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] Operation "Mirror Blocks" defect (http://www.cfd-online.com/Forums/ansys-meshing/98297-operation-mirror-blocks-defect.html)

despaired student March 7, 2012 12:34

Operation "Mirror Blocks" defect
 
3 Attachment(s)
Hi,

today I tried to mirror some blocks but it didn't work out as I wanted it to ...
Some of the mirrored blocks edges do not fit as they should. (figure below)

How can I fix this problem. Later on I'll create O-grids around all inner cylinders and I would like to mirror them instead of creating O-grids at every 73 cylinder.

Another problem appeared when I "pre-meshed" the blocks - just for fun. (figure below)

Maybe there are some key-procedures I didn't do/ don't know. Or my strategy is bad. If someone knows a better strategy to block my geometry please tell me. I'd be happy to better understand this program. (geometry below)
It's a membrane reactor; a large cylinder with several small (membrane)cylinders inside plus in- and outlet.


Thanks in advance!

Sören

energy382 March 7, 2012 14:02

no need to create blocking for all small cylinders. just make the blocking for one single passage and rotate. there are of course other (maybe better or more efficient) ways to do the blocking. you've to try ;)

it would be helpful, if you can provide .blk and .tin files

despaired student March 8, 2012 08:14

3 Attachment(s)
Thanks for the hint Kai. For this model it is a good blocking-strategy - at least I, as an absolute beginner, think so. But the "Mirror Blocks" problem remains.
Later I want/must test two other models and compare them with my reference-model. For my other models (see attached figures) I might need the "mirror Blocks" option.
In one of my test-models I separate the large cylinder with circular plates, maybe 5-20 from bottom to top. In this model I'd like to mirror the blocks from one level down to the next.
.tin and .blk files of my reference model are attached. I can't attach the other models because ICEM imports the geometry incorrect from AutoCAD; Inner structure (helix and plates) is missing. And the autoCAD files are too big to attach.


Best regards

Sören

despaired student March 8, 2012 08:42

2 Attachment(s)
...ok, I just managed to create the .tin files from my two test-models...

energy382 March 8, 2012 08:58

first of all, before you start with your blocking, you have to prepare your geometry and define your parts (like inlet, outlet....).

second step is to think about the best blocking strategy. it makes no sense to start without that.

how does this reactor work? really don't understand. seems that there should be a closed volume, at least at the bottom!?

despaired student March 8, 2012 12:48

3 Attachment(s)
The functioning of the reactor is pretty simple. Inside there is e.g. water and in the small cylinders are hollow-fibre-membranes through which e.g. oxygen gases into the water without much turbulence like in other reactors. The inner "medium-large" cylinder is an auxiliary cylinder to support the global structure.
The inlet is the pipe at the bottom, outlet the one at the top.
So the fluid is streaming from the bottom to the top is being enriched with gas. The entire cylinder is supposed to be filled with the fluid. Membranes and inner cylinder are solid.
--> a .tin-file with the defined parts is attached.
--> closed volume is correct

Considering your suggestion for the blocking-strategy I would create a block around one of the outer small cylinder, construct an O-grid, delete the inner rectangular block, associate the inner rectangular edges to the small cylinder-curve, enhance the number of nodes, make the 4 remaining "O-grid-blocks" rotate. The same procedure for the other 3 rings of membranes. (see attached figure and attached .prj-file)

But now a problem appears:
how can I associate the blocks between the inner and outer circles to each other so that they fill the empty space?


Greetings

Sören

energy382 March 8, 2012 14:27

..................

despaired student March 9, 2012 08:13

1 Attachment(s)
...ok the answer was quite easy, I admit; "merge vertices".


Now my blocking has improved a little bit but isn't ready yet.
I tried to mesh (Pre-Mesh) it but it seems that an error or so appears. ICEM starts to work and collapses, meaning ICEM works and works and works... and nothing happens. I waited for about 15 Minutes.
Maybe the computer-power is too low or I created too many nodes. I don't know where the problem is. It would be nice if someone can tell me what I'm doing wrong.

Far March 9, 2012 09:01

attach blk

energy382 March 9, 2012 09:10

i'm stressed out at the moment, so far is your man :D

despaired student March 9, 2012 09:47

1 Attachment(s)
.blk - file is attached.

Far March 9, 2012 10:17

did you start from 2d blocking?`

despaired student March 9, 2012 10:28

No, I started with 3D-Blocking (3D Bounding Box to be exact).

Far March 9, 2012 10:41

Quote:

Originally Posted by energy382
first of all, before you start with your blocking, you have to prepare your geometry and define your parts (like inlet, outlet....).

this is really missing , you should work on this step first.

Quote:

No, I started with 3D-Blocking (3D Bounding Box to be exact).
I am thinking to start with 2-D blocking and then I shall create the 3D using extrude face along the curve. Using same command I shall also create blocking for two pipes. At the end O-grid wherever deemed necessary.

Far March 9, 2012 10:49

energy382 more ideas are welcome. Lets start the competition and lets see who makes the best blocking. ;). I shall be sharing my ideas frequently.

despaired student March 9, 2012 13:01

5 Attachment(s)
I tried to mesh with auxiliary curves and also started with 2D Blocking. Constructing one circle of blocks was no problem but when I wanted to make another block with the option "create block -> initialize blocks" it didn't work. Two messages occurred (pictures below). When I click "no" nothing happens, when I click "yes" and "yes" my other blocks disappear and one new block appears.

2D-Blocking and extrusion seems to be an easy way but it's not working...???

Attached is also my last try to mesh the geometry. Maybe you can tell me if it will lead to a successful end or if it's just a waste of time...or if I could win the competition for best blocking...:)


Best regards and have a nice weekend!

Far March 11, 2012 07:24

tin and blk :http://www.4shared.com/zip/5RTz8E-C/3D-ME0M.html
This is the my rough working on model. Worked extensively on geometry and more work still needed to be done on this aspect. Helps in making blocking.

Right now I took 90 deg sector and created 2-d block. My idea is to covert it to 3d (fixed distance option) and then using the extrude along option to create blocking for two in and out pipes.

Regarding the 90 deg sector you have two options: either use periodic boundary condition and save computational cost, or you can copy the blocking for remaining 270 deg sector.

http://img197.imageshack.us/img197/4104/56767971.png
http://img404.imageshack.us/img404/5526/28288331.png
http://img20.imageshack.us/img20/7475/79461549.png
http://img35.imageshack.us/img35/630/30809186.png
http://img594.imageshack.us/img594/8535/61418506.png

despaired student March 11, 2012 12:51

Thank you very much for your help Far.

I can't have a look on your data right now because the computer-pool at university is only open during the week from 8-18 o'clock.
Your blocking looks to me more complex than I expected the work to be...but I'm still a beginner with ICEM..:)
Thanks for your advise and the time you spent on blocking.
I wish I could say that from now on I can work with ICEM on my own but I really doubt that...:(



Best Greetings

Sören

Far March 11, 2012 12:54

I will guide you at each step. The blocking is very simple. The blocking you think is complex is the magic of o-grid in ICEM, I have done nothing except pushing the o-grid button more often.

Far March 11, 2012 13:34

Meanwhile practice four to five hexa tutorials (Version 12) and also grid fin (to have flavour of bottom-up approach) http://mecanica.eafit.edu.co/~sorrego/icemtutorial.pdf


All times are GMT -4. The time now is 10:05.