CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Advice on a blocking strategy (

pipolaki July 17, 2012 13:06

Advice on a blocking strategy
Good morning.

I'm a complete beginner in ICEM and I've been working on different tutorial since last week in order to understand the philosophy of this high level meshing tool.

Since yesterday I'm involved in a new project and I have to mesh a swirl injector. Here is a view from behind of the swirl chamber with its four intake ducts which are connected to the main manifold. Orange part is the inlet and blue parts are actually the walls.

My question is : what could be the block topology of a such geometry ? :confused:
The central cylinder can be modelized by a block with a O-grid but I can't figure out what could be the shape of the block that can be used for each intake duct and especially for the main manifold which is all around.

Any suggestions on the way that I can cut and arrange my blocks would be appreciate.
Thank you in advance for any help and have a nice day.

diamondx July 18, 2012 00:01

although complex. the first thing coming to my mind is to try to mesh on one quarter your geometry. it should be a way so you can rotate it so it can have this final geometry. i will start by an initial block of the central cylinder, then performing some split and extruding along the tube . i don't have any idea about what's next... not sure if it's even possible to mesh the rest... need to improvise. give it a try.
Sorry i can't acess icem right now... i would have asked you to share you geometry so i can give it a try.

pipolaki July 18, 2012 03:32

Thanks for your comment. If somebody needs the geometry to give a try, tell me want kind of file I have to post. Is the STEP file enought ?

Anyway, I'm going to work on this injector all day long and if I realize some progress, I will post the results here.

I'm still open to any suggestions.


BrolY July 18, 2012 03:39

Post the .tin file so we can have a look ;)

pipolaki July 18, 2012 13:27

Here is the tin file.

No progress from my side... Tomorrow I will show you different topologies I tried without success.
Despite its very simple geometry, this problem is really difficult for me.

BrolY July 19, 2012 05:19

1 Attachment(s)
1) Your geometry is not good. You need to trim surfaces to let the flow go through the different parts.
2) Attached an example of blocking (with an "open" geometry) which is NOT completed. You need to move vertices to optimize the blocking and maybe add O-grid. It's just to give you some indications.

Let see if others users (they will recognize themselves) can share their experience ;)

pipolaki July 19, 2012 10:00

Dear BrolY,

Thank you very much for your answer.
1) I'm going to work on the geometry as you said.
2) Your blocking strategy is quiet impressive and I will need time to completely analyze it. In fact, I was far, far away from this solution !!! Thanks.

BrolY July 19, 2012 10:07

1 Attachment(s)
I took another look to the blocking and I made a mistake to handle the geometry between the part 3, 4 and 5. I attached the new files.
I forgot to trim the surface at the connection between the pipe and the rectangle.
And I changed the blocking at this connection and add an O-grid to improve a little bit the quality.

Do not hesitate to ask question ;)

pipolaki July 20, 2012 13:01

I think I begin to understand the principle of the blocks in ICEM... Even if I'm still far away from the solution proposed by BrolY, please, have a look on the solution I propose and tell me what you think.

1) First of all, the initial geometry was too ambitious, so I simplified it in order to focus on the difficult part of the problem. Here is the new geometry whith only the swirl chamber, only one inlet manifold and the main manifold all around.

2) Now here is my blocking strategy. I tried to keep it simple because when there are vertices everywhere I'm lost. It's certainly not the best blocking strategy but it helps me to understand the main features of ICEM.

3) And now here is the mesh. It's a fist shot so the mesh quality is not very good. I think I can improve it. We'll see...

Let me know what you think about this solution and thank you very much to BrolY and his fruitful comments.

BrolY July 23, 2012 03:54

I can't see your images ...

pipolaki July 23, 2012 05:13

It should be OK now.

BrolY July 23, 2012 05:39

Your strategy looks good !
You can add an o-grid inside your only inlet manifold to improve the quality.

If you are lost when there are too many vertices, you can play with the index blocking (right click on blocking -> index blocking).

pipolaki July 23, 2012 12:26

Now, I'm going to work on the complete geometry with O-grid inside the inlet manifolds.

Concerning the use of the index blocking, I don't know this feature : I'm going to look inside the Help Manual and I think that I'm going to learn something new today. :)

I will post the final result with the graph of the mesh quality in few days.

Far July 23, 2012 17:46

nice work. good job

All times are GMT -4. The time now is 17:02.