CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Merging Hexa Meshes in ICEM (

screech1987 April 7, 2011 20:21

Merging Hexa Meshes in ICEM
To model an aerodynamic body I have made the flow domain out of three seperate blockings. One being the majoirty of the flow, being rotated from a 2D blocking.
Then to improve quality in the centre of this freestream region, eradicating wedge shaped cells, I blocked this 'core' separately using a 3D O-Grid.

Now I have come to merge them, by loading in the separate blockings and merging the topologies by using the merge option by right-clicking on the blocking 'tree'.

The end result is essentially what I want but with the following problems:

1) Non-Manifold Vertices and Multiple Edges are detected during mesh checking. I assume this is due to there being two surface meshes at the blocking interfaces? If so how do I merge the meshes completely/fix the problem.

2) More of a preference, but is there a way to make the merged blockings output a mesh as a single region, i.e. the surface meshes found at the interfaces between the separate blockings are not produced.

Thanks in advance,


jeevankumarb April 8, 2011 02:08

if you post some images it will be much more clear to understand.

Thank you

BrolY April 8, 2011 07:39

For your first point, associate all your curves with the edges at the interfaces of your 2 mesh. I guess you are looking for a conformal merge, so you must have the same number of nodes on each edges of your interface, and the same node repartition.

As jeevankumarb said, pictures would help !

screech1987 April 8, 2011 07:53

Sorry, will try to upload some pics as soon as, just not in front of my normal PC today.

With regards to the assigning to curves and mesh spacings. I have associated the both interface edges to the same curves where relevant. Also have checked the number of nodes and spacings are the same on each edge before merge. and also after. The merged mesh looks as if the nodes are conforming across the interfaces matching up.

BrolY April 8, 2011 08:03

I have juste re-read your first topic : have you tried to merge your blocking ??

You should convert your 2 meshes to "unstructured mesh", and then do the merging (merge nodes with an adequate tolerance, deelte the 2 interfaces). I think it should be better.

PSYMN April 8, 2011 09:06

Remove Association to Surface...
It sounds to me like you are saying the merge worked (you had matching topologies and you merged all the verts). It is not actually necessary to match edge distributions as that is taken care of automatically...

So the remaining issue is that you have the interface faces still projected to something...

1) Check that the blocks in all three sections are in the same blocking material (like FLUID). If not, and you have different volume zones on purpose, your solver will want shells, even if they are later marked as internal walls.

2) If you are sure that you don't want/need internal shells, just go into Blocking => Associations and delete the association on those faces... Without association to surface, no shells will form.

Best regards,


screech1987 April 8, 2011 13:13

3 Attachment(s)
Thankyou for the replies so far.

I have attached some pictures as requested. A geometry Screen grab, A blocking screen grab, and a second blocking pic zoomed in on the hexa core.

Follwing on form the sugestions made, I created a body, and reset appropriate asscoaitions.
However, when creating a new body, i attempted to delete all previous ones, but none were present, so went ahead and defined one anyway.
Furthermore when restting associations I noticed duplications of blocking edges in the interface locations. I assume this is the cause of the errors encountered above in the previous post.
Also to note the resetting of associations and a new body did not solve my shell mesh at interfaces problem.



screech1987 April 8, 2011 13:19

3 Attachment(s)
A couple more screen grabs of resulting mesh, and shells at interface and a region showing the duplicate edges (hard to tell)

galap August 24, 2012 02:34

I have struggeld the last days trying to merge two hexa meshes having the same node distribution at the interface plane. PSYMN suggestion was helpful regarding the face issue but it was not the solution for the uncovered faces issue.

What I have done now:

I deleted the face associations at the interface section


Edit Mesh -> Check Mesh -> Uncovered Face -> found elements in subset

only visible elements in this subset -> Edit Mesh -> Merge Nodes by Tolerance -> set in a proper tolerance -> Apply

Edit Mesh -> Check Mesh -> Fix duplicate elements

that worked perfect for me!

PSYMN, If you read this: In the ansys support section there is a FAQ regarding the merge of two hexa meshes using "Merge volume meshes". I think you should distinguish between conform and not conform meshes.. No one wants pyramid elements when you work with conform meshes.

ozgulo August 28, 2012 02:18

did you try t-merge? it is a utility of fluent. if you have two seperate mesh files with nearly the same distribution on the interface, and they are unstructured (i mean mesh method is unstructured not tetras), you can merge them with t-merge as far as i know.

galap August 28, 2012 02:22

No I haven't, but now I have spoken with ANSYS support. The usual way is to merge different topologies. You do this in your blocking tree. Right click on root topology -> Merge. Blockings have to be in the same part.

Works perfect for me!

hemant.ghule March 13, 2014 12:45

help me to mesh shell block geometry in icem
help me to mesh shell block geometry in icem

All times are GMT -4. The time now is 21:37.