...zA&usp=sharing


Attachment 23805

Attachment 23806

Thanks,
Ankur

ankur_kr August 6, 2013 16:02

Hi Ali,

I wonder if you got a chance to see my files that I uploaded.
Sorry for the inconvenience but please let me know if it would be worth to go for the blocking strategy (structured mesh) and whether for my geometry blocking would be straight-forward.

Thanks,
Ankur

diamondx August 6, 2013 19:12

I have a blocking strategy for this in my mind, but I can't apply it. I can't associate edge to curve because you don't have curve, and i need four curves for each tubes. I tried creating them but ICEM fails...you want me to share the topology in paint ???

ankur_kr August 6, 2013 20:02

Simpler geometry files
 
Hi Ali,

Thanks for giving a look at my files. Maybe U-shaped tubes are causing problems in creating curves. I am uploading a simpler geometry (single and un-meshed straight tube in meshed furnace)

https://drive.google.com/folderview?...G8&usp=sharing

My current status is that with unstructured mesh, I have been facing convergence issues in Fluent (after having played with all other simulation settings). I don't know whether a structured mesh might help me solve the convergence problem. I will have to learn blocking from scratch.

If the structured mesh gives better results for this simple geometry, then I can go ahead, learn and create structured mesh for my more complex geometries.

Please give a look at this simpler geometry file.

Thanks,
Ankur

diamondx August 6, 2013 21:48

how many elements/nodes your pc can support ?

ankur_kr August 6, 2013 21:59

I use a high performance computing cluster and generally run Fluent in parallel mode (4 CPUs). Currently the models I had been testing with had about 4,000,000 cells and similar number of faces. But smaller the number of elements it would be better, though quality of mesh would be the first priority ofcourse.

diamondx August 7, 2013 00:08

I noticed that the number of nodes is quite high but there is nothing you can do about it.
Quality can be better by moving vertices. It's up to you to play with the edge spacing depending on your problem. To sum up, it should help you determine whether it is a grid problem or the setup of you simulation:

https://dl.dropboxusercontent.com/u/...etra-tubes.rar

ankur_kr August 7, 2013 17:39

Couple of points
 
Hi Ali,

Thanks for the files. Mesh quality seems to be good and I an hopeful this mesh will give me better results. Just a couple of concerns,

1) My geometry had 2 rectangular boxes at the bottom (bounded by surfaces named 'coffin' and 'outlet'). These simply act as wall (coffin) + pressure-outlet (outlet). The volume bounded by these boxes were not supposed to be meshed but it seems these have been meshed in the files that you sent me.
Can you please remove the mesh in these 2 volume regions or tell me how to do it.

2) I see a volume part named 'Fluid' has been created. Is it simply a material point ? I already had a material point named 'Body'. Should I just delete the 'Body' material point ?

Thanks,
Ankur

diamondx August 9, 2013 00:02

Sorry i lost track on this thread, i deleted the block inside the coffins
https://dl.dropboxusercontent.com/u/...etra-tubes.rar

I created the blocks with the material FLUID, you can delete body...

ankur_kr August 11, 2013 23:35

Thanks Ali. My simulation behaves much better now and converges to a nice solution :)

findtheinvisiblecow July 16, 2014 08:25

Compute mesh not progressing
 
Hi folks,

I'm trying to compute a volume-mesh with ICEM using the 'Quick (Delaunay)' method, based on an existing Quad Dominant, Patch Dependant surface mesh.

The computation seems to be stuck at 'Tetrahedra from Existing Mesh' according to the progress bar, with '835 un-meshed faces' in the status window. (Please see the screenshot in the link below). ICEM has been stuck here for a couple of hours. While the program itself is responding just fine and the rest of the computer works fine as well.

Here are the program and system specs:
ANSYS ICEM CFD 14.5
Intel(R) Core(TM)2 Duo CPU @2.0GHz, Memory 2.00 GB

Please find links to the related files below.

Because I'm not entirely sure whether ICEM is actually still actively working on the mesh or that the system stopped, my question is what to do next?

https://www.dropbox.com/s/4cin2smjmh...screenshot.jpg
https://www.dropbox.com/s/kcjf6seg6p...ACA23012-2.tin
https://www.dropbox.com/s/ym2v8ba1hy...2-2%20surf.uns

findtheinvisiblecow July 16, 2014 10:20

I've just checked again with the settings of my Density field. Those were a bit too fine, so I've increased the size, but still keep stuck at the 'Tetrahedra from Existing Mesh' with '835 un-meshed faces'.

https://www.dropbox.com/s/vbd4nq2i5ehgi6v/NACA23~2.prj

akhme July 2, 2015 20:59

Quote:

Originally Posted by FJSJ (Post 436384)
Ratio size must be 1.3

Try again changing hexa and tetra size to get something similar to this relation.

Excuse me, what ratio size is this?


All times are GMT -4. The time now is 13:55.