ICEM's Block strategy needs high IQ..
 

Hi guys,

I wanna make hex mesh on this mixer. all is fine to me if there is no this rod in it. but I dont know how should I build block on this whole geometry. any one can show me some hints?

Thanks,

Could you share the geometry file?

https://www.dropbox.com/s/kceyaoyzzo...imple.rar?dl=0

pls see this link. Thank u bro

Does the mixer move separately from the stators? I would guess so.

I suggest you add a cylinder around the mixer to separate things and treat this as two simpler blockings. Each blocking forms the mesh for a zone. The outer zone is fixed and the inner zone will be rotating.

Both blockings should start with an Ogrid of the cylinder and then just split from there to capture the boundary baffles... Pretty straight forward once you know how. ;)

Thanks man,

I saw u were extremely professional in ICEM field. Actually I already did as u said, all is fine if I delete the rod, see this:
http://www.cfd-online.com/Forums/att...1&d=1418290040
http://www.cfd-online.com/Forums/att...1&d=1418290040

u can see there is no mesh on this rod. Because Im really confused how to build o-block on these blocks. see:

http://www.cfd-online.com/Forums/att...1&d=1418290040

http://www.cfd-online.com/Forums/att...1&d=1418290040

Splitting blocks is tough....

If I were you I would create bloks something like this:
https://www.dropbox.com/s/s0631zd8ly...nline.zip?dl=0
:)
And It was not bad mention about two domains from PSYMN was done.

Sure... The red pipe is not a problem. OGrids are the secrete here also. But first, does the red pipe inject fluid into the mixer (is it open)? Or is it really just an obstacle? It would cause a slight change to how I would block it.

Basically, you should split to create a block around the horizontal portion of the pipe. This block would be "radial" if you are thinking about the whole model. Basically, It would be within one of those outer blocks from your first OGrid.

Put an Ogrid inside this block with a face where the pipe exists the domain.

Then Split the OGrid to capture the step in the pipe radius. If the pipe is closed to the flow, you just need the one OGrid split. If the pipe is open, you should make sure the OGrid is outside the pipe, so you will need two spits, one for each cylinder radius. Associate the edges, etc.

You can limit your splits with your index control so they don't go all the way around.

Then isolate a vertical block(s) for the vertical portion of the pipe up and into the horizontal pipe. Put Ogrids in these blocks with faces where the pipe exits the domain.

Apply the OGrid and fit it to the pipe... Actually, I guess even if the pipe is just an obstacle, you will have a better boundary layer around it if you split the OGrid so you have some outside and some inside...

Finally I succeed! I love ICEM!

Sorry guys, I did not reply this thread, cause last year Im quite newbie on ICEM. Now I have some progress. haha. But newbie too.

http://www.cfd-online.com/Forums/att...1&d=1430477621

Hi guys,
I'm trying to generate a Hybrid mesh for a Lobed-Mixer in ICEM. I used SolidWorks to come up with the geometry. The computational domain consists of three lobes and a quarter of the conical plug. The final mesh should be as follow: Hexa cells on the lobes and the plug with some inflation layers around the lobes and Tetra cells for the rest of the domain. The problem is that when I try to block the lobes I get lost pretty quickly.There are so many splittings and it gets really messy. I tried to go with 2D surface blocking and the MultiZone Fill to generate the volume mesh but the mesh has problems and the hexa cells are unstructured and I need structured hexa cells. So I should go with blocking. Has anybody done something similar to this problem?
Attach please see the geometry and the domain

Thanks in advance

Do you really need hexa ???? why not tri ??
sometimes it is not worth it to go hexa... It will generate too many elements.

Dear Ali,
Thanks for your reply, If it was in my hand I definitely would use Tri cells but my advisor wants quad cells on the lobes and plug, we are trying to capture the shear layers and I should generate hexa cells. Its been a week that I'm involved with it without any progress. It's becoming frustrating

Hello,
I also have a problem with bloking in ICEM.
I created mesh in Ansys Mechanical APDL. It's easiest wa for me to draw geometry and then generate mesh, but this program cannot create inflation.
I'm simulating free surface and wall swirls in a cylinder. I need hexa all the way. I got nice hexa from Ansys but now it's time for inflation near the wall.
Icem offers such thing but it needs bloking, as it is in the Bloking-> Pre mesh parameters -> edge params.
I've watch many tutorials on bloking but it doesn't work, same outcome "no body was created".
One very important detail: I load MESH form Ansys, from def file.
I can't even create points (first step for blocking) as they end up outside the mesh.
Can you suggest anything? Some video tutorial or where I could find answers or what should I look for.
I'd very grateful for answer!

Hi Guys,

I decided to use tetra cells with inflation layers for the geometry that I described above. I tried to but there are some problems. when I display color by counts all of the curves are red. No yellow or blue curves. But when I generate the volume mesh and check the mesh it gives me so many problems.
There are holes in the mesh and there isn't any mesh on some surfaces. I used the build diagnostic topology but it didn't fix the problem. Does anybody no what the problem is??

Thanks